MICROFLUIDIC DEVICES IN NANOTECHNOLOGY

MICROFLUIDIC DEVICESIN NANOTECHNOLOGY

Applications

Edited by

CHALLA S. KUMAR


Applications

Edited by

CHALLA S. KUMAR

Copyright � 2010 by John Wiley & Sons, Inc. All rights reserved

Published by John Wiley & Sons, Inc., Hoboken, New Jersey

Published simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by

any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted

underSection107or108of the1976UnitedStatesCopyrightAct,without either thepriorwrittenpermission

of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright

Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or

on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the

Permissions Department, JohnWiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011,

fax (201) 748-6008, or online at http://www.wiley.com/go/permission

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in

preparing this book, theymakeno representationsorwarrantieswith respect to the accuracyor completeness

of the contents of this book and specifically disclaim any impliedwarranties ofmerchantability or fitness for

a particular purpose. No warranty may be created or extended by sales representatives or written sales

materials. The advice and strategies contained herein may not be suitable for your situation. You should

consult with a professionalwhere appropriate. Neither the publisher nor author shall be liable for any loss of

profit or any other commercial damages, including but not limited to special, incidental, consequential, or

other damages.

For general information on our other products and services or for technical support, please contact our

Customer Care Department within the United States at (800) 762-2974, outside the United States at (317)

572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not

be available in electronic formats. For more information about Wiley products, visit our web site at www.

wiley.com.

Library of Congress Cataloging-in-Publication Data:

Microfluidic devices in nanotechnology. Applications / edited by Challa S. Kumar.

p. cm.

Includes bibliographical references and index.

ISBN 978-0-470-59069-0 (cloth)

1. Microfluidic devices. 2. Nanofluids. 3. Nanotechnology. 4. Fluidic

devices. I. Kumar, C. S. S. R. (Challa S. S. R.)

TJ853.4.M53M5325 2010

620.1006–dc22 2009051009

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

http://www.wiley.com/go/permission

www.wiley.com

www.wiley.com

CONTENTS

Preface vii

Contributors xi

1 Microfluidics For Nanoneuroscience 1

Pamela G. Gross and Emil P. Kartalov

2 Nanoporous Membrane-Based Microfluidic Biosensors 47

Shalini Prasad, Yamini Yadav, Manish Bothara, Vindhya Kunduru, and

Sriram Muthukumar

3 Nanoparticle-Based Microfluidific Biosensors 91

Giovanna Marrazza

4 Microfluidic Enzymatic Reactors Using Nanoparticles 125

Chunhui Deng and Yan Li

5 Microfluidic Devices for Nanodrug Delivery 187

Clement Kleinstreuer and Jie Li

6 Microchip and Capillary Electrophoresis Using Nanoparticles 213

Muhammad J. A. Shiddiky and Yoon-Bo Shim

7 Pillars and Pillar Arrays Integrated in Microfluidic Channels:

Fabrication Methods and Applications in Molecular

and Cell Biology 255

Jian Shi and Yong Chen

v

8 Nanocatalysis in Microreactor for Fuels 281

Shihuai Zhao and Debasish Kuila

9 Microfluidic Synthesis of Iron Oxide and Oxyhydroxide

Nanoparticles 323

Ali Abou-Hassan, Olivier Sandre, and Val�erie Cabuil

10 Metal Nanoparticle Synthesis in Microreactors 361

Peter Mike G€unther, Andrea Knauer, and Johann Michael K€ohler

Index 395

vi CONTENTS

PREFACE

I hope you had an opportunity to go through the first volume. It gives me immense

satisfaction in placing the second volume of the two-volume book series—

Microfluidic Devices for Nanotechnology: Applications—in your hands. The second

volume is the first book ever to be published that covers nanotechnology applications

using microfluidics in a broad range of fields, including drug discovery, biosensing,

catalysis, electrophoresis, enzymatic reactions, and synthesis of nanomaterials.While

the first volume,Microfluidic Devices for Nanotechnology: Fundamental Concepts,

in its combined formprovides readers anup-to-date knowledgeof thefluid andparticle

kinetics, spatiotemporal control, fluid dynamics, residence time distribution, and

nanoparticle focusing within microfluidics, the second volume primarily captures up-

to-date applications. The book fills in a long-term gap that existed for the real-time

measurement of biomolecular binding in biosensors and justification for incorporating

nanoporousmembranes into “lab-on-a-chip” biosensing devices. Focusing on lab-on-

a-chip systems for drug delivery (also called bio-MEMS), separating bioanalytes

using electrophoresis, genomics, proteomics, and cellomics, the book is a must for

biologists and biochemists. Highlighting the importance of nanoneuroscience, the

book educates the reader on the discipline ofmicrofluidics to study the nervous system

at the single-cell level and decipher physiological processes and responses of cells of

neural origin. For a nanomaterials chemist interested in novel approaches for synthesis

of nanomaterials, this book is an excellent source of information covering a wide

variety of microfluidic-based approaches for synthesis of metallic and nonmetallic

nanomaterials. Finally, opening a window for the next-generation alternative energy

portable power devices, nanocatalyst development for industrially useful reactions in

silicon-basedmicroreactors is discussed especially in the contextof syngas conversion

to higher alkanes, which could solve current difficulties of storage and transportation

vii

by converting natural gas into liquid fuels. Overall, the book contains reviews by

world-recognized microfluidic and nanotechnology experts providing strong scaf-

folding for futuristic applications utilizing synergy between microfluidics and

nanotechnology.

Chapter 1 byDrs. Pamela G. Gross and Emil P. Kartalov focuses on the application

of microfluidic devices to study the nervous system at single-cell level using

nanotechnologies. This chapter describes various aspects of microfluidic chips

used to decipher physiological processes and responses of cells of neural origin

with examples of novel research not previously possible. Continuing on a similar

theme, Chapter 2 by Professor Shalini Prasad et al. provides a detailed account of real-

time biomolecular sensing through incorporation of nanoporous membranes, man-

made as well as natural, into “lab-on-a-chip” biosensing devices. In addition to

nanoporousmembranes, simple spherical nanoparticles are finding novel applications

when incorporated within the microchannels. Chapter 3 by Professor Giovanna

Marrazza reviews the most recent applications of nanoparticles within microfluidic

channels for electrochemical and optical affinity biosensing, highlighting some of

their technical challenges and the new trends. Chapter 4 by Professors Chunhui Deng

and Yan Li presents the recent advances in the field of immobilized microfluidic

enzymatic reactors (IMERs), which constitutes a new branch of nanotechnology. In

viewof the increasinguse of lab-on-a-chip systems in thehealthcare industry, there is a

growing demand for discovery, development, and testing of active nanodrug carriers

within the microfluidic environment for controlled drug delivery. Chapter 5 by

Professor Clement Kleinstreuer and Jie Li provides a comprehensive treatise on

fundamentals and applications of microfluidics and bio-MEMS with respect to

nanodrug targeting and delivery.

Capillary electrophoresis (CE) and microchip electrophoresis (MCE) are two

promising separation techniques for analyses of complex samples, in particular,

biological samples. Not surprisingly, these techniques have been profoundly influ-

enced by the advances in nanotechnologies. Chapter 6 by Muhammad J. A. Shiddiky

and Professor Yoon-Bo Shim covers the recent developments and innovative applica-

tions of nanomaterials as stationary and/or pseudostationary phases in CE and MCE.

This chapter illustrates the importance of various types of nanomaterials, including

metal and metal oxide nanoparticles, carbon nanotubes, silica nanoparticles, and

polymeric nanoparticles, in enhancing the separation of biological samples using CE

and MCE. The examples we have seen so far involve externally fabricated nanoma-

terials, which are later on utilized for a number of applications within themicrofluidic

channels. Chapter 7 by Drs. J. Shi and Yong Chen discusses pillars and pillar arrays

integrated into microfluidic chips in the fabrication process itself. This chapter

demonstrates how such an approach provides a large variety of functionalities for

molecule and cell biology studies.

The applications we have seen so far in the first seven chapters range from biology

to drug delivery. Chapter 8 by Shihuai Zhao and Professor Debasish Kuila is uniquely

placed in thebookas it bringsout the recent recognition formicroreactor as anovel tool

for chemistry and chemical process industry, such as fuel industry. This chapter

presents silicon-based microreactors for the development of nanocatalysts for

viii PREFACE

industriallyuseful reactions. For example,methanol steam reformer toproduceH2and

CO purifier is described in detail for potential microreactor applications in the next

generation of alternative energy for portable power devices.

The last example that the book provides is the application of microfluidic reactors

for the synthesis of nanomaterials. With the increase in the demand for high-quality

metal nanoparticles with narrow size, shape distribution, and homogeneous compo-

sition, the continuous-flow microfluidic processes are gaining attention as they are

particularly suited for realizing constant mixing, reaction, and quenching conditions

necessary for production of high-quality metallic nanomaterials. Chapter 9 by Dr. Ali

Abou-Hassan et al. reviews the recent scientific literature concerning the use of

microfluidics for the synthesis of the iron oxides nanomaterials. Chapter 10 by

Professor J. Michael K€ohler and coworkers is a fitting conclusion to the book

delineating a number of promising opportunities and challenges for the application

of microreaction technology for the synthesis and manipulation of metallic nano-

particles. In combination with the Chapter 9 in Volume 1, this will provide a strong

platform from both theoretical and experimental perspectives on synergism between

microfluidics and nanotechnology for automated microreactor-based controlled

synthesis and engineering of nanomaterials for a number of applications.

In conclusion, the two volumes bring out a clear understanding of theoretical and

experimental concepts of microfluidics in relation to nanotechnology in addition to

providing a seamless transition of knowledge between and micro- and nanofluidics.

The contributors for both the volumes are world-renowned experts exploiting the

synergy between microfluidics and nanotechnology. I am very much grateful to all of

them for sharing my enthusiasm and vision by contributing high-quality reviews, on

time, keeping in tunewith the original design and theme of both the volumes. Youwill

not be having this book in your hand but for their dedication, perseverance, and

sacrifice. I am thankful tomy employer, the Center for AdvancedMicrostructures and

Devices (CAMD), who has been supporting me in all my creative ventures. Without

this support, it would be impossible to make this venture of such magnitude a reality.

No words can express the understanding of my family in allowing me to make my

home a second office and bearing with my spending innumerable number of hours in

front of the computer. It is impossible to thank everyone individually in this preface;

however, I must make a special mention of the support fromWiley in general and the

publishing editorAnita Lekhwani in particular,whohas beenworking closelywithme

to ensure that this project becomes a reality. I am grateful for this support.

Note: Additional color versions of selected figures are available on ftp://ftp.wiley.

com/public/sci_tech_med/microfluidic_devices_concepts

CHALLA S. S. R. KUMARBaton Rouge, LA, USA

November 15, 2009

PREFACE ix

CONTRIBUTORS

Ali Abou-Hassan, Laboratoire de Physicochimie des Electrolytes Colloydes etSciences Analytiques (PECSA), UMR 7195, Equipe Colloydes Inorganiques,UniversitO Paris 6, Paris Cedex 5, France

Manish Bothara, Department of Electrical and Computer Engineering, PortlandState University, Portland, OR, USA

Valerie Cabuil, Laboratoire de Physicochimie des Electrolytes Colloydes etSciences Analytiques (PECSA), UMR 7195, Equipe Colloıdes Inorganiques,Universit�e Paris 6, Paris Cedex 5, France

Yong Chen, Institute for Integrated Cell-Material Sciences, Kyoto University,Kyoto, Japan

Chunhui Deng, Department of Chemistry, School of Pharmacy, Fudan University,Shanghai, China

Pamela G. Gross, Student Health and Wellness Center, University of Nevada atLas Vegas, Las Vegas, NV, USA

Peter Mike Gunther, Department of Physical Chemistry and Microreaction Tech-nology, Institute of Micro- and Nanotechnologies, Ilmenau University of Tech-nology, Ilmenau, Germany

Emil P. Kartalov, Keck School of Medicine, University of Southern California,Los Angeles, CA, USA

xi

Clement Kleinstreuer, Department of Mechanical and Aerospace Engineering andDepartment of Biomedical Engineering, North Carolina State University,Raleigh, NC, USA

Andrea Knauer, Department of Physical Chemistry and Microreaction Technol-ogy, Institute of Micro- and Nanotechnologies, Ilmenau University of Technol-ogy, Ilmenau, Germany

Johann Michael Kohler, Department of Physical Chemistry and MicroreactionTechnology, Institute of Micro- and Nanotechnologies, Ilmenau University ofTechnology, Ilmenau, Germany

Debasish Kuila, Institute for Micromanufacturing, Louisiana Tech University, Rus-ton, LA, USA; Department of Chemistry, North Carolina A&T State University,Greensboro, NC, USA

Vindhya Kunduru, Department of Electrical Engineering, Arizona State Univer-sity, Tempe, AZ, USA

Jie Li, Department of Mechanical and Aerospace Engineering, North CarolinaState University, Raleigh, NC, USA

Yan Li, Department of Chemistry, School of Pharmacy, Fudan University,Shanghai, China

Giovanna Marrazza, Dipartimento di Chimica, UnivesitA di Firenze, Via dellaLastruccia, Sesto Fiorentino, Italy

Sriram Muthukumar, Intel Corporation, Chandler, AZ, USA

Shalini Prasad, Department of Electrical Engineering, Arizona State University,Tempe, AZ, USA

Olivier Sandre, Laboratoire de Physicochimie des Electrolytes Colloıdes etSciences Analytiques (PECSA), UMR 7195, Equipe Colloıdes Inorganiques,Universit�e Paris 6, Paris Cedex 5, France

Jian Shi, Ecole Normale Sup�erieure, Paris, France

Muhammad J. A. Shiddiky, Department of Chemistry and Institute of BiophysioSensor Technology, Pusan National University, Busan, South Korea

Yoon-Bo Shim, Department of Chemistry and Institute of Biophysio Sensor Tech-nology, Pusan National University, Busan, South Korea

Yamini Yadav, Department of Electrical and Computer Engineering, Portland StateUniversity, Portland, OR, USA

Shihuai Zhao, Institute for Micromanufacturing, Louisiana Tech University,Ruston, LA, USA; Tianjin University, Tianjin, China

xii CONTRIBUTORS

1MICROFLUIDICS FORNANONEUROSCIENCE

PAMELA G. GROSS

Student Health and Wellness Center, University of Nevada at Las Vegas, Las Vegas,

NV, USA

EMIL P. KARTALOV

Keck School of Medicine, University of Southern California, Los Angeles, CA, USA

1.1 INTRODUCTION

The nervous system of an organism is like the information technology department of

an organization. Each of the billions of building blocks of the nervous system, called

neurons, is amultistate device similar to the transistors of amicroprocessor. But while

transistors are binary state devices, neurons are capable of being in many thousands

of states, and this adds many orders of magnitude to the complexity of possible

connections within a nervous system. In addition, each neuron has multiple connec-

tions with other neurons, and some of these connections are bundled into tracts and

nerves that travel within brain and spinal cord, and out to peripheral locations. In

computers, disconnection of one network cable, or disabling of the electronic circuits

in the server, can seriously compromise the function of the organization. Similarly,

traumatic injuries or neurodegenerative processes such as multiple sclerosis,

Alzheimer’s disease, or Parkinson’s disease can significantly impair the functionality

of an individual by damaging the neurons, tracts, and nerves. However, unlike

computer systems, medical repair processes do not yet exist because we do not yet

understand how the system operates in the healthy state. This may change in the near

future as cell biologists pursue stem cell interventions to regenerate or remodel

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. KumarCopyright � 2010 John Wiley & Sons, Inc.

1

damaged areas of the nervous system. Simultaneously, engineers are teaming up

with biologists to design electronic implants and prostheses that can interface with

functioning tissue on either side of a damaged connection and act as a bridge to allow

restoration of injured neuronal circuits. Pharmaceutical researchers are using nano-

technologies to create novel systems capable of delivering targeted drugs and other

agents across the previously impenetrable blood–brain barrier,1,2 a feature of nervous

systems that chemically separates the system from the rest of the organism.

All these advances may be accelerated by knowledge derived from studies of

cellular physiology using tools designed to study biological processes at the single

cell level. As our ability to fabricate tools on the micro- and nanoscale levels has

progressed, we can now study cellular processes at a scale compatible with cell size,

and this is revealing new information about their operational responses, including how

they respond to physical and chemical cues from their immediate environment. It is

important that neuroscience researchers be aware of these new technologies, so that

their use can be optimized.

Recent advances in biological applications of micro- or nanotechnology have

included novel micro- or nanoscaled carriers for drug delivery,3–6 quantum dots that

operate as nanoscaled sensors at the cellular level,7–11 and nanoelectrodes.12 In

addition, self-assembled monolayers and scaffolding, as well as carbon nanotubes,

have been used as artificial nanotechnology matrices for cell culture.13–19 In neuro-

science specifically, nanoparticles have been used for free radical scavenging in

ischemic and neurodegenerative diseases.20 Scaffolds made of self-assembling

nanofibers are being developed to enhance neuroregeneration.21 The blood–brain

barrier has been successfully breached by drugs attached to special nanoparticles.22

High-resolution studies of the topography and material properties of live nervous

systemcells are beingcarriedout byatomic forcemicroscopy (AFM) (Figure 1.1).23,24

Single-molecule tracking using quantum dots has revealed details about the structure

and function of membrane receptors.10,25,26 Finally, nanotubes, nanowires, and

nanoneedles are being developed for use as relatively nontraumatic intracellular

electrodes.12,27,28 On a slightly larger scale, microfabrication technology has been

used to create microfluidic platforms that have been employed for a variety of

nanoneuroscience studies, and these platforms will now be discussed.

Microfluidics refers to a technology that utilizesmicroscale channels tomanipulate

fluid and suspended objects in a controlled manner at the nanoliter scale. Most

microfluidic chips are designed and constructed using the same techniques as used in

the development of microelectronic circuitry. Microfluidics has been advancing

rapidly over the past decade and has progressed from basic devices, for example, a

channel,29 a valve,30 and a pump,30 to large-scale two-dimensional integration of

components,31 three-dimensional architectures,32 and nonlinear autoregulatory sys-

tems.32 Simultaneously, the development of the fundamental technology has enabled

the advent of a plethora of specialized devices that have miniaturized important

macroscale applications such as protein crystallization,33,34 DNA sequencing,35 and

PCR (polymerase chain reaction), a technique for DNA detection and amplifica-

tion.36,37 The same development has also enabled the advent of novel techniques

to conduct fundamental research in a scale that was never previously possible.

2 MICROFLUIDICS FOR NANONEUROSCIENCE

More recently, some microfluidic chips incorporate other microtechnology and

nanotechnology hardware, such as electrodes,38–44 magnetic coils,45,46 and sur-

face-emitting lasers,47 to enhance their capabilities beyond fluid handling.

Manyof the first applications ofmicrofluidic chips involved studying the physics of

fluid dynamics at the microscale (characterized by low Reynolds numbers, laminar

flow, and fast diffusion), which is quite different from the flow characteristics of bulk

fluid at the macroscale (characterized by higher Reynolds numbers, turbulence, and

slow diffusion). The unusual behavior of fluid traversing microchannels has allowed

creation of new methodologies to manipulate molecules, in order to synthesize novel

nanomaterials and chemical/pharmaceutical moieties, and this has been described in

other chapters. For biologists,microfluidic platformshave emergedas invaluable tools

to study biology at small scales, even down to the single cell level. For neuroscientists,

these “lab-on-a-chip” platforms have enabled a novel approach for experiments on

the cellular physiology of the nervous system. Their usefulness in deciphering the

complicated interactions involved in the differentiation, growth, and maintenance of

neurons in health and in disease has become increasingly apparent within the past

5 years, asmore research in this field continues to be reported.Asmore neuroscientists

become familiar with this technology, we anticipate a rapid evolution of the field. This

chapter will review pertinent contributions in the use of microfluidics to study the

physiology and pathophysiology of neurons and their support cells and will hopefully

serve as a primer for neuroscientists unfamiliar with this technology, inspiring some

to develop new applications of microfluidics to the field of neuroscience.

FIGURE 1.1 Atomic force microscopy images of neural lineage cells. (a) Three-dimen-

sional rendering of an oligodendrocyte differentiated from a murine neural stem cell. Fixed

sequentiallywith 100%ethanol and 4%PFA, air dried, and then imaged on anAsylumResearch

MFP3DAFMusing anOlympusAC160 cantilever inACmode in air. Note the detailed process

formation. Scan size is 90mm� 90mm. (b) Three-dimensional rendering of a portion of a

living astrocyte derived from a human embryonic stem cell on a polyornithine/laminin-coated

substrate, imaged inmedia, inACmodewith anOlympusBiolever, and on anAsylumResearch

MFP 3D AFM. The image shows cytoskeletal fibrous elements visible through the cell

membrane in the proximal thicker area of the cell as they enter a broad, flat attachment

area. Scan size is 30mm� 30mm (unpublished data, Pamela G. Gross).

INTRODUCTION 3

Microfluidic platforms typically contain a series of chambers and channels that

each measure in the range of 1mm to a few hundred microns and are used to process

fluid at amicroscopic scale. For in vivo applications, microfluidic technology has been

integratedwith neural implants for precise delivery of solutions.48 Three-dimensional

electrodes with bundled microfluidic channels that can be implanted into severed

nerves to guide and monitor their regeneration while allowing infusion of drugs are

also under development.49 However, the most common biological application of

microfluidics has been for in vitro studies, such as the delivery and processing of

biochemical reactants for DNA sequencing35 and protein analysis,50 the sorting,

counting, and analysis of cells by flow cytometry,51 the delivery of cell adhesives

and cells for substrate micropatterning of cell populations,52,53 the development of

biomimetic three-dimensional tissues, complete with stromal support molecules,18,19

and the isolation and nurtured maintenance of individual cells to study basic cell

physiology and cell–cell interactions on a single (or near-single) cell basis.54–59

In addition, microfluidic platforms have been used to study the effect of laminar flow

and shear forces on the function of endothelial and other types of cells,60,61 to provide

artificial circulation through various organ-simulating cell culture chambers in

order to determine the pharmacokinetics of prospective pharmaceutical agents,62

and to deliver test samples containing potential toxins to cells acting as biosensors

(also known as “lab-in-a-cell” technology).63–65 Finally, microfluidics can be used to

study physiology within small organisms, such as the effects of anesthetics on the

regrowth of severed axons, or the recovery of axonal synapses after laser ablation

in Caenorhabditis elegans nematodes that have been captured and immobilized in

microfluidic chips.66,67

Microfluidic-based cell studies are a useful adjunct to conventional in vitro

techniques or mini culture systems68 because microfluidic chambers have the ability

to control both the amount of material (media, growth factors, etc.) used for cell study

and their exact distribution over well-defined periods. This can permit better control

of the experiment by limiting unanticipated extraneous factors and diffusion con-

straints that can occur in larger systems. The effects of cell population variability will

also be more limited in smaller systems and therefore individual differences among

similar cells will be less likely to influence results. From an economic standpoint,

small culture volumes allow cost savings since the required volume of expensive

media, hormones, and growth factors is orders of magnitude less than that used in

typical culture flasks. These platforms can also be designed for high throughput and

compatibility with automated laboratory equipment such as plate readers. In addi-

tion, the hardware is portable and it can be mass produced so inexpensively that it can

be very cost-effective to perform massively parallel microfluidic platform-based

experiments, in order to confirm results or test the effects of numerous agents

simultaneously.

These parallel experiments are necessary to verify results obtained on individual

cells since it is known that there can be significant variation in the behavior of

particular cells, even if they are cloned from the sameprecursor cell.69 Similarly, itwill

be imperative that the effect of microenvironment parameters such as mechanical

forces, shear stress, effective culture volume, and material interfaces be well


understood and controlled before interpreting single cell study results so that these

factors do not contribute to misleading conclusions.70 Nevertheless, observations

derived from studies of individual cells in a controlled microenvironment may be

muchmore likely to reveal true cellular physiology responses than those derived from

studying the responses of populations of cells simultaneously, as is done with

conventional in vitro studies.

Although the development of this technology has progressed significantly over the

past 5–7 years, its utility as a tool is just beginning to be appreciated by biologists.

There are many published reviews on the general topic of microfluidics for biological

applications,69–85 but there have only been a few that have focused on microfluidic

applications in neuroscience.80–82 This chapterwill update the reader on the discipline

ofmicrofluidics to study the nervous system at the single cell level. Specifically, it will

report on microfluidic chips used to decipher physiological processes and responses

of cells of neural origin, and it will also focus on examples of systems that combine

microfluidic chambers with other technologies for novel research not previously

possible.

Section 1.2 will begin with a description of current microfluidic chamber con-

struction techniques, starting with a discussion of the characteristics of the PDMS

polymer used in many microfluidic chambers and then moving on to cover step-

by-step fabrication processes. Various architectural designs of use for cellular studies

will then be introduced, followed by a description of alternative applications of PDMS

to create tools that are useful in customizing the substrate of microfluidic chips for

specific experiments. Practical limitations of microfluidic techniques will then be

discussed to present a balanced view of the topic.

In Section 1.3, gradient-generating designs will be reviewed, along with examples

of how they have been used to study cellular responses. Methods of incorporating

electrophysiological measurements into chip design, including patch clamping, will

be examined and then use of other integratedmicro- and nanoscaled analytical devices

will be considered. The theory andmethodology used for in vivo tissue simulationwill

be evaluated, since the natural behavior of cells is ultimately what most biological

research is attempting to discern.

Following this, a literature review of neuroscience research involvingmicrofluidic

platformswill bedetailed inSection1.4, startingwith cell identification and separation

tools, which is essential for researchers requiring specific subpopulations of neural

lineage cells. Studies on microfluidic analysis of neuropeptide release will follow,

which is of interest to individuals studying synapse formation and function. The use of

microchips to study the effects of physical and chemical guidance cues on single cells

will then be considered since this is a key to understanding how neural cells interact

with their environment and with each other.

Section 1.5 will focus on electrophysiology studies that use multielectrode arrays

(MEAs) as microfluidic chamber substrates. This is a popular field of endeavor since

these two technologies seem to be complementary and can allow studies on action

potential characteristics and propagation in single axons. The effect of growth factors

on neuronal responses of microfluidically cultured and isolated cells will be covered

after this, given its significance in understanding cell differentiation and maturation.

INTRODUCTION 5

The use of microfluidic chambers for gene therapy studies on neural cells will be

subsequently discussed. Although this is a relatively new area of study, preliminary

results are very promising and future research will likely take advantage of the unique

capabilities that microfluidic chips offer to this field. The final area of research to be

covered involves studies based on the microfluidic isolation of axons and neural cell

bodies. This approach to neural research is gaining great interest, given the potential

applications for those studying neural degeneration and regeneration processes, in

addition to those interested in axonal transport mechanisms, and synapse formation

and physiology. A general discussion with consideration of future perspectives will

complete the chapter. It is hoped that the readerwill gain an appreciation for the future

potential of these platforms to uncover previously hidden cell-based interactions in

the nervous system, and this will stimulate new applications of microfluidics for their

specific research programs.

1.2 PDMS MICROFLUIDIC DESIGN AND FABRICATION

1.2.1 Characteristics of PDMS

Initially, most microfluidic chamberswere constructed on siliconwafers using “hard”

lithography. Since those early studies, “soft” lithography has been developed and

various polymers and fabrication techniques have been investigated.86 Now, soft-

sided chambers made of polydimethylsiloxane (PDMS) are gaining increased popu-

larity, especially for biological applications. PDMS is a silicon-type elastomer and

can be purchased commercially as Sylgard� 184 by DowCorning or RTV by General

Electric. It can be molded into many different shapes to form valves, chambers, and

channels. PDMS is advantageous for biological studies since it is biocompatible,

optically transparent down to wavelengths as low as 280 nm, permeable to gases

needed for cellular respiration, autoclavable, and naturally inhibitory to cellular

adhesion.87,88 PDMS has therefore provenvery handy for cellular studies by allowing

long-term cultures, optical microscopy, and fluorescent/chemiluminescent studies,

while the cells are still in situ in the chip.64 A final advantage of this material for use

with cell culture systems is that PDMS has been shown to be an excellent protective

coating for on-chip solid-state analytical devices (such as surface-emitting lasers),

since PDMS is optically transparent yet prevents the detrimental effects of ions

migrating from the culture medium into sensitive electrical junctions.47

Native PDMS is hydrophobic, and this influences many of its surface properties,

including its interactions with fluid and molecules that are in contact with it. These

properties can be altered by physical and chemical treatments that can change

the hydrophobicity of the surface of the PDMS channels and change its adhesive

properties if this is desired.74 For example, the pretreatment of the PDMS channels

with bovine serumalbumin (BSA)will assist in blocking cell adhesion to its surface.54

Alternatively, PDMS can be made hydrophilic and supportive of cell growth by

treatmentwith oxygen plasma,88 orUV/ozone,89 that acts by changing themoieties on

the PDMS surface to increase the number of silanol groups and decrease the number


of siloxane groups. Polyethylene glycol (PEG) can also be used to alter the surface

chemistry of PDMS.35

The surface interactions of PDMS with adjacent molecules will also depend on

local flowconditions. Experimentally, proteins such as collagen andfibrinogen adhere

to both hydrophobic and hydrophilic (oxygen plasma-treated) PDMS. But under flow

conditions, the oxygen plasma-treated hydrophilic surfaces experienced only tempo-

rary adhesion, followed by rapid detachment of any adherent cells, whereas the

hydrophobic PDMS channels became permanently clogged with protein and cells.88

Therefore, systems that are designed to have continuous exposure to protein-laden

media and cellswill likely benefit frompretreatment of the PDMSwith oxygenplasma

to increase the functional lifetime of the channels.

1.2.2 PDMS Chip Fabrication Protocol

Microfluidic chip fabrication uses many of the same techniques used in electronic

circuit production. The process begins with the creation of an architectural design

using a computer-aided design (CAD) software program. The design is printed on a

transparency using a high-resolution printer, since the feature size on the final chip

will be determined by the resolution of features on the transparency. This transparency

acts as a photomask during the next step, in which it is placed over a substrate (silicon

wafer or glass) that is precoatedwith a thin layer of photoresist, a photocurable epoxy.

UV exposure polymerizes exposed areas for photoresists such as SU-8, so the

developer solution can strip away the unexposed areas because they are not poly-

merized, while the polymerized exposed structures remain. This type of photoresist is

called negative photoresist because the result is the reverse, or “negative,” of themask.

On the other hand, photoresists such as 5740, SPR-220, and the AZ family are called

“positive photoresists” because the result corresponds to the mask; that is, the result

is “positive” to the mask. UV exposure makes a chemical change in positive

photoresists that results in the material becoming more soluble, for example, in a

strong base. Thus, the developer solution removes the material from the exposed

areas, while the structures in the unexposed areas remain. In both cases, the result is

a mold where the features are built in photoresist. Since photoresist is softer than

silicon, the resulting mold is softer than traditional molds, and so the technique has

been named “soft lithography.”

In the next step, PDMS is combined with its catalyst in a 10:1 proportion and the

mixture is degassed in a vacuum chamber to remove bubbles. It is then poured onto the

master, allowed to cure, and then peeled off the mold. Access ports are punched after

casting (or silicon tube ports are placed during casting) to create connections to input

and drainage tubes. The PDMS slab is then placed onto a substrate such as a silicon

wafer or a glass slide to create the final microfluidic platform. The PDMS forms a

reversible conformal seal to the substrate, but optional treatment of the PDMS with

plasma oxidation of the PDMS surface after curing will render the surface more

hydrophilic and allow the PDMS to irreversibly bond to the substrate. After steriliza-

tion bymeans of autoclaving, UV treatment, or immersion in 70% ethanol, the system

is ready for use.

PDMS MICROFLUIDIC DESIGN AND FABRICATION 7

Although PDMS can reproduce features down to 10 nm in size,75 actual fabricated

channels in PDMS have not yet achieved a cross-sectional area smaller than 1mm2.74

This is because the feature size of the PDMS is determined by the printed resolution on

the photomask, which is determined by the printer used to create it. For example,

standard printers that have a resolution of 5080 dpi can reproduce features on the

photomask down to 25mm resolution, whereas photoplotters that print at 20,000 dpi

can achieve a resolution down to 8 mm.75,90 To reproduce smaller features, chrome

masks may be used, which are created with e-beam or laser writing and are much

more costly. In addition, the relative softness of PDMS makes it difficult to maintain

uniformly high quality of the reproduced features when the linear scale is decreased

below a few microns.

1.2.3 Architectural Designs of Microfluidic Platforms

The physical behavior of a fluid flowing throughmicroscale channels is very different

from the flow characteristics of the same fluid flowing through larger channels. For

example, fluid flow throughmicrofluidic channels is laminar, somixingdoes not occur

between solutes placed at different locations in the channel cross section, except by the

slow process of diffusion. Without turbulence, solute gradients will remain relatively

intact as fluid traverses downchannels of uniformwidth. If cells are localized at certain

areas of this channel, their exposure to specific concentrations of solute can be tightly

controlled. In fact, different parts of the cell can even be exposed to different and

controlled concentrations of the solute. This laminar flow behavior can also be used to

pattern and deposit specific solute concentrations onto the substrate or to pattern cell

adhesives and repellents next to each other onto the substrate prior to introduction of

cells. Alternatively, if mixed patterns are desired, deliberate oblique grooving of the

floor of the channel can be employed to create turbulence in order to mix solutes,91–93

and nanotopographic features can also be added to the platform substrate to influence

cell adhesion.94

Most microfluidic chips use some type of dynamic flow conditions, with flow

achieved by the use of syringe pumps, gravity-driven reservoirs,95 electrokinetic

control,96 or other more complicated functional PDMS valve structures. These valves

are designed by layering “control channels” that act as bladders across flow channels.

Application of pneumatic pressure in the “control channels” can then cause controlled

collapse of the underlying fluidic lumen, and this controlled deformation of the flow

channel’s lumen creates a functional valve.30 Digital control and sequential coordi-

nation of these valves can create peristaltic pumps. Rotary pumps based on similar

mechanisms have been designed and used for applications that require repeated

cycling of fluid for mixing, such as on-chip PCR, used in amplification and identifi-

cation of DNA strands in genetic engineering.97

Various structures have been devised to immobilize cells withinmicrofluidic chips.

These architectures must be able to catch and retain a cell from a passing stream of

media, while minimizing damage to the cell. Sieves have been used within the culture

chamber to retain cells while also producing a nutrient gradient.98 Channel walls can

be constructed at partial height to create a dam that allows flow from one channel to


another while gently transporting and immobilizing cells for later analysis.99 Inverted

T junctions that have small docks with tiny drain channels at the junction have been

used to immobilize single cells and then perform rapid on-chip calcium flux assays.54

Curved docking areas that can balance the forces exerted on cells (fluid flow versus

gravitation) have also been used to isolate individual cells for culture and study of

calcium mobilization.56,57 Gravity-induced flow has been combined with dielectro-

phoresis to trap and sort cellswithout physically contacting them.100 For the capture of

cells with variable dimensions such as pancreatic islets (used in diabetes research),

designs have combined one semiellipsoidal wall and one movable wall to create an

adjustable holding area that will allow studies on the regional effect of infused glucose

and drugs.59 Studies on pairs of cells have used intersecting channels that have been

designed to trap pairs of cells from different populations to study intercellular

communicationvia gap junctions between their cellmembranes.55 Finally, the surface

of PDMS has been microstructured with arrays of wells and coupled to a microfluidic

system to create a test platform for parallel experiments on single cells or small

groups of cells.101 As described above, a significant advantage of microfluidic

chambers is that the architectural design and the dimensions of the channels and

chambers can be customized for the morphology of the cells to be studied and to the

task to be accomplished. As new researchers enter this field, we expect to see awealth

of new designs for novel applications.

1.2.4 PDMS Tools

In many cases, it is desirable to have a microfluidic chip substrate that is patterned

with different molecules prior to assembly of the chip. This can be easily achieved by

creating a separate PDMS tool that contains the substrate pattern and can be used as a

stencil or a stamp. This tool is fabricated using the same techniques as outlined above.

After completion of the tool, it can be used for microcontact printing by dipping the

patterned area into a fluid with the desired concentration of solute molecules and

then transferring this pattern to the substrate. PDMS can also be formed into a two-

dimensional stencil sheet that allows patterned deposition of selected proteins or

agents onto the underlying substrate, and this PDMSstencil has the advantage of being

useful on irregular or curved substrate surfaces. Once the protein pattern has been

created on the substrate, the remaining platform can be constructed by applying the

matched PDMS chip so that its channels are complementary to it. With this arrange-

ment, future cell attachment and differentiation can be guided, and cocultures can be

created in controlled geometric patterns.102–104 Since this technique can help control

the exact position of neurons on a substrate, the resultant controlled neuronal patterns

can be very helpful in studying neural networks and interactions occurring within

synapses.69,102,103,105 These techniques have also been combined with selective

oxygen plasma treatment to create long-term and short-term cell repellent areas to

coculture cells in controlled geometric patterns.104 In this case, cell repellent polymers

were homogeneously deposited on a substrate, and a PDMS stencil was used to

selectively protect the repellent from plasma treatment in certain areas. Unprotected

areas lost their repellent nature and could then be treated with adhesives like


fibronectin and short-term repellents like BSA to create patterns of relative adhesivity

over time that could then be seeded as desired with different cell types.

1.2.5 Practical Considerations and Limitations

As with all new technologies, there are certain practical design considerations and

limitations that must be recognized before planning a microfluidic chamber for cell

studies. Cell viability has definitely been correlatedwith channel size and proportions,

closed versus open-channel configurations, and static versus dynamicmedia flow. For

example, it has been shown that in contrast to cells grown inconventional tissue culture

flasks, the proliferation rates of cells grown in microfluidic channels without media

flow depend on the height of the channel.106 This is likely due to loss of convective

movement of cell-expressed inhibitory factors away from cells, rather than lack of

nutrients or change in osmolarity or pHof the culturemedium. In a static systemwhere

there is no flowofmedia or connection of themedia in the channels to a bulk container,

secreted factors can only be dissipated by diffusion, and this can be insufficient to

remove their often deleterious effects.106 If continuous or intermittent flow is designed

into the system, the flow rate must be optimized to provide nutrients and remove

wastes, without producing excessive shear stress that can change morphology or

migration of the cells or even detach the cells.98,107 Similarly, certain secreted factors

maybeessential for cellular health, and if theflow is toohigh, then these factorsmaybe

washed out.

Pretreatment of the PDMS prior to the introduction of cells can have significant

effects on cell culture success. For example, Matsubara et al. showed that different

treatments to make the PDMS hydrophilic affected both the morphology and the

density of mast cells.64 Similarly, Prokop et al. found that extracellular matrix

deposition and plasma treatment of the PDMS improved subsequent cell cultures.98

Other important considerations include recognition of the fact that the tiny volume

ofmicrofluidic culture systems confers amuch less stable homeostatic system in terms

of temperature, carbon dioxide concentration, and humidity control compared to

standard Petri or tissue culture flasks. These chips equilibrate muchmore rapidly with

their environment than larger systems given their larger surface area to volume ratio,

so each time these chips are removed from the incubator, they are prone to more rapid

alteration of their temperature, atmosphere, and humidity. Temperature alone is

known to directly influence gene expression, biochemical reactions, and diffusion

speed. To maintain a stable system, steps must be taken to minimize losses of

environmental stability. Similarly, if media is fed into these chips via tubing that is

outside the incubator (i.e., if connected to a syringe pump), or if the tubing is part of

a “mini” culture system, it is important that media temperature and CO2 content do

not change during transport through the tubing.68 Therefore, although these micro-

scale systems are technically portable, control systems for their ambient environment

may be necessary if they require transport outside the incubator for time lapse imaging

or other interventions. We witnessed this effect directly when an isolated axon in

a microfluidic chip was observed to shrink back significantly within a few minutes

of removal from the incubator and placement onto the cold microscope stage


(unpublished data). Control of humidity is also critical because water diffuses into

PDMS according to Fick’s law of diffusion.108 From there, it can evaporate and lead to

increased osmolarity of the cell culture medium and premature cell death. Methods

to minimize evaporation, such as coating the PDMS with a thin layer of parylene,

have been successfully implemented and shown to prolong cell viability.108

The last practical consideration in usingPDMSas a culture chamber inmicrofluidic

systems involves the high ratio of chamber surface area to the volume of culture

medium and number of cells compared to standard culture flasks. This increased

surface area to volume ratio can lead to increased interactions between chamber

contents (media, cells) and chamber walls. For example, it is known that small

hydrophobic molecules may partition into the PDMS and therefore be less bioavail-

able when studying their effects on cells.109 This extraction of media solutes by the

PDMS can significantly change the concentration of some agents in media within

microfluidic channels by many orders of magnitude. The magnitude of this change

depends on the partition coefficient of the substance, the pH of the culture medium,

and the counterion pairing in the media. Major decreases in the media concentration

of neurotransmitters, hormones, and growth factors could change experimental

outcomes, and the cost savings for using small volumes of these agents inmicrofluidic

platforms can be lost if much larger quantity of the substance has to used to achieve

the same effect. Therefore, many individuals are now experimenting with different

surface treatments to decrease the porosity of the PDMS and avoid some of these

effects.

In addition to taking up biomolecules, the PDMSmay also release potentially toxic

agents from its polymer matrix.110 These can then be concentrated in a proportionally

smaller volume of culture medium and may affect more sensitive cells. Certainly,

neurons from different sources vary in their hardiness, and culturing sensitive neurons

at low density and in serum-free conditions can be difficult, even in the most tightly

controlled environments. For example, we have personally had difficulty maintaining

the viability of human neural stem cells in PDMS chips, whereas rat dorsal root

ganglion cells thrived in the same conditions (unpublished data).

One possible explanation of this phenomenon may be found in the work of Millet

et al., who hypothesized that there might be seepage of toxins from the PDMS.110 In

their research, they tried to improve neuron survival in open- and closed-channel

microfluidic chips by treating the PDMS with serial solvent-based extraction

processes (to remove potentially cytotoxic uncross-linked oligomers and residual

platinum catalyst in the PDMS) or with autoclaving (to drive cross-linking and outgas

solvents). They found that treatment with extraction improved neuron survival,

increased the development of neurites, and lowered platinum levels in the PDMS

more than did autoclaving. Specifically, the ratio of neuron survival was 3:28:51 for

native PDMS, autoclaved PDMS, and extracted PDMS, respectively. Overall cell

viability in low density, small volume, serum-free studies in closed-channel devices

was improved from less than 2 days in native PDMS to over 7 days for extracted

PDMS. If gravity-driven flow was added, survival could be further increased to over

11 days by improving nutrient delivery and waste removal compared to static

systems. These extraction processes will be imperative for future studies on


individual cells or cell-to-cell interactions in low-density cultures in microscale

culture volumes.

Perhaps the biggest impediment to the general acceptance of microfluidic plat-

forms as tools for biological investigation will be the initial need for interdisciplinary

teamsof researchers that include both engineerswhocandevise and fabricate the chips

andbiologistswhoknowwhenandwhere to best apply the technology.Aswith all new

technology, the developers may not be able to recognize its most useful niche, and the

users may not be aware of the technology or have the auxiliary tools and expertise

to operate it correctly. However, as more individuals take the steps to experiment with

the technology, it would become more commonplace and better utilized.

On the other hand, the novelty of the underlying technology and of the general

approach of combining neuroscience with microfluidics offers unique and exciting

opportunities to address fundamental problems with new tools in new ways. These

technologies thereby carry the immense promise of important breakthroughs and new

insights both in fundamental neuroscience and in its extensions to biomedical practice

in improving the treatment of many neurological diseases.

1.3 DESIGNS AND DEVICES FOR NEUROSCIENCE APPLICATIONS

1.3.1 Gradient-Generating Designs

As discussed above, the laminar flow that occurs in microscale channels can be used

advantageously to create high-resolution gradients of solutes and special factors

within the cell culture chamber, in order to assess the effects of these gradients on the

behavior of individual cells. These designs employ two or more inputs—one for the

studied factor and one for the dilutingmedium,with each connected to its own syringe

pump. The inputs connect to a network of serpentine, interconnected channels that

repeatedly split and remix, with each generation of splitting channels increasing in

number, until they finally coalesce back into a single larger channel (Figure 1.2). At

each branch point, some mixing occurs so that there is a gradient of concentrations

of the studied factor(s) that is oriented perpendicular to the flow direction at the final

exit channel, and this gradient has a range of resolution spanning from severalmicrons

to hundreds of microns. After exiting the gradient-generating device, the established

gradient is maintained by laminar flow. By varying the flow rate into one input,

dynamic and asymmetric gradients of variable shape (smooth, step, ormultiple peaks)

can also be created.52,111,112 These devices have been used to study the effects of IL8

(interleukin 8) on neutrophil chemotaxis,113EGF (epidermal growth factor) on breast

cancer cell chemotaxis,114 and various growth factors on neural stem cells.115

These gradient-generating devices are commonly used for research on the physi-

ology of neural lineage cells since the devices provide precise control over exposure

of growth and inhibitory factors to these cells. In addition, these same devices have

been used to etch a controlled gradient into the surface topology of chip substrates

by injecting etching reagents, or to lay down gradients of adhesives, self-assembled

monolayers (SAMs), and dyes.111 Finally, these gradient generators have been


combined with large chip-based arrays of cell culture chambers (10� 10) to simulta-

neouslyperform100parallel tests on the effect of anagent’s variousdilutions.116 Since

these chambers each had four individual access ports, repeated growth/passage cycles

of the cells could be performed on-board by microfluidic control, so the cell cultures

could be maintained over long periods. As demonstrated with the studies described

above, these gradient devices are very useful tools for investigations into cellular

responses to varying concentrations of specific factors, whether these factors are

substrate bound or dissolved in media.

FIGURE1.2 (a) Photograph showing amicrofluidic device used for generating gradients of

green and red dyes in solution. The three incoming channels (top part of the photograph) were

connected to syringes via tubings (not visible). After combining the streams into a single, wide

channel (bottom of the photograph shown in (a)), a gradient was formed across the channel,

perpendicular to the direction of flow. (b) Schematic explaining the nomenclature used for the

mathematical description of the network. (c) Schematic demonstrating the application of the

formulas governing the splitting ratios at the branching points. The dotted lines indicate the

boundary between the two combined streams. The concentrations at the end of the serpentine

channels can be calculated bymultiplying the concentration of the incoming streams (cp, cq, cr)

with the corresponding numbers of the splitting ratio (Vp þ 1)/B, (B�Vq)/B, (Vq þ 1)/B, and

(B�Vr)/B, as indicated). Reprinted with permission from Ref. 112. Copyright 2001 American

Chemical Society.

DESIGNS AND DEVICES FOR NEUROSCIENCE APPLICATIONS 13

1.3.2 Integrated Electrophysiology

For electrically active cells such as neurons and muscle cells, integrated electrical

recording is a very valuable addition to microfluidic platforms. An early example of

this was a unique system that was designed as a self-contained, portable unit for field

use as a cell-based biosensor. The unit incorporated a hybrid glass/PDMS/silicon

chamber for cell culture with integrated microfluidics, a microelectrode array

substrate modified with fibronectin and gelatin for cell growth, a temperature

regulation system, on-chip electronics for acquisition, analysis and display of action

potentials, and a transparent cover that makes the unit amenable to microscopic

inspection.65 This approach of creating a stand-alone unit with its own environmental

controls may eventually be required of many platforms in the future; however, most

neuronal studies have employed much simpler hardware, typically using commercial

MEAs as the substrate for a PDMSmicrofluidic chip. Thesewill be discussed inmore

detail in Section 1.4.4.

The “gold standard” for electrophysiological studies has always been patch

clamping, and many microfluidic platforms incorporating arrays of patch clamp

electrodes have been engineered and successfully demonstrated. Conventional patch

clamps use suction to attach the tip of a glass micropipette to a cell membrane, and

then break the membrane and record the intracellular potential using a conductive

fluid in the micropipette. In early microfluidic designs, these systems used PDMS

microfluidic channels to guide cells to pores micromachined into silicon wafers,117 or

they used cell-trapping pores in a horizontal PDMS substrate.118–120 These pores

simulated the tip of a conventional glass micropipette and were used to create a high-

resistance seal to the cell wall for subsequent electrical recording.

Ionescu-Zanetti et al. improved on this design by incorporating pores on a vertical

channelwall of the PDMS to facilitate the use of optical and fluorescentmicroscopy to

monitor the procedure.121 This vertical approach allows both the cell and the capillary

tube leading to the pore to be in the sameplaneof focus, and, therefore, it permits easier

guidance of the selected cell to the pore (using a combination of flow in the cell

chamber and suction from the pore). It also permits visual monitoring of the cell

condition and position during the recording. Each pore is connected to a capillary tube

that applies negative pressure (suction) to attach the cell and break the membrane

and to a silver/silver chloride electrode that then connects to a multiplexer circuit to

process the recorded signals. Using CHO (Chinese hamster ovary) cells, the seal

resistance between pores and cells was an average of 300megaohms, and the system

wasable to recordcurrentsdown to20 pA. Individual cell trappingcouldbeachieved in

less than 3 s, and the seal was stable for 20–40min (Figure 1.3). This group further

updated this system by raising the trapping pore above the chamber floor to avoid

deformation of the trapped cell. They also opened the ceiling of main chamber to

ensure rapid fluidic access for high-throughput drug profiling on the clamped cell. For

these authors, this microfluidic approach to patch clamping represented a much more

efficient system for pharmaceutical analysis than traditional patch clamp technology.122

Unfortunately, patch clamp resistance seals in the megaohm range as reported

above are not ideal, and other groups have been modifying their techniques to


improve this.ChenandFolchusede-beam lithography to create a1mmcell attachment

aperture in their patch clamp chip. This method was combined with standard

photolithography using high-resolution photomasks to create larger suction channels.

They also used O2 plasma treatment of their master and PDMS chip to smooth the

edges on the aperture, and they achieved reliable gigaohm seals and signal quality

that was similar to that obtained with traditional glass pipette patch clamps.123

Commercial forms of these microfluidic patch clamp technologies will likely be

available in the near future.

1.3.3 Other Integrated Sensors and Microfluidic Capabilities

Microfluidic chips have employed many other complementary microtechnologies in

recent years for application to biological studies. Although they have not all been used

specifically for neuroscience studies, theydo have this potential and they are presented

here for the interested and motivated reader. For example, as an alternative to using

electrodes, chargedmembrane-permeable, potential-sensitive dyes have been used in

a microfluidic device to determine the membrane potential of cells in a rapid, highly

FIGURE 1.3 Patch clamp array on a microfluidic platform. (a) Cell trapping is achieved by

applying negative pressure to recording capillaries that open into a main chamber containing

cells in suspension. Attached cells deform, protruding into the capillaries. Patch clamp

recordings are obtained by placing AgCl electrodes in each of the capillaries, as well as in

the main chamber. Signals are fed through a multiplexing circuit and into the data acquisition

system. (Multiplexer setup and microscope objective are not to scale.) The device is bonded to

a glass coverslip for optical monitoring. (b) Scanning electron micrograph of three recording

capillary orifices as seen from the main chamber. The capillary dimensions are 4mm� 3mm,

with a site-to-site distance of 20mm. (c) Dark-field opticalmicroscope image of cells trapped at

three capillary orifices. Trapping was achieved by applying negative pressure to the recording

capillaries. The device consists of 12 capillaries arrayed 6 along each side of the main chamber

fluidic channel along a 120mm distance. Reprinted with permission from Ref. 121. Copyright

2005 National Academy of Sciences USA.


sensitive manner, with minimal consumption of reagents.124 Various electrical para-

meters including amperometry, impedance measurement, and potentiometry have

also been used to analyze cells and their ionic secretions inmicrofluidic chambers (see

the excellent review on this topic by Bao, Wang, and Lu).85

The relative acidity of the contents ofmicrofluidic channels has beenmonitored by

pH-sensitive fluorescently tagged monolayers (SAMs) that are bound to the substrate

of the microfluidic platform.125 An alternative pH meter with higher sensitivity has

used a different technology in microliter flow chambers to measure pH changes down

to 0.5� 10�3 pH units. This device has been useful in the study of cellular processes

that alter ATP levels, such as receptor activation and signal transduction.126 The

oxygen content in microscale cell cultures is another important parameter that can be

monitored by an on-chip oxygen sensor based on fluorescent quenching of ruthenium

dye particles encapsulated in the PDMS of the microfluidic culture device.62 Silicon

chips containingmultiplemicrosensors for bulk detection of extracellular pH, oxygen

consumption rates, and cell morphological alterations have also been developed.127

and although not yet applied to microfluidic single cell studies, it is reasonable to

expect that they might be adaptable to this purpose in the future.

More advanced technology has also been miniaturized for on-chip use. For

example, single nonperfused neurons have been studiedwithNMR (nuclearmagnetic

resonance) microcoils, and NMR spectroscopy has been used to determine their

metabolite content, but the need for continuous perfusion to prevent cell death was

noted.45 To address this issue, planar NMR probes have been incorporated into

microfluidic platforms and preliminary studies on their functionality are underway.46

Other advanced technologies such as surface-enhanced Raman scattering and confo-

calmicroscopy have been combinedwithmicrofluidics to study real-time intracellular

chemical dynamics of single live cells with high spatial and temporal resolution.128

Apoptosis (programmed cell death) is an important cellular process that is well

studied by both biologists and pharmaceutical companies since it is critical to

understanding how to control cancer and cell growth in general. Microfluidics

have beenused to study themultiplemorphologic andbiochemical changes associated

with apoptosis at the single cell level.72 Tamaki et al. noninvasively monitored the

change in cytochrome c distribution that occurred during apoptosis of single neuro-

blastoma–gliomahybrid cells confined in quartz glassmicrofluidic chambers by using

scanning thermal lens microscopy without the need for any labeling materials.129

Finally, on-chip single cell genetic evaluation and manipulation will be useful for

neuroscience cellular studies. One successful technique for this involved a combined

microfluidic/microelectroporation chip that could isolate and temporarily immobilize

individual prostate cancer cells in a channel, prior to application of a 10V, 100ms

electric pulse to puncture the cellmembrane and insert green fluorescent protein genes

into them.130 Inmore recentwork, a lower applied voltage of only 0.8V for 6.5mswas

focused at one location on the membrane of individually trapped HeLa cells and

resulted in successful electroporation.131 Because of the lower voltage requirement

and because this design was able to monitor the permeation of the membrane by

recording accompanying jumps in electrical current across the cell membrane, it

represented a definite improvement over previous technology.


Another example of a platform for genetic studies usedmicrofluidics to isolate cells

and then lyse them prior to purifying and recovering their mRNA, the genetic

instruction codes that cells use to synthesize proteins.132 Further development of

this technology from this same laboratory has yielded chips capable of performing

72 parallel, 450 pL reverse-transcriptase PCR reactions that could detect mRNA

levels down to 34 RNA templates.133 They have also used microfluidics to synthesize

cDNA from subpicogram mRNA templates isolated from single cells134 and per-

formed gene ligation with plasmids and successfully transformed the plasmid DNA

into competent cells.135 This technology has potential utility for neuronal studies.

1.3.4 Simulating In Vivo Tissues with Microfluidics

To draw reasonable conclusions about in vivo processes using data derived from our

in vitro experimental models, these models must simulate real tissues as closely as

possible. This is why many researchers have tried to consolidate multiple cell types

and extracellularmatrix proteins into a three-dimensional architecture.Otherwise, the

data may be misleading and oversimplified. For example, most muscle cells grown

in vitro have different morphology and function compared to those grown in vivo.

However, one group of researchers found that they could culture cardiac muscle

myocytes with more typical morphology if they cocultured fibroblasts alongside

of them on linear, intersecting patterns of collagen deposited within microfluidic

channels.136 Similarly, a perfused microfluidic platform built upon a substrate with

alternating cell adhesive (matrigel on poly-D-lysine) and cell repellent stripes (poly-

acrylamide and polyethylene glycol) was used to grow and fuse myoblasts into

realistic multinucleated myotubes.91

Based on studies such as these, it is now believed that many environmental

parameters can directly influence a cell’s cytoskeleton and subsequently alter cell

behaviors such as proliferation, motility, and migration, to name a few. To better

understand these interactions, the effect of variably sized and shaped microfabricated

cell culturewells has been studied. Preliminarywork shows that these single cell wells

(treated with cell adhesive material via PDMS microcontact printing) have been

successful in altering the three-dimensional shapeof the cell containedwithin them.137

Further study in this field may reveal how cell shape in vivo alters cell function.

In addition to controlling individual cell shape, environmental cues from cell

attachmentmatrices containing self-assembling proteins and gel-like substances have

influenced the three-dimensional shape of a group of cells in vitro.17 Microfluidics

have been used to build up realistic vascular tissue by sequentially depositing layers

containing different cell types and extracellular matrices (collagen, matrigel, etc.)

within a platform.18,19 These types of microfluidic platforms have the capability of

creatingmore biomimetic in vitro systems, but theymay also encounter the same type

of limitations on study that using live tissue sections do. Rather than trying to simulate

real tissue, it may bemore important in the in vitro study of individual cell physiology

to create a substrate with the right characteristics. For example, to be truly

“physiologic,” in vitro substrates must re-create several characteristics of normal

in vivo extracellular matrices, including mechanical properties (such as elasticity,


rigidity, and strain), chemical properties (such as ligand density and orientation), and

topographic properties (such as surface curvature and fibrous contact guidance).

These characteristics control cell distribution in tissues and guide cell morphology,

behavior, gene expression, proliferation, differentiation, and apoptosis, presumably

through interactions with transmembrane integrin receptors.17,138,139 This theory is

reinforced by studies that show that neurons grow better on a “soft” bed of astrocytes

thanonglass.139As research in this field progresses,weexpect to seemorephysiologic

substrates being incorporated into microfluidic systems.

Researchers investigating neural prosthesis development are very interested in

trying to optimize substrate topography and chemistry. Their goal is to find the best

substrate thatwill prevent astrocytes fromovergrowing the implanted electrodes since

these cells can interfere with signal transmission by insulating the electrodes from

the neurons. Results from research on microfluidic substrates for neural growth and

on prosthesis optimization will likely benefit both fields. For example, studies have

shown that certain cells do prefer certain nanotopographies, as demonstrated by the

finding that astroglial cells preferentially attach to pillars over wells and respond to

the topography by changing their expression of cytoskeletal proteins such as actin and

vinculin.140

Independent chemical cues also have different effects on different cell types. For

example, substrate-bound peptides with the amino acid sequence IKVAV preferen-

tially promote neural adhesion, whereas the sequence RGD promotes fibroblast and

glial cell adhesion.141 In contrast to the goals of prosthesis development to limit

astrocyte attachments, neural stem cell researchers desire a controlled bed of astro-

cytes to generate a permissive environment for the differentiation of neural stem cells.

Research shows that combining topographic cues (in the form of substrate grooves)

with chemical cues (in the formof adsorbed lamininmolecules) can orient over 85%of

astrocytes in the direction of thegrooves.142 This can help to control the differentiation

of neural stem cells cultured over the astrocytic bed and hopefullywill permit directed

axon regeneration in future studies.

A final limitation of developing truly three-dimensional cultures in vitro has been

the difficulty of maintaining cell viability when cell density approaches the order of

magnitude seen in live tissues. Cullen et al. theorized that this might be due to lack of

adequate perfusion to supply nutrients and remove waste products.143 They used

microfluidics to create a cylindrical PDMSculture plate that hadmultiple inlet ports at

the base and peripheral (circumferential) outlet ports along the edge. The plate was

covered with FEP (fluorinated ethylene propylene) membrane to minimize evapora-

tive losses while allowing gas exchange, and the ports were connected to a syringe

pump to keep the culture volume constant, while using forced interstitial convection

at various perfusion rates. Cells were loaded into a 500 mm thick 3D matrigel matrix

preloaded in theplate.The researcherswere able todemonstrate that aperfusion rate of

10–11mLmin�1 allowed greater than 90% viability in neuronal or neuronal/astrocyte

cocultures, with cell densities that more closely matched the density of the brain than

prior successfulmodels (although still lower than that found in the brain cortex). Itwill

likely be imperative that three-dimensional systems incorporate excellent perfusion

systems to truly simulate an in vivo experience.


Rowe et al. approached the perfusion problem from a structural engineering

standpoint and actually engineered a three-dimensional scaffolding system made

of SU-8 photoresist and gold electrodes.144 The design included a system of integral

microchannels and ports within the major support struts, to simulate microvascular

perfusion, and the entire structurewas encased in a PDMSmanifold for fluid delivery.

Their preliminary results found that their 3D substrate successfully supported

neuronal cultures and the resultant neural networks that developed showed more

complexity than those grown on two-dimensional electrode arrays. Future studies

planned to include electrical recordings from the gold electrodes (Figure 1.4).

Regardless of whether future in vivo-type culture systems employ three-dimensional

support structures built up chemically or structurally, or using a combination of these

techniques, theywill likelyfigure prominently in research that is trying to decipher cell

behavior in complex geometries and communities.

1.4 NEUROPHYSIOLOGY EXPERIMENTS USINGMICROFLUIDIC CHIPS

1.4.1 Cell Separation Tools

One prerequisite for in vitro studies on cells of neuronal lineage is starting with the

right cells.Whenharvested from thenervous systemofdonor animals, samples of cells

contain a mixture of large neurons and generally smaller glial support cells. A

separation process must then be used to isolate the desired cell type. Wu et al. devised

FIGURE 1.4 Schematics showing the microscaffold system. (a) An 8� 8 array of hollow

microtowers with functional cross-connects along the x-direction and structural cross-connects

along the y-direction. The microtowers in this schematic project 1.5mm out the topside and

backside of the Si orifice plate. (b) A 3D cross section of the microscaffold device encased in

the PDMS fluid manifold. The fluid manifold allows continuous perfusion of the cells growing

within the active microscaffold. Reproduced with permission from the Royal Society of

Chemistry.144

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS 19

a method of microfluidic cell separation using the change in spatial distribution that

occurs when fluid streams of different viscosity are mixed across an expanding

channel.145 They suspended themixture of cell types in an aqueous solution of sodium

alginate that supported neuronal survival rates of greater than 90% during the

separation process. They then mixed a stream of this suspension with an eluent

stream traveling at a different flow rate. As the flow rate increases, the viscosity

decreases since the alginate polymers become more linearly aligned with the micro-

fluidic channel walls. As the flow rate decreases, the viscosity increases since the

polymer chain bunches and intertwines with itself. If the eluent stream has a higher

viscosity than the stream in which the cells are suspended, then the interface between

the two streams is moved in such a way that the cell suspension stream becomes

narrower and more closely applied to the channel wall. When this cell suspension

stream width approaches a certain minimal value, the larger cells are picked up by

the higher viscosity eluent stream and the smaller cells are left behind in the original

cell stream. When these streams then encounter a branching point, the cells can be

separated. The researchers were able to successfully separate neurons and glia at a

rate nearing 100% when flow rates and interface distributions were stringently

controlled.

Stem cells that are fated to differentiate into neurons or astrocytes cannot be

separated based on their size since they are indistinguishable by their morphology and

size. Separationbyflowcytometry is based on the elaboration of different antigens that

can be fluorescently tagged, but these may not be present early in the differentiation

process. Flanagan et al. devised a microfluidic chip that used the dielectric properties

of stem cells to characterize their fate bias.146 Electrodes applied an alternating

electric field to cells in a microfluidic channel. At low frequencies, the cells are

repelled from the electrodes, but as the frequency of the switching field is increased,

they are attracted to the electrodes. The frequency atwhich they are “trapped” between

repulsion and attraction is characteristic for the cell’s fate bias, and these differentiat-

ing dielectric properties can be demonstrated prior to the development of protein

markers. In this study, live cells could be distinguished from dead cells that do not

“trap.” Astrocytes were trapped at 0.3MHz, neurons trapped at 5MHz, and neural

stem cells trapped at 1MHz. In addition, neural stem cells harvested from embryos at

earlier developmental ages trap at a higher frequency than those harvested from

embryos at later developmental ages, consistentwith the fact thatmore stemcells from

young embryos differentiate into neurons and more stem cells from older embryos

differentiate into astrocytes. This technique can be used to measure heterogeneity in

cell cultures and has several advantages over flow cytometry and FACS since it can be

done on small numbers of cells, it can exclude dead cells, it is sensitive to minor

differences in cells, and it needs no antibody.

1.4.2 Neuropeptide Release

Neuropeptides are endogenous chemicals that play a major role in the differentiation,

maturation, and communication between cells of neuronal lineage. A significant

portion of neurophysiology research involves deciphering the factors involved in their


release, uptake, and induced responses in local and distant cells. The effect of

neuropeptides on developing neuromuscular junctions has been studied using micro-

fluidic chips. This research is exemplified by the work of Tourovskaia et al. who used

PDMS masks to apply micropatterns of cell adhesive and cell repellent molecules to

isolate singlemyotubes in thin parallel lines.147Amicrofluidic devicewas then placed

over this, myoblasts were allowed to attach to the substrate, and streams of media

containing agrin were microfluidically perfused perpendicular to the axis of the cells.

Agrin is a molecule released by the tip of growing axons when they contact muscle

cells. The study showed that focal application of agrin stabilized the acetylcholine

receptor aggregations in the myotubes, consistent with its presumed actions involved

in the early development of neuromuscular synapses (Figure 1.5).

Indirect measurement of neurotransmitter release from isolated single cells was

demonstrated by Huang et al.148 They used PDMS microchannels to guide single

pheochromocytoma PC12 cells into a chamber etched into glass. When this cell was

stimulated with nicotine, they could detect dopamine release by amperometric

monitoring using a carbon fiber microelectrode. Gold-covered single cell wells in

silicon chips were also used to record catecholamine release from adrenal chromaffin

cells.149 PDMS microchannels were used to immobilize PC12 cells and use ampero-

metry to record calcium-induced dopamine and norepinephrine release.150 In similar

work, Sun and Gillis were able to record quantal exocytosis of catecholamines after

stimulation of chromaffin cells in microfluidic channels with potassium solution.151

They recorded amperometric spikes using indium tin oxide (ITO) electrodes when the

catecholamines were oxidized on the electrode surface.

When trying to differentiate the complex interactions that occur between different

cells in mixed-type or pure-type cultures, it is important to determine the cell’s

chemical response to stimulation, in addition to its electrical and morphological

responses. Although the latter responses have been more readily studied at the single

cell or near-single cell level using microfluidics, the analysis of released chemicals at

this scale has been more difficult to achieve, given the extremely minute concentra-

tions of the chemicals to be studied, in addition to the fact that the exact chemical

speciesmay not be known in advance. Jo et al. designed amicrofluidic chip that would

allow off-line analysis of neuropeptides released in response to chemical stimulation

of neuronswith potassiumchloride (KCl).152 Theyusedmultichamberedmicrofluidic

chips that contained a cell culture chamber functionalized with poly-L-lysine to allow

attachment of Aplysia bag cell neurons. Valves were used to selectively connect this

chamber to three other chambers, each functionalized with a SAM that could adsorb

molecules (released neuropeptides) by hydrophobic interactions. The cell culture

chamber was exposed to the KCl that activated release of the neuropeptides, and

the cell culture fluid was sequentially flushed from the culture chamber into each of

the SAM-containing chambers before, during, and after KCL stimulation. Once the

connecting valves were closed and the solutions were allowed time to adsorb to the

SAM, the PDMS was peeled from the chip, exposing the SAMs for MALDI mass

spectrometry measurements and subsequent imaging. The results of the research

confirmed that this methodology could successfully detect two different released

neuropeptides and that the majority of the released peptides adsorbed onto the SAM


FIGURE1.5 Synaptogenesis on a chip. (a) During development, neurons release agrin at the

site of contact between nerve and muscle. (b) Fluorescence micrograph of a portion of the

myotube microarray after staining the AChRs with Alexa Fluor 488-conjugated a-bungar-otoxin (BTX�). Scale bar is 50mm. (c) Three high-magnification fluorescence micrographs of

myotubes stained with BTX�, showing that aneural AChR clusters display intricate shapes

similar to those found in vivo. (d) Phase contrast micrograph of the microfluidic device

containing a ladder micropattern of myotubes during stimulation by a laminar stream of agrin

(spiked with Allura red dye for visualization). The black dashed box corresponds to the area

shown in (b). Reproduced from Ref. 147. Copyright 2008 Elsevier.


layer and not onto the PDMS walls. This innovative combination of microfluidic

control of cell bathing solution, with off-line mass spectrometry, has significant

potential to study known and previously unknown chemical responses of neurons

spatially and temporally.

Since both the concentrations and the volumes of released neuromediators are so

low, microfluidic-based on-chip analysis of these agents can represent an attractive

alternative to standard laboratory techniques. Mourzina et al. devised and optimized

an on-chip capillary electrophoresis system to separate neuromediators.89 They

experimented with various PDMS treatments and separation buffers to improve

electroosmotic pumping and decrease adsorption of the neuromediators onto the

PDMS surface. With the addition of field-amplified sample stacking, they were able

to achieve separation of fluorescently labeled neuropeptides (including oxytocin,

serotonin, glutamic acid, and others) within tens of seconds at 110 pL volume. Field-

amplified sample stacking utilizes a principlewhere the analyte is dissolved in a dilute

ionic solution that is sandwiched between higher concentration ionic solutions.

Application of fluid flow and electrical current causes formation of a stepped electric

field, resulting in migration of the analyte into the boundary area between these

solutions for easier separation. This work demonstrates the utility of using micro-

fluidic chips to process biologically relevant samples at minute scales.

Along similar lines, unpublished data fromother researchers (Phillips andWellner)

has demonstrated the adaptation of commercial microfluidic micromixer chips to

detect proteins such as proinflammatory cytokines in tiny samples of blood or

perspiration from patients with depression using recycling immunoaffinity chroma-

tography (online communication at http://www.nibib.nih.gov/HealthEdu/eAdvances/

30Jan09). Further refinement of these techniques will likely be forthcoming.

1.4.3 Physical and Chemical Guidance Cues

The ultimate goal ofmost neuroscience research involves learningways of preventing

degeneration and promoting repair and regeneration of neurons and their processes.

One key prerequisite for this is an understanding of how physical and chemical

guidance cues affect neurite growth. This is also of great interest to individuals who

study neural network design. Many studies in this field use microchannels to isolate

and observe axonal responses. An early study demonstrated the relative importance of

chemical guidance when neurites from chick spinal neurons that were otherwise

physically confined in “v”-shaped channels and pits on a silicon nitride substratewere

able to grow out of the channels if the substrate was pretreated with polylysine.53 The

effects of physical cues alone were demonstrated by studying neurite elaboration

by cells confined to square bottomed channels (constructed from polyimide walls

placed on a glass substrate).153 Narrow channels (20–30 mm in width) caused fewer

neurites to be elaborated from each cell, and each neurite was longer and more likely

to be oriented parallel to the channel wall. These changes might be due to inflexibility

of the cytoskeleton.

The effect of isolated chemical cues is the subject of many ongoing studies. In

one such study, a PDMS device with parallel stripes of channels was used to deliver


poly-L-lysine (PLL) or collagen to a substrate pretreated with sequential 3-amino-

propyltriethoxysilane (APTES) and glutaraldehyde.154 This pretreatment allowed

PLL or collagen to bind covalently to the substrate, so it was more stable and

structurally homogeneous than those attached by simple protein adsorption. The

PDMS was then removed and Aplysia neurons were applied and monitored as their

neurites developed. Standard electrophysiology and mass spectrometry were used to

investigate any differences between neurons cultured on patterned versus uniform

layers of protein. The results showed that patterned substrates caused shorter, thicker,

less branched, and slower growing neurites and caused changes in the cell’s electrical

activity compared to neurons grown on uniform proteins.

The effect of substance gradients on neurite extensionwas also specifically studied

by Whitesides and colleagues.52 This group utilized the serpentine network of

microfluidic channels discussed in Section 1.3.1 to study and quantify the effect of

different laminin concentrations on neurite sprouting, differentiation (into dendrite

or axon), and directionality from hippocampal neurons. They deposited a layer of

poly-L-lysine onto channels in a plasma-treated PDMSchip and then created a gradual

gradient that ranged from pure BSA to pure laminin. This treated PDMS channel was

then cut from its substrate, inverted in a Petri dish, and seeded with hippocampal

neurons. Cells that grew alone in the center of the gradient (center of channel) were

observed and the length of their neurites was measured to determine axonal specifi-

cation (axons were typically four times the length of dendrites). The results showed

that approximately 60% of axons were oriented within a 120� arc in the direction of

increasing laminin concentration, and although the laminin concentration did not

guide the axonal growth, it did specify which early neuritewould become the axon. In

addition, the concentration of laminin required to influence axonal specification was

determined. It is of particular importance to understand the in vitro effects of laminin

since it has been shown to play a critical role in axonal pathfinding in the embryonic

CNS in vivo.155

The effect of the absolute concentration of a substance on axonal growth is just

one parameter that has been studied inmicrofluidic chips. The slope of the substance’s

concentration gradient also exerts effects on axonal growth on specific types of

neurons. Lang et al. demonstrated this with ephrin A5, which is a repulsive axon

guidance molecule.156 They used silicon wafers with etched microfluidic channels to

create multiple stripes of varying concentrations of ephrin A5. Themolecules in these

solutions were transferred to a PDMS stamp and ultimately into polystyrene culture

dishes into which chick retinal ganglion cells originating from nasal or temporal

locations on the retina were cultured. They tested the effects of both steep gradient

variations and shallow gradient variations on axonal growth. They found that axons

growing from neurons originating on the nasal portion of the retina do not respond to

ephrinA5 gradients at all. However, axons growing fromneurons originating from the

temporal portion of the retina are inhibited, grow farther into shallow gradients than

steep gradients, and halt their growth at a lower total ephrinA5 concentration and total

exposure in shallow gradients of this molecule than in steep gradients. This research

indicates the complicated inputs involved in the growth of axons of differentiated cells

and the usefulness of microfluidics in deciphering the signals controlling axonal


responses (Figure 1.6). These researchers published a detailed protocol on their

methodology, including the use of a second active protein to set up an overlapping or

countergradient.157

Similarly, gradient mixers have been used to create substrate-bound gradients with

multiple agents of defined concentrations and slopes.112 Li et al. used a syringe pump

connected to a PDMS gradient mixer to combine laminin, chondroitin sulfate

proteoglycans (CSPGs), and/or BSA to form various patterns on a glass substrate.158

After 12 h of adsorption, the PDMS superstructure was removed and DRG neurons

were applied. The study showed that cells adhered more strongly to higher laminin

and lowerCSPGconcentrations, neurites grew towardhigher laminin and lowerCSPG

concentrations, and double opposing gradients provided the strongest guidance cues.

This research confirmed that neurites can detect and respond to both the slope and the

fractional concentration change of substrate-bound gradients.

FIGURE 1.6 Temporal RGC axons stop in substrate-bound gradients produced by mFN.(a) Fluorescence images of a stepwise gradient of ephrin A5 spanning a distance of 320mm,

and the corresponding countergradient of Fc and temporal axons stained with phalloidin

invading the gradient. In the original articlewith color images, in (b–e) phalloidin-stained axons

are shown in black and antibody-stained Fc in green. (b) Temporal axons stopping in a steep

ephrin A5 gradient. (c) The stop zone shifts further into the gradient in a shallow gradient.

(d) Nasal axons in a steep gradient do not stop. (e) Temporal axons growing on laminin lanes

without underlying gradient. For scale, see (a).With kind permission from Springer Science þBusiness Media.156


In reality, it may be an oversimplification to study the effect of individual agents on

neurons since it is more likely that they are programmed to respond to complex,

interacting, and synergistic forces, including topography, electromagnetic fields,

and chemical/biological cues. For example, when adult rat hippocampal progenitor

cells were cocultured with astrocytes aligned on laminin-coated substrate grooves,

there was enhanced neuronal differentiation and alignment of neurites parallel to

the astrocyte processes and substrate grooves. It is presumed that the astrocytes align

themselves on the grooves and then secrete soluble factors that are concentrated

locally by the topography, resulting in facilitated neuronal differentiation of the

progenitor cells.159

1.4.4 Electrophysiology and Microfluidics

Studies on neurons frequently involvemonitoring ofmultiple physiologic parameters,

including their electrical activity, as well as their morphology and expressed proteins.

Research in this field has been made easier by the availability of commercial MEAs,

which typically employ a glass substrate with electrically conductive microcontacts

and leads made of gold, platinum, or transparent ITO. Heuschkel et al. used an ITO

MEA and engineered a microfluidic chamber on its surface by layering negative

photoresist and using photolithography to pattern buriedmicrochannels in the resist.38

Once it was processed and baked, laminin and polyornithine were applied and chick

embryonic motoneurons were introduced via the microchannels. As the neurons

grew, their electrical activity was monitored via the microelectrodes. Although the

experiment was successful and the photoresist was found to be biocompatible, most

investigators now use PDMS to fabricate microfluidic channels intended for cell

culture.

For example, Morin et al. aligned the wells and channels of a PDMS microfluidic

chip to commercial and custom planar microelectrode array substrates.39 Poly-

L-lysine or laminin was applied to the chips prior to addition of chick or murine

cortical neurons. These cells remained viable and electrically active for weeks, as

demonstrated by optical microscopy and electrical responses to stimulation. Despite a

few problemswith isolating potentials from single cells, inhomogeneity of the cells in

eachwell (neuronversusglial), andPDMSadhesion to the commercialmicroelectrode

substrate, the authors felt that this system showed potential for the development of

neuronal networks. In similar work, Claverol-Tinture et al. used PDMS chips with

channels and wells over poly-L-lysine-coated planar ITO microelectrode arrays and

manually placed individual neurons in the wells.40 Once the axons grew, they were

able to record single cell spikes from the soma (contained in the well), or the axon

(contained in the channels), depending on how themicrofluidic channelswere aligned

on the electrodes (Figure 1.7). In follow-up studies, theywere able to achieve a signal-

to-noise ratio of 20 dB when recording electrical spike activity of up to 300mVamplitude from multiple sites on single neurites extending in microchannels.41

Thiebaud et al. also developed a PDMS microfluidic chamber that incorporated

microelectrode arrays.42 The first step in the fabrication process involved using a

PDMS microcontact stamp to deposit laminin onto an MEA substrate in parallel


stripes whose width was consistent with the size of microelectrode arrays. A PDMS

microfluidic apparatus with aligned microchannels then delivered culture medium

containing neuronal cells to the laminin stripes to establish the culture. Once the cells

attached, agents were injected into the parallel channels and delivered to the

established lines of cells via laminar flow to study the electrophysiological effects

of various pharmaceutical agents on the neurons.

Commercial microelectrode arrays have also been integrated into microfluidic

devices and used to study the effect of temperature changes on the electrical activity

of a subpopulation of cold-sensitive cells derived from the dorsal root ganglion.44

In this research, a microfluidic chip was used to deliver polylysine over the array prior

to the application of the cell suspension. There were two inlets to the chip—one

FIGURE 1.7 (a) Phase contrast image of a confined bipolar neuron after 12 days in vitro

sprouting two neurites along a microchannel. (b) Extracellular potentials associated with an

action potential and recorded by the eight microelectrodes shown in (a). (c) Raster plot

showing KCl and glutamate dose responses. Reproduced with permission from Ref. 41.

Copyright 2007 IEEE.


providing culturemediumat 45�C, and the other providingmediumat 4�C.Byvaryingthe flow rate of each input, they could rapidly (<1 s) switch the culture medium

temperature flowing over cells from 35 to 16�C. Using amultichannel multiprocessor

recording system, they recorded action potentials from the array of electrodes and

found that certain cells consistently changed their firing rate from a mean of

0.028 spikes per second to a mean of 0.94 spikes per second in response to a switch

to a colder temperature. These cells had the same morphology as the others that were

not sensitive to temperature changes. This systemdemonstrated that the combined use

of two unique microtechnologies (microfluidics and microelectrodes) could charac-

terize and identify a special subpopulation of cells based on their electrophysiological

responses.

In a follow-up study, these researchersmodified theirmicrofluidic design by adding

small reservoirswith flexiblemembrane covers to the side channels.43 This essentially

introduced a “switch” that could rapidly perturb the relative flow from each input

and vary the temperature flowing over the cells over a 50ms pulse. Future physiologi-

cally relevant researchwill benefit from this potential to rapidly and transiently deliver

agents to cultured cells on a timescale that more closely matches a cell’s innate

responsiveness capabilities.

1.4.5 Growth Factor Effects

Asdiscussed in Section 1.3.1,microfluidic channels can be arranged to create gradient

generators to study the effects of special factors on cell physiology.One research study

examined the effect of various concentrations and combinations of growth factors

(epidermal growth factor, fibroblast growth factor 2, and platelet-derived growth

factor) on neural stem cell proliferation and differentiation.115 Chung et al. first

demonstrated that they could successfully culture human neural stem cells in

chambers thatwere precoatedwith poly-L-lysine and laminin, and kept under constant

low flow (0.1mLmin�1) of culture medium for at least 7 days. This low flow would

helpminimize autocrine and paracrine effects of secreted factors.When growth factor

gradients were added, they demonstrated that neural stem cells proliferated in direct

proportion to the growth factor concentration, whereas astrocytes differentiated in

inverse proportion to the growth factor concentration, and all cells demonstrated

increased migration toward areas with higher concentrations. The use of microfluidic

gradients in this study presented a significant advantage over routine in vitro culture

techniques and helped to elucidate the action of these specific factors on these cells.

Wittig et al. also investigated the use of microfluidic channels to deliver graded

concentrations of growth factors.160 They applied a reusable PDMS microfluidic

culturemediumand a factor delivery apparatus onto a standard poly-L-lysine/laminin-

coated Petri dish. Two channels delivering different additives in media coalesced in

a “Y” shape that was placed adjacent to a neonatal spiral ganglion explant culture

area. The base of the “Y” allowed growing neurites to sample two different media

choices and then to decide which arm of the “Y” they preferred to grow into. They

demonstrated that neurites preferred to grow toward culture medium containing

neurotrophin-3 and they anticipated that this approach could beveryuseful in studying


the various effects of growth and inhibitory factors on proliferation, neurite extension,

and cell migration.

In addition to allowing precise delivery of controlled concentrations of growth

factors spatially, microfluidic platforms with control valves have also been used to

control the timing of the delivery of these substances to cells to determine the

differentiation of cells. Nakashima and Yasuda used a microfluidic control valve

to release nerve growth factor (NGF) through nanopores to control the differentiation

and axonal growth of adrenal pheochromocytoma cells.161 They were able to switch

the microvalve on and off with controlled frequency and duty cycles to guide cellular

differentiation. They anticipated that the addition of an electrode to monitor real-time

cellular response to their pulsed release of growth factors would be very useful in

studying the physiological responses of cells during axonal regeneration processes.

1.4.6 Gene Therapy

An alternative method of delivering growth factors to cells involves the use of

microfluidic methods to genetically manipulate neurons by directly delivering the

DNAcodes for a givengrowth factor to them.Houchin-Ray et al. cultured neuronson a

substrate that had been microfluidically patterned with a mixture of lipoplexes and

plasmid DNA that coded for NGF and green fluorescent protein.162 They were able to

achievea transfection efficiencyof25%usingavector concentration 10 times less than

typically used in culturemedia. Neurons cultured on the patterned areas had improved

survival and enhanced neurite outgrowth, indicating that the NGF DNA had been

incorporated into the cells and expressed. They also determined that pretreatment of

the PDMSmicrofluidic chipwith pluronic (an amphiphilic copolymer with surfactant

properties) improved transfection rates since it made the PDMS less likely to bind the

cationic lipid/DNA complexes. The NGF concentration gradient could be adjusted

by changing the plasmid density in the solution and the size of the microchannels in

the PDMS chip used to pattern the substrate. The potential application of patterned

gene delivery and expression to specific cells in culture has significant promise for

studies of neural physiology, cell-to-cell interactions, and regenerative medicine.

Microfluidic chips that allow fluidic isolation of parts of neurons (Section 1.4.7)

have enabled selective delivery of nonviral DNA to either neurites or the cell body of

neurons to study the differences in processing that occurs at these locations. This is

important for eventual in vivo studies since access to both the soma and the axon may

not be possible given the long lengths ofmany axons. Since many studies are aimed at

axon regeneration after spinal cord injury, treatments may need to be specialized for

the site that is accessible. Bergen and Pun illustrated this point with their research that

microfluidically delivered DNA in culture medium to the isolated neurites or to the

isolated soma of PC12 neuron-like cells.163 The DNAwas attached to either a lipid-

based carrier (lipofectamine) or PEI (poly(ethylenimine)), a polymeric nanoparticle.

In general, both had 4–5 times the uptake if delivered at the soma, compared to the

neurite, although the lipid-based carriers had an uptake that was 5–7 times that for

thePEI carrier.Uptake seemed tobemediated byvesicle formation.Whendelivered to

the isolated neurites, the lipoplexed DNA could be taken up, but not transported,


whereas the PEI-based DNAwas taken up and retrogradely transported in a saltatory

fashion, but never made it into the soma. Therefore, gene expression occurred only if

the gene-complexed carrier was delivered directly to the soma, and the transfection

efficiencywasmuch higher for the lipofectamine than for the PEI. This research sheds

light on the difficulties inherent in repairing a damaged spinal cord.

1.4.7 Axonal Isolation

As demonstrated in the above literature review,microfluidic chambers have been used

to isolate small numbers of cells of neuronal lineage and to study their individual

responses to stimuli, often with the aid of on-board analytical devices. However, none

of these studies were able to fluidically isolate cell segments (like axons) from their

soma, or from neighboring cells of various lineages. The only method to achieve this

in the past involved the use of Campenot chambers that used nerve growth factor or

brain-derived neurotrophic factor to “artificially” stimulate axonal growth from

macro-scaled cultures across grease layers.

However, Jeon’s group at the University of California at Irvine has successfully

designed and implemented a microfluidic platform that incorporates tiny grooves of

adequate size and length to allowfluidic isolation of axons from the regular cell culture

chamber, so that their physiology can be studied independently.164 Specifically, they

designed a PDMS chip with two chambers that were separated by a series of grooves

that each had dimensions of 10mm width, 3mm height, and 150mm length. These

grooves were used to guide neurite growth from a cell culture chamber into a second

chamber that could be used to study the isolated axons. The narrowness of the grooves

prevented cell bodies from penetrating them, and the length of the grooves prevented

the typically shorter dendrites from emerging into the axonal isolation chamber.A key

design element of the tiny grooves was their high resistance to fluid transport that

allowed them to achieve temporary (15 h) fluidic isolation of the somal compartment

from the axonal compartment by applying a slightly higher hydrostatic pressure in

the somal compartment. On doing this, there was one-way flow only, going from the

somal to the axonal compartment, in a very slow and restricted manner. This allows

independent chemical manipulation of the axon, without any direct effects on the

soma, unless the axondirectly transports the agents to the soma in a retrograde fashion.

These researchers used a patterned poly-L-lysine substrate to keep the axons

aligned in parallel stripes, so that they could be more easily identified along with

their respective cell body.164–166 They used several alternative tools for the patterning,

including a PDMS mold with tiny channels that could wick the agent in by capillary

action (micromolding in capillaries), a PDMS stamp to transfer the agent by micro-

contact printing, and a PDMSmask to selectively protect and preserve a uniform, pre-

applied, dried agent during plasma etching.

In follow-up research, Jeon’s group lengthened the microgrooves to 450mmto allow longer studies (14 days) that still isolated axons from dendrites

(Figure 1.8).164,165 They used this design to investigate axonal injury, regeneration,

and interactions with cocultured oligodendrocytes during myelination. In addition,

they proved that their isolated compartments could permit detection of purely axonal


FIGURE1.8 Themicrofluidic-based culture platform directs axonal growth of CNS neurons

and fluidically isolates axons. (a) The culture chamber consists of a PDMS mold containing a

relief pattern of somal and axonal compartments (1.5mm wide, 7mm long, 100mm high)

connected bymicrogrooves (10mmwide, 3mmhigh). The optically transparent PDMS adheres

to a polylysine-coated coverslip. Rat CNS neurons (medium gray spots) are added to the somal-

side reservoir and are drawn into the somal channel (black) by capillary action.Within 3–4 days,

axonal growth is guided into the axonal side (light gray) through the microgrooves. (b) A

volume difference between the somal side and the axonal side (�50mL) allows chemical

microenvironments to be isolated to axons for over 20 h owing to the high fluidic resistance of

the microgrooves. Similarly, the volume difference can be reversed to isolate a chemical

microenvironment to the somal side. (c) Fluidic isolation of Texas red dextran (top panel) to the

axonal compartment demonstrates that axonal or somatic microenvironments can be indepen-

dently manipulated using this culture platform. Axonally restricted application of CellTracker

Green (middle panel) backtracked neurons from their isolated axons. The bottom image is the

merged figure. Scale bar, 100mm. (d) Counts of radioactivity in samples from somal and

axonal compartments after [35S]methionine was localized to the axonal compartment for over

20 h. Counts in the somal compartment (3.7 c.p.m.� 1.5 s.e.m.) were similar to background

levels. Error bars, s.e.m. (n¼ 3). Reprinted with permission from Ref. 165. Copyright

2005Macmillan Publishers Ltd.


mRNA, and they demonstrated changes in gene expression in the soma in response to

axonal chemical and physical manipulations. This chip has now become an important

tool used in research programs at different academic settings, given its potential to

reveal details of neuronal pathophysiology in neurodegenerative diseases and trau-

matic nerve damage.

When neuronal axons are damaged, the distal portion experiences a process called

Walleriandegeneration, inwhich the proximal portionof the axon shrinks back toward

the cell body. Regeneration of damaged axons is possible but seems to be hampered

by many complicated and interacting factors, including glial scar tissue that prevents

the regrowing axons fromfinding their target cells.One active inhibitorycomponent of

the scar tissue is a group of molecules composed of core proteins with carbohydrate

side chains, known as CSPGs. Neuroscientists studying axonal regeneration have

postulated that treatments that limited the production of these molecules at sites of

nerve damage might permit axonal regrowth and reestablishment of normal function.

Jeon’s group tested this hypothesis in their microfluidic chip by isolating axons in

a chamber with pre-applied stripes of alternating inhibitory and permissive mole-

cules.167 The inhibitory stripes contained these aggrecan proteins to mimic glial scars

that are generated after spinal cord injury. These inhibitory areas were effectively

neutralized when chondroitinasewas added to the axonal chamber of themicrofluidic

chip. This molecule acts to dissolve the carbohydrate glycosaminoglycan side chains

of the CSPGs, but left the core protein intact. Since the addition of the chondroitinase

allowedaxons to cross onto theCSPGstripe, theyconcluded that the core proteinof the

CSPG was not inhibitory, but the carbohydrate side chains were the active inhibitory

component of the CSPG molecule (Figure 1.9).

Other proteins inhibitory to axon growth have also been tested by this research

laboratory. Park et al. used their axon isolation chip to test the effects of two myelin-

associated proteins on axon regeneration.168 After the axons grew into the isolation

chamber (about 7 days), they were severed with vacuum aspiration and their

regeneration was monitored in the presence and absence of various concentrations

of NOGO-66 and MAG protein. Each of these proteins decreased the length of

regenerated axons by 75–80% compared to controls. Higher concentrations ofNOGO

lead to increased inhibition of regeneration, with the effect saturating at a concentra-

tion of 10 nM.

The effect of toxins on the electrophysiology and morphology of neurons has been

studied using similar axonal isolationmicrofluidic chips that are combinedwithMEA

substrates. Ravula et al. mated a PDMS superstructure onto a glass substrate with a

patternedMEAand recorded the spontaneous and stimulated electrical activities from

neurons that they cultured in this platform.169 They were able to record action

potentials from both the soma and the isolated axons when they added potassium

chloride. When the sodium channel blocker tetrodotoxin was added to the axonal

compartment, only axonal action potentials ceased. In follow-up work, Ravula et al.

tested the effects of other chemicals on the electrophysiology of fluidically isolated

neuronal soma and axons. They found that low-dose vincristine caused no effect if

directly applied to the soma of the neuron, but axonal application sequentially caused

a decreased excitability of the axon, an initial increase in excitability of the soma, and


an eventual degeneration of the axon.170 Higher dose vincristine caused degeneration

if applied to either compartment. The change in electrophysiology occurred approxi-

mately 6 h after vincristine exposure andwas an earlier indicator of degeneration than

were morphological changes. The degradation of the electrical response to a depolar-

izing dose of potassium chloride is first noted in the distal axon and then progresses

proximally. Eighteen hours after exposure, morphological degeneration begins and

progresses at a rate of 1–2mmper day. All the above studies indicate that microfluidic

axonal isolation chips, with or without electrical monitoring ability, hold promise for

high-throughput screening of many pharmaceutical agents that have effects on

neuronal health and might be useful in enhancing axon regeneration.

1.5 DISCUSSION AND FUTURE PERSPECTIVES

Neuroscientists have typically beenveryopen to the use of cutting edge technology for

the studyof neuronal physiologyandhave incorporated this technology into their daily

research. The challenges of applying novel micro- and nanofabricated hardware to

classic neurophysiology experiments will hopefully be matched by their potential

FIGURE 1.9 Effect of chondroitinase ABC treatment on axons growing in microfluidic strip

assay device. 10 DIV cortical neurons growing on aggrecan–PLL patterned strips on glass.

Axons avoid inhibitor-coated areas until a drug is applied to help axons overcome the inhibition

or break down the surface-bound inhibitors. Axons avoid aggrecan strips and are confined to

PLL strips during entire experimental period. Striking changes in growth conemorphology and

axonal projections were observed when the pattern was treated with chondroitinase ABC, an

enzyme that removes the CSPGs’ side chains without degradation of the core protein. After

30min ChABC treatment, axons are observed to randomly extend across the pattern. Higher

resolution micrographs indicate that aggrecan inhibition is overcome not only in the growth

cones but also in the middle of the stripes. Reprinted from Ref. 167. Copyright 2008 Elsevier.

DISCUSSION AND FUTURE PERSPECTIVES 33

yield. The strength of microfluidics is in the highly efficient utilization of the

reductionist approach in well-defined, tightly focused environments and problems.

Another advantage of microfluidics is the integrability with multiple methods of

interrogation, including chemical, optical, and electrical methods. Microfluidic plat-

forms have the ability to deliver individual cells to specific locations, and then allow

study of the effects of temporally and spatially controlled environmental perturbations

on isolated parts of the cells. Arrays of these platforms can be arranged to allow large

parallel experiments that can multiply experimental yield. When analytical compo-

nents (such as NMR coils, pH, and electrical and chemical sensors) are incorporated

into the microfluidic chips, real-time information on individual live cells can be

recorded and followed. It is possible that the conclusions drawn from studies of

conventional cell culture populations may not be borne out at the single cell level.

Determiningwhich in vitro cellular behaviors aremost consistentwith in vivo realities

will then become imperative since it is likely that the choice of culture techniquesmay

influence experimental results.

Initiating a microfluidic-based research program is much simpler than one might

think.Most clean room fabrication facilities at universities already have the technolo-

gy to create themaster for the PDMSplatform since it is the same as that used to create

silicon wafer-based electronics. If there is no access to these services locally,

commercial businesses and some universities (such as the University of California

at Irvine Integrated Nanosystems Research Facility Foundry171) do offer these

services for a fee. Once the master is available, the supplies and the equipment to

cast the PDMS chips are common and inexpensive. For example, PDMS (sold as

Sylgard� 184) is available from Fisher Scientific for about $60 for a 1.1 pound kit,

which will be sufficient for dozens of chips. Vacuum pumps (to degas the PDMS prior

to pouring), and ovens (to bake the mold), will improve the quality of the casting,

but these are not necessary for chips that lack very small features. The only other tools

that may be necessary are punches to gain access to the channels in the chip, and

syringe pumps, if called for in the design. Therefore, once a master is obtained, the

ability to produce multiple chips is within easy reach to life scientists.

For those who prefer to purchase their PDMS chips, complete microfluidic

fabrication foundries (such as those located at Stanford University and California

Institute of Technology) offer commercial services to build chips based on original

designs sent by researchers. Plasma treatment equipment confers the advantage of

temporarily altering the surface of the PDMS to make it more hydrophilic and able to

irreversibly bond to the substrate. However, reversible bonds are sufficient for most

low-pressure cell culture applications, and, if desired, the PDMS surface can be made

hydrophilic by simple exposure to media containing 10–15% serum.

Commercialmultielectrode arrays and supporting electronicmodules are available

frommany sources and can be used as the substrate for microfluidic chips. Substrates

with specialized nanotopographic features can be produced by most clean room

fabrication facilities. Patterned cell adhesives are easily obtained using dedicated

PDMS stencils or stamps. Three-dimensional scaffolding such as collagen sponges

from BD Biosciences or Inamed Biomaterials can be used within PDMS growth

chambers for studies on tissue engineering within microfluidic chips.


Interdisciplinary collaborations between neuroscientists and engineers may offer

the best chance for success for life scientists who plan to initiate a microfluidic

research program. In this setting, the engineers are often happy to provide technical

advice and hardware in exchange for biological applications for their technology. This

may be especially fruitful for eachmember of the team since federal funding agencies

are recognizing the importance of interdisciplinary work for innovative groundbreak-

ing fundamental research and industrial applications.

As we look toward the future, material science issues may become a critical factor

for the progress of microfluidic neuroscience. The literature cited above validates

the potential for microfluidics, but issues of possible chemical interactions between

the PDMSand the cells or the culturemediumwill have to be resolved. It is known that

many materials that are used in standard in vitro studies may interact with culture

contents. For example, proteins are known to adsorb onto polystyrene culture flasks

but the relative ratio of the culture medium volume to the polystyrene surface area

usually makes the impact of this phenomenon negligible for most applications.

However, molecular and fluid interactions with PDMS on tiny culture volumes

will be more likely to play a role in influencing results. Potential toxic effects of

the PDMS on sensitive neurons will also need to be controlled. It may be that PDMS

will have to be layered with other polymers, like parylene, to limit its interaction with

media.Or PDMSmayhave to be replacedwith other polymers that are not as “porous”

to water and solutes.

Finally, environmental controls may have to be built into the chips so that their

temperature, osmolarity, and atmosphere are preserved during manipulation and

imaging of the chip. Given the platform’s small volume and lower homeostatic

reserve, the heat, humidity, and carbon dioxide levels established in an incubator may

otherwise change very quickly as the chip is removed for microscopic examination or

media changes. The larger ratio of surface area to volume of these small culture

chambers may also lead to a greater tendency toward evaporation and osmotic

concentration of the media. All these issues can result in significant effects on the

cell’s physiological responses and viability. So, to take advantage of the platform’s

ability to precisely control the milieu of the cells it contains, we may first need to

engineer a controllable environment for the platform itself. Despite these challenges,

the application of microfluidic technology to appropriate in vitro neuroscience

research has the potential to offer new insights and to augment ongoing conventional

in vitro and in vivo investigations.

As with all technology, there are certain limitations and challenges to microfluidic

studies, as outlined in Section 1.2.5. The “ground-up” approach to studying cellular

interactions may yield truly novel results, but caution must be exercised before

generalizing information from single cell studies to either homogeneous or heteroge-

neous populations of cells. Moreover, as demonstrated in the field of tissue engineer-

ing, generalizing conclusions from any two-dimensional cell culture (whether

microfluidic or conventional in scale) to in vivo settings may also be misleading.

But, as long aswekeep our perspective,we can combine the results obtained fromboth

conventional and alternative methodologies to obtain a greater understanding of the

processes we are trying to decipher.

DISCUSSION AND FUTURE PERSPECTIVES 35

In conclusion, the science of studying individual neuronal cells is still in its infancy,

but we now have the tools needed for this endeavor in the form of microfabricated

microfluidic channels and electronic sensors that provide the platform for cell

proliferation, separation, differentiation, and monitoring. Newer designs also allow

isolationof individual axonson these cells,whichwill enable studyof localized axonal

physiology and its effects on the cell body. Combining microfluidics with other

nanotechnologies will enable truly novel experiments never previously possible. As

this chapter outlines, there are many applications of microfluidics to neuroscience,

and individuals who can make use of this technology may pioneer entire new areas of

research, including studiesof individual axons, individual synapses, or the interactions

of single neurons with other isolated CNS cells (microglia, oligodendrocytes, and

astrocytes).

Advances in other complementary microtechnologies, such as noninvasive single

cell electrical recording, monitors for other key markers of cellular physiology (such

as pH and ionic currents), tools for NMR spectroscopy, mass spectrometry, and

designs to allow on-chip genetic manipulation and evaluation, should allow rapid

advances to bemade toward understanding neuronal physiology. Oncewe understand

how these cells operate under controlled conditions, we may be able to use this

information to gain control of cell responses in disease, and design treatments and

possible cures for patients suffering from neurologic diseases.

ACKNOWLEDGMENT

We thank Raymond W. Glover, MD, for reviewing and editing the manuscript.

REFERENCES

1. Silva, G.A. Nanotechnology approaches for the regeneration and neuroprotection of the

central nervous system. Surg. Neurol. 2005, 63, 301–306.

2. Silva, G.A. Neuroscience nanotechnology: progress, opportunities and challenges. Nat.

Rev. Neurosci. 2006, 7, 65–74.

3. Moshfeghi, A.A.; Peyman, G.A.Micro- and nanoparticulates.Adv. DrugDeliv. Rev. 2005,

57, 2047–2052.

4. Rosi, N.L.; Giljohann, D.A.; Thaxton, C.S.; Lytton-Jean, A.K.R.; Han, M.S.; Mirkin, C.A.

Oligonucleotide-modified gold nanoparticles for intracellular gene regulation. Science

2006, 312, 1027–1030.

5. Svenson, S.; Tomalia, D.A. Dendrimers in biomedical applications—reflections on the

field. Adv. Drug Deliv. Rev. 2005, 57, 2106–2129 (commentary).

6. Yokoyama, M. Drug targeting with nano-sized carrier systems. J. Artif. Organs 2005, 8,

77–84.

7. Gomez, N.; Winter, J.O.; Shieh, F.; Saunders, A.E.; Korgel, B.A.; Schmidt, C.E.

Challenges in quantum dot-neuron active interfacing. Talanta 2005, 67, 462–471.


8. Michalet, X.; Pinaud, F.F.; Bentolila, L.A.; Tsay, J.M.; Doose, S.; Li, J.J.; Sundaresan, G.;

Wu, A.M.; Gambhir, S.S.; Weiss, S. Quantum dots for live cells, in vivo imaging, and

diagnostics. Science 2005, 307, 538–544.

9. Pinaud, F.; Michalet, X.; Bentolila, L.A.; Tsay, J.M.; Doose, S.; Li, J.J.; Iyer, G.; Weiss, S.

Advances in fluorescence imaging with quantum dot bio-probes. Biomaterials 2006, 27,

1679–1687.

10. Vu, T.Q.; Maddipati, R.; Blute, T.A.; Nehilla, B.J.; Nusblat, L.; Desai, T.A. Peptide-

conjugated quantum dots activate neuronal receptors and initiate downstream signaling

of neurite growth. Nano Lett. 2005, 5, 603–607.

11. Winter, J.O.; Liu, T.Y.; Korgel, B.A.; Schmidt, C.E. Recognition molecule directed

interfacing between semiconductor quantum dots and nerve cells. Adv. Mater. 2001,

13, 1673–1677.

12. Esplandiu, M.J.; Bittner, V.G.; Giapis, K.P.; Collier, C.P. Nanoelectrode scanning probes

from fluorocarbon-coated single-walled carbon nanotubes.Nano Lett. 2004, 4, 1873–1879.

13. Albrecht, D.R.;Underhill, G.H.;Wassermann, T.B.; Sah,R.I.; Bhatia, S.N. Probing the role

of multicellular organization in three-dimensional microenvironments. Nat. Methods

2006, 3, 369–375.

14. Hu, H.; Ni, Y.C.; Montana, V.; Haddon, R.C.; Parpura, V. Chemically functionalized

carbon nanotubes as substrates for neuronal growth. Nano Lett. 2004, 4, 507–511.

15. Mattson, M.P.; Haddon, R.C.; Rao, A.M.Molecular functionalization of carbon nanotubes

and use as substrates for neuronal growth. J. Mol. Neurosci. 2000, 14, 175–182.

16. Mohammed, J.S.; DeCoster, M.A.; McShane, M.J. Micropatterning of nanoengineered

surfaces to study neuronal cell attachment in vitro. Biomacromolecules 2004, 5,

1745–1755.

17. Stevens, M.M.; George, J.H. Exploring and engineering the cell surface interface. Science

2005, 310, 1135–1138.

18. Tan, W.; Desai, T.A. Layer-by-layer microfluidics for biomimetic three-dimensional

structures. Biomaterials 2004, 25, 1355–1364.

19. Tan, W.; Desai, T.A. Microscale multilayer cocultures for biomimetic blood vessels.

J. Biomed. Mater. Res. A 72A, 2005, 146–160.

20. Dugan, L.L.; Lovett, E.G.; Quick, K.L.; Lotharius, J.; Lin, T.T.; O’Malley, K.L. Fullerene-

based antioxidants and neurodegenerative disorders.Parkinsonism Relat. Disord. 2001, 7,

243–246.

21. Yang, F.; Murugan, R.; Ramakrishna, S.; Wang, X.; Ma, Y.X.; Wang, S. Fabrication of

nano-structured porous PLLA scaffold intended for nerve tissue engineering.Biomaterials

2004, 25, 1891–1900.

22. Tiwari, S.B.; Amiji, M.M. A review of nanocarrier-based CNS delivery systems. Curr.

Drug Deliv. 2006, 3, 219–232.

23. McNally, H.A.; Ben Borgens, R. Three-dimensional imaging of living and dying neurons

with atomic force microscopy. J. Neurocytol. 2004, 33, 251–258.

24. McNally, H.A.; Rajwa, B.; Sturgis, J.; Robinson, J.P. Comparative three-dimensional

imaging of living neurons with confocal and atomic force microscopy. J. Neurosci.

Methods 2005, 142, 177–184.

25. Dahan, M.; Levi, S.; Luccardini, C.; Rostaing, P.; Riveau, B.; Triller, A. Diffusion

dynamics of glycine receptors revealed by single-quantum dot tracking. Science 2003,

302, 442–445.

REFERENCES 37

26. Pathak, S.; Cao, E.; Davidson, M.C.; Jin, S.H.; Silva, G.A. Quantum dot applications to

neuroscience: new tools for probing neurons and glia. J. Neurosci. 2006, 26,

1893–1895.

27. Guillorn, M.A.; McKnight, T.E.; Melechko, A.; Merkulov, V.I.; Britt, P.F.; Austin, D.W.;

Lowndes, D.H.; Simpson, M.L. Individually addressable vertically aligned carbon nano-

fiber-based electrochemical probes. J. Appl. Phys. 2002, 91, 3824–3828.

28. Patolsky, F.; Timko, B.P.; Yu, G.H.; Fang, Y.; Greytak, A.B.; Zheng, G.F.; Lieber, C.M.

Detection, stimulation, and inhibition of neuronal signals with high-density nanowire

transistor arrays. Science 2006, 313, 1100–1104.

29. Duffy, D.C.; McDonald, J.C.; Schueller, J.A.; Whitesides, G.M. Rapid prototyping of

microfluidic systems in poly(dimethylsiloxane). Anal Chem. 1998, 70, 4974–4984.

30. Unger, M.A.; Chou, H.P.; Thorsen, T.; Scherer, A.; Quake, S.R. Monolithic microfabri-

cated valves and pumps by multilayer soft lithography. Science 2000, 113–116.

31. Thorsen, T.; Maerkl, S.J.; Quake, S. Microfluidic large-scale integration. Science 2002,

298, 580–584.

32. Kartalov, E.P.; Walker, C.; Taylor, C.R.; Anderson, W.F.; Scherer, A. Microfluidic vias

enable nested bioarrays and autoregulatory devices in Newtonian fluids. PNAS 2006, 103,

12280–12284.

33. Hansen, C.L.; Skordalakes, E.; Berger, J.M.; Quake, S. A robust and scalable microfluidic

meteringmethod that allows protein crystal growth by free interface diffusion.PNAS 2002,

99, 16531–16536.

34. Hansen, C.L.; Sommer, M.O.A.; Quake, S. Systematic investigation of protein phase

behavior with a microfluidic formulator. PNAS 2004, 101, 14431–14436.

35. Kartalov, E.P.; Quake, S.R. Microfluidic device reads up to four consecutive base pairs in

DNA sequencing-by-synthesis. Nucleic Acids Res. 2004, 2873–2879.

36. Liu, J.; Enzelberger,M.; Quake, S. A nanoliter rotary device for polymerase chain reaction.

Electrophoresis 2002, 23, 1531–1536.

37. Liu, J.; Hansen, C.L.; Quake, S. Solving the “world-to-chip” interface problem with a

microfluidic matrix. Anal. Chem. 2003, 75, 4718–4723.

38. Heuschkel, M.O.; Guerin, L.; Buisson, B.; Bertrand, D.; Renaud, P. Buried microchannels

in photopolymer for delivering of solutions to neurons in a network. Sens. Actuators B

Chem. 1998, 48, 356–361.

39. Morin, F.; Nishimura, N.; Griscom, L.; LePioufle, B.; Fujita, H.; Takamura, Y.; Tamiya, E.

Constraining the connectivity of neuronal networks cultured on microelectrode arrays

with microfluidic techniques: a step towards neuron-based functional chips. Biosens.

Bioelectron. 2006, 21, 1093–1100.

40. Claverol-Tinture, E.; Ghirardi, M.; Fiumara, F.; Rosell, X.; Cabestany, J. Multielectrode

arrays with elastomeric microstructured overlays for extracellular recordings from pat-

terned neurons. J. Neural. Eng. 2005, 2, L1–L7.

41. Claverol-Tinture, E.; Cabestany, J.; Rosell, X. Multisite recording of extracellular poten-

tials produced by microchannel-confined neurons in vitro. IEEE Trans. Biomed. Eng.

2007, 54, 331–335.

42. Thiebaud, P.; Lauer, L.; Knoll, W.; Offenhausser, A. PDMS device for patterned applica-

tion of microfluids to neuronal cells arranged by microcontact printing. Biosens.

Bioelectron. 2002, 17, 87–93.


43. Pearce, T.M.;Williams, J.J.; Kruzel, S.P.; Gidden,M.J.;Williams, J.C. Dynamic control of

extracellular environment in in vitro neural recording systems. IEEE Trans. Neural. Syst.

Rehabil. Eng. 2005, 13, 207–212.

44. Pearce, T.M.; Wilson, J.A.; Oakes, S.G.; Chiu, S.Y.; Williams, J.C. Integrated microelec-

trode array and microfluidics for temperature clamp of sensory neurons in culture. Lab

Chip 2005, 5, 97–101.

45. Grant, S.C.;Aiken,N.R.; Plant, H.D.;Gibbs, S.;Mareci, T.H.;Webb,A.G.; Blackband, S.J.

NMR spectroscopy of single neurons. Magn. Reson. Med. Sci. 2000, 44, 19–22.

46. Massin, C.; Vincent, F.; Homsy, A.; Ehrmann, K.; Boero, G.; Besse, P.A.; Daridon, A.;

Verpoorte, E.; de Rooij, N.F.; Popovic, R.S. Planar microcoil-based microfluidic NMR

probes. J. Magn. Reson. 2003, 164, 242–255.

47. Gourley, P.L.; Copeland, R.G.;Cox, J.D.;Hendricks, J.K.;McDonald,A.E.; Peterson, S.L.;

Sasaki, D.Y. Biocompatible semiconductor optoelectronics. J. Biomed. Opt. 2002, 7,

546–554.

48. Lee, K.; He, J.P.; Clement, R.; Massia, S.; Kim, B. Biocompatible benzocyclobutene

(BCB)-based neural implants with micro-fluidic channel. Biosens. Bioelectron. 2004, 20,

404–407.

49. Suzuki, T.; Kotake, N.;Mabuchi, K.; Takeuchi, S. Bundledmicrofluidic channels for nerve

regeneration electrodes. Proceedings of the 3rd International IEEE EMBS Conference on

Neural Engineering, 2007, pp. 17–18.

50. Kartalov, E.P.; Zhong, J.F.; Scherer, A.; Quake, S.R.; Taylor, C.R.; Anderson, W.F. High-

throughput multi-antigen microfluidic fluorescence immunoassays. Biotechniques 2006,

40, 85–90.

51. Fu, A.Y.; Chou, H.P.; Spence, C.; Arnold, F.H.; Quake, S.R. An integratedmicrofabricated

cell sorter. Anal. Chem. 2002, 74, 2451–2457.

52. Dertinger, S.K.W.; Jiang, X.Y.; Li, Z.Y.; Murthy, V.N.; Whitesides, G.M. Gradients of

substrate-bound laminin orient axonal specification of neurons.Proc. Natl. Acad. Sci. USA

2002, 99, 12542–12547.

53. Martinoia, S.; Bove, M.; Tedesco, M.; Margesin, B.; Grattarola, M. A simple microfluidic

system for patterning populations of neurons on silicon micromachined substrates.

J. Neurosci. Methods 1999, 87, 35–44.

54. Wheeler, A.R.; Throndset, W.R.; Whelan, R.J.; Leach, A.M.; Zare, R.N.; Liao, Y.H.;

Farrell, K.; Manger, I.D.; Daridon, A. Microfluidic device for single-cell analysis. Anal.

Chem. 2003, 75, 3581–3586.

55. Lee, P.J.; Hung, P.J.; Shaw, R.; Jan, L.; Lee, L.P. Microfluidic application-specific

integrated device formonitoring direct cell–cell communication via gap junctions between

individual cell pairs. Appl. Phys. Lett. 2005, 86.

56. Peng,X.Y.; Li, P.C.H.A three-dimensional flowcontrol concept for single-cell experiments

on a microchip. 2. Fluorescein diacetate metabolism and calcium mobilization in a single

yeast cell as stimulated by glucose and pH changes. Anal. Chem. 2004, 76, 7400–7400.

57. Peng, X.Y.; Li, P.C.H. A three-dimensional flow control concept for single-cell experi-

ments on a microchip. 1. Cell selection, cell retention, cell culture, cell balancing, and cell

scanning. Anal. Chem. 2004, 76, 5273–5281.

58. Peng, X.Y.; Li, P.C.H. A three-dimensional flow control concept for single-cell experi-

ments on a microchip. 2. Fluorescein diacetate metabolism and calcium mobilization in

REFERENCES 39

a single yeast cell as stimulated by glucose and pH changes. Anal. Chem. 2004, 76,

5282–5292.

59. Rocheleau, J.V.; Walker, G.M.; Head, W.S.; McGuinness, O.P.; Piston, D.W. Microfluidic

glucose stimulation reveals limited coordination of intracellular Ca2þ activity oscillations

in pancreatic islets. Proc. Natl. Acad. Sci. USA 2004, 101, 12899–12903.

60. Ma, N.; Koelling, K.W.; Chalmers, J.J. Fabrication and use of a transient contractional flow

device to quantify the sensitivity of mammalian and insect cells to hydrodynamic forces.

Biotechnol. Bioeng. 2002, 80, 428–437.

61. Gray, B.L.; Lieu, D.K.; Collins, S.D.; Smith, R.L.; Barakat, A.I.Microchannel platform for

the study of endothelial cell shape and function. Biomed. Microdevices 2002, 4, 9–16.

62. Sin, A.; Chin, K.C.; Jamil, M.F.; Kostov, Y.; Rao, G.; Shuler, M.L. The design and

fabrication of three-chamber microscale cell culture analog devices with integrated

dissolved oxygen sensors. Biotechnol. Prog. 2004, 20, 338–345.

63. Rothert, A.; Deo, S.K.; Millner, L.; Puckett, L.G.; Madou, M.J.; Daunert, S. Whole-cell-

reporter-gene-based biosensing systems on a compact disk microfluidics platform. Anal.

Biochem. 2005, 342, 11–19.

64. Matsubara, Y.; Murakami, Y.; Kobayashi, M.; Morita, Y.; Tamiya, E. Application of on-

chip cell cultures for the detection of allergic response. Biosens. Bioelectron. 2004, 19,

741–747.

65. DeBusschere, B.D.; Kovacs, G.T.A. Portable cell-based biosensor system using integrated

CMOS cell-cartridges. Biosens. Bioelectron. 2001, 16, 543–556.

66. Guo, S.X.; Bourgeois, F.; Chokshi, T.; Durr, N.J.; Hilliard, M.A.; Chronis, N.; Ben-Yakar,

A. Femtosecond laser nanoaxotomy lab-on-a-chip for in vivo nerve regeneration studies.

Nat. Methods 2008, 5, 531–533.

67. Allen, P.B.; Sgro, A.E.; Chao, D.L.; Doepker, B.E.; Edgar, J.S.; Shen, K.; Chiu, D.T.

Single-synapse ablation and long-term imaging in live C. elegans. J. Neurosci. Methods

2008, 173, 20–26.

68. Ho, C.L.; Mou, T.Y.; Chiang, P.S.; Weng, C.L.; Chow, N.H. Mini chamber system for

long-term maintenance and observation of cultured cells. Biotechniques 2005, 38,

267–273.

69. Chen, C.S.; Jiang, X.Y.; Whitesides, G.M. Microengineering the environment of mam-

malian cells in culture. MRS Bull. 2005, 30, 194–201.

70. Walker, G.M.; Zeringue, H.C.; Beebe, D.J. Microenvironment design considerations for

cellular scale studies. Lab Chip 2004, 4, 91–97.

71. Li, N.; Tourovskaia, A.; Folch, A. Biology on a chip: microfabrication for studying the

behavior of cultured cells. Crit. Rev. Biomed. Eng. 2003, 31, 423–488.

72. Qin, J.H.; Ye, N.N.; Liu, X.; Lin, B.C. Microfluidic devices for the analysis of apoptosis.


73. Vilkner, T.; Janasek, D.; Manz, A. Micro total analysis systems. Recent developments.

Anal. Chem. 2004, 76, 3373–3385.

74. Kartalov, E.P.; Anderson,W.F.; Scherer, A. The analytical approach to PDMSmicrofluidic

technology and its biological applications. J. Nanosci. Nanotechnol. 2006, 6, 2265–2277.

75. Sia, S.K.; Whitesides, G.M. Microfluidic devices fabricated in poly(dimethylsiloxane) for

biological studies. Electrophoresis 2003, 24, 3563–3576.

76. Park, T.H.; Shuler, M.L. Integration of cell culture and microfabrication technology.

Biotechnol. Prog. 2003, 19, 243–253.


77. Andersson, H.; van den Berg, A. Microtechnologies and nanotechnologies for single-cell

analysis. Curr. Opin. Biotechnol. 2004, 15, 44–49.

78. Andersson, H.; van den Berg, A. Microfluidic devices for cellomics: a review. Sens.

Actuators B Chem. 2003, 92, 315–325.

79. Andersson, H.; van denBerg,A.Microfabrication andmicrofluidics for tissue engineering:

state of the art and future opportunities. Lab Chip 2004, 4, 98–103.

80. Pearce, T.M.; Williams, J.C. Microtechnology: meet neurobiology. Lab Chip 2007, 7,

30–40.

81. Gross, P.G.; Weiner, L.P.; Kartalov, E.P.; Scherer, A. Microfluidic techniques for studying

the nervous system. Crit. Rev. Neurobiol. 2005, 17, 119–144.

82. Gross, P.G.; Kartalov, E.P.; Scherer, A.; Weiner, L.P. Applications of microfluidics for

neuronal studies. J. Neurol. Sci. 2007, 252, 135–143.

83. Meyvantsson, I.; Beebe, D.J. Cell Culture Models in Microfluidic Systems. Annu. Rev.

Anal. Chem. 2008, 1, 423–449.

84. Paguirigan, A.L.; Beebe, D.J. Microfluidics meet cell biology: bridging the gap by

validation and application of microscale techniques for cell biological assays.

Bioessays 2008, 30, 811–821.

85. Bao, N.; Wang, J.; Lu, C. Recent advances in electric analysis of cells in microfluidic

systems. Anal. Bioanal. Chem. 2008, 391, 933–942.

86. Becker, H.; Gartner, C. Polymer microfabrication methods for microfluidic analytical

applications. Electrophoresis 2000, 21, 12–26.

87. De Silva,M.N.; Desai, R.; Odde,D.J.Micro-patterning of animal cells on PDMS substrates

in the presence of serum without use of adhesion inhibitors. Biomed. Microdevices 2004,

6, 219–222.

88. Peterson, S.L.; McDonald, A.; Gourley, P.L.; Sasaki, D.Y. Poly(dimethylsiloxane) thin

films as biocompatible coatings for microfluidic devices: cell culture and flow studies with

glial cells. J. Biomed. Mater. Res. A 72A, 2005, 10–18.

89. Mourzina, Y.; Kalyagin, D.; Steffen, A.; Offenhausser, A. Electrophoretic separations of

neuromediators on microfluidic devices. Talanta 2006, 70, 489–498.

90. Linder, V.; Wu, H.K.; Jiang, X.Y.; Whitesides, G.M. Rapid prototyping of 2D structures

with feature sizes larger than 8 mu m. Anal. Chem. 2003, 75, 2522–2527.

91. Tourovskaia, A.; Figueroa-Masot, X.; Folch, A. Differentiation-on-a-chip: a microfluidic

platform for long-term cell culture studies. Lab Chip 2005, 5, 14–19.

92. Stroock, A.D.; Dertinger, S.K.; Whitesides, G.M.; Ajdari, A. Patterning flows using

grooved surfaces. Anal. Chem. 2002, 74, 5306–5312.

93. Stroock, A.D.; Dertinger, S.K.W.; Ajdari, A.; Mezic, I.; Stone, H.A.; Whitesides, G.M.

Chaotic mixer for microchannels. Science 2002, 295, 647–651.

94. Martines, E.;McGhee, K.;Wilkinson, C.; Curtis, A. A parallel-plate flow chamber to study

initial cell adhesion on a nanofeatured surface. IEEE Trans. Nanobioscience 2004, 3,

90–95.

95. Zhu,X.; YiChu, L.; Chueh, B.H.; Shen,M.;Hazarika, B.; Phadke,N.; Takayama, S.Arrays

of horizontally-oriented mini-reservoirs generate steady microfluidic flows for continuous

perfusion cell culture and gradient generation. Analyst 2004, 129, 1026–1031.

96. Erickson, D.; Liu, X.Z.; Venditti, R.; Li, D.Q.; Krull, U.J. Electrokinetically based

approach for single-nucleotide polymorphism discrimination using a microfluidic device.

Anal. Chem. 2005, 77, 4000–4007.

REFERENCES 41

97. Liu, J.; Enzelberger,M.; Quake, S. A nanoliter rotary device for polymerase chain reaction.


98. Prokop, A.; Prokop, Z.; Schaffer, D.; Kozlov, E.; Wikswo, J.; Cliffel, D.; Baudenbacher, F.

NanoLiterBioReactor: Long-term mammalian cell culture at nanofabricated scale.

Biomed. Microdevices 2004, 6, 325–339.

99. Yang, M.; Li, C.W.; Yang, J. Cell docking and on-chip monitoring of cellular reactions

with a controlled concentration gradient on a microfluidic device. Anal. Chem. 2002, 74,

3991–4001.

100. Boettcher, M.; Jaeger, M.; Kischbaum, M.; Mueller, T.; Schnelle, T.; Duschl, C.

Gravitation-driven stress-reduced cell handling. Anal. Bioanal. Chem. 2008, 390,

857–863.

101. Lovchik, R.; von Arx, C.; Viviani, A.; Delamarche, E. Cellular microarrays for use with

capillary-driven microfluidics. Anal. Bioanal. Chem. 2008, 390, 801–808.

102. Chang, J.C.; Brewer, G.J.; Wheeler, B.C. A modified microstamping technique enhances

polylysine transfer and neuronal cell patterning. Biomaterials 2003, 24, 2863–2870.

103. Cornish, T.; Branch, D.W.; Wheeler, B.C.; Campanelli, J.T. Microcontact printing: a

versatile technique for the study of synaptogenic molecules. Mol. Cell. Neurosci. 2002,

20, 140–153.

104. Tourovskaia, A.; Barber, T.; Wickes, B.T.; Hirdes, D.; Grin, B.; Castner, D.G.; Healy,

K.E.; Folch, A.Micropatterns of chemisorbed cell adhesion-repellent films using oxygen

plasma etching and elastomeric masks. Langmuir 2003, 19, 4754–4764.

105. Nam,Y.; Chang, J.C.;Wheeler, B.C.; Brewer, G.J.Gold-coatedmicroelectrode arraywith

thiol linked self-assembled monolayers for engineering neuronal cultures. IEEE Trans.

Biomed. Eng. 2004, 51, 158–165.

106. Yu, H.M.; Meyvantsson, I.; Shkel, I.A.; Beebe, D.J. Diffusion dependent cell behavior in

microenvironments. Lab Chip 2005, 5, 1089–1095.

107. Gu, W.; Zhu, X.Y.; Futai, N.; Cho, B.S.; Takayama, S. Computerized microfluidic cell

culture using elastomeric channels and Braille displays. Proc. Natl. Acad. Sci. USA 2004,

101, 15861–15866.

108. Heo, Y.S.; Cabrera, L.M.; Song, J.W.; Futai, N.; Smith, G.D.; Takayam, S.

Characterization and resolution of evaporation-mediated osmolality shifts that constrain

microfluidic cell culture in poly(dimethylsiloxane) devices. Anal. Chem. 2007, 79,

1126–1134.

109. Toepke, M.W.; Beebe, D.J. PDMS absorption of small molecules and consequences in

microfluidic applications. Lab Chip 2006, 6, 1484–1486.

110. Millet, L.J.; Stewart, M.E.; Sweedler, J.V.; Nuzzo, R.G.; Gillette, M.U. Microfluidic

devices for culturing primary mammalian neurons at low densities. Lab Chip 2007, 7,

987–994.

111. Jeon, N.L.; Dertinger, S.K.W.; Chiu, D.T.; Choi, I.S.; Stroock, A.D.; Whitesides, G.M.

Generation of solution and surface gradients usingmicrofluidic systems. Langmuir 2000,

16, 8311–8316.

112. Dertinger, S.K.W.; Chiu, D.T.; Jeon, N.L.; Whitesides, G.M. Generation of gradients

having complex shapes using microfluidic networks. Anal. Chem. 2001, 73, 1240–1246.

113. Lin, F.; Nguyen, C.M.C.; Wang, S.J.; Saadi, W.; Gross, S.P.; Jeon, N.L. Effective

neutrophil chemotaxis is strongly influenced by mean IL-8 concentration. Biochem.

Biophys. Res. Commun. 2004, 319, 576–581.


114. Wang, S.J.; Saadi, W.; Lin, F.; Nguyen, C.M.C.; Jeon, N.L. Differential effects of EGF

gradient profiles on MDA-MB-231 breast cancer cell chemotaxis. Exp. Cell Res. 2004,

300, 180–189.

115. Chung, B.G.; Flanagan, L.A.; Rhee, S.W.; Schwartz, P.H.; Lee, A.P.; Monuki, E.S.; Jeon,

N.L. Human neural stem cell growth and differentiation in a gradient-generating

microfluidic device. Lab Chip 2005, 5, 401–406.

116. Hung, P.J.; Lee, P.J.; Sabounchi, P.; Lin, R.; Lee, L.P. Continuous perfusion microfluidic

cell culture array for high-throughput cell-based assays. Biotechnol. Bioeng. 2005,

89, 1–8.

117. Pantoja,R.;Nagarah, J.M.; Starace, D.M.;Melosh,N.A.; Blunck,R.; Bezanilla, F.;Heath,

J.R. Silicon chip-based patch-clamp electrodes integrated with PDMS microfluidics.

Biosens. Bioelectron. 2004, 20, 509–517.

118. Sigworth, F.J.; Klemic, K.G. Microchip technology in ion-channel research. IEEE Trans.

Nanobioscience 2005, 4, 121–127.

119. Klemic, K.G.; Klemic, J.F.; Reed, M.A.; Sigworth, F.J. Micromolded PDMS planar

electrode allows patch clamp electrical recordings from cells.Biosens. Bioelectron. 2002,

17, 597–604.

120. Klemic, K.G.; Klemic, J.F.; Sigworth, F.J. An air-molding technique for fabricating

PDMS planar patch-clamp electrodes. Pflugers Arch. 2005, 449, 564–572.

121. Ionescu-Zanetti, C.; Shaw, R.M.; Seo, J.G.; Jan, Y.N.; Jan, L.Y.; Lee, L.P. Mammalian

electrophysiology on a microfluidic platform. Proc. Natl. Acad. Sci. USA 2005, 102,

9112–9117.

122. Lau, A.Y.; Hung, P.J.; Wu, A.R.; Lee, A.P. Open-access microfluidic patch-clamp array

with raised lateral cell trapping sites. Lab Chip 2006, 6, 1510–1515.

123. Chen, C.; Folch, A. A high-performance elastomeric patch clamp chip. Lab Chip 2006, 6,

1338–1345.

124. Farinas, J.; Chow, A.W.; Wada, H.G. A microfluidic device for measuring cellular

membrane potential. Anal. Biochem. 2001, 295, 138–142.

125. Mela, P.; Onclin, S.; Goedbloed, M.H.; Levi, S.; Garcia-Parajo, M.F.; van Hulst, N.F.;

Ravoo, B.J.; Reinhoudt, D.N.; van den Berg, A. Monolayer-functionalized microfluidics

devices for optical sensing of acidity. Lab Chip 2005, 5, 163–170.

126. Hafner, F. Cytosensor� microphysiometer: technology and recent applications. Biosens.

Bioelectron. 2000, 15, 149–158.

127. Brischwein, M.; Motrescu, E.R.; Cabala, E.; Otto, A.M.; Grothe, H.; Wolf, B.

Functional cellular assays with multiparametric silicon sensor chips. Lab Chip 2003,

3, 234–240.

128. Zhang, X.; Yin, H.; Cooper, J.M.; Haswell, S.J. Characterization of cellular chemical

dynamics using combined microfluidic and Raman techniques. Anal. Bioanal. Chem.

2008, 390, 833–840.

129. Tamaki, E.; Sato, K.; Tokeshi, M.; Aihara, M.; Kitamori, T. Single-cell analysis by a

scanning thermal lens microscope with a microchip: direct monitoring of cytochrome c

distribution during apoptosis process. Anal. Chem. 2002, 74, 1560–1564.

130. Huang, Y.; Rubinsky, B. Flow-through micro-electroporation chip for high efficiency

single-cell genetic manipulation. Sens. Actuators A Phys. 2003, 104, 205–212.

131. Khine, M.; Lau, A.; Ionescu-Zanetti, C.; Seo, J.; Lee, L.P. A single cell electroporation

chip. Lab Chip 2005, 5, 38–43.

REFERENCES 43

132. Hong, J.W.; Studer, V.; Hang, G.; Anderson, W.F.; Quake, S.R. A nanoliter-scale nucleic

acid processor with parallel architecture. Nat. Biotechnol. 2004, 22, 435–439.

133. Marcus, J.S.; Anderson, W.F.; Quake, S. Parallel picoliter RT-PCR assays using micro-

fluidics. Anal. Chem. 2006, 78, 956–958.

134. Marcus, J.S.; Anderson, W.F.; Quake, S.R. Microfluidic single cell mRNA isolation and

analysis. Anal. Chem. 2006, 78, 3084–3089.

135. Hong, J.W.; Chen, Y.; Anderson, W.F.; Quake, S.R. Molecular biology on a microfluidic

chip. J. Phys. Condens. Matter 2006, 18, 1–11.

136. Camelliti, P.; McCulloch, A.D.; Kohl, P. Microstructured cocultures of cardiac myocytes

and fibroblasts: a two-dimensional in vitromodel of cardiac tissue.Microsc. Microanal.

2005, 11, 249–259.

137. Dusseiller, M.R.; Schlaepfer, D.; Koch, M.; Kroschewski, R.; Tester, M. An inverted

microcontact printingmethod on topographically structured polystyrene chips for arrayed

micro-3-D culturing of single cells. Biomaterials 2005, 26, 5917–5925.

138. Schwarz, U.S.; Bischofs, I.B. Physical determinants of cell organization in soft media.

Med. Eng. Phys. 2005, 27, 763–772.

139. Discher, D.E.; Janmey, P.;Wang, Y.-l. Tissue cells feel and respond to the stiffness of their

substrate. Science 2005, 310, 1139–1143.

140. Turner, A.M.P.; Dowell, N.; Turner, S.W.P.; Kam, L.; Isaacson, M.; Turner, J.N.;

Craighead, H.G.; Shain, W. Attachment of astroglial cells to microfabricated pillar

arrays of different geometries. J. Biomed. Mater. Res. 2000, 51, 430–441.

141. Massia, S.P.; Holecko, M.M.; Ehteshami, G.R. In vitro assessment of bioactive coatings

for neural implant applications. J. Biomed. Mater. Res. 68A, 2004, 177–186.

142. Recknor, J.B.; Recknor, J.C.; Sakaguchi, D.S.; Mallapragada, S.K. Oriented astroglial

cell growth onmicropatterned polystyrene substrates.Biomaterials 2004,25, 2753–2767.

143. Cullen, D.K.; Vukasinovic, J.; Glezer, A.; Laplaca, M.C. Microfluidic engineered high

cell density three-dimensional neural cultures. J. Neural Eng. 2007, 4, 159–172.

144. Rowe, L.; Almasri, M.; Lee, K.; Fogleman, N.; Brewer, G.J.; Nam, Y.; Wheeler, B.C.;

Vukasinovic, J.; Glezer, A.; Frazier, A.B. Active 3-D microscaffold system with fluid

perfusion for culturing in vitro neuronal networks. Lab Chip 2007, 7, 475–482.

145. Wu, Z.; Hjort, K.; Wicher, G.; Svenningsen, A.F. Microfluidic high viability neural cell

separation using viscoelastically tuned hydrodynamic spreading. Biomed. Microdevices

2008, 10, 631–638.

146. Flanagan, L.A.; Lu, J.; Wang, L.; Marchenko, S.A.; Jeon, N.L.; Lee, A.P.; Monuki, E.S.

Unique dielectric properties distinguish stem cells and their differentiated progeny.

Stem Cells 2008, 26, 656–665.

147. Tourovskaia, A.; Li, N.; Folch, A. Localized acetylcholine receptor clustering dynamics

in response to microfluidic focal stimulation with agrin. Biophys. J. 2008, 95,

3009–3016.

148. Huang, W.-H.; Cheng, W.; Zhang, Z.; Pang, D.-W.; Wang, Z.-L.; Cheng, J.-K.; Cui, D.-F.

Transport, location and quantal release monitoring of single cells on a microfluidic

device. Anal. Chem. 2004, 76, 483–488.

149. Chen, P.; Xu, B.; Tokranova, N.; Feng, X.; Castracane, J.; Gillis, K.D. Amperometric

detection of quantal catecholamine secretion from individual cells on micromachined

silicon chips. Anal. Chem. 2003, 75, 518–524.


150. Li, M.W.; Spence, D.M.; Martin, R.S. A microchip-based system for immobilizing PC12

cells and amperometrically detecting catecholamines released after stimulation with

calcium. Electroanalysis 2005, 17, 1171–1180.

151. Sun, X.; Gillis, K.D. On-chip amperometric measurement of quantal catecholamine

release using transparent indium tin oxide electrodes. Anal. Chem. 2006, 78,

2521–2525.

152. Jo, K.; Heien, M.L.; Thompson, L.B.; Zhong, M.; Nuzzo, R.G.; Sweedler, J.V. Mass

spectrometric imaging of peptide release from neuronal cells withinmicrofluidic devices.

Lab Chip 2007, 7, 1454–1460.

153. Mahoney, M.J.; Chen, R.R.; Tan, J.; Saltzman, W.M. The influence of microchannels on

neurite growth and architecture. Biomaterials 2005, 26, 771–778.

154. Romanova, E.V.; Fosser, K.A.; Rubakhin, S.S.; Nuzzo, R.G.; Sweedler, J.V. Engineering

the morphology and electrophysiological parameters of cultured neurons by microfluidic

surface patterning. FASEB J. 2004, 18, 1267–1269.

155. Cohen, J.; Burne, J.F.; McKinlay, C.; Winter, J. The role of laminin and the laminin/

fibronectin receptor complex in the outgrowth of retinal ganglion cell axons. Dev. Biol.

1987, 122, 407–418.

156. Lang, S.; von Philipsborn, A.C.; Bernard, A.; Bonhoeffer, F.; Bastmeyer,M. Growth cone

response to ephrin gradients produced by microfluidic networks. Anal. Bioanal. Chem.

2008, 390, 809–816.

157. vonPhilipsborn,A.C.; Lang, S.; Jiang, Z.; Bonhoeffer, F.; Bastmeyer,M. Substrate-bound

protein gradients for cell culture fabricated by microfluidic networks and microcontact

printing. Sci. STKE 2007, 414, 1–11.

158. Li, G.N.; Liu, J.; Hoffman-Kim, D. Multi-molecular gradients of permissive and

inhibitory cues direct neurite outgrowth. Ann. Biomed. Eng. 2008, 36, 889–904.

159. Recknor, J.B.; Sakaguchi, D.S.; Mallapragada, S.K. Directed growth and selective

differentiation of neural progenitor cells on micropatterned polymer substrates.

Biomaterials 2006, 27, 4098–4108.

160. Wittig, J.H.; Ryan, A.F.; Asbeck, P.M.A reusablemicrofluidic platewith alternate-choice

architecture for assessing growth preference in tissue culture. J. Neurosci. Methods 2005,

144, 79–89.

161. Nakashima, Y.; Yasuda, T. Cell differentiation guidance using chemical stimulation

controlled by a microfluidic device. Sens. Actuators A Phys. 2007, 139, 252–258.

162. Houchin-Ray, T.; Whittlesey, K.J.; Shea, L.D. Spatially patterned gene delivery for

localized neuron survival and neurite extension. Mol. Ther. 2007, 15, 705–712.

163. Bergen, J.M.; Pun, S.H. Analysis of the intracellular barriers encountered by nonviral

gene carriers in amodel of spatially controlled delivery to neurons. J.GeneMed. 2007, 10,

187–197.

164. Taylor, A.M.; Rhee, S.W.; Tu, C.H.; Cribbs, D.H.; Cotman, C.W.; Jeon, N.L.Microfluidic

multicompartment device for neuroscience research. Langmuir 2003, 19, 1551–1556.

165. Taylor, A.M.; Blurton-Jones, M.; Rhee, S.W.; Cribbs, D.H.; Cotman, C.W.; Jeon, N.L.

A microfluidic culture platform for CNS axonal injury, regeneration and transport. Nat.

Methods 2005, 2, 599–605.

166. Rhee, S.W.; Taylor, A.M.; Tu, C.H.; Cribbs, D.H.; Cotman, C.W.; Jeon, N.L. Patterned

cell culture inside microfluidic devices. Lab Chip 2005, 5, 102–107.

REFERENCES 45

167. Vahidi, B.; Park, J.W.; Kim, H.J.; Jeon, N.L.Microfluidic-based strip assay for testing the

effects of various surface-bound inhibitors in spinal cord injury. J. Neurosci. Methods

2008, 170, 188–196.

168. Park, J.W.; Vahidi, B.; Joon, K.H.; Rhee, S.W.; Jeon, N.L. Quantitative analysis of CNS

axon regeneration using a microfluidic neuron culture device. BioChip J. 2008, 2, 44–51.

169. Ravula, S.K.; McClain, M.A.; Wang, M.S.; Glass, J.D.; Frazier, A.B. A multielectrode

microcompartment culture platform for studying signal transduction in the nervous

system. Lab Chip 2006, 6, 1530–1536.

170. Ravula, S.K.; Wang, M.S.; McClain, M.A.; Asress, S.A.; Frazier, B.; Glass, J.D.

Spatiotemporal localization of injury potentials in DRG neurons during vincristine-

induced axonal degeneration. Neurosci. Lett. 2007, 415, 34–39.

171. Minteer, S.D., Ed. Microfluidic Techniques: Reviews and Protocol, Humana Press:

Totowa, NJ, 2006.


2NANOPOROUS MEMBRANE-BASEDMICROFLUIDIC BIOSENSORS

SHALINI PRASAD AND VINDHYA KUNDURU

Department of Electrical Engineering, Arizona State University, Tempe, AZ , USA

YAMINI YADAV AND MANISH BOTHARA

Department of Electrical and Computer Engineering, Portland State University,

Portland, OR, USA

SRIRAM MUTHUKUMAR

Intel Corporation, Chandler, AZ, USA

2.1 INTRODUCTION

There has been a tremendous interest in the past decade in the use of nanoporous

membranes to develop sensors. These membrane-based technologies offer multiple

advantages in sensing: due to the nanoporosity of membranes that enables size-based

trapping of biomolecules, they can be manufactured using standard processes with

tremendous quality control in membrane specifications, they are biocompatible, and

they can be easily surface functionalized to ensure specificity in immobilization. This

chapter reviews the advances and current state of the art in nanomembrane technology

in the context of biosensing; key membrane manufacturing technologies have been

reviewed and the methodologies of incorporating them into microelectronic/micro-

fabricated platforms for developing biosensor devices have been reviewed.

This chapter explains in detail the need for real-time measurement of biomolecule

binding in biosensors and justification for incorporating nanoporous membranes into

“lab-on-a-chip” biosensing devices. Prior to understanding the complex procedure of


47

construction of a membrane-based biosensor, the basic concept of a biosensor and

different techniques to embed the nanoporous membrane have been reviewed. The

techniques of biomolecule entrapment including encapsulation, covalent bonding,

and adsorption that play an important role in determining the functionality and

performance of a nanoporous membrane-based biosensor have been looked at. We

have evaluated a number of applications for the nanoporous membrane-based

biosensors in healthcare and environmental monitoring. Awide range of nanoporous

membranes made from organic, inorganic, and hybrid nanomaterials have been

examined. Methods such as anodization, ion track etching, phase separation, sol–gel

rapid thermal annealing, focused ion beam, and lithography used for synthesis of

nanoporousmembrane have been described in this chapter.We have reviewed a range

of electrochemical detection methodologies that have been most commonly adopted

to perform detection with these membranes. The incorporation or integration of these

nanoporous membranes into microscale platforms has been reviewed. We have also

compared themultiple detectionmechanisms based on constraints such as sensitivity,

selectivity, time, and cost. Nanoporous membrane-based biosensors for detection of

microorganism such as bacteria, pathogens, and viruses and analysis of concentration

for body fluids such as glucose and cholesterol have been discussed. Finally, prospects

of the development of natural and biomimetic nanomembranes have been discussed.

The chapter has been organized to demonstrate the need for integration of

nanoporous membranes into microfluidics with specific applications in biosensing.

Oneof the emergingneeds in thedomainofbiosensing is the real-timemeasurement or

the kinetics of detection. This has a far-reaching impact on diagnostics in the areas of

healthcare and the environment. The next section articulates the motivation for

developing nanoporous microfluidic biosensors. The next section addresses the types

of biosensors and how label-free biosensors can be developed using nanoporous

membranes in conjunctionwithmicrofluidics. The next section focuses on addressing

the need for nanoporous membranes for detection of both small molecules and

molecular entities. The next section identifies the types of nanoporous membranes

that have been used for biosensing and analyzes the merits and disadvantages of the

multiple material systems. The next section focuses on methods of fabrication and

processing of the nanoporous membranes and the methods of integrating them into

platforms relevant to biosensing applications. The last section focuses on a number of

healthcare and environmental applications that employ nanoporous microfluidic

biosensors. The most distinguishing feature of this chapter is that it evaluates a

newclass ofminiaturized lab-on-a-chip platforms—the nanoporousmembrane-based

microfluidic biosensors. Contemporary leading reviews focus either on microfluidics

oron application for biosensors.This chapter is thefirst attempt to lookat anewclass of

biosensors that have been quietly emerging in the domain of lab-on-a-chip devices.

2.2 NEED FOR REAL-TIME MEASUREMENTS

Monitoring human health for early detection of disease conditions or health disorders

is vital for maintaining a healthy life. Many tissues, microorganisms, organelles, cell

receptors, enzymes, antibodies, nucleic acids, biomolecules, and so on help determine

48 NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

the physiological state of a disease condition. In addition, analysis of food and

environment for perturbants such as pesticides and river water contaminants with

harmful biomolecules has also become invaluable for health diagnosis. Thus, there is

an ongoing need for rapid analysis, active continuous-time monitoring systems with

substantial accuracy for detecting biomolecules. A “real-time” biosensor that detects

the analytes of interest in a near-continuous-time manner plays an important role in

effective data generation and data processing, supporting real-time decision making,

and rapid manipulation. In order to meet these multiple requirements from multiple

environments, one of the standard approaches that have been adopted is developing

hybrid biochemical analysis systems. These multiscale biosensors are versatile

because they can monitor specific analytes from a wide range of environments at

ultralow concentrations. They comprise a combination of nanomaterials integrated

with microfluidic capabilities. This approach is similar to that followed by the

semiconductor industry in integrated circuits. Microfluidic research involves the

study of several fluid manipulation, detection, and separation techniques. Often,

these different components are integrated into essential electronics to develop a

complete “on-chip” analysis system. Well-established fabrication techniques are

adapted from the semiconductor industry such as micromachining, injection and

replica molding, soft lithography, wet etching, and photolithography. These techni-

ques enable miniaturization of fluid handling systems to palm-held “micrototal

analysis systems” or “lab-on-a-chip” devices that can perform a myriad of diagnostic

and analysis tasks associated with a standard clinical laboratory assay. Hence, a lot

of the present research in this area focuses on the integration of these complex

requirements of real-time measurements with on-chip detection capabilities to build

multifunctional biosensors.

Oneof the innovativeapproachesadopted toaddress these requirementshasbeen the

use of nanoporous membranes embedded into microfabricated structures to generate

multiscale platforms. These platforms are integrated into the appropriate microfluidics

to achieve size-based trapping of analytes of interest, which are then immobilized and

interrogated using a number of label-free methods that have been discussed in the

following sections of the chapter. Hence, nanoporous membrane embedded with a

microfluidic device provides a powerful tool for real-time measurement of “lab-on-a-

chip” biosensors. The following section explains the basic concepts of biosensors.

2.3 BASIC CONCEPTS OF BIOSENSORS

A label-free biosensor is a means of detecting biological agents such as antibodies,

nucleic acids, tissues, cells,microbes, andmetabolites. Theworkingprinciple consists

of binding bioanalytes of interest to bioreceptors, which in turn modulate the

physiochemical signal associated with binding. Later, the electrochemical or optical

transducer captures and converts the physiochemical signal into an electrical signal.

The variation in signal such as electric potential, current, conductance, impedance,

intensity and phase of the electromagnetic radiation, mass, temperature, and viscosity

is monitored. The analysis of variation in one or more of these parameters quantifies

the presence or absence of bioagents. This quantification is achieved without using

BASIC CONCEPTS OF BIOSENSORS 49

fluorescent tagsor labels.Figure2.1demonstratesaschematicdiagramofbiosensors. In

nanoporous membrane-based biosensors shown in Figure 2.2, the porous nanomem-

brane acts as an intermediate layer between biological agents and the physicochemical

detector component, or biological agents and transducer are combinedwith a nanopor-

ousmembrane toconstruct abiosensor. In theareaofnanomaterial-basedbiosensor,one

of thegenresofnanomaterials that are incorporated in thesensors is thenanomembrane.

The huge interest in nanomembranes is driven by their many desirable properties,

particularly the ability to tailor the size and structure and thereby optimize signal

transduction properties for sensing systems. These nanomembrane-based electrical

biosensors produce signal amplification, leading to lower limit of detection.

Applications in both clinical and nonclinical environments are discussed below.

2.4 APPLICATIONS OF NANOPOROUS MEMBRANE-BASED

MICROFLUIDIC BIOSENSORS

Most common applications of nanoporous membranes are self-regulated drug deliv-

ery, biomolecular separation devices, and biosensors. In biosensors, often the

t

fTransductionTransducer

Electrical signal

Physiochemical signal

Biomolecular complex

Recognition Bioreceptors

Bioanalytes

Signal processing

EnzymesAntibody

Nucleic acidTissue

MicrobialPolysaccharide

Electric potentialElectric curren

Electric conductanceElectric impedance

Intensity and phase oEM radiation

MassTemperature

Viscosity

FIGURE 2.1 Schematic prototype of a standard biosensor.

FIGURE 2.2 Integration of bioelements and transducer to construct a nanoporous mem-

brane-based biosensor.3


nanoporous membrane is incorporated for biomolecular separation. These nanopor-

ous membrane-based microfluidic biosensors are significantly smaller in size than

conventional fluid manipulation systems, rendering them portable and extremely

useful in the areas of nanobiotechnology, bioanalysis, pharmaceuticals, medicine, and

diagnostics. Advances in nanotechnology have impacted research in biotechnology

with the development of “smart devices” capable of molecular manipulation.

Microfabricated bioanalytical devices offer highly efficient platforms for genomic,

proteomic, andmetabolic studies. The nanoporous membrane plays an important role

in fabrication of biocompatible and cost-effective lab-on-a chip devices.

Moreover, analogous to the miniaturization of computer chips, bioanalytical

sensors are also undergoing reduction in their size. This offers the benefit of

shorter reaction speeds with faster analysis times. Multiple active sites on the

microfluidic chip offer parallel operation modes, thereby resulting in multiplexed

analysis and higher throughput. These membrane-based biosensors can be cost-

effectively produced on a large scale with a promise of higher analysis rates and

better efficiency, owing to the compactness and better process control; lower

analyte consumption lowers the cost incurred on expensive reagents and is environ-

ment friendly during disposal. Figure 2.3 illustrates the flowchart for different

applications of biosensors. The important part of membrane-based biosensor is their

thin porous membrane. The need for this nanoporous membrane is explained in

following sections.

2.4.1 Need for Nanoporous Membrane

Overall sensitivity of a sensor depends on signal transduction andmass transport effect.

Miniaturization of a sensor increases signal-to-noise (S/N) ratio, an inherent advantage

for signal transduction. It has been reported that mass transport of analytical solution

through the sensor surfaceplaysan important role indeterminingsensitivity.4Detection

limit for bioassays depends upon the amount of biomolecule interactionwith the sensor

surface. Whitman’s group at Naval Research Laboratory, Washington, DC, reported

FIGURE 2.3 Flowchart representation of a biosensor.3

APPLICATIONS OF NANOPOROUS MEMBRANE-BASEDMICROFLUIDIC BIOSENSORS 51

that femtomolar detection limits for bioassays are likely.Theypredicted that the limit of

detection will be due to analyte transport limitation, not due to signal transduction

limitation, and without directed transport of biomolecules, individual nanoscale

sensors will be limited to picomolar-order sensitivity for practical timescales.4

Total flux to the sensor was studied as a function of sensor geometry and volumetric

flow. Enhancing mass transport by conventional methods of decreasing height de-

creased volumetric flow rate, which in turn decreased total flux of the sensor. On the

other hand, directly injecting the analyte into the sensor rather thanmerely streaming it

past increasedmass transport effects, which in turn increased total flux of the sensor.4 It

was found that the sensor flux could be increased by using a nanoporous membrane.

Thus, the sensitivity of the sensor can be enhanced by incorporating nanoporous

membrane into the microfluidic biosensors.

2.4.2 Efficient Size Sorting

Membrane-based technology has been identified as a usefulmethod for the separation

of biomaterials including viruses, owing to its efficiency, ease of implementation, and

cost effectiveness. Due to nanometer-sized domain, and relatively thin membrane

thickness and narrow size distribution, the nanoporous membrane shows high flux,

mechanical strength, and high selectivity for biomaterial separation. Comparative

studies were performed between track-etched polycarbonate (PC) and anodized

aluminum oxide (AAO) membranes with a uniform pore size for virus separation;

it was found that the nanoporous AAO membrane yields an excellent selectivity and

high solution flux.

2.4.3 Design Considerations for Molecular Sorting

One of the important aspects of the porous membrane typically exploited in micro-

fluidic devices is the sorting of biomolecules on the basis of size from different

bioanalytes for downstream analysis of the bioagent on biosensor platform.

Depending on size and diffusion rate, pore size is determined. Adiga et al. explained

that the nanoporous membrane is evaluated in terms of porosity of the membrane that

is defined by solvent flux through unit area of the membrane under a unit pressure

difference.1

Lp ¼r2p

8h

!Ak

Dx

� �

where rp is thepore radius, 8h is the solvent viscosity, and (Ak/Dx) is the ratio of surfaceporosity to the pore length.

Earlier, electrophoresis was used as a mechanism for the separation of biomole-

cules.5 Increasing demand of real-time measurement and miniaturization of biosen-

sors to build a lab-on-a-chip device has led to the integration of the nanoporous

membrane into the microfluidic device. Kuo et al. have built a simple, rapid proto-

typing of porous nanostructures inside the microchannels for the separation of


DNA molecules.6 Periodic porous nanostructures with a cavity size of 300 nm and

interconnecting pore of 30 nm were fabricated inside the microchannels. The fabri-

cation process step using negative photoresist SU-8 is explained in Figure 2.4.

First, silica colloidal is grown inside the SU-8 microchannel, then the space between

colloidal crystals is filled with SU-8, which is then cured using ultraviolet light.

The silica nanoparticles are finally removed using BOE, creating pores in the

SU-8 structure.

2.5 TYPES OF NANOPOROUS MATERIALS

Porous membranes can be broadly classified depending upon their intrinsic materials

such as organic, inorganic, polymer, and composite membranes. Depending upon the

barrier structure, membranes are nonporous or have micropores, mesopores, and

macropores with pore diameters of <2 nm, 2–50 nm, and 50–500 nm, respectively.7

Lehrstuhl f€ur Technische Chemie II explained that porous barriers could be used for

very precise continuous permselective separations based on subtle differences in size,

shape, and/or functional groups.7 Many scientists are developing novel composites

andpolymers to tailorwell-definedporousmembranes in termsof pore size to increase

their functionality and selectivity. Furthermore, depending on the cross section,

membranes are classified as isotropic, anisotropic, bimultilayer, thin layer, or matrix

of mixed composite. Yang et al. synthesized a nanoporous membrane for virus

filtration with good dimensional stability under high pressures maintaining high

selectivity.8 Themembrane consists of a double layer. The upper layer is a nanoporous

film with a pore size of 17 nm and a thickness of 160 nm, which was prepared by

polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) copolymer. The lower

layer consists of conventional microfiltration membrane to enhance mechanical

strength. Phase separation, sol–gel method, interface reaction, stretching extrusion,

track etching, andmicrofabrication technique can also be used for the classification of

porous membranes. Inorganic membranes are made from oxides, ceramics, and

metals, while the most commonly used polymeric materials are Nafion, polycarbon-

ate, polyethylene terephthalate, or polysulfone. In comparison to polymer mem-

branes, inorganicmembranes are versatile, can be used at very high temperatures, and

are much more resistant to chemical attack. Different strategies for fabricating

and integrating the reviewed nanoporous membranes into sensing devices are

explained briefly in the following section.

FIGURE 2.4 Schematic diagram of a nanoporous structure in a microfluidic channel.6

TYPES OF NANOPOROUS MATERIALS 53

2.6 FABRICATION AND INTEGRATION OF NANOPOROUS

MEMBRANES INTO MICROFLUIDIC DEVICE

2.6.1 Anodization

Anodization is an “electrochemical etching” process. Traditionally, nanoporousAAO

membranes are fabricated by two ways: mild anodization (MA) or hard anodization

(HA).TheMAmethod produces self-ordered pore structureswith a limitation in terms

of slow process; the processing time requires several days, the oxide growth rate is

2–6mmh�1, and the process requires a narrow range of processing conditions called

self-ordering regimes. Being a slow process, MA is not suitable for mass production

and industrial processes.Another technique that iswidely found in the literature isHA.

HA is performed at a high voltage by using sulfuric acid at relatively low temperatures

and high current densities resulting in the rapid growth of a thick porous oxide layer of

approximately50–100mmh�1 for various applications.Comparative study shows that

porous oxide films formed by HA are disoriented compared to the pores formed by

MA. Moreover, they are mechanically unstable due to a strong tendency to develop

cracks under the influence of evenweakmechanical forces. These aspects have caused

serious problems in practical application of anodic films, especially in nanotechnolo-

gy research. A limitation with HA process is in controlling important structural

parameters, such as pore size, interpore distance, and the aspect ratio of the nanopores

of the resulting alumina membranes. A comparison between MA and HA based on

porosity, interpore distance, pore diameter, and pore density is provided in Table 2.1.

A schematic diagramprovided in Figure 2.5 illustrates the hexagonal structural layout

of AAO by the anodization method.

Researchers at Max Planck Institute of Microstructure Physics in Halle have

developed a novel approach for structural engineering of nanoporous alumina using a

pulse anodization method with oxalic acid.9 The process sequence of the pulse

anodization method is shown in Figure 2.6. In the pulse anodization method, low

and high potential pulses were applied alternatively to achieve mild and hard

anodization conditions, respectively.Thenewprocess is an effectivewayof improving

mechanical stability with a thickness of >100 mm of hard anodized alumina. In

addition, it also provides a unique opportunity for producing individual alumina

nanotubes with uniform diameter and length by taking advantage of cracking

phenomena in the hard anodization process. The interpore distances of the AAO

are approximately Dint¼ 200–300 nm that can be achieved only by mild anodization

TABLE 2.1 Comparison Between MA and HA in 0.3N H2C2O49

Parameter MA HA

Porosity (P, %) 10 3.3–3.4

Interpore distance (Dint, nm) 100 220–300

Pore diameter (Dp, nm) 40 49–59

Pore density (p; pores cm�2) 1.0� 1010 1.3–1.9� 109


FIGURE 2.5 Schematic diagram demonstrating a porous oxide film produced by the

anodization method.

FIGURE 2.6 Scheme for the fabrication of porous alumina with modulated pore diameters

by a combination of MA and HA of a prepatterned aluminum substrate.9

FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES 55

processes. It offers substantial advantages over conventional anodization processes

in terms of processing time, allowing 2500–3500% faster oxide growthwith improved

ordering of nanopores. Perfectly ordered alumina membranes with high aspect

ratios (>1000) of uniform nanopores with periodically modulated diameters have

been realized.

In addition to sulfuric acid (H2SO4), oxalic acid (H2C2O4) and phosphoric acid are

also used for making AAO. Figure 2.7 illustrates different conditions for the synthesis

of hard and mild anodization nanoporous membranes. Shalini Prasad’s group from

Portland State University demonstrated uniform formation of pores onto an alumina

membrane fabricated using the anodization method.10 A combination of oxalic

(0.1M) and sulfuric acid (0.3M) was used as an oxidizing acid. Alumina membrane

fabricatedhada thicknessof 250 nmwithadistanceof separationbetweeneachporeof

10–15 nm, and the diameter of each pore was 200mm. The volume of each pore was

8� 10�21mL. The SEM characterization of AAO is shown in Figure 2.8.

2.6.2 Ion Track Etching

Track etching technologyconsists of chemical andphysicalmodifications of thinfilms

induced by energetic ion irradiation.11 The ion tracking process consists of irradiation

of polymeric thin foilswith heavy ions and subsequent chemical etching of the particle

tracks. Cylindrical, conical, funnel-like, or cigar-like pores with diameters ranging

from tens of nanometers to the micrometer range can be obtained.12 To obtain

nanopores, the etching is performed in an ion accelerator and controlled bymonitoring

etchable ion tracks of swift heavy ions. The energy of these specific ion tracks should

FIGURE 2.7 Graph illustrating different conditions for synthesis of hard and mild anodiza-

tion nanoporous membranes.9


be greater than 1MeV per nucleon. Each ion passing though the thin polymeric film

produced single ion track, which subsequently represents one pore. By controlling the

number of ion tracks, one can produce 109 pores cm�2 to 1 pore per sample.13 A single

ion is hit into a membrane to produce single pores.

Ametallic platewith a thickness of 0.2mm and an aperture of 0.2mm is inserted in

front of the track-etched templates for support. A semiconductor detector is placed

behind the sample to detect each ion passing through the aperture and then through the

sample. At such low fluxes, the probability of an ion passing through the aperture is 1

event s�1.14 This gives enough time to the automated system to switch off the beam

using a fast chopper after an ion hit is detected by the detector. The next step after

irradiating the polymer is chemical etching of the ion track to produce pores of desired

shape, with size up to 5–100 nm. Length to diameter ratio in the range 10–1000 can be

easily achieved.Chemical composition of the chemical etchants, optimum irradiation,

postirradiation treatments, and etching temperature conditions determine shape, size,

and density of the pores.

There are two parameters that determine the shape and size of the pore:

1. vb: the bulk etch rate (etching rate for nonirradiated material).

2. vt: the track etch rate (the etching rate along the ion track).

Typically, the etching results in a conical or double conical pore (depending on

whetheroneorbothfacesofthefoil, respectively,areexposedtotheetchingbath)withan

FIGURE 2.8 Scanning electron micrographs showing the alumina nanomembrane. (a)

Density of the pores is clearly visible from these micrographs. Panel (b) shows the uniformity

of the pores and the size distribution of each pore. (c and d)The pore goes all theway through the

membrane and creates uniform wells.


opening angle a: tga¼ vb/vt. Alternatively, in the case of high-selectivity etching

conditions, that is, vt� vb, the shape of the pores can be approximated to be cylindrical.

Polycarbonate and polyethylene terephthalate are some of the polymers used for

synthesis of nanoporous membrane. Even though former polymers have lower

chemical, mechanical, and thermal stabilities, both cylindrical and conical pores

can be easily etched in comparison to other polymers. PC also has advantages in terms

of dissolving capability into several organic solvents such as dichloromethane. Using

this method, 100 nm pores were created in a PC membrane. The group of Voss,

Germany, investigated asymmetric nanopores in PET and polyimide (Kapton) mem-

branes.15 It was observed that the transient properties of the pores depend both on the

chemical structure of the polymer and on the irradiation and etching procedures used.

However, thismethod cannot be applied tometal foils, except for some special cases of

amorphous “glassy” metals, because swift heavy ions passing through a metal do not

generate tracks that could be etched to form holes.

2.6.3 Phase Separation

The phase separation method is the most common method for preparation and

production of polymeric porous membranes. There are three types of phase

separation techniques that are typically used to generate a polymeric membrane,

which include the wet phase separation method, the thermally induced phase

separation method, and the nonsolvent-induced phase separation method. In the first

method, a cast thin layer of a polymer solution is immersed in a liquid nonsolvent,

which is miscible with solvent.16 The exchange of the solvent from a thin layer of

polymer solution with a nonsolvent from the coagulation bath produces thermody-

namic instability in ternary systems. The thermodynamic instability is resolved by

separation into polymer-rich and polymer-lean phases.17 The polymer-rich phase

forms a solid membrane matrix, while the polymer-lean phase leaves a porous

structure by leaching out of the system. By varying the polymer concentration, the

thickness of the solution and coagulation medium, and temperature, a wide variety of

asymmetrical porous membranes with a very large variety of properties can be

synthesized. The ternary systems most commonly comprise polymers such as cellu-

lose acetate, polysulfone, poly(methyl methacrylate), polyamide, and polyurethane,

solvents such as acetone, N,N-dimethylformamide, and acetamide, and nonsolvent

such as water.18

Kawakami’s group from the Tokyo Metropolitan University fabricated three-

dimensional fluorinated polyimide microporous membranes of cylindrical structures

by the wet phase inversion process, which is formed by a ternary system of poly-

imide–solvent–water.19 Moreover, microporous polystyrene and polycaprolactone

(PCL) porousmembraneswere also synthesized by the phase separation process. PCL

has several advantages, including low cost, biocompatibility, and biodegradability.

Moreover, PCL is an FDA-approved material for implantable devices. Thus, it is a

superior material to fabricate affordable devices.

In polymeric systems, phase separation can also be induced by solvent evaporation,

temperature, or addition of a nonsolvent. The process where optimization of porous


property is performed by controlling the thermal conditions is called the thermally

induced phase separation (TIPS) process.20 The TIPS technique has been utilized for

making microporous materials such as porous membranes and foams from semicrys-

talline polymers. In the nonsolvent-induced phase separation method, the concentra-

tions of polymer, solvent, and nonsolvent are highly critical for fabrication of the

nanoporous membrane. Kuo’s group from the Center of Membrane Technology at

Taiwan prepared a polyvinylidene fluoride (PVDF) microporous membrane by using

alcohol as nonsolvent (coagulant).21 To prepare the PVDF membrane, the wet phase

inversion process was carried out by using water and n-propanol as nonsolvents.

An increase in the pore size of the PVDFmembrane was observed with an increase in

the immersion time in the n-propanol bath.

Comparative studies showed that TIPS could help a homogeneous solution reach

liquid–solid phase separation more quickly.22 Using this technique, phase-separated

solutions spend less time in the liquid–liquid separation region. Fast solidification can

avoid pore coalescence. Thus, porous membranes with high porosity can be prepared

with TIPS. Moreover, nanoscale pore size can be obtained via NIPS. Therefore,

nanoporous membranes with high porosity can be prepared via the combination of

TIPS and NIPS.

Recently, Kurt et al. developed a polymerization-induced phase separation

(PIPS) technique.23 Porous membranes were developed from a monomer and

solvent mixture using PIPS. This novel membrane has a potential application in

flow-through biosensors based on protein or DNA microarrays. Depending on

the requirements of the application, often high demands are put on specific properties

such as the control over pore size, pore size distribution, morphology, and surface

functionality, which is controlled by adjusting monomer concentration and UV

intensity.

2.6.4 Lithography

At present, novel nanoporous micromachined membranes using lithography techni-

ques are being developed for biosensor applications. Ferrari’s group from the

University of Illinois fabricated porous silicon membranes using simple standard

lithography techniques. These membranes are highly reproducible, extremely stable,

and have the ability to be integrated into the silicon-based platform technology. In

addition, these membranes exhibit selective permeability and low biofouling.

Filtration of biomolecules is desirable for biosensors. Membranes with pore sizes

as small as 20 nm are available. Even so, the filtration at these dimensions is far from

absolute. The use of ion track etching for the synthesis ofmembranes yieldsmillipores

with low porosities (<109 pores cm�2) and limited pore sizes, and the pores are

randomly distributed across the surface. In comparison, the anodization technique

yields alumina pores of higher densities, such as 1010 pores cm�2, but pore sizes

greater than 20 nm, and the pore configurations and arrangements are difficult to

control. The technology of micromachining allows fabrication of membranes with

multiple pore configurations and arrangements. Several microfabrication methods

have been used to create pore sizes in the tens of nanometers on silicon substrates using


photolithography and deposition/selective removal of sacrificial layers.24 Figure 2.9

demonstrates a schematic diagram for themicrofabrication of nanoporousmembranes

with a pore thickness of 24.5 nm.

In the above example of the lithography process, a nitride layer is grown onto a

silicon wafer, which functions as an etch stop layer. On the top of the nitride layer, a

base layer of polysilicon is deposited that acts as a structural support. The holes are

then defined into the base layer by chlorine plasma. Later, thin sacrificial oxide is

grown using thermal oxidation. The thickness of the sacrificial layer determines the

final pore size. These pores are then completely filled with the polysilicon plug layer,

which is planarization through plug layer to base. The final step is to deposit nitride

layer and backside patterning. The protective, sacrificial, and etch stop layers are

removed by etching inHF. Similarly, the process sequence shown abovewas also used

byDesai’s group from theBostonUniversity for fabricationofporousmembraneswith

pore sizes ranging from 5 to 100 nm.25

In anothermethod for fabrication ofmicro- and nanoporousmembranes, pillar-like

templates are used. These pillar templates are fabricated using photolithography and

e-beam lithography.Grant’s group from theUniversity ofTexas atAustin built organic

polystyrene (PS) and inorganic (ZnO) pillar arrays as a template andpolysulfone (PSf)

porous films were fabricated.26 Figure 2.10 illustrates the complete process sequence

of template-based porous membrane fabrication.

FIGURE 2.9 Cross-sectional schematic depiction of the microfabrication process for

nanoporous silicon membrane fabrication.24


2.6.5 Focus Ion Beam Etching

A nanopore is most effective as a single-molecule detector when the diameter of the

pore is close to the diameter of the molecule, typically 2–10 nm, being detected.27

Nanoporous membranes have pore sizes ranging from<1 to 100 nm in diameter. The

focused ion beam (FIB) technique has been conventionally used for preparing

nanopores in thin films.28 This technique offers very good resolution and is able to

directly pattern arrays of well-defined pores. The anodization method has disadvan-

tages in termsof fabrication of pore sizes greater than20 nmand it is nearly impractical

to etch a pore below 30 nm in diameter reproducibly in terms of size and shape using

commercial FIB systems. It is possible to etch nanopores using a high-energy focused

ion beam.However, Losic’s group from theUniversity of SouthAustralia at Adelaide,

SA, overcame the system and pore size limitations and nanofabricated pore size less

than 10 nm onto AAO porous membranes using the FIB technique.29 The FIBmilling

technique can be successfully used for removing the oxide barrier film and controlled

pore opening of AAO to form a single nanopore or nanopore arrays. Figure 2.11

demonstrates the process sequence for fabricating the nanoporous membranes.

Bezryadin’s group from the University of Illinois at Urbana-Champaign has

demonstrated the fabrication of symmetric sub-5 nm nanopores using focused ion

and electron beams.27 FIB scans reduce the size of the pore.During the sequence of ion

scans, the gradual shrinking of the pore size and the change in pore geometry in a

regular fashion are observed. The beam provides some mobility to the material of the

membrane around the pore and allows the pores to shrink in diameter due to surface

tension. This sculpting process also allows an array or a pattern consisting of multiple

nanopores to be fine-tuned at the same time.

FIGURE 2.10 Schematic diagram for fabrication of a porous membrane using the template

method.26


Similar to ion beam milling, electron beams are also used for micromachining

finely precise nanopore arrays in the nitride membrane. Meller from the Boston

University fabricated nanopores in thin Si3N4 films using the intense e-beam of

field emission.30 Si3N4 films were formed using standard lithography techniques and

low-pressure chemical vapor deposition (LPCVD). Different irradiation conditions

enabled nanopore fabrication in the range of 2–20 nm with exceptional size control

and greater than 0.5 nmvariability, thus resulting in an effective nanopore thickness of

17 nm. Similarly, David C. Bell from the Center for Nanoscale Systems, Harvard

University at Cambridge, fabricated small nanopores in SiN on the order of 1 nm

diameter with great accuracy and reproducibility.31,32

2.6.6 Rapid Thermal Annealing

Using the rapid thermal annealing (RTA) technique, Striemer et al. were able to

fabricate ultrathin porous nanocrystalline silicon (pnc-Si)membranes.33 These pnc-Si

porous membranes are approximately 10 nm in thickness. These nanomembranes are

highly fragile and the technique is very complex; rapid thermal annealing is not

commercially used for large-scale production. With RTA, the average pore size

created varies approximately from 5 to 25 nm. The pore size distributions in pnc-

Si membranes can be controlled through adjustment and by varying the temperature.

In the RTA process, during crystallization, voids are often spontaneously formed as

nanocrystals nucleate and it grows into a 15 nm thick amorphous silicon (a-Si)

film. The process sequence for the fabrication of the ultrathin pnc-Si membrane is

shown in Figure 2.12.

FIGURE 2.11 Schematic diagram demonstrating fabrication of a single pore to an array of

nanopores onto the AAO membrane using the FIB technique.29


2.6.7 Sol–Gel Technique

Sol–gel technology offers a cheap and quick alternative for producing bioactive

porous surfaces for variousbiosensors.The sol–gel thinfilm techniqueoffers a number

of advantages including low-temperature processing, ease of fabrication, and precise

microstructural and chemical control.34 Tailoring the thickness of themembrane, pore

size, and density provides an additional advantage to biocompatibility.Highly porous,

supercritically dried sol–gel low-density membranes formed using the sol–gel

technique are called aerogels. These biocompatible membranes are commonly

used in biosensors and electrochemical biosensors. Various types of inorganic

aerogels such as silica, carbon, and alumina are fabricated. Similarly, inorganic

aerogels such as SEAgel can also be synthesized.

The startingmaterials used in the preparation of the sol are usually inorganicmetal

salts or metal organic compounds such as metal alkoxides [M(OR)n], where M

represents a network forming element such as Si, Ti, Zr, Al, B, and so on and R is

typically an alkyl group.35,36 The most commonly used precursors are tetramethyl

orthosilicate (TMOS) and tetraethyl orthosilicate (TEOS) in the sol–gel process.37

The basic sol–gel reaction begins when metal alkoxide is mixed with water and a

mutual alcoholic solvent in the presence of acid/base catalyst. Thin films can be

produced on a piece of substrate by dip, spin, and spray coating. During the sol–gel

transformation, the viscosity of the solution gradually increases as the sol becomes

interconnected to forma rigid, porousnetworkof gel.38During thedryingprocess at an

ambient pressure, the solvent liquid is removed and substantial shrinkage occurs.

Extracting the liquid component of a gel through supercritical drying, a highly porous

FIGURE 2.12 Schematic block diagram representing formation of a nanoporous membrane

using rapid thermal annealing.33


network known as an aerogel is produced. Silica aerogels are biocompatiblematerials

mostly used for bioapplications. Power et al. from the University of Virginia at

Charlottesville prepared sol–gel-based silica macroporous membranes. These porous

membranes were on the order of 10–100mm.39 These silica membranes were used to

trap viral bacteria Escherichia coli (pET-gfp). Similarly, Tiwari et al. prepared a

biopolymer–SiO2nanocomposite aerogel. The resulting composite aerogel consists of

a mesoporous material that was controlled by thermal curing.40 Recently, Li et al.

synthesized aprotein-encapsulatedbioaerogel, inwhich a recombinant redfluorescent

protein, DsRed, was chosen as a model protein.41 It was prepared using sol–gel

polymerization of TEOS with an ionic liquid as the solvent and pore-forming agent.

A bioaerogel formed showed high porosity. Recently, scientists from theUniversity of

Leeds,UK, for thefirst time explored the use of self-assembledpeptide organogels and

hydrogels as starting materials for the creation of new nanostructured aerogels.42 The

novel aerogels impart biological-like functionality for sensing applications. Peptides

such as P7-2, P9-2, P11-2, P24-2, and K2 were studied.

2.7 FUNCTIONALITY OF MEMBRANE IN BIOSENSORS

Membranes with various pore sizes, lengths, morphologies, and densities have been

synthesized from diverse materials for size exclusion-based separation. Specific

bioagent immobilization and detection remains a great technical challenge in

many biosensors. To achieve this, a material with controllable pore diameter, length,

and surface chemistry is needed. Selective capture requires two steps: collection and

immobilization. Membranes are well suited for this because of their enhanced

probability of interaction of the surface with the liquid being analyzed. In order to

create an excellent biosensor, the biological component has to be attached to

transducers. This process is known as immobilization. Membrane entrapment,

physical absorption, matrix entrapment, and covalent bonding are four ways to couple

a biosensing element to the membrane (Table 2.2). The bioagents are bonded to the

sensing element by one of these four ways. Figure 2.13 illustrates schematic block

diagram demonstrating functionality of the membrane in the biosensor.

The adsorption method is the simplest and involves minimal preparation. In this

method, although the bonding between bioagent and sensing transducer is weak, it is

extremelyuseful for exploratory research.Thephenomenonofphysical adsorptionvia

TABLE 2.2 Different Immobilization Techniques

Immobilization technique

Physical Adsorption,

entrapment,

confinement, and

encapsulation

Chemical Covalent binding and

cross-linking


van derWaals bonds is demonstrated by Prasad and coworkers from the Portland State

University in nanomonitor protein biosensors.10 The AAO membrane was used to

adsorb enzymes onto the sensing microelectrode array surfaces. Similarly, an anti-

body-based conductometric biosensor using porous filter membranes, developed by

Muhammad-Tahir and Alocilja, has been shown to detect bacteria and bovine viral

diarrhea virus antigens.43 In addition, Song et al. incorporated absorption techniques

inbiosensors to buildmembrane-based assay devices.44Other forces such as hydrogen

bonds, hydrophobic forces, and ionic forces are used to attach a biomaterial to the

surface of the sensor. At a single instance, multiple forces can also be used in a single

biosensor. However, the adsorption technique has a disadvantage: under mild con-

ditions, through the pores the enzyme drains from the carrier and alters the change in

pH or ionic strength.

Entrapment of the bioagent in the nanomembrane creates a good biorecognition

layer close to the transducer, which is mainly a gold microelectrode. The membrane

separates the analyst and the bioelement. This facilitates a better biosensor. Kueng et

al. immobilized a biorecognition element using an enzyme entrapment technique to

integrate amperometric ATP microbiosensors.45 A dual microdisk electrode configu-

ration was integrated to immobilize the enzymes at one of the microdisk electrodes.

Shan et al. constructed a novel glucose biosensor by electrochemical entrapment of

glucose oxidase (GOD) into porous poly(acrylonitrile-co-acrylic acid), which was

FIGURE 2.13 Schematic block diagram demonstrating physical and chemical functionality

of a nanoporous membrane in biosensors.3

FUNCTIONALITY OF MEMBRANE IN BIOSENSORS 65

synthesized via radical polymerization of acrylonitrile and acrylic acid.46 Similarly,

scientists from theKyushu Institute of Technology in Japan used nanoporous template

electrodes to enable efficient enzyme entrapment by simple physical adsorption.47

Template electrodes made of porous carbon were efficient nanomaterials for entrap-

ment of bioagents. The integrated biosensor using entrapment provides an increased

surface area and sensitivity to sense the biochemical reaction; however, the technique

has also demonstrated problems such as an easy leakage, serious diffusion constraints,

and lower stability48.

In the microencapsulation technique, the porous entrapment scheme forms a

porous encapsulation matrix around the biological analytes that helps binding to

the sensor transducer. The porous membrane facilitates transfer of electrons and ions.

Darder et al. demonstrated encapsulation of enzymes by alumina membranes of

controlled pore sizes.49 These AAO porous membranes created by anodization were

used for immobilization of the biologically active elements. Hexagonally structured

AAOmembranes allowedhigher amounts of glucose oxidase (GOx) enzymeuptake in

a thin film. This idea facilitated the construction of amperometric biosensors with the

nanoporous Al2O3 membrane. Similarly, in choline biosensors, enzyme choline

oxidase (ChOx) was immobilized and encapsulated in a hybrid mesoporous mem-

branewith 12 nmpore diameter.50Moreover,Kimet al. developed a glucose biosensor

based on a sol–gel-derived zirconia/Nafion composite film as an encapsulation

matrix.51 To conclude, in all these biosensors, the nanoporous structure film greatly

enhances the active surface area available for protein immobilization.

In covalent bonding techniques, biological elements are linked to biosensing

membranes by strong covalent bonds. Sensing surfaces are treated with reactive

chemical groups. De Stefano et al. created a covalent bond between the porous silicon

surface and the biomolecules, which specifically recognize the unknown analytes.52

It has been demonstrated that porous silicon-based optical microsensors help detect

L-glutamine from E. coli using the covalent bonding method. However, covalent

bonding and cross-linking produce more stable immobilization, but may degrade the

activity of enzyme by including drastic synthesis environment in the immobilization

process.48

2.8 DETECTION MECHANISM

Diagnostic biosensors are probably the largest area of research in the field of

bionanotechnology. Over the past few years, many new ideas and technologies

have been proposed in the literature. In this section, a comparative study of these

detection mechanisms will be discussed. Each detection mechanism significantly

differs in the design of the device, the fabrication methodology adopted for such a

design, and the area of application. Each mechanism has the same goals in mind that

include reducing the sensing elements to be equivalent to the size of the target species,

improving the sensitivity, reducing the reagent volumes, offsetting the costs of the

reagents, moving toward a real-time system to acquire the results and simultaneously

getting a lower detection limit, miniaturizing the entire system, and improving


portability. Based on the basic principles of detection, biosensor devices can be

broadly classified into two classes, namely, labeled detection and label-free detection.

2.8.1 Membrane-Based Labeled Detection

A label is generally a chemical such as fluorophore, which is tagged along with the

biochemical under observation for detection. In the labeled detection mechanism, as

demonstrated in Figure 2.14, a biomolecule that does not react to incident light, but

the chemical (label) reactive to light, is tagged to the biomolecule of interest.When the

complex matrix of biomolecule and label is irradiated with light, the complex gets

excited to give an output signal at the photodetector. Hence, detection of biomolecule

of interest by indirect detection of the label is called labeled detection. The most

popular labeled detection method is fluorescence detection and ELISA (enzyme-

linked immunosorbent assay).

Immobilized biomolecules are emerging as popular analytical tools given their

reusability and sensitivity. In a nanomembrane-based labeled detection biosensor,

immobilization of the antibody is performed by simply inserting the porous nano-

membrane into the reaction chamber.Membrane-based immunochemicalmethods are

gaining wide acceptance as they offer the advantages of sensitivity, specificity,

rapidity, simplicity, and cost effectiveness, which is important for routine testing.

However, they have significant disadvantages that label-free detection has been

striving to overcome and replace ELISA over the past decade.

2.8.2 Membrane-Based Label-Free Detection

When the biomolecule can be directly detectedwithout the help of external labels, it is

called label-free detection. In the label-free mechanism, detection and quantification

of the properties of interest of the biomolecule are achieved without the help of

external labels. This has a huge advantage as it is possible to measure the direct

interaction of the biomolecule with the substrate rather than its interaction with an

FIGURE 2.14 Schematic diagram demonstrating the labeled detection technique.

DETECTION MECHANISM 67

external label. It also eliminatesmultiple stepsof attaching the label to thebiomolecule

and the control steps to eliminate nonspecific binding of labels. It also eliminates

contamination due to external chemicals. The electrochemical or electrical method of

detection is a good example of label-free detectionwherein the electrical properties of

the biomolecules are observed and correlated for analysis.

Label-free detection has many advantages over labeled detection. It has lower cost

per assay, lower contamination, higher sensitivity, and significantly shorter detection

time. However, one of the important issues faced by label-free detection is the lack of

throughput.Most biomolecules have different yet very similar physical characteristics

when it comes to label-free detection. Because of the nature and size of the

biomolecules, it was difficult to distinguish them using detection systems that

were many orders of magnitude bigger. A large amount of sample volume is also

required to achieve high sensitivity, which also proved impractical, as it is not possible

to have tens of milliliters of blood every time the patient needs to be tested. These

issues were solved with the advent of nanotechnology. With the advent of nanofab-

rication techniques and nanomaterials, device setups around the same size of the

biomolecules could be designed. Much higher surface area of interaction with the

biomolecules is possible,which enables high sensitivity in detection. Sizematching of

the measuring arrangement to the biomolecules significantly reduces the sample

volume required and increases detection limit to very low analyte concentrations.

Hence, the label-free detection technique has become themost active area of research

in the field of proteomics in the past decade because of its promising trend to replace

ELISA as the predominantly used technique for protein detection.

Label-free nanoporous membrane-based biosensors can be classified on the basis

of detection mechanism, namely, electrical detection, optical detection, and mechan-

ical detection. Electrical biosensors can be further classified on the basis of the

electrical measurement, which includes voltammetric, amperometric/coulometric,

potentiometric, and impedance.

2.8.2.1 VoltammetryThe current–potential relationship of an electrochemical cell provides the basis for

voltammetric sensors. Amperometric sensors are also based on the current–potential

relationship of the electrochemical cell, which can be considered a subclass of

voltammetric sensors.53 In amperometric sensors, a fixed potential is applied to

the electrochemical cell, and a corresponding current due to the reduction or oxidation

reaction is obtained. This current can be used to quantify the species involved in the

reaction. The key consideration of an amperometric sensor is that it operates at a fixed

potential. However, a voltammetric sensor can operate in othermodes such as linear or

cyclic voltammetric modes. Consequently, the respective current–potential response

for each mode will be different. This technique measures the current associated with

electrons, which are on the surface of biomolecules during redox processes.

Shimomura et al. using amperometric mechanism detected choline with enzyme

immobilized in a hybrid mesoporous membrane.50 Such an ability of the hybrid

mesoporousmembraneF127Msuggests great promise for effective immobilization of

enzyme useful for electrochemical biosensors. Liu et al. developed a sensitive


amperometric biosensor based on gold nanoelectrode array (NEA). The gold nanoe-

lectrode array was fabricated in the template of PC membranes.54 A conventional

three-electrode system was used in all the above measurements. NEA and NEA/GOx

were used as a working electrode, Ag/AgCl was used as a reference electrode, and a

spiral platinum wire acted as a counter electrode. The enzyme electrode exhibits an

excellent response performance to glucose with a linear range from 10�5 to 10�2M

and a fast response time of 8 s. Figure 2.15 demonstrates immobilization of the PC

membrane on the Au electrode.

2.8.2.2 PotentiometryPotentiometric sensors are fundamentally based on the potential developed between

two electrodes.When a redox reaction takes place at an electrode surface, a potential is

developed at the electrode–electrolyte surface. This potential is found and used to

characterize the activity of the species involved in the reaction. Electrodes in

potentiometric sensors can be inert or active. An inert electrode merely provides

the surface for an electron transfer process. However, an active electrode either

donates or accepts ions in the reaction. One electrode is always used as the reference

electrode to complete the circuitry for the potentiometric sensor. A noninterference

half-cell reaction occurs on this electrode. Potentiometric or voltammetric techniques

are generally not used for protein detection because most proteins cannot be detected

based on oxidation states and redox reactions. It is quite difficult to exchange charges

between proteins and a medium for a change in oxidation state. These methods are

used to signify the importance of metals in detection, but are not actively used for

detection in biosensors in general. Shishkanova et al. demonstrated functionalization

of the PVC membrane with single-stranded oligonucleotides for a potentiometric

biosensor.55 Reddy et al. estimated triglycerides by porous silicon-based potentiomet-

ric biosensors.56 Lipase, an enzyme that hydrolyzes triglycerides, was immobilized on

PS and was thermally oxidized. Upon hydrolysis, the triglycerides result in the

formation of fatty acids, which changes the pH of the solution. The enzyme

Teflon cap

PC membrane

Au electrode

FIGURE 2.15 Integration of the PC membrane with the gold electrode.54


solution-oxidized PS–crystalline silicon structure was used to detect changes in pH

during the hydrolysis of tributyrin as a shift in the capacitance–voltage (C–V)

characteristics.

2.8.2.3 Impedance/CapacitanceThismethod of detection employs true electrical parameters for detection and is based

more on perturbations in the electrical components than on charge transfers or redox

reactions. It calculates the amount of perturbation introducedby thebiomolecule to the

system more than the inherent change in the properties of the biomolecule. The most

popular methodology used is electrochemical impedance spectroscopy (EIS). EIS is

the plot of the overall impedance and phase between two electrodeswith respect to the

frequency applied to the circuit. This method of detection requires at least a two-

electrode arrangement for a closed circuit. Interdigitated electrodes or working

electrode–counter electrode setups are the most popular and efficient electrode

arrangements. Interdigitated electrodes offer much higher surface area of interaction

than normal planar electrodes and therefore are more preferred. In a working

electrode–counter electrode setup, the counter electrode is built with much higher

surface area than theworking electrode. This difference in surface area creates a large

difference in the interaction between the electrodes and the biochemicals, thereby

providing a highly sensitive detection.Using one of such two-electrode arrangements,

the impedance or capacitance is measured between those electrodes. Capacitance is

preferred as a parameter of detection due to its high accuracy and reliability. It can be

greatly enhancedwith an increase in surface area of interaction and plays a crucial role

in determining selectivity. Thismethod iswell suited for detecting proteins, as it needs

tomeasure the amount of perturbation introduced by the proteins to the system and not

charge transfers or half-cell reactions. For such small protein biomolecules to perturb

an electrical component, a large surface area for interaction is required. For just planar

electrodes, such as planar interdigitated electrodes, theremight not be enough surface

area for the proteins to bind. The surface also needs to be highly sensitive to protein

interaction for significant perturbation.

Bothara and coworkers designed a membrane-based nanomonitor protein biosen-

sorusingan impedanceandcapacitivemeasurement technique.57Figure2.16 showsan

electric double-layer electrode on which the nanoporous structure exists. Electrons in

the electrode cause the free ions in the solution to adsorb at the metal–liquid interface

creating the double layer. Some of the solvated ions occupy these spaces and the

proteins, which are also charged, reach close to these surfaces, which modify

the double layer causing changes to the frequency at which the peaks in energy

occur. Figure 2.17 illustrates a schematic diagram of an impedance/capacitance

porous membrane-based biosensor.

The measurement is achieved by the redox reactions at the surface for optimal

charge transfer. The nonfaradaic conductance of the electrical double layer formed at

the electrode surface is sensitive to reactions and is the basis of nanomonitors. This

technique is advantageous since it does not require addition of any redox probes.

Furthermore, conductance measurements at different bias voltages can reveal much

information about dielectric and charge environment at the interface. The


nanomonitors comprise multiple sensing sites with each sensing site containing

approximately a quarter million nanowells. Using the nanomonitor immunoassay

technique, we were able to detect CRP and MPO with the present lower limit of

detection at 10 and 20 ngmL�1, respectively. The upper limit of detection for both the

antigens was 100 mgmL�1. The dynamic range for CRP was 100mgmL�1 to

10 ngmL�1 and that for MPO was 100 mgmL�1 to 20 ngmL�1.

2.8.2.4 Optical DetectionOptical detection techniques have been widely used in the field of biosensors for their

high sensitivity. The most common examples include ELISA and fluorimetry, as

discussed earlier. The detection mechanism involved is usually fluorescence or

chemiluminescence. Fluorescence detection techniques are based on fluorescent

markers that emit light at specific wavelengths when light is incident on it. The

change in intensity of light emitted or its absolute value, as in fluorescence resonance

energy transfer (FRET), determines occurrence of the binding reaction. Certain

advances in this technique have been able to identify as low as single-molecule

detection levels. Fluorescence-based detection has been used in chips designed for

FIGURE2.16 Schematic diagram illustrating charge distribution across the liquid–electrode

interface forming a double layer.57


the purpose of microarrays. These techniques are true lab-on-a-chip devices with

integrated fluidic channels to direct the biomolecules. Chemiluminescence is the

generation of energy in the form of light when a chemical reaction takes place. In

synthetic compounds, such chemiluminescence takes place when a highly oxidized

species such as peroxide emits energy during a chemical reaction. Sometimes,

synthetic compounds are added to the biomolecule for detection, which in turn forms

a conjugate that releases energy in the form of light. Bioluminescence is another type

of luminescent technique, which has been reported by using the firefly luciferase/

luciferin as the synthetic compound. Surface plasmon resonance (SPR) is a technique

that looks at the surface activity. In this technique, a longitudinal wave of a certain

charge density is propagated along the surface of the metal and the dielectric. Due to

the total internal reflection of light against materials such as gold or silver, the

evanescent wave or field created at the surface penetrates the interface into the dense

medium (metal). This evanescent light is able to couple with the free electrons on

the surface, called plasmons, and this creates a resonance wave, which is recorded.

Hence, in the presence of anybiomolecule layer located on the surface of themetal, the

SPR adsorption profile is obtained. Due to the properties of biomolecules, each SPR

spectrum is different, hence making it a valid detection technique. The biggest asset

FIGURE 2.17 The schematic diagram of an impedance/capacitance porous membrane-

based biosensor.57


and the biggest drawback of optical detection system is light. Optical detection

requires huge infrastructures, light sources, and other bulky instruments, making

it very expensive and not portable. It is also very difficult to integrate and

calibrate the light for detection. A tiny jitter in the system could make it completely

inoperable and take hours for a highly skilled personnel to fix it. Specificity is also an

issue with many optical detection techniques, but ELISA is considerably well

established in this respect.

De Stefano et al. demonstrated single-stranded DNA sensor using porous silicon

surface.58 Using photochemical functionalization process, the porous silicon passiv-

ated the surface of optical biosensors. Fluorescence measurements have been used to

investigate the stability of the DNA single strands bound to the nanostructured

material. A dose–response curve in the 6–80mM range for an optical label-free

biosensor has been realized. Similarly, Bonanno and DeLouise demonstrated a

membrane-based optical biosensor.59 The sandwich assay scheme incorporated in

the sensor comprises a linking biotin/streptavidin to attach biotinylated anti-rabbit

IgG receptor to detect rabbit IgG. The schematic diagram illustrating the working

principle is shown in Figure 2.18. The final detection range of rabbit IgG was

0.07–3mgmL�1 (0.23–9.8mgmm�2). In the detectionmechanism, a normal incident

beam of white light (spot size of <13mm2) was exposed to the sensor surface.

The optical reflectance spectrum was measured. The optical shift due to specific

binding interactions within the porous matrix was measured.

2.8.2.5 Mechanical DetectionMechanical detection for biochemical entities and reactions is generally through

the use of nanomembrane and micro- or nanoscaled cantilever sensors. They can

be further divided into mass sensors or stress sensors. Microgravimetric transducers

BB

B

B

BB

B

Biotin

APTMS SiO2

Streptavidin

Bioreceptor Biotin-α-rabbit IgG

Target Rabbit IgG

PSi

FIGURE 2.18 Schematic diagram of a silicon porous membrane-based optical label-free

biosensor.59


monitor mass changes that occur during the binding of target analytes to the surface-

confined recognition layer. Xue and Cui micropatterned carbon nanotube (CNT) and

cantilever arrays fabricated with layer-by-layer nano-self-assembly that has applica-

tion toward biosensors.60 In recent years, the areas of biomicroelectromechanical

systems (BioMEMS) and nanotechnology have gained a high level of prominence and

have become almost inseparable from biological applications including detection,

diagnostics, therapeutics, and tissue engineering. In the nanoporousmembrane-based

mechanical biosensor, the porous nanomembrane acts as a filter for a bioagent in the

microfluidic channel; later, the biomolecules are detected using a cantilever or a

piezoelectric crystal. In a mass biosensor, change in the mass caused by chemical

binding to small piezoelectric crystals is detected.Payen et al. demonstrated anMEMS

rf-interrogated biosensor. Figure 2.19 demonstrated the step-by-step sequence for

measuring the levelof pHandeventuallyglucose concentration (Brix) ingrapes.61The

biosensing structure consists of a microneedle, a functionalized gel, a tuned tank

oscillator circuit, and a miniature antenna.

A microneedle used was to puncture the organism. A two-membrane matrix was

integrated into the membrane–filter system. The membrane system was used to

separate the wet part of the hydrogel from the dry portion in which the tank oscillator

circuit is located.Thefirstmembranewas used to allowdiffusionof thegrape juice into

the hydrogel. The second compliant membrane was impermeable. Its function was to

separate the wet region of the hydrogel from integrated circuit and antenna. A rigid

porous membrane was incorporated to protect the hydrogel from biological fluids

and to prevent contamination by molecules. The change in pH induces a swelling of

the hydrogel. The hydrogel displacement depends on the hydrogel sensitivity and the

flexible membranes. The membrane displacement induces a change in capacitance.

The LC tank circuit frequency interrogation system determines the optimal frequency

for data collection.

Bioelements (crop)

Grapes, fruits, roots, etc.

Fluidic delivery Microchannel

Biosensor Hydrogel

Membrane matrix

Filter system

Transducer

Transduction mechanism Capacitor

FIGURE 2.19 Process sequence for the detection of pH and glucose in grapes.61


2.9 POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION

OF LIVING ORGANISM

2.9.1 E. coli Biosensor

Meat products are increasingly contaminated by foodborne pathogens, thereby

increasing product recalls in the United Stated. These products are contaminated

from a number of sources, including the environment and the animal itself.62 E. coli

serotype O157:H7 are harmful and deadliest bacteria found in meat and dairy

products. These O157:H7 bacteria cause food poisoning, gastroenteritis, urinary tract

infections, and neonatal meningitis. Thus, the microorganisms need to be detected at

an early stage of infection in human beings. Many researchers are conducting

experiments and spending billions of dollars to fabricate biosensors to detect the

disease in a short time frame andwith a high degree of accuracy.Wang et al. fabricated

a novel nanoporous biosensor based on single-stranded DNA (ssDNA) probe func-

tionalized AAO nanopore membranes for E. coli O157:H7 DNA detection.63 These

membrane-based biosensors offered low detection limit for DNA in picomoles and

rapid label-free and easy-to-use bacteria detection that holds the potential for future

lab-on-a-chip devices. Similarly, disposable porous filter membranes were used by

Abdel-Hamid et al. to develop a flow-through amperometric immunofiltration assay

system for rapid detection of E. coliO157:H7.64 In the experiment, nylon membranes

were used as a solid support for immobilization of antibodies. Two types of mem-

branes, Biodyne B and C, were tested on which antibodies to E. coli O157:H7 were

immobilized. These porous membranes were used for filtering the antigen-containing

solution through the antibody-coated filter membrane. This results in excellent

antigen–antibody binding, thereby significantly reducing the assay time. The effect

of membrane pore sizes of 0.45, 1.2, and 3mm determined the amount of immobilized

anti-E. coli antibodies. The detection limit of E. coli O157:H7 cells is 100 cellsmL�1

and theworking range is 100–600 cellsmL�1. A complete immunoassay is carried out

in 30min.

Again, Liu et al. conjugated CdSe/ZnS core–shell dendron nanocrystals with the

corresponding antibodies and then passed through the microporous membrane where

they attached to the membrane antigen–antibody.65 The membrane antigen–antibody

conjugated with the nanocrystals facilitated an efficient and stable photolumines-

cence. The biosensor built using this technique was used to detect not only E. coli but

also hepatitis B with a limit of detection as low as 2.3 CFUmL�1 and 5 ngmL�1,

respectively. Due to unique optical properties, nanocrystals were used as fluorescent

labels. A microporous membrane was chosen for its mechanical strength and

biocompatibility. The biosensor system comprises a flow chamber with an immuno-

filter interface with sensitive and robust dendron nanocrystals as the detection

indicators. By using a matrix of the membrane, the scientists observed an increase

in the efficiencyof the immunoreactionbetweenantibodies andpathogens, decrease in

the detection time, and reduction in the detection limit. The pore size was optimized

such that therewas an excellent immunoreaction between antibodies and pathogens in

a small hole and the liquid solution could be drained out. In the biosensor system, as

POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION OF LIVING ORGANISM 75

shown inFigure 2.20, the antibodies are immobilized onto themicroporousmembrane

using covalent bonding. Later, theE. coli antigens are capturedonto antibodies to form

immunocomplexes. The immunocomplexes formed on the surface of the membrane

continue to react with the dendron nanocrystal-conjugated antibodies and form

“sandwich” immunocomplexes. By measuring the photoluminescence from the

dendron nanocrystals, the targets, that is, E. coli antigens, are detected. Then, the

nanocrystal-labeled antibody solution was injected to form a sandwich immunocom-

plex of immobilized antibody–E. coli O157:H7–nanocrystal-labeled antibody.

The complex immunostructure using photoluminescence was measured by a

spectrofluorometer.

A common practice for the E. coli pathogen detection is to use an antibody-coated

filter porous membrane to immobilize E. coli O157:H7, and different detection

mechanism are used to detect the immunoreactions.

2.9.2 Salmonella enteritidis

A commercially produced foodstuff containing raw eggs such as ice cream often gets

contaminated with one of the deadliest pathogens S. enterica. These pathogens

suppress the human immunity system. Therefore, scientists are developing and

fabricating biosensors for detection of S. enteritidis. Zhang and Alocilja investigated

a label-free DNA electrochemical biosensor for the detection of S. enteritidis.66

A nanoporous silicon-based DNA biosensor consists of a porous silicon surface,

which was functionalized with DNA probes specific to the gene of S. enteritidis.

Electrical property of DNA, redox indicators, and cyclic voltammetry were used for

FIGURE 2.20 Schematic pictorial representation of a porous membrane-based biosensor for

detection of E. coli bacteria.65


the characterization of the biosensor. Porous silicon was fabricated using an anodiza-

tion process in an electrochemical Teflon cell. The anodization was carried out by

hydrofluoric acid. A uniform pore structure, with pore sizes ranging from 10 to 30 nm

pores,was connected, and an interpore space of 10–30 nmwasobserved.These porous

membranes were functionalized in two steps, which involved silanization of the chip

surface followed by DNA probe immobilization. By immobilizing the specific DNA

probe onto the PS layer, the biosensor had the ability to capture complementary DNA

(cDNA). Finally, a cyclic voltammogram of target DNA concentrations on porous

silicon chips was obtained. This DNA probe has high selectivity and affinity for the

target DNA. The redox marker has a greater affinity for dsDNA, and therefore, a

greater electrochemical response was observed when hybridization occurred.

When the concentration of target DNA increased, the charge transfer between the

redoxmarker and thePSelectrodewas enforced so that the peak current increasedwith

DNA concentration. The detection limit of the PS-based label-free DNA biosensor

was 1 ngmL�1.

2.9.3 Virus Detection

Multiple diseases are caused by a lethal submicroscopic infectious agent known as

virus. In themodernworld, serious diseases such as Ebola, AIDS, avian influenza, and

SARS are caused by viruses. Therefore, all around the world many researchers are

developing biosensors to detect viruses. Riccardi et al. designed a novel label-free

electrochemical detection system of DNA hybridization for detecting hepatitis C

virus.67 A porous polycarbonate membrane was integrated to detect short sequence

(18-mer) target DNA after diffusion. A voltammetric microbiosensor based on

immobilization of the 18-mer HCV-1 DNA probe was applied in combination with

SECM line scans to evaluate hybridization of DNA fragments diffusing through a

porous polycarbonate membrane. Rossi et al. were able to detect bacteriophage virus

MS2 using a porous silicon biosensor.68 For immobilization of the virus, covalent

bioconjugation of antibodies inside porous silicon films was carried out. By fluores-

cence, 2� 107 plaque-forming units per milliliter (pfumL�1) were detectable. The

nanoporousmembrane had an average pore size of approximately 50 nm. The internal

surfaces of porous silicon film are hydrogen terminated for bioconjugation chemistry.

The hydrogen is then replaced by a functional organic group, which can be linked by a

desired proteinmolecule using a cross-linker.Onto this linker, the unlabeled anti-MS2

antibodies are binded. The conjugation of either Alexa 488 to the antibodies or Alexa

532 fluorophores to MS2 viruses was carried out depending on the functionalization

methods. Once again, the porous silicon surfaces were used to immobilize unlabeled

anti-MS2 antibodies. The amounts of antibody and virus bound to the porous silicon

surface were evaluated by fluorescence intensity at the emission maximum.

Suggestions were made to improve the sensor performance by further optimization

of porous layer structure and thickness. Gyurcs�anyi mentioned about “passive”

nanopore counters for detecting icosahedral chlorella virus viruses.69 A real-time

monitoring of the antibody–virus binding was able to detect concentrations as low as

5� 107 particlesmL�1 by a label-free technique. Reichmuth developed a lab-on-a-


chip device for rapid and portable diagnostics for detecting zoonotic diseases.70

A microchip-based electrophoretic immunoassay with an integrated nanoporous

membrane incorporated into an open-channel electrophoresis and laser-induced

fluorescence detection with a labeled antibody was carried out to detect influenza

virus. The functionality of polymer membrane filtration eliminates the need for

washing, commonly required in surface-based immunoassays, increasing the speed of

the assay.Yanget al.were able tofilter human rhinovirus type 14major pathogen of the

common cold in humans using nanoporous block copolymers.8 These nanoporous

membranes showed excellent resistance to all organic solvents.

2.9.4 Glucose Detection

For the treatment of diabetes mellitus, the amount of glucose present in a mammal’s

blood is analyzed. Saha et al. fabricated a bioelectronic biosensor for the detection of

glucose.71 A nanoporous cerium oxide (CeO2) thin film deposited on a platinum (Pt)-

coated glass plate using pulsed laser deposition (PLD) has been utilized for immobili-

zation of glucose oxidase (GOx). Differential pulsed voltammetry (DPV) and optical

measurements show that the GOx/CeO2/Pt bioelectrode exhibits linearity in glucose

concentration ranging from 25 to 300mg dL�1. Immobilization of GOx onto CeO2

matrix was achieved by electrostatic interaction of positively charged CeO2 and

negatively chargedGOx enzyme at pH 7.0. Fourier transform infrared (FTIR), atomic

force microscopy (AFM), and DPV techniques helped investigation of the CeO2/Pt

electrode and GOx/CeO2/Pt bioelectrodes. Wei et al. had a novel approach for

detection of glucose oxidase.72 Glucose oxidase is entrapped into a complex nano-

composite film of chitosan/nanoporous ZrO2/multiwalled carbon nanotubes

(MWNTs). Nanoporous ZrO2 helped to enhance the stability of the immobilized

enzyme. Awide linear response range from 8mmol L�1 to 3mmol L�1 was obtained

by an amperometric glucose biosensor. The ability of CNTs to promote the electron

transfer of hydrogen peroxide (H2O2) suggested a promising idea for the construction

of oxidase-based amperometric biosensors. The organic–inorganic CHIT/ZrO2/

MWNT nanocomposite had an added advantage in terms of toughness of CHIT

and chemical and thermal stability of nanoporous ZrO2.

Another approach by Wang et al. demonstrated a nonenzymatic electrochemical

glucose sensor based on nanoporous PtPb networks.73 A reproducible one-step

hydrothermal method helped grow PtPb networks on Ti substrates. Voltammetry

and amperometric methods are used to evaluate the electrooxidation activities of the

synthesized electrodes toward nonenzymatic glucose oxidation in neutralmedia in the

absence and presence of chloride ions. The nanoporous PtPb electrodes have strong

and sensitive current responses to glucose. The excellent performance of the PtPb

electrode was achieved at an optimal PtPb composition of 50%. Not only thin films

with pores are used as a platform in a sensing device but also beads with pores are used

as a mechanism to trap GOx. Vamvakaki and Chaniotakis utilized porous silica beads

with pore sizes of 10 nm for the immobilization and stabilization of the GOx with

diameters on the order of 7 nm.74 The confinement of the GOx leads to enhanced

enzyme stability. Silica beads and porous polymer beads were embedded for the


development of novel and highly stable glucose biosensor systems. In order to

construct the GOx biosensor, the beads with the immobilized enzymes were placed

on platinum electrodes through Nafion membranes. The glucose biosensor used the

electrochemical measurement technique, which consists of a three-electrode system

of silver/silver chloride double junction reference electrode and a platinum counter

electrode. Silica beads have a well-defined pore size of 10 nm and a particle size of

60–80 mesh, while polymer beads (PLRP-S) with a spherical porous structure have

pore sizes of 10 and 30 nm and an average particle diameter of 50–70 mm. The

comparative study showed that the response time of the free enzyme biosensors was

between 30 and 60 s and that of the bead-adsorbed enzyme biosensorswas between 60

and 90 s. The free GOx biosensor had a remaining activity of 60% after 70 h of

continuous operation, while bead-adsorbed enzyme biosensors did not lose any of

their initial activity even after 70 h of continuous operation. In addition, it was evident

in the experiments that the size matching between the pore size and the molecular

diameter of the enzymes is very important to achieve high enzymatic activity and

prevent enzyme leaching.

Ekanayake et al. enhanced the adsorption of glucose oxidase by introducing

artificial porosity into polypyrrole-based glucose biosensors.75 It also enhanced

the performance of the sensor in terms of increasing high enzyme loading, stability,

sensitivity, reproducibility, and repeatability. The immobilization was done by

physical adsorption. Glutaraldehyde was used for cross-linking, while in enzyme

adsorption, the response current is increased. The reaction that produces H2O2 for

sensing is shown below.

GlucoseþO2 �!GOx Gluconic acidþH2O2

H2O2 !O2 þ 2Hþ þ 2e�

PF6� dopant helps produce microporous PPy films. This improves the porous

structure and the amperometric response of the electrodes. The thickness of the

nanoporous film is about 30–40mm. PF6� dopant introduced to the PPy led to a

significant improvement in sensor characteristics. Most recently, Fink et al. were able

to detect glucose using a reusable enzyme-modified ion track membrane sensor.76

The enzyme glucose oxidase covalently linked to nanopores by covalent linking. The

detection range of glucose concentrations is between10mMand1M.Theprinciple for

glucose sensor is described in Figure 2.21.

The glucose biosensor consists of PET foils of 4� 106 cm�2 conical nanopores

with a pore size of approximately 1.0� 0.1 mm at the wide end and approximately

30� 20 nm at the tip. The measurement compartment was separated by a sensing

membrane and electrodes were inserted into both parts of the vessel. The voltage and

transmission ion current were measured by applying sinusoidally shaped AC voltage

of up to 5Vpeak–peak operating at 1Hz. On the basis of the current–voltage character-

istics obtained, a diode-like rectification curve was obtained due to ion flow. The

soluble reaction products produced due to enzyme reaction diffuse away from pore tip

and avoid blocking of the pore, thereby rendering these sensors reusable. Using the


enzymeglucose oxidase (GOx) as the catalyst, the oxidation of glucose in the presence

of oxygen is performed for glucose biosensing purposes. The glucosewas oxidized by

GOx and not by any of the other components of the device. The linearity between the

measured ion current through the sensor and the applied glucose concentration was

recorded.

2.9.5 Cholesterol

Understanding and knowing cholesterol level in our bloodstream is extremely

essential to reduce the risk of a heart attack or stroke. Li et al. have fabricated a

cholesterol biosensor. Cholesterol oxidasewas entrapped in a silicic sol–gel matrix.77

The half-life of the biosensor is about 35 days. The cholesterol biosensor has a high

sensitivity and selectivity and can determine cholesterol oxidase ranging from

1� 10�6 to 8� 10�5mol L�1 with a detection limit of 1.2� 10�7mol L�1. They

also fabricated a selective cholesterol biosensor based on the composite film-modified

electrode for amperometric detection.78 A concentration range of approximately

10�510�4mol L�1 with a detection limit of 6� 10�7mol L�1 was determined.

The excellent sensitivity and selectivity were attributed to the PB/PPy layer on the

biosensor. Singh et al. immobilized cholesterol oxidase (ChOx) onto zinc oxide (ZnO)

nanoporous thin films grown on gold surface.79 The porous thin film was fabricated

using rf magnetron sputtering. A cyclic voltammetric measurement method was used

for detectionofChOxand the sensitivity of detectionwas in range 25–400mg dL�1. In

this voltammetric method, the ChOx/ZnO/Au bioelectrode was found to detect

cholesterol. Arya developed a ChOx biosensor using optical measurement. The

ChOx molecule was covalently bonded to the sensing ODT electrode. The life of

the ODT electrode-based biosensor was 2 months.80 Ansari et al. derived a sol–gel

nanoporous cerium oxide film.81 A nanostructured cerium oxide (NS-CeO2) film

deposited on the indium tin oxide (ITO)-coated glass substrate was used to detect

cholesterol oxide.

FIGURE 2.21 The pictorial representation of a glucose sensor.76


2.9.6 New Diverse Sensors

Recently, biomolecules were immobilized onto the sensing porous microelectrodes,

which are used as a means of detecting bioagents. Thesemicroelectrodes are porous in

nature, which help localize the bioagent for sensing. Song et al. developed a biosensor

for diagnosis and monitoring of liver disease. An electrochemical array of nanoporous

silicon electrodes is integrated for constructing these sensors.82 The liver plays amajor

role in metabolism, digestion, detoxification, and elimination of various substances

from the body. Biomarkers such as cholesterol, bilirubin, and aminotransferases

present in the serum help analyze the status of the liver. Using the silanization

technique, sensitivities of the device were recorded to be 0.2656mAmM�1 for

cholesterol, 0.15354mAmM�1 for bilirubin, 0.13698mA (U l�1)�1 for alanine ami-

notransferase (ALT), and 0.45439mA (U l�1)�1 for aspartate aminotransferase (AST).

Compared to traditional analyte measurement procedures, the novel analytical device

demonstrated high level of sensitivities for the analyses of multiple samples and

analytes without a marked cross-interference effect. A single device containing the

multiarray electrodes for sensing three different analytes is shown in Figure 2.22.

A readout within minutes of application of microvolumes of a sample, reduced

physical dimensions of the device, relative stability of the reagents used, and simple

electronic component assembly useful for point-of-care biomarker liver analyses are

some of the advantages of the liver biosensor discussed above.83

Measurement of urea in real urine samples is performed by a urea biosensor

developed by Yang et al.84 The urease was immobilized onto nanoporous alumina

membranes prepared by the two-step anodization method. Simple physical adsorption

andcross-linking techniqueused for immobilizationof theseenzymesandpiezoelectric

mechanical detection mechanism were used for analysis. In the urea biosensor, urease

was immobilizedontonanoporousaluminamembranes, and the frequency responsesof

the ESPS/FIA detection system with the alumina electrode and the alumina/urease

electrode were measured after the injection of 1.0 and 0.1mM urea solutions. The

catalytic reaction of the urease–urea system can be described by the equation below,

thereby increasing the conductivity due to change of uncharged urea molecule to three

ions. The frequency decrease is thereby attributed to the increase in conductivity.

NH2CONH2þ 2H2O�!urease 2NH4þ þCO3

2�

FIGURE 2.22 Schematic representation of a biosensor for liver diagnosis.83


The fabricated urea biosensor presented high-selectivity monitoring of urea,

better reproducibility (SD¼ 0.02, n¼ 6), 30 s shorter response time, wider linear

range from 0.5mM to 3mM, lower detection limit of 0.2 mM, and good long-

term storage stability with about 76% of the enzymatic activity retained after 30

days. The activity of enzyme increases with increasing pore length for large pore size,

while for correspondingly small pore size, enzymatic activity slightly depends on

pore length.85

A nanoporous membrane can also be used to interface biological materials with a

biohybrid system.Wolfrum et al. suspended nanoporous membranes as interfaces for

neuronal biohybrid systems. Specification of the porosity parameters showed change

in the transconductance of the nanopores and therefore helped control diffusion of

molecules through themembranes. The alumina nanoporous membranes are biocom-

patible with both primary vertebrate and insect neurons. The thin aluminum filmwith

500 nm nanopores was integrated onto silicon nitride/silicon oxide to create stable

suspended nanoporous Si3N4/SiO2 membranes by simple lithography anisotropic

chemical silicon etching from the backside of the wafer in combination with

anodization of thin aluminum films. Cells are genetically engineered and are placed

on alumina membrane, which serves as a voltage-gated potassium ion channel. The

cells showed adhesion and grew on the surface. The nanoporous membranes act as a

cell interface and facilitate control of the cell environment with minute quantities of

chemicals.

2.10 MICROFLUIDIC BIOSENSOR SYSTEMS

A long-term goal in the field of microfluidics is to create integrated, portable

clinical diagnostic devices for home and bedside use. Nanoporous membrane-based

biosensors are integral components of a microfluidic system. This section describes

the applications of microfluidic systems that have demonstrated the incorporation

of nanoporous membranes to develop sensor systems. The chip used for this system

is composed of an inexpensive and biocompatible polydimethylsiloxane (PDMS)

layer. This overlays a Pyrex glass substrate that contains arrays of microelectrodes,

which are used to detect the electrical signal in the biological environmental

system. These chips consist of a microchannel to dispense or flow enzymes and

analytes into the sensing area. These microfluidic biosensor systems provide low

cost, portable, and high-throughput analytical systems. The approach applied for

the construction of the microfluidic module provided a precise sample handling in

terms of volume and flow rates, minimal dead volume at the inlet and outlet holes,

that is, no sample losses during the analysis and approximately 100% waste

disposal, and the ability for quick interchange of channels with different geometries

and dimensions.86 Joo performed integration of a thin nanoporous platinum film

into a microfluidic system for nonenzymatic electrochemical glucose sensing.87

The schematic diagram of a biosensor embedded into a microfluidic is shown in

Figure 2.23.


Glucose sensors cause intrinsic problems in manufacturing, storage, and distribu-

tion due to a serious, unacceptable dependence on temperature and humidity.

Moreover, quality control difficulty and a short lifetime ultimately mean sizable

costs. Therefore, the glucose biosensor was integrated into microfluidic systems.

Programmed fluidic control of multiple reservoirs yields high-throughput analysis,

automatic calibration, and multiple uses.

Similarly, Metz et al. using micromachining and ion track technology fabricated

polyimide microfluidic devices with integrated nanoporous filtration systems.88

Figure 2.24 illustrates a schematic block diagram of a microfluidic biosensor with

embedded porous membrane.

The device consists of inlet and outlet openings with the buried channels. The

porous area is the filtration platform. One of the key advantages of a biosensor-based

microfluidic is the limited contamination from outer atmosphere. Maeng et al.

invented a novel microfluidic biosensor based on an electrical detection system for

alpha-fetoprotein,89 whereas Goral et al. detected nucleic acid sequences using an

electrochemical microfluidic biosensor.90 The microfluidic biosensor system was

made up of PDMS andwas integrated into an interdigitated ultramicroelectrode array,

and microchannels were fabricated in a glass chip. Encapsulation of biosensor area

into the closed system eliminated background signals to absolute nil and the IDUA

responded in a highly reversible manner to the injection of various volumes and

various concentrations of the electrochemicalmarker. Thus, the limit of detectionwas

pushed to 1 fmol per assay and the dynamic range was 1–50 fmol.

Similarly, Zaytseva et al. developed a microfluidic biosensor module with fluores-

cence detection for rapid and reliable identification of pathogenic organisms and

FIGURE 2.23 Schematic representation of a microfluidic system.87

FIGURE 2.24 Schematic diagram demonstrating a polyamide microfluidic device for

biosensor application.88

MICROFLUIDIC BIOSENSOR SYSTEMS 83

viruses.86 Themicrofluidic biosensor included a network ofmicrochannels fabricated

using soft lithography.Asexplained in theprevious example,PDMS,which is themost

commonly used polymeric substance, is used for fabricating the biomodule.

Integration of the micro total analysis system into a sensing element enabled sample

preparation and detection steps onto a single platform. A network of microchannels

comprising the microbiosensor is modeled using computer-aided design software,

nanofabricated in a silicon wafer by photolithography and etching techniques and

molded into a PDMS elastomer by replica molding. In addition, Kwakye et al.

developed an electrochemical microfluidic biosensor for the quantification of

RNA.Theconceptwas todemonstrate detectionofdenguevirusRNA.91Thedetection

mechanism was based on nucleic acid hybridization and liposome signal amplifica-

tion. In addition, integration of a minipotentiostat, an interdigitated ultramicroelec-

trode array, and a microfluidic biosensor was successfully exhibited. Microchannel

fluidic devices were fabricated using standard photolithography and dry etching.

Using soft lithography, the microchannels were realized in PDMS covered with a

glass slide consisting of interdigitated ultramicroelectrode arrays and packaged in a

Plexiglas housing. Incorporating microfluidic systems yields a potentially inexpen-

sive, portable, and disposable biosensor device that is easy to assemble and use.

Yu et al. demonstrated a nanoporous impedance sensing transducer for fast bacteria

patterning and detection at a low-frequency spectrum.92 A poly(ethylene glycol)

(PEG) hydrogel microfluidic chip included patterned nanoporous aluminum oxide

membrane (AOM) that allowed the detection limit to improve from 104 to around

102CFUmL�1, in comparison to the conventional impedance measurement system.

Vargas-Bernal presented multiple topologies based on the microfluidic device in

order to optimize the biosensor design for evaluation and detection of pesticides.93

Theaimwas to reduce the sizeof sample, reduceboth the analytical systemsize and the

test time, automate the operation, and finally ease the transportation. These require-

ments were met by using smart matrices for enzyme immobilization, microfluidic

systems, more sensitive enzymes, enzymes with more specific roles, and different

detecting methods.

2.11 SUMMARY AND FUTURE PERSPECTIVE

In conclusion, we believe that nanoporous membrane-based microfluidic systems

have a wide range of applications from development of biosensors for detection of

living microorganisms that cause harmful diseases to fabrication of membrane-based

drug delivery systems. These nanoporous membranes have opened a new field in the

area of membrane-based biosensors. These nanoporous microfluidic membrane-

based biosensors not only are tested in scientific laboratory but are also being tested

in clinical environments. Due to these reasons, there is a continuous need for rapid

improvement in sensitivity, stability, ease of formation of solvent-free membranes,

robustness, arrays, and cost. These requirements have led to various scientists

exploring new methods and strategies in membrane fabrication and development.

Natural membranes are now being used for fabrication of biosensors. Most common,


naturally found nanopores membranes are lipid membranes,94 and recently, micro-

scopic diatoms are being investigated in biosensors.

These biomimetic lipid membranes consist of natural greasy molecules that are

found in real membranes and some molecules that are not usually found in the body.

These membranes stick tightly to metal surfaces, and this makes them rugged enough

to provide a good home for membrane proteins even in industrial applications.

Diatoms are microscopic, single-celled algae that possess rigid cell walls composed

of hard and porous amorphous silica. Diatom biosensors are devices incorporating a

biological molecular recognition component connected to a transducer capable of

outputting a signal proportional to the concentration of the molecule being sensed.95

These low-cost and largely available natural materials found in microorganisms are

now being investigated as transducer elements for optical biosensors or as targeting

microcapsules for drug delivery.96 In addition, instead of nanoporous films, porous

electrodes are being designed and fabricated for building biosensors that possess

high sensitivity, outstanding selectivity, repeatability, and cost effectiveness.

Researchers are discovering porous structures of any form that are biocompatible

and can be embedded into a transducer for forming biosensing systems. In conclusion,

there is a paradigm shift from manufactured membranes toward using natural and

biomimetic membranes.

REFERENCES

1. Adiga, S.; Curtiss, L.; Elam, J.; Pellin, M.; Shih, C.-C.; Shih, C.-M.; Lin, S.-J.; Su, Y.-Y.;

Gittard, S.; Zhang, J.; Narayan, R. Nanoporous materials for biomedical devices.

JOM J. Miner. Met. and Mater. Soc. 2008, 60, 26–32.

2. Whitesides, G.M. The origins and the future of microfluidics.Nature 2006, 442, 368–373.

3. Mohanty, S.P.; Kougianos, E. Biosensors: a tutorial review. IEEE 2006, 6, 35–40.

4. Sheehan, P.E.; Whitman, L.J. Detection limits for nanoscale biosensors. Nano Lett. 2005,

5, 803–807.

5. Cheng, K.-L.; Sheng, Y.-J.; Jiang, S.; Tsao, H.-K. Electrophoretic size separation of

particles in a periodically constricted microchannel. J. Chem. Phys. 2008, 128, 101101.

6. Kuo, C.-W.; Shiu, J.-Y.; Wei, K.H.; Chen, P. Monolithic integration of well-ordered

nanoporous structures in the microfluidic channels for bioseparation. J. Chromatogr. A

2007, 1162, 175–179.

7. Ulbricht, M. Advanced functional polymer membranes. Polymer 2006, 47, 2217–2262.

8. Yang, S.Y.; Jinhwan, J.P.; Moonhor, Y.; Key, R.S.; Kim, J.J.K. Virus filtration membranes

prepared from nanoporous block copolymers with good dimensional stability under high

pressures and excellent solvent resistance. Adv. Funct. Mater. 2008, 18, 1371–1377.

9. Lee, W.; Ji, R.; Gosele, U.; Nielsch, K. Fast fabrication of long-range ordered porous

alumina membranes by hard anodization. Nat. Mater. 2006, 5, 741–747.

10. Reddy, R.K.; Bothara, M.G.; Barrett, T.W.; Carruthers, J.; Prasad, S. Nanomonitors:

protein biosensors for rapid analyte analysis. IEEE Sens. J. 2008, 8, 720–723.

11. Apel, P. Track etching technique in membrane technology. Radiat. Meas. 2001, 34,

559–566.

REFERENCES 85

12. Awasthi, K.; Kulshrestha, V.; Acharya, N.K.; Singh, M.; Vijay, Y.K. Ion transport through

track etched polypropylene membrane. Eur. Polym. J. 2006, 42, 883–887.

13. Dobrev, D.; Neumann, R.; Angert, N.; Vetter, J. Formation of metal membranes by direct

duplication of etched ion-track templates. Appl. Phys. A: Mater. Sci. Process. 2003, 76,

787–790.

14. Enculescu, I. Nanowires and nanotubes prepared using ion track membranes as templates.

Digest J. Nanomater. Biostruct. 2006, 1, 15–20.

15. Siwy, Z.; Apel, P.; Dobrev, D.; Neumann, R.; Spohr, R.; Trautmann, C.; Voss, K. Ion

transport through asymmetric nanopores prepared by ion track etching. Nucl. Instrum.

Methods Phys. Res. B: Beam Interact. Mater. Atoms 2003, 208, 143–148.

16. Vogrin, N.; Stropnik, C.;Musil, V.; Brumen,M. Thewet phase separation: the effect of cast

solution thickness on the appearance of macrovoids in the membrane forming ternary

cellulose acetate/acetone/water system. J. Membr. Sci. 2002, 207, 139–141.

17. Lee, H.J.; Jung, B.; Kang, Y.S.; Lee, H. Phase separation of polymer casting solution by

nonsolvent vapor. J. Membr. Sci. 2004, 245, 103–112.

18. Stropnik, C.; Kaiser, V. Polymeric membranes preparation by wet phase separation:

mechanisms and elementary processes. Desalination 2002, 145, 1–10.

19. Taketani, Y.; Nagaoka, S.; Kawakami, H. Fabrication of three-dimensionally ordered

microporous membrane by wet phase separation. J. Appl. Polym. Sci. 2004, 92,

3016–3021.

20. Li, D.; Krantz, W.B.; Greenberg, A.R.; Sani, R.L. Membrane formation via thermally

induced phase separation (TIPS): model development and validation. J. Membr. Sci. 2006,

279, 50–60.

21. Kuo, C.-Y.; Lin, H.-N.; Tsai, H.-A.; Wang, D.-M.; Lai, J.-Y. Fabrication of a high

hydrophobic PVDF membrane via nonsolvent induced phase separation. Desalination

2008, 233, 40–47.

22. Fu, X.Y.; Sotani, T.; Matsuyama, H. Effect of membrane preparation method on the outer

surface roughness of cellulose acetate butyrate hollowfibermembrane.Desalination 2008,

233, 10–18.

23. Kurt, R.; Simon, L.; Penterman, R.; Peeters, E.; de Koning, H.; Broer, D.J. Control over the

morphology of porous polymeric membranes for flow through biosensors. J. Membr. Sci.

2008, 321, 51–60.

24. Leoni, L.; Boiarski, A.; Desai, T.A. Characterization of nanoporous membranes for

immunoisolation: diffusion properties and tissue effects. Biomed. Microdevices 2002,

4, 131–139.

25. Lopez, C.A.; Fleischman, A.J.; Roy, S.; Desai, T.A. Evaluation of silicon nanoporous

membranes and ECM-based microenvironments on neurosecretory cells. Biomaterials

2006, 27, 3075–3083.

26. Yan, X.; Liu, G.; Dickey, M.; Willson, C.G. Preparation of porous polymer membranes

using nano- or micro-pillar arrays as templates. Polymer 2004, 45, 8469–8474.

27. Lo, C.J.; Aref, T.; Bezryadin, A. Fabrication of symmetric sub-5 nm nanopores using

focused ion and electron beams. Nanotechnology 2006, 17, 3264–3267.

28. Patterson, N.; Adams, D.P.; Hodges, V.C.; Vasile, M.J.; Michael, J.R.; Kotula, P.G.

Controlled fabrication of nanopores using a direct focused ion beam approach with

back face particle detection. Nanotechnology 2008, 19, 1–10.


29. Lillo, M.; Losic, D. Ion-beam pore opening of porous anodic alumina: the formation of

single nanopore and nanopore arrays. Mater. Lett. 2009, 63, 457–460.

30. Kim, M.J.; McNally, B.; Murata, K.; Meller, A. Characteristics of solid-state

nanometre pores fabricated using a transmission electron microscope. Nanotechnology

2007, 18, 1–5.

31. Kim, M.J.; Wanunu, M.; Bell, D.C.; Meller, A. Rapid fabrication of uniformly sized

nanopores and nanopore arrays for parallel DNA analysis. Adv. Mater. 2006, 18,

3149–3153.

32. Bell, D.C. Fabrication and application of nanopores using TEM, STEM and ion beams.

Microsc. Microanal. 2008, 14, 244–245.

33. Striemer, C.C.; Gaborski, T.R.; McGrath, J.L.; Fauchet, P.M. Charge- and size-based

separation of macromolecules using ultrathin silicon membranes. Nature 2007, 445,

749–753.

34. Balamurugan, A.; Kannan, S.; Rajeswari, S. Structural and electrochemical behaviour

of sol–gel zirconia films on 316L stainless-steel in simulated body fluid environment.

Mater. Lett. 2003, 57, 4202–4205.

35. D�ıaz-Garc�ıa, M.E.; La�ınno, R.B. Molecular imprinting in sol–gel materials: recent

developments and applications. Microchim. Acta 2005, 149, 19–36.

36. Kumar, A.; Gupta, R. Bioactive materials for biomedical applications using sol–gel

technology. Biomed. Mater. 2008, 3, 1–16.

37. Bhatia, R.B.; Brinker, C.J.; Gupta, A.K.; Singh, A.K. Aqueous sol–gel process for protein

encapsulation. Chem. Mater. 2000, 12, 2434–2441.

38. Collinson, M.M. Recent trends in analytical applications of organically modified silicate

materials. Trends Anal. Chem. 2002, 21, 31–39.

39. Power, M.; Hosticka, B.; Black, E.; Daitch, C.; Norris, P. Aerogels as biosensors: viral

particle detection by bacteria immobilized on large pore aerogel. J. Non-Crystalline Solids

2001, 285, 303–308.

40. Tiwari, A.; Mishra, A.P.; Dhakate, S.R.; Khan, R.; Shukla, S.K. Synthesis of electrically

active biopolymer–SiO2 nanocomposite aerogel. Mater. Lett. 2007, 61, 4587–4590.

41. Li, Y.K.; Chou, M.J.; Wu, T.-Y.; Jinn, T.-R.; Chen-Yang, Y.W. A novel method for

preparing a protein-encapsulated bioaerogel: using a red fluorescent protein as a model.

Acta Biomater. 2008, 4, 725–732.

42. Scanlon, S.; Aggeli, A.; Boden, N.; Koopmans, R.; Brydson, R.; Rayner, C.M. Peptide

aerogels comprising self-assembling nanofibrils. Micro Nano Lett. 2007, 2, 24–29.

43. Yueming, Z.; Chakrabartty, S.; Muhammad-Tahir, Z.; Pal, S.; Alocilja, E.C. Spatio-

temporal processing for multichannel biosensors using support vector machines. IEEE

Sens. J. 2006, 6, 1644–1651.

44. Song, X.; Kaylor, R. Membrane-Based Assay Device, Kimberly-Clark Worldwide, Inc:

Neenah, WI, 2002; p 24.

45. Kueng, A.; Kranz, C.; Mizaikoff, B. Imaging of ATPmembrane transport with dual micro-

disk electrodes and scanning electrochemical microscopy. Biosens. Bioelectron. 2005, 21,

346–353.

46. Shan, D.; He, Y.;Wang, S.; Xue,H.; Zheng,H.A porous poly(acrylonitrile-co-acrylic acid)

film-based glucose biosensor constructed by electrochemical entrapment. Anal. Biochem.

2006, 356, 215–221.

REFERENCES 87

47. Ekanayake, E.M.I.M.; Preethichandra, D.M.G.; Kaneto, K. Fabrication and characteriza-

tionof nanostructured conductingpolymer electrodes forglucosebiosensor applications. In

Second International Conference on Industrial and Information Systems, 2007, pp. 63–66.

48. Park, J.; Kim, H.K.; Son, Y. Glucose biosensor constructed from capped conducting

microtubules of PEDOT. Sens. Actuators B: Chem. 2008, 133, 244–250.

49. Darder, M.; Aranda, P.; Hern�andez-V�elez, M.; Manova, E.; Ruiz-Hitzky, E. Encapsulation

of enzymes in alumina membranes of controlled pore size. Thin Solid Films 2006, 495,

321–326.

50. Shimomura, T.; Itoh, T.; Sumiya, T.; Mizukami, F.; Ono, M. Amperometric determination

of cholinewith enzyme immobilized in a hybridmesoporousmembrane. Talanta 2009, 78,

217–220.

51. Kim, H.-J.; Yoon, S.H.; Choi, H.N.; Lyu, Y.-K.; Lee, W.-Y. Amperometric glucose

biosensor based on sol–gel-derived zirconia/Nafion composite film as encapsulation

matrix. Bull. Korean Chem. Soc. 2006, 27, 63–70.

52. De Stefano, L.; Rotiroti, L.; Rendina, I.;Moretti, L.; Scognamiglio, V.; Rossi,M.; D’Auria,

S. Porous silicon-based optical microsensor for the detection of L-glutamine. Biosens.

Bioelectron. 2006, 21, 1664–1667.

53. Bronzino, J.D.TheBiomedical EngineeringHandbook, BocaRaton, FL:CRCPress, 2000.

54. Liu, Y.; Zhu, Y.; Zeng, Y.; Xu, F. An effective amperometric biosensor based on gold

nanoelectrode arrays. Nanoscale Res. Lett. 2009, 4, 210–215.

55. Shishkanova, T.V.; Volf, R.; Krondak, M.; Kr�al, V. Functionalization of PVC membrane

with ss oligonucleotides for a potentiometric biosensor. Biosens. Bioelectron. 2007, 22,

2712–2717.

56. Reddy, R.R.K.; Basu, I.; Bhattacharya, E.; Chadha, A. Estimation of triglycerides by a

porous silicon based potentiometric biosensor. Curr. Appl. Phys. 2003, 3, 155–161.

57. Reddy, R.K.K.; Prasad, S.; Barrett, T.; Carruthers, J. Electrical immunoassays toward

clinical diagnostics: identification of vulnerable cardiovascular plaque. JALA 2008, 33–39.

58. De Stefano, L.; Arcari, P.; Lamberti, A.; Sanges, C.; Rotiroti, L.; Rea, I.; Rendina, I. DNA

optical detection based on porous silicon technology: from biosensors to biochips. Sensors

2007, 7, 214–221.

59. Bonanno, L.; DeLouise, L. Steric crowding effects on target detection in an affinity

biosensor. Langmuir 2007, 23, 5817–5823.

60. Xue, W.; Cui, T. Carbon nanotube micropatterns and cantilever arrays fabricated with

layer-by-layer nano self-assembly. Sens. Actuators A: Phys. 2007, 136, 510–517.

61. Payen, S.;Walther, D.C.; Pisano, A.P. AMEMSRF-interrogated biosensor. In Third IEEE/

EMBS Special Topic Conference on Microtechnology in Medicine and Biology, 2005, pp.

76–79.

62. Alocilja, E.C. Biosensors for detecting pathogenic bacteria in the meat industry. Meat

Biotechnol. 2008, 335–359.

63. Wang, L.; Liu, Q.; Hu, Z.; Zhang, Y.; Wu, C.; Yang, M.; Wang, P. A novel electrochemical

biosensor based on dynamic polymerase-extending hybridization for E. coli O157:H7

DNA detection. Talanta 2009, 78, 647–652.

64. Abdel-Hamid, I.; Ivnitski, D.; Atanasov, P.; Wilkins, E. Flow-through immunofiltration

assay system for rapid detection of E. coli O157:H7. Biosens. Bioelectron. 1999, 14,

309–316.


65. Liu, Y.; Brandon, R.; Cate, M. Sensitive and rapid detection of pathogens using lumines-

cent CdSe/ZnS dendron-nanocrystals and a porous membrane immunofilter. Anal. Chem.

2007, 79, 8796–8802.

66. Zhang, D.; Alocilja, E.C. Characterization of nanoporous silicon-basedDNAbiosensor for

the detection of Salmonella enteritidis. IEEE Sens. J. 2008, 8, 775–780.

67. Riccardi, C. d.S.; Kranz, C.; Kowalik, J.; Yamanaka, H.; Mizaikoff, B.; Josowicz, M.

Label-free DNA detection of hepatitis C virus based on modified conducting polypyrrole

films at microelectrodes and atomic force microscopy tip-integrated electrodes. Anal.

Chem. 2008, 80, 237–245.

68. Rossi, A.M.; Wang, L.; Reipa, V.; Murphy, T.E. Porous silicon biosensor for detection of

viruses. Biosens. Bioelectron. 2007, 23, 741–745.

69. Gyurcs�anyi, R.E. Chemically-modified nanopores for sensing. Trends Anal. Chem. 2008,

27, 627–639.

70. Reichmuth, D.S.;Wang, S.K.; Barrett, L.M.; Throckmorton, D.J.; Einfeld,W.; Singh, A.K.

Rapid microchip-based electrophoretic immunoassays for the detection of swine influenza

virus. Lab Chip 2008, 8, 1319–1324.

71. Saha, S.; Arya, S.K.; Singh, S.P.; Sreenivas, K.; Malhotra, B.D.; Gupta, V. Nanoporous

cerium oxide thin film for glucose biosensor. Biosens. Bioelectron. 2009, 24, 2040–2045.

72. Wei, W.-Z.; Zhai, X.-R.; Zeng, J.-X.; Gao, Y.-P.; Gong, S.-G. New amperometric glucose

biosensor by entrapping glucose oxidase into chitosan/nanoporous ZrO2/multiwalled

carbon nanotubes nanocomposite film. J. Central South Univ. Technol. 2007, 14, 73–77.

73. Wang, J.; Thomas, D.F.; Chen, A. Nonenzymatic electrochemical glucose sensor based on

nanoporous PtPb networks. Anal. Chem. 2008, 80, 997–1004.

74. Vamvakaki, V.; Chaniotakis, N.A. Immobilization of enzymes into nanocavities for the

improvement of biosensor stability. Biosens. Bioelectron. 2007, 22, 2650–2655.

75. Ekanayake, E.M.I.M.; Preethichandra, D.M.G.; Kaneto, K. Polypyrrole nanotube array

sensor for enhanced adsorption of glucose oxidase in glucose biosensors. Biosens.

Bioelectron. 2007, 23, 107–113.

76. Fink, D.; Klinkovich, I.; Bukelman, O.; Marks, R.S.; Kiv, A.; Fuks, D.; Fahrner, W.R.;

Alfonta, L. Glucose determination using a re-usable enzyme-modified ion trackmembrane

sensor. Biosens. Bioelectron. 2009, 2418, 2702–2708.

77. Li, J.; Peng, T.; Peng, Y. A cholesterol biosensor based on entrapment of cholesterol

oxidase in a silicic sol–gel matrix at a Prussian Blue modified electrode. Electroanalysis

2003, 15, 1031–1037.

78. Li, J.-P.; Gu, H.-N. A selective cholesterol biosensor based on composite film modified

electrode for amperometric detection. J. Chin. Chem. Soc. 2006, 53, 575–582.

79. Singh, S.P.; Sunil, K.A.; Pratibha, P.; Malhotra, B.D.; Shibu, S.; Sreenivas, K.; Vinay, G.

Cholesterol biosensor based on rf sputtered zinc oxide nanoporous thin film. Appl. Phys.

Lett. 2007, 91, 063901.

80. Arya, S.K.; Solanki, P.R.; Singh, S.P.; Kaneto, K.; Pandey,M.K.; Datta,M.;Malhotra, B.D.

Poly-(3-hexylthiophene) self-assembled monolayer based cholesterol biosensor using

surface plasmon resonance technique. Biosens. Bioelectron. 2007, 22, 2516–2524.

81. Ansari, A.A.; Kaushik, A.; Solanki, P.R.; Malhotra, B.D. Sol–gel derived nanoporous

cerium oxide film for application to cholesterol biosensor. Electrochem. Commun. 2008,

10, 1246–1249.

REFERENCES 89

82. Song, M.-J.; Yun, D.-H.; Min, N.-K.; Hong, S.-I. Electrochemical biosensor array for liver

diagnosis using silanization technique on nanoporous silicon electrode. J. Biosci. Bioeng.

2007, 103, 32–37.

83. Song, M.-J.; Yun, D.-H.; Min, N.-K.; Hong, S.-I. Electrochemical biosensor array for

liver diagnosis using a silanization technique on nano-porous silicon electrode.

In Nanotechnology Materials and Devices Conference, NMDC 2006, IEEE, 2006,

pp. 362–363.

84. Yang, Z.; Si, S.; Dai, H.; Zhang, C. Piezoelectric urea biosensor based on immobilization

of urease onto nanoporous alumina membranes. Biosens. Bioelectron. 2007, 22,

3283–3287.

85. Yang, Z.; Si, S.; Zhang, C. Study on the activity and stability of urease immobilized

onto nanoporous alumina membranes. Microporous Mesoporous Mater. 2008, 111,

359–366.

86. Zaytseva, N.V.; Goral, V.N.; Montagna, R.A.; Baeumner, A.J. Development of a micro-

fluidic biosensor module for pathogen detection. Lab Chip 2005, 5, 805–811.

87. Joo, S.; Park, S.; Chung, T.D.; Kim, H.C. Integration of a nanoporous platinum thin film

into a microfluidic system for non-enzymatic electrochemical glucose sensing. Anal. Sci.

2007, 23, 277–281.

88. Metz, S.; Trautmann, C.; Bertsch, A.; Renaud, P. Polyimide microfluidic devices with

integrated nanoporous filtration areas manufactured by micromachining and ion track

technology. J. Micromech. Microeng. 2004, 14, 324–331.

89. Maeng, J.-H.; Lee, B.-C.; Ko, Y.-J.; Cho,W.; Ahn, Y.; Cho, N.-G.; Lee, S.-H.; Hwang, S.Y.

A novel microfluidic biosensor based on an electrical detection system for alpha-fetopro-

tein. Biosens. Bioelectron. 2007, 23, 1319–1325.

90. Goral, V.N.; Zaytseva, N.V.; Baeumner, A.J. Electrochemical microfluidic biosensor for

the detection of nucleic acid sequences. Lab Chip 2006, 6, 414–421.

91. Kwakye, S.; Goral, V.N.; Baeumner, A.J. Electrochemical microfluidic biosensor for

nucleic acid detection with integrated minipotentiostat. Biosens. Bioelectron. 2006, 21,

2217–2223.

92. Yu, J.; Liu, Z.; Liu, Q.; Yuen, K.T.;Mak, A.F.T.; Yang,M.; Leung, P. A polyethylene glycol

(PEG) microfluidic chip with nanostructures for bacteria rapid patterning and detection.

Sens. Actuators A: Phys. 2009, 154(2), 288–294.

93. Vargas-Bernal, R. Topologies based on microfluidics of pesticides biosensors. In

Proceedings of the Electronics, Robotics and Automotive Mechanics Conference,

IEEE Computer Society, 2007.

94. Nikolelis, D.P.; Raftopoulou, G.; Nikoleli, G.-P.; Simantiraki, M. Stabilized lipid mem-

brane based biosensors with incorporated enzyme for repetitive uses. Electroanalysis

2006, 18, 2467–2474.

95. Parkinson, J.; Gordon, R. Beyond micromachining: the potential of diatoms. Trends

Biotechnol. 1999, 17, 190–196.

96. De Stefano, L.; Lamberti, A.; Rotiroti, L.; De Stefano, M. Interfacing the nanostructured

biosilica microshells of the marine diatom Coscinodiscus wailesii with biological matter.

Acta Biomater. 2008, 4, 126–130.


3NANOPARTICLE-BASEDMICROFLUIDIFIC BIOSENSORS

GIOVANNA MARRAZZA

Dipartimento di Chimica,Univesit�a di Firenze, Via della Lastruccia, SestoFiorentino, Italy

3.1 INTRODUCTION

Over the past decade, many important advances have been made in the use of

nanotechnology for biomolecular detection. The use of nanoscale materials for

biosensing has seen explosive growth in recent years following the discovery of

carbon nanotubes (CNTs) by Sumio Ijima in 1991. Recently, many advances were

achieved in the electrochemical and optical detection of DNA and immunoreactions,

through the use of innovative detection schemes (e.g., microfluidic platform) and new

materials, particularly the use of nanoparticles (NPs), nanotubes and nanowires. This

advanced technology has been extended throughout the field of biosensors and

biochips. Specifically, nanoparticles, made from metals, semiconductor, carbon,

and polymeric materials, have been widely investigated to enhance the reaction

signal of bioreceptors such as enzymes, antibodies, and oligonucleotides.

Use of nanoparticle labels has proved to be particularly advantageous in sensing

andbiosensing applications due to the fact that single biorecognition events (i.e., DNA

hybridization or immunoreaction) are typically translated into a significant effect on

its optical (change of the light absorption or emission) or electrochemical properties

(oxidationor reductioncurrent) onto a transducingplatform, offeringnovel options for

bioanalysis. Moreover, the application of nanoparticles in biosensors strongly relates

to their properties that derive to a certain extent from synthesis and later modifications

(chemical and biological).


91

Although the use of NPs in bioanalysis is a recent area of research, there are many

publications on their use as immobilization platform or labels for detecting numerous

analytes.

In the past few years, several excellent reviews have been published on the

application of nanoparticles1,2 and particularly on the use of metal nanoparticles

(gold nanoparticles)3,4 for the improvement of biosensing performance.

Tamanaha et al.5 report an overview of the various approaches developed for

magnetic labeling and detection as applied to biosensing.

Moreover, nanoparticles have been applied widely to microanalytical systems

(lab on a chip). For a biosensor to be used outside the laboratory, it either has to be as

simple or automated with regard to sample processing and reagent addition.

Automation and miniaturization of biological analytical techniques, as well as the

development of on-line and remotemeasurement equipment, can be achieved through

biosensor technology. Microfluidic systems are well recognized for their ability to

move small volumes of fluids through different processes and over a sensing surface.

The use of microfluidic technology presents many advantages, for example, cost

reducing because of reduced fabrication expense and decreased requirements for

costly reagents, reducing process and assay times that are proportional to liquid

volumes. Moreover, the microfluidic devices can be realized in different steps of

analytical procedure: (1) the microfluidic devices can be used for target preconcen-

tration and target separation from other sample components, (2) the sample pretreat-

ment (e.g., cell disruption or sample homogenization) can be performed in a

microfluidic system, (3) both active and passive mixers are available for combining

sample and assay reagents, (4) techniques for using solid-phase materials for

separations in microflow have been identified, and finally (5) on-chip temperature

control is available for temperature-dependent reactions such as PCR or simply for

maintaining stability of the system in harsh environments.

Therefore, microfluidic devices have found great application in the fields of

biosensor technology. An overview has been published by Merkoci and coworkers

on important aspects of microfluidic chip platforms as new materials for electro-

chemical sensing. It describes important trends in constructing detectors and

their electronics on microfluidic chip platforms, the importance of selecting appro-

priate detector materials, and the different detection modes in on-chip amperometry,

conductometry, and potentiometry.6 Rios et al. have presented a general overview on

the potential of analytical microsystems. They discuss the issues involved in the

analytical process and the different steps involved in chemical analysis. Moreover,

they identify challenges in applying analytical microsystems to these uses.7

Synergism between nanoparticle-based and microfluidics technologies may bring

to new miniaturized analytical devices for multiple targets in bioanalytical assays.8

Choi et al. have reviewed important nanotechnologies such as the application of

nanoparticles for the detection of biomolecules, the immobilizationof biomolecules at

nanoscale, nanopatterning technologies, and the microfluidic system for molecular

diagnosis.9

In a recent study, Ligler has discussed emerging science and technology that

will enable the creation of more efficient application-specific optical biosensors.

92 NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

Biological recognition and signal amplification strategies, nanotechnology for

geometric control of the biochemistry and signal enhancement, microfluidic for

automated reagent delivery and reaction control, and emergence of optical elements

amenable to improved systems integration will play a critical role in this evolution.

The use of nanoparticles therefore allows miniaturization of biosensor, develop-

ment of microfluidic systems, and increase in the sensitivity of bioassays.10

Microfluidic approaches provide one of the most promising strategies to interface

nanoengineered biosensors in a wide spectrum of clinical and biomedical applica-

tions. Most biomedical samples are naturally in a liquid environment, so the sensors

must be combined with compelling fluid handling systems. Appropriate microfluidic

delivery systemscanbeused to eliminate contamination,minimize analysis times, and

enable portable systems.

Here, without pretending to being exhaustive, the most recent applications of

nanoparticles in microfluidic format for electrochemical and optical affinity biosen-

sors have been reported, highlighting some of their technical challenges and the new

trends by means of a set of selected recent applications.

3.1.1 Microfluidic in Bioassays

In recent years, the need for microfluidic devices has driven their continuous

development. Microfluidic-based microchips can improve analytic efficiency by

reducing analysis time and sample volume, increase both sensitivity and selectivity,

and allow for miniaturization of analytic devices. Because of these advantages,

microchips are widely used in clinical diagnosis, environmental toxicant detection,

bimolecular separation, and cell handling systems.

Microfluidic involves the manipulation, transport, and analysis of fluids in

micrometer-sized channels.

Integration into flow injection systems, capillary electrophoresis, and microfluidic

platforms is just the latest, logical step in the direction of automation. All these trends

are important and should occur in parallel in the future development of bioassay

methods for medical and clinical applications.

One of themost important issues to developmicrofluidic devices is the selection of

material because surface effects become enlarged as the device is miniaturized.

Thematerials for a microfluidic biodevice ought to be considered by the following

issues as well as process ability: (1) chemical stability; (2) price, for disposable

device to minimize contamination; (3) surface properties for biofouling; and

(4) thermal stability for nucleic acid amplification such as PCR. Other properties,

like optical transparency, can be preferred for testing modules. For example, for

optical transducers, glass-utilizing devices are preferred over silicon devices owing to

their optical properties.

Silicon, glass, and polymers are the three main types of materials used for

microfluidic fabrication. Although metals are one of the most widely used materials

in industries, many limitations in micromachining prevented the extensive use of

metal. The micro- and nanodimensions required by these devices can only be easily

fabricated with semiconductor technology; thus, silicon became one of the first

INTRODUCTION 93

materials to be used in the early 1990s. However, silicon is opaque and this prevented

the use of fluorescent labels for detection, which are very popular with immunoassays.

Biological molecules also tend to adsorb to silicon surfaces and these limitations

prompted the search of other fabrication materials. Naturally, glass became the next

material as it is transparent to nearly all absorption and emission wavelengths of

fluorescent labels. However, the difficult fabrication techniques and toxic chemicals

involved did not make glass a popular choice among researchers and manufacturers.

Recently, researchers have turned their attention to the use of polymers. Polymers

offer the advantages of being optically clear, non-toxic, and low cost. In addition,

easy fabrication techniques and a variety of surface modification methods are

available to improve the efficiency of these devices. Polycarbonate, polymethylmetha-

crylate (PMMA), polyethylene, polypropylene, and polystyrene are some examples of

polymers used widely in all fields of research and industries. One of the most

extensively used polymers in the past few years is polydimethylsiloxane, also known

as PDMS.

The elastomer PDMS has attracted attention as a material suitable for the easy and

rapid fabrication of microfluidic devices using soft lithography. PDMS has a number

of advantages: (1) features on the micrometer scale can be reproduced with high

accuracy, (2) it is optically transparent down to 280 nm, (3) it cures at low tem-

peratures, (4) it is not toxic, and (5) it can seal reversibly to itself and a range of other

materials bymakingmolecular (van derWaals) contact with the surface, or it can seal

irreversibly after exposure to an air plasma by formation of covalent bonds. These

characteristics made PDMS very compatible with biological studies.

Lim and Zhang have reviewed into some fabrication materials and techniques

available for microfluidic and have elaborated on the advantages of these devices for

immunoassays.11

3.1.2 Nanoparticles Used in Bioassays

Disease biomarkers and biological agents are often present at ultralow levels and

require ultrasensitivemethods for detection. Different strategies have been employed

for amplifying the transducing signals of bioassays. Most conventional amplification

strategies have relied on the use of labels, such as enzymes, electroactive molecules,

redox complexes, and metal ions. The emergence of nanotechnology is opening new

horizons for the use of nanomaterial labels in signal amplification. The applications of

nanomaterials in bioassays can be classified into two groups according to their

functions: (1) nanomaterial modified electrochemical transducers to facilitate

antibody/acid nucleic immobilization or improve properties of transducers and

(2) nanomaterial–bimolecular conjugates as labels for bioassays. In particular,

nanomaterial labels are showing the greatest promise for developing ultrasensitive

bioassays. Antibodies or nucleic acids labeled with nanomaterials can retain their

bioactivity and interact with their counterparts, and based on the electrochemical and

optical detection of those nanomaterials, the amount or concentration of analytes can

be determined. The enormous signal enhancement associated with the use of

nanomaterial amplifying labels provides the basis for ultrasensitive bioassays.11,12


Therefore, NPs have been used extensively in affinity biosensors for the detection

of nucleic acids and proteins. These particles are unique because their nanometer size

gives rise to a high reactivity and beneficial physical properties (electrical, electro-

chemical, optical, and magnetic) that are chemically tailorable.

The composition of NPs determines the compatibility and the suitability of the

probes with analytes and many assays are possible. Table 3.1 shows some of the NPs

that have been more frequently described in analytical biosensors, together with the

detection systems used.

The noble metal NPs, mainly gold nanoparticles (AuNPs), have been the most

extensively used for this purpose involving optical or electrochemical detection. They

exhibit bright colors due to the presence of a plasmon absorption band that is not

present in the spectrum of the bulk metal, which is a result of the resonance of the

incident photon frequencywith the collective excitation of the conductive electrons of

the particle. This effect is termed localized surface plasmon resonance (LSPR) and

depends on the size, shape and composition of the nanoparticles, the distance between

nanoparticles, and the refractive index of the environmental medium. Another feature

of these nanoparticles is their capability to produce surface-enhanced Raman scatter-

ing (SERS) effects. The different and interesting properties of AuNPs have been

widely explored in bioassays using a variety of detection systems.

The redox properties of AuNPs have led to their widespread use particularly as

electrochemical labels in protein and nucleic acid detection, with numerous config-

urations being explored.3

The application of underpotential deposition of Ag monolayer as a means of

enhancing the electrochemical sensitivity of biomolecular reaction has been success-

fully applied to the development of a new bioaffinity platform via metal-enhanced

electrochemical detection (MED sensors). This is based on the discovery that

immobilized metal layer, as continuous film, particle, colloids, or monolayer, signifi-

cantly amplifies the electrochemical signals,while reducing the reorganization energy

TABLE 3.1 Nanoparticles and Detection Systems Commonly Used in Analytical

Bioassays

Nanoparticles Detection systems

Noble metals (Au, Ag, Pt) Photometry, fluorimetry, Rayleigh and Raman scattering,

surface plasmon resonance, potentiometry, amperometry,

conductimetry, stripping voltammetry, quartz crystal

microbalance

Quantum dots Photometry, fluorimetry, FRET, stripping voltammetry

Silica or polystyrene

Dye doped Fluorimetry, phosphorimetry

Lanthanide chelate doped Fluorimetry, FRET

Ruthenium chelate doped Electrogenerated chemiluminescence

Carbon nanotubes Electrochemical

Dendrimers Fluorimetry

Source: Adapted from Ref. 2.

INTRODUCTION 95

following molecular recognition at sensor. Silver enhancement scheme has been

utilized in the recent biobarcode approach and provided the lowest detection limit to

date for both DNA (500 zM, PCR less) and protein targets ((3–30 aM).

Quantum dots (QDs) are inorganic semiconductor nanocrystals with interesting

luminescent and electrochemical properties extensively used in numerous bioassays.

Briefly, these NPs show broad excitation profiles and narrow emission peaks and can

emit in a range ofwavelengths by changing their size and composition.Also, they lack

photobleaching and have long fluorescence lifetimes. However, QDs can show

blinking characteristics when they are excited with high-intensity light, which could

be a limiting factor for fast scan systems, such as flow cytometry. Other limitations are

toxicity, size variation, agglomeration, and nonspecific binding. Surface oxidation of

QDs can occur under combined exposure to aqueous/UV light excitation, which can

lead (e.g., in CdSe-based QDs) to the release of cadmium ions, so that these NPs are

inadequate for in vivo applications, such as in vivo drug delivery assays.However, they

offer better imaging results than those achievedbyorganic dyes in cell-based or tissue-

based drug studies.

A large number of bioassays have used dye-doped silica NPs that consist of

luminescent organic or inorganic species dispersed inside a silica matrix. These NPs

enable significant amplification of the analytical signal due to the numerous dye

molecules inside each NP.

The silica-based NPs functionalized for coupling and containing stable lanthanide

and silica NPs containing ruthenium(II) chelates, mainly tris(2,20-bipyridyl)dichlor-

oruthenium(II) (RuBpy), have also been used in fluorescence resonance energy

transfer (FRET)-based assays andfluorescenceor chemiluminescence (CL)detection.

Carbon nanotubes represent an important group of nanomaterials with attractive

geometrical, electronic, and chemical properties. The structure of CNTs comprises

concentric cylinders, with a diameter of a few nanometers to hundreds of lumens in

length, that have interlinked hexagonal carbon rings. In addition to favorable

electronic properties, they show a large surface area and an electrocatalytic effect

that have been used in constructing electrochemical biosensors.

Conducting polymer nanowires (CPNWs) are attractive alternatives to silicon

nanowires and carbon nanotubes because of their tunable conductivity, flexibility,

chemical diversity, and ease of processing. CPNWs can be prepared using a variety of

protocols, such as chemical synthesis, template electrochemical synthesis, and

electrospinning, and some chemical and biological sensors based on CPNWs have

been reported. Wang et al. introduced a new approach for the in situ electrochemical

fabrication of an individually addressable array of CPNWs positioned within an

integrated microfluidic device and also demonstrated that such an integrated device

can be used as a chemical sensor immediately after its construction.14

There are other NPs that have so far found fewer applications in bioassays. This is

the case for dendrimers that are hyperbranched, tree-like structures having three

different regions (i.e., core, branches, and surface). They have been used in some

bioassays as reagents by adsorbing, caging, or covalently binding active molecules,

such as fluorescent dyes, inside or onto their surface.

Magnetic particles, which respond to an external magnetic field, have been used

extensively for separation and preconcentration purposes and in electrochemical


biosensors. Their unique properties allow magnetic particle-conjugated molecules to

be quickly agglomerated or resuspended in the medium according to the external

magnetic force, thus making them suitable for purifying biologically active com-

pounds, such as nucleic acids and proteins.5 They comprise a metal or metallic oxide

core, encapsulated in an inorganic or polymeric coating that renders the NPs

biocompatible and stable and that may serve as a support for biomolecules. Also,

the development of hybrid NPs, such as magnetic AuNPs or magnetic QDs, which

combine samplemanipulation and sensitive detection, is a very promising recent field

of research.

Another nanoparticles used in biosensor technology are liposomes. Liposomes

are composed of a lipid bilayer with the hydrophobic chains of the lipids forming the

bilayer and the polar headgroups of the lipids oriented toward the extravesicular

solution and the inner cavity. The sizes of the liposomes vary, ranging from

nanometers to several micrometers, which depends on the synthesis conditions.

Owing to its high surface area, large internal volume, and capability to conjugate

bilayer lipids with a variety of biorecognition elements, liposomes have beenwidely

used as bioassay labels by encapsulating enzymes, fluorescent dyes, electrochemi-

cal and chemiluminescent markers, DNA, RNA, ions, and radioactive isotopes. An

excellent review of the uses of liposomes in immunoassays is available in the

literature.15

There is a clear trend toward developing multiplexed bioassays using nanobar-

codes, which are based on the synthesis of particles that contain a mixture of NPs

functionalized with the corresponding recognition agents. These agents can be

antibodies or oligonucleotide sequences that recognize the targets of interest in

protein or nucleic acid detection, respectively. In fact, it is often necessary to monitor

or quantize several components in a complex system. For example, due to the limited

specificity and sensitivity of biomarkers for clinical diagnosis, the measurement of a

single biomarker is usually insufficient for diagnostic purpose. Some studies have

shown that the measurement of biomarkers panel can avoid false positive or false

negative results to improve their diagnostic value. Traditionally, bioassay of analytes

panel is performed as discrete tests, that is, one analyte per assay run, and several runs

are needed to detect all components in a complex system.Great consumptions of time,

reagent, and labor limit the application. To overcome these limitations, multiplexed

bioassays can measure two or more analytes in a single run. Compared to parallel

single analyte methods, multiplexed bioassay offers some remarkable advantages,

such as high sample throughput, improved assay efficiency, low sample consumption,

and reduced overall cost per assay.16

3.2 FUNDAMENTALS OF BIOSENSORS

A biosensor can be defined as an integrated receptor transducer device that is capable

of providing selective quantitative or semiquantitative analytical information using a

biological recognition element. A biosensor converts a biological event into

a detectable signal by the action of a transducer and a processor. The usual aim of

a biosensor is to produce either discrete or continuous digital electronic signals that are

FUNDAMENTALS OF BIOSENSORS 97

proportional to a single analyte or a related group of analytes. The principal

components and functions of biosensors are shown in Figure 3.1.

In a biosensor, the physical–chemical transformation due to the interaction

between the biological element and the analyte target is converted into a usable

signal by the transducer. The main purpose of the recognition system is to

provide specificity to the biosensor, thus creating a device capable to detect either

a specific target or a related family of compounds.

Immobilization of the biological component can be performed using a variety of

methods such as chemical or physical adsorption, physical entrapment within a

membrane or gel, cross-linking of molecules, or covalent binding.

In general, biosensors are classified by either their biological element or the

transducer used.

Biosensors can be subdivided into two classes based on the type of biorecognition

molecule. Catalytic biosensors employ enzymes and microorganisms as the biore-

cognition molecule that catalyses a reaction involving the analyte to give a product.

Common analytes for catalytic biosensors are small organic molecules such as

glucose.

The other category of biosensors is affinity biosensors. Biorecognition molecules

commonly used in affinity biosensors include antibodies, DNA, peptides, and lectins.

Affinity biosensors are characterized by a binding event between the biorecognition

molecule and the analyte (the affinity reaction), often with no further reaction

occurring. Hence, the challenge then becomes transducing the biorecognition event.

As this class of biosensor is compatible with the detection of virtually all biological

agents, it is this challenge that faces researchers attempting todevelopportabledevices

for detecting toxins, microbes, and viruses.

Transduction of affinity biosensors has been achieved using labeled species and

label-free approaches. If transduction is achieved using labeled species, the principles

are very similar to immunoassay, with the amount of analyte detected being inferred

from the amount of label that binds to the interface. Label-free methods most

frequently involve evanescent wave-based optical methods or using mass-sensitive

acoustic wave devices that monitor molecules binding to, or desorbing from, a

transducer surface.

The most common transducers for detecting labeled species are optical, where an

optically active label is detected, or electrochemical, where the label is electroactive.

FIGURE 3.1 Scheme of biosensors.


Electrochemical transduction transforms the effect of electrochemical interaction

between analyte and electrode into a primary signal. Such effects may be stimulated

electrically or result in a spontaneous interaction at the zero current condition.

Voltammetric, amperometric, potentiometric, solid electrolyte, gas sensor, and

chemically sensitized field effect transistor (CHEMFET) are distinct subgroups.

Optical transducers are basedonvarious technologies involvingoptical phenomena

that are the result of an interaction of analytewith receptor. This groupmay be further

subdivided according to the optical properties that have been applied in sensing

(i.e., absorbance, reflectance, luminescence, fluorescence, refractive index, surface

plasmon resonance (SPR), optothermal effect, and light scattering).A large number of

optical transduction techniques can be used for biosensor development.

In the case of both electrochemical and optical methods, one of the key factors that

limits biosensor performance is nonspecific binding. The problem of nonspecific

binding highlights the importance of interfacial design in a biosensor.

3.2.1 Affinity Biosensors

Affinitybiosensors are a subclassof biosensors: theyare analytical devices comprising

a biological or biomimetic affinity element (receptor). The sensing element is a highly

specific receptor, and it is generally biologic, for example, a receptor of natural origin

(bioreceptors): enzymes, antibodies, and nucleic acids. In the past few years, a great

interest in biomimetic biosensors has risen: they use artificial or semiartificial

receptors. This class of synthetic receptors includes PNA (peptide nucleic acid),

LNA (locked nucleic acid),MIPs (molecular imprinted polymers), oligopeptides, and

aptamers.

In this chapter, immunosensors and DNA biosensor using electrochemical and

optical transducers have been considered.

3.2.2 Immunoassay

Immunoassays are currently the predominant analytical technique for the quantitative

determination of a broad variety of analytes in clinical,medical, biotechnological, and

environmental significance.

The recognition elements are immunochemical antibody–antigen (Ab–Ag) inter-

actions. This type of device combines the principles of solid-phase immunoassaywith

physical–chemical transduction elements (electrochemical, optical, piezoelectric,

EW, and SPR).

3.2.2.1 AntibodiesThe use of highly specific antibodies is very popular not only in biosensor research but

also in bioanalytical chemistry. The most important applications of antibodies are

immunoassay, immunosensor, and immunoaffinity columns.

Antibodies are serum molecules produced by B lymphocytes; they represent the

soluble form of specific receptors for the antigen expressed by B lymphocytes.


In plasma subjected to electrophoresis, antibodies are found as proteins related to

the gamma fraction, the highest in molecular weight (globulins) portion. Hence, the

protein fraction that contains antibodies is usually called immunoglobulin or Ig.

Different types of Ig exist: IgG, IgA, IgM, IgD, and IgE. All the antibodies have the

same base structure, but they differ in the region that binds the antigen.

Immunoglobulin G (IgG) is the most abundant immunoglobulin species in

serum and also the most commonly used antibody in sensor applications. The

molecule consist of four polypeptide chains, two identical heavy (H) chains and

two identical light (L) chains joined to form a “Y”-shaped unit (Figure 3.2a). The

length of the two chains is 450 amino acids for the H chain and 212 amino acids

FIGURE 3.2 (a) Scheme of a conventional IgG antibody. (b) Different bioassay formats:

(A) direct label-free detection of the target protein binding to immobilized antibodies,

(B) detection of labeled target proteins, (C) competitive assay, and (D) sandwich assay.

Adapted from Ref. 28.


for the L chain. The two identical H chains are connected via disulphide bridges.

The connection between the L chain and the H chain also consists of disulphide

bonds. Since all these bonds connect two chains, they are named interchain

disulphide bridges. Both chains, L and H, also have intrachain disulphide bridges.

The globular structure of the protein that is responsible for the name immuno-

globulin is a result of these intrachain bonds.

The binding site (paratope) is located within the VH and VL domains and each arm

contains one binding site. In the variable regions, amino acid sequences canvary from

antibody to antibody and allow the specific adaptation to certain antigens.

The exact regions within these variable regions that have very high amino acid

variability are called hypervariable regions, also known as complementary determin-

ing regions (CDRs). Three CDRs are integrated into the L chain and three into the H

chain, resulting in six CDRs for each arm.

Amolecule used to induce an immune response is called an immunogen. Anantigen

is a molecule that can bind to an antibody. The region of the antigen that binds to an

antibody is called epitope or antigenic determinant, and the corresponding part in the

antibody is named paratope. Antigens can be big or small molecules, but the small

moleculesareantigeniconlywhentheyarecoupledtoaproteincarrier.Suchcompounds

are namedhaptens. Themost used protein carriers are thebovine serumalbumin (BSA)

and ovalbumin (OVA).A spacer bridge is used to conjugate the hapten to the protein, in

order tohave thecorrectdistancebetweenthemandtoallowthecorrect immunizationof

the organismagainst the target. The spacer bridgeneeds to be amolecular chain lacking

in substituents groups to avoid immunization against the bridge.

The use of carrier proteins is not restricted to small nonimmunogenic compounds;

in fact, small proteins of low native immunogenicity have been coupled to carrier

proteins and successfully used to generate antibodies.

3.2.2.2 Antigen/Antibody InteractionAntigenic molecules are surrounded and trapped inside a pocket formed by the light

and heavy chains of an antibody. This is called the combinatory site and is where a

protein can be captured. The union between an antigen and an antibody is the result of

noncovalent interactions between the amino acidic residuals of the antigens and of the

combinatory site of the antibody. These bonds areweak interactions of different types

such as hydrogen bonds, electrostatic forces, van der Waals forces, and hydrophobic

interactions, but they all contribute to a binding of relevant energy. The strength of

these bonds depends critically on the distance d between reagent groups. This strength

is proportional to 1/d2 in the case of electrostatic forces and 1/d7 in the case of van der

Waals forces.

The affinity of an antibody toward an antigen can be expressed by the strength of

repulsive and attractive forces. An antibody with high affinity for an antigen fits

perfectly to its specific antigen. The law of mass action can be used to calculate the

affinity given by the equilibrium constant K (equation (3.1)).

AbþAg $ Ab�Ag

K ¼ ka

kd

ð3:1Þ


where [Ab] represents the antibody concentration, [Ag] the antigen concentration

and [Ab–Ag] the antigen–antibody complex concentration. When the antibody is

added in a solution containing the antigen, binding sites of antibody are involved in

the antigen interaction; this reaction is regulated by antigen concentration in the

sample and by the equilibrium constants. This assumption can be used when

antibody and antigen are homogeneous, when the antigen has only one epitope

and the antibody has only one binding site, and the separation between the bound

and free forms is complete.

Affinity constants are a quantitative measurement of the affinity between the two

reagents and range from 106–1012M�1.

3.2.2.3 Conventional ImmunoassaysImmunoassaymakesuseof the specificand sensitiveantibody–antigen interactionand

can be used to determinate either antibody or antigen concentration. To monitor the

interaction, one of the species is conjugated with one of the many available labels. In

general, in an immunoassay, one of the reaction partners, antibody or antigen, is bound

to a surface and the binding of the second immunoreagent is detected bymeasuring the

concentration of the attached label. Various assay formats are used for different

applications.

Immunoassays can be catalogued based on the type of label in these categories:

(1) enzyme immunoassay (EIA); (2) radioimmunoassay (RIA); (3) fluorescent

immunoassay (FIA); (4) enzyme-linked immunosorbent assay (ELISA).

ELISA test is the most commonly used format for analysis. The factors

contributed to the success of this format are speed, sensitivity, selectivity, and

no use of radioactive materials. A typical ELISA is carried out in 96-well

microtiter plate automated plate; shakers, washers, and readers are available.

Immunoassays are well established in many fields of analytical interest for

screening purposes or accurate results; in fact, they are cheap, sensitive, and

allow many simultaneous analyses. Immunoassay procedures have been optimized

for numerous analytes, and commercial kits are available for a wide variety of

analytical compounds.

Immunoassays can be set up in a variety of formats (Figure 3.2b) and the most

important are competition, sandwich, and displacement formats. Themain difference

between the formats is the immobilized species, such as antibody, antigen, or hapten

conjugate, and the number of layers used. The decisionwhich of the formats is used for

a particular analytical problem is influenced by the nature of the analyte, the cost and

availabilityof antigen andantibody, and the required sensitivity.Most of thedeveloped

immunosensors are based on either competitive or sandwich assaywhen applied to the

detection of low (herbicides, toxins) and high (proteins, cells) molecular weight

molecules, respectively.

The sandwich assay is used when the molecular target has multiple epitopes that is

able to bindmore than one antibody at the same time; this is possiblewhen the epitopes

are spatially well separated. In a direct sandwich assay, the first antibody (called

capture antibody) is immobilized on the solid phase, and then the antigen (analyte) is

added. The solid phase is washed to remove the unreacted components and then


incubated with a secondary labeled antibody, able to react with the other epitope.

After the washing step, the amount of secondary antibody bound to solid phase is

determined.

A competitive assay is generally used when the target has only one epitope and, as

the name indicates, a competition reaction between two reagents for a third one exist.

In an indirect competitive assay, the solid phase is modified by the immobilization

of a spacer-linked antigen and labeled antibody and free antigen (analyte) are added.

The solid-phase-bound and free antigens, present in the tested solution, compete for

the antibody binding site. The extent of the affinity reaction is detected by adding a

secondary labeled antibody able to bind the first one. When the labeled antibody

concentration is kept constant, an increasing analyte concentration occupiesmore and

morebinding sites.Thismeans that less labeled antibodycanbind to the surface-bound

antigen. Increasing analyte concentrations cause decreasing signals.

In a direct competitive assay, the solid phase is modified by immobilization of the

specific antibody and free and labeled antigens are added. The concentration of

the labeled antigen is kept constant. Free antigen is added in a dilution series or as an

unknown sample. Free and labeled antigens compete for the binding site of the

immobilized antibody. High analyte concentrations occupy most of the binding sites

and results in a low signal. Low analyte concentration allows the labeled antigen to

bind. Increasing analyte concentrations cause a decrease in signal.

3.2.3 Immunosensor

Competitive binding assay formats are most used in immunosensors using optical or

electrochemical transducers.

The main limitation of electrochemical techniques is the detection of the

immunoreactions because it is necessary to use enzymes that will generate electro-

chemically active compounds.

In general, a vast number of optical transduction techniques can be used

for biosensor development. These may employ linear optical phenomenon (e.g.,

adsorption, fluorescence, phosphorescence, and polarization) or nonlinear phenome-

na (e.g., second harmonic generation). The choice of a particular optical method

depends on the analyte and the sensitivity needed.

Optical detection, absorbance, fluorescence, chemiluminescence, or evanescent

wave monitoring, is the most common form of detection used with immunoassays.

Evanescent field-coupled refractometric optical sensors have increased in popularity,

as they monitor immunological hybridization reactions in real time.

In a recent review, the state of the art and the recent developments in immu-

nosensor have been described.17 Homogeneous immunosensor, heterogeneous

immunosensor, integrated immunosensor, and biochip format immunosensor are

presented based on optical, electrochemical, magnetic, or mechanical detection/

transduction systems.

Most of the developed immunosensors include a sensing layer supporting a

particular immobilized antigen or antibody. The solid support used is generally in

closecontactwitha transducerneeded for thedetectionof the formed immunecomplex.


In protein-sensing devices, the immobilized compound determines the specificity

of the device, and the immobilizationmethod frequently influences parameters such as

lower detection limit, sensitivity, dynamic range, reusability, or liability for unspecific

binding.Thus, varieties of immobilizationmethods are described that are applicable to

different supports onto which the compound has to be immobilized (immobilization

substrate).18 The choice of the immobilization substrate depends on the chosen assay

format and detection principle.

Many immunosensors in recent years have used different NPs as labels and have

given rise to the return of interest in metalloimmunoassays, taking into account the

metallic character ofmost of theseNPs. There is a trend to developmultiplexed assays,

but so far most of the immunoassays described only allow individual determinations,

partly because of the cross-reactivity limitations of the antibodies. However, some

recent examples show that NPs are useful for simultaneous determinations, which are

described below. Although other methods have been reported, the assays selected can

give an overview of the usefulness, as well as the versatility and the applicability of

NPs in immunoassays.

3.2.3.1 Electrochemical ImmunosensorElectrochemical immunosensors have been widely used in amperometric, potentio-

metric, and conductimetric configurations.

Over the past few years, DiagnoSwiss has developed polymer devices with the

advantages of microfluidic for bioanalytical applications.19 A new microfluidic

biosensor platform dubbed GRAVI is commercially available for running assays

with paramagnetic nanoparticles. Capture antibody is immobilized on nanoparti-

cles, which can be preincubated with sample, in tube. After incubation, the mixture

is flowed through the microchannels, and the paramagnetic nanoparticles are

trapped near the electrodes by virtue of a magnet array. The biological reactions

(occasionally requiring longer incubation times) can thus be freely adjusted in

function of the assay. Reactions leading to the formation of the immune complex can

be performed in tube or in the microchip, while washing steps and detection of the

enzymatic reaction take place in the microchannels. In this manner, the microchip

merely serves as physical biosensor, in which tests with a variety of microbeads can

be performed in successive runs. The flexibility to consecutively use nanobeads

functionalized with different capturing moieties translates into a full random access

solution.

Characterized by dramatically reduced time to result (<10min) and significantly

decreased sample/reagent consumption, the cost-efficient biosensor instrumentation

allows performing multimenu analysis with minimal laboratory infrastructure.

Coupled to a robotic liquid handler, the system dispenses samples and reagents

from conventional plates or tubes into microchannels of a microchip in which assays

are processed and results readout. As in conventional 96-well microtiter plates, the

microchannels have a standard spacing of 9mm to facilitate automation. With solely

gravity- and capillary force-driven fluidics within the microchannels, liquids are free

to flow, while magnetic beads, functionalized with the antibody of choice, are trapped

near incorporated electrodes byvirtue of amagnet array. Following assayperformance


and electrochemical signal detection in the parallel microchannels, chips are regen-

erated by magnet release and rinsing of beads out from the microchannels.20

Biosensors have emerged as anew technique formonitoring cancerous cells or their

specific interaction with different analytes. Microfluidic-based microchips have

become the focus of research interest for immunoassays and biomarker diagnostics.

This is because theconventional immunoassays require relatively longassay times and

large, complicated detection devices. Some important cancer biomarkers, alpha-

fetoprotein (AFP), carcinoembryonic antigen (CEA), and prostate-specific antigen

(PSA), have been detected by microfluidic-based immunosensing microchips for the

diagnosis of liver, colon, and ovarian cancers, respectively. An excellent review has

been published on the biosensor technology available today, areas that are currently

being developed and researched for cancer markers diagnosis, and a consideration of

future prospects for the technology.21

Maeng et al. have developed and characterized an immunoassay methodology

comprised of microbeads and microbiochips. In this method, microbeads are used to

filter and immobilize antibodies, and an immunogold silver staining (IGSS)method is

then used to amplify electrical signals that correspond to the bound antibodies. The

chip used for this system is composed of an inexpensive and biocompatible PDMS

layer over a Pyrex glass substrate that contains a platinum (Pt) microelectrode used to

detect the electrical signal in this system. The microelectrode is fabricated on the

substrate and a microchannel and pillar-type microfilter is formed in the PDMS layer.

A sandwich immunoassay approach was applied to detect AFP, a cancer biomarker,

using this system.The results of this study showed that the time required for a complete

assaywas reducedby1 handadetection limit as lowas1 ng/mLwasattainedwhen this

system used, which indicates that similar bead-based electrical detection systems

could be used for the diagnosis of many forms of cancer.22

In clinical diagnosis, detection of one biomarker cannot provide sufficient clinical

information for various cancer-related diseases, and the clinical information obtained

from biomarkers is often related to the stage of tumorigenesis, monitoring of

treatment, and the state of the patient. Therefore, it is important to develop the

microchip system with high multiplexing capabilities as well as an efficient detection

method. Researches related to immunosensing microchips have achieved efficient

multiplex detection of biomarkers.

A biobarcode assay (BCA) capable of achieving low detection limits and high

specificity for both protein and DNA targets was developed by Goluch et al.23 The

BCA utilizes AuNPs functionalized with oligonucleotides (the so-called biobarcodes

that serve as surrogate targets and amplifying agents) and a target recognition element

that may be an antibody for protein detection or an unique oligonucleotide sequence

for nucleic acid detection. The BCA also uses functionalizedmagnetic microparticles

(MMPs) adorned with antibodies that bind to the target. In the presence of targets

(protein or oligonucleotide molecules) in solution, the MMPs form a sandwich

complex with targets and gold nanoparticles, which can be localized and collected

under an applied magnetic field. The barcode oligonucleotide molecules are then

chemically released, identified, and quantified. The realization of a BCA in a

microfluidic format presents unique opportunities and challenges. A modified


form of the BCA called the surface immobilized biobarcode assay (SI-BCA) was

developed by the same research group.24 The SI-BCA employs microchannel walls

functionalized with antibodies that bind with the intended targets (Figure 3.3).

Compared to the conventional BCA, it reduces the system complexity and results

in shortened process time, which is attributed to significantly reduced diffusion times

in the microscale channels. Raw serum samples, without any pretreatment, were

evaluatedwith this technique. Prostate-specific antigen in the samples was detected at

concentrations ranging from 40 pM to 40 fM. The entire assay, from sample injection

to final data analysis, was completed in 80min.

Wilson and Nie have developed a microchip for seven cancer biomarkers using an

electrochemical detection method. The above-mentioned microchip-based multiplex

immunosensing devices require a small quantity of sample and are time saving and

convenient. Nevertheless, some shortcomings, such as requiring high power source,

poor reproducibility, and no real-time monitoring, still remain.25 Yoomin and

coworkers have described the development of a microchip-based multiplex

FIGURE 3.3 Schematic diagram of the surface immobilized biobarcode assay protocol.24

(a) The walls of the capture region are coated with antibodies, (b) samples are flowed through

the capture region, (c) the target molecules attach to the antibodies on the channel walls, and

(d) the target proteins are tagged with cofunctionalized nanoparticles containing polyclonal

antibodies and unique barcode DNA oligonucleotides. (e) The barcode DNA is then released

from the nanoparticles and transferred to the detection region where the complementary

sequence is patterned. Steps (f)–(h) illustrate a scanometric detection protocol. (f) The barcode

molecules attach to the complementary sequences in the appropriate regions, (g) universal

nanoparticle probes are attached to the barcode DNA, and (h) the universal probes are silver

stained to facilitate visualization in the visible spectrum. The upper channels represent the

target capture region while the lower channels mark the barcode detection region of the device.

Pneumatic control channels are inserted for directing the flow of fluid.


electroimmunosensing system for simultaneous detection of cancer biomarkers using

gold nanoparticles and silver enhancer. The microchip is composed of biocompatible

poly (PDMS) and glass substrates. To fix the antibody immobilized microbeads, they

used pillar-type microfilters within a reaction chamber. An IGSS method was used to

amplify the electrical signal that corresponded to the immune complex. To demon-

strate this approach, the authors simultaneously assayed three cancer biomarkers,

AFP, CEA, and PSA, on themicrochip. The electrical signal generated from the result

of the immunoreaction was measured and monitored by a PC-based system. The

overall assay time was reduced from 3–8 h to about 55min when compared to

conventional immunoassays. The working range of the proposed microchip was

from 10�3 to 10�1mg/mL of the target antigen.26

Tang et al. have reported a novel method for the detection of tumor markers,

such as a-fetoprotein, CEA, cancer antigen 125 (CA 125), and CA 15-3 can

be found in the body (usually blood or urine) when cancer is present. They

synthesized magnet core/shell NiFe2O4/SiO2 nanoparticles and fabricated an

electrochemical magnetic controlled microfluidic device. The immunoassay sys-

tem consisted of five working electrodes and an Ag/AgCl reference electrode

integrated on a glass substrate. Each working electrode contained a different

antibody immobilized on the NiFe2O4/SiO2 nanoparticle surface and was capable

of measuring a specific tumor marker using noncompetitive electrochemical

immunoassay. Under optimal conditions, the multiplex immunoassay enabled

the simultaneous detection of four tumor markers. The sensor detection limit

was <0.5 mg/L for most analytes.27

The affinity biosensors have been widely applied for foodborne pathogen detec-

tions with the goal to overcome problems associated with traditional microbiological

detection techniques. These are very elaborate, time-consuming, and have to be

completed in a microbiology laboratory and are therefore not suitable for on-

site monitoring. Rapid and reliable detection methods of pathogenic, toxin

producing bacteria suchas Salmonella spp.,Listeriamonocytogenes,Escherichia coli

0157:H7, or Staphylococcus aureus that are responsible of some of the major

worldwide foodborne outbreaks are required. Biosensor-based tools offer the most

promising solutions and address some of the modern-day needs for fast and sensitive

detection of pathogens in real time or near real time. In particular, electrochemical

biosensors for the detection of food pathogens have the advantage of being highly

sensitive, rapid, inexpensive, and amenable toward microfabrication and were re-

viewed by Tamiya and coworkers.28

Asaprinciple of transduction, the impedance techniquehasbeenapplied in thefield

of microbiology as a means to detect and/or quantify foodborne pathogenic bacteria.

The integration of impedance with biological recognition technology for detection of

bacteria has led to the development of impedance biosensors that are finding

widespread use in the recent years. Yang and Bashir have reviewed the progress

and applications of impedance microbiology, particularly the new aspects that have

been added to this subject in the past few years, including the use of interdigitated

microelectrodes, the development of chip-based impedancemicrobiology, and the use

of equivalent circuits for analysis of the impedance systems.29


Some researchers have continuously improved the impedance biosensor methods

by integrating newly developed nanoparticles and microfuidics with interdigitated

microelectrodes.

Varshney and Li30 have realized an IDA-based impedance biosensor coupled with

magnetic nanoparticle–antibody conjugates for rapid and specific detection of E. coli

O157:H7 in ground beef samples. Instead of immobilizing antibodies directly on the

electrode surface, antibodies were immobilized onmagnetic nanoparticles. Magnetic

nanoparticles (Fe3O4, 145 nm diameter) were conjugated with anti-E. coli antibody

through biotin–streptavidin chemistry. The conjugates were then used to separate and

concentrate E. coli cells from ground beef samples. The nanoparticle–cell complexes

in 0.1 Mmannitol solution were measured by impedance using IDA microelectrodes

with 50 pairs of finger electrodes, each measuring 15mm in width and space. When

2mL of the complexes solution was spreading on the IDA electrode surface, nano-

particle–cell complexes were concentrated into the active layer of the IDAwith the

assistance of a magnet field. The lowest detection limits of this biosensor system for

detection of E. coli O157:H7 in pure culture and ground beef were 7.4� 104 and

8.0� 105 cfu/Ml, respectively. This biosensor method has been later refined into a

microfluidic chip-based biosensor by the same group.31 The microfluidic chip had a

small detection chamber (60 nL) formed by a PDMS with embedded gold interdigi-

tatedmicroelectrodes on thebottomof the chamber.Magnetic particle–cell complexes

inmannitol solutionwere injected into the detection chamber for sensitive impedance

measurement. This microfluidic impedance biosensor was able to detect as low as

1.6� 102 and 1.2� 103 cfu/mL ofE. coli cells present in pure culture and ground beef

samples, respectively.

Boehmet al. have developed an on-chipmicrofluidic biosensor forE. coli detection

and identification. In this microfluidic biosensor, anti-E. coli antibodies were

immobilized on the glass surface that served as the bottom of the microfluidic

chamber; the impedance detection electrodes were however on the top cover of

the chamber. Bacteria in suspension passing through the microfluidic chamber were

selectively recognized and captured by the immobilized antibodies, thereby increas-

ing the measured impedance within the chamber. This biosensor was able to detect

about 104 cfu/mL of E. coli when a shallow chamber (2 mm) was used.32

The use of nanoparticles can improve the capture efficiency of antibodies to target

cells.31 The microfluidic-based sensors allow continuous injection/perfusion of

bacteria samples and accumulation/concentration of bacterial cells inside the imped-

ance detection chamber over time, which can enhance the detection sensitivity and is

particularly useful for detecting low concentrations of bacteria.32

These studies have brought attention to the impedance techniques suitable for

label-free detection and have important advantages such as speed, de-skilled analysis,

fewer numbers of steps, and the potential for the multianalyte detection.

Baeumner and coworkers33 have developed microfluidic biosensors for detecting

cholera toxin subunit B (CTB) as a model analyte using electrochemical and optical

transducers. They employed liposome-based signal amplification, encapsulating

labels within the liposomes. The microfluidic devices were made from PDMS using

soft lithography fromsilicon templates.Thepolymer channelswere sealedwith aglass

plate and packaged in a polymethylmethacrylate housing that provided leakproof


sealing and a connection to a syringe pump. In the electrochemical format, an

interdigitated ultramicroelectrode array (IDUA) was patterned onto the glass slide

usingphotolithography, gold evaporation, and lift-off processes. ForCTBrecognition,

CTB-specific antibodies were immobilized onto superparamagnetic beads and gan-

gliosideGM1was incorporated into liposomes. The fluorescence dye sulforhodamine

B (SRB) and the electroactive compounds potassium hexacyanoferrate (II)/hexacya-

noferrate(III) were used as detection markers that were encapsulated inside the

liposomes for the fluorescence and electrochemical detection formats, respectively.

The limits of detection (LOD) of both assay formats for CTBwere found to be 6.6 and

1.0 ngmL�1 for the fluorescence and electrochemical formats, respectively.Changing

the detection system was very easy, requiring only the synthesis of different marker-

encapsulating liposomes, as well as the exchange of the detection unit. It was found

that in addition to a lower LOD, the electrochemical format assay showed advantages

over the fluorescence format in terms of flexibility and reliability of signal recording.

Electrochemical biosensors have been studied formanyyears at a development and

research level, successfully applied in industries in the past few years, and now

accepted as a standard method for screening some bacterial cells in food samples.

Nanomaterials are claimed to improve electrochemical biosensor sensitivity. In

respect to other transducing principles, electrochemical techniques aremuch easier to

use and allow the miniaturization for the integration in handheld devices.

3.2.3.2 Optical ImmunosensorOptical biosensors have been used widely over the past decade to analyze biomolecu-

lar interactions providing detailed information on the binding affinity and kinetics of

interaction.

A novel interference localized surface plasmon resonance (iLSPR) biosensor for

the label-free detection of biomolecules in an arbitrary solution is reported by Tamiya

and coworkers. The experimental and simulation analysis of an original nanostructure

design constructed with plasmonic gold nanoparticles and photonic thin-film multi-

layers of silicondioxide (500 nmin thickness) and siliconona substratewaspresented.

The nanostructure substrate showed a high sensitivity for various refractive index

solutions and a prominent capacity for functionalizing alkanethiol molecules on the

gold surface and demonstrates great potential in the development of a microfluidic-

based biosensor for monitoring biotin–avidin interactions in real time.34

A reflection-based localized surface plasma resonance fiber-optic probe for

chemical and biochemical sensing, called fiber-optic localized plasma resonance

(FO-LPR), has been proposed.35 Biomolecular recognition has detected the unique

optical properties of self-assembled gold nanoparticles on the unclad portions of an

optical fiber whose surfaces aremodifiedwith a receptor. To enhance the performance

of the sensing platform, the sensing element is integrated with microfluidic chips to

reduce sample and reagent volume, to shorten response time and analysis time, and to

increase sensitivity. The main purpose of the present study is to simulate the

biochemical assays in the FO-LPR microfluidic chip and to investigate the effects

of parameters, such as inlet concentrations of analyte or the flow rate on the

biochemical binding kinetics. The geometry of the grooved channel is also proposed

to enhance the biochemical binding on the unclad optical fiber. The results reveal that


the chaotic mixing generated by the grooves enhances the biochemical binding when

the injected flow rate is high and, because of this, limits the performance of the

molecular mixing. The enhancement of biochemical binding performance was

significant, especially at the low injected concentration of analyte.

Xu and coworkers have described a rapid and ultrasensitive detectionmethod using

a microfluidic chip for analyzing benzodiazepines. Benzodiazepines are mainly used

in the treatment of insomnia in clinical cases. Clonazepam (CZP) is a benzodiazepine

derivative, and 7-aminoclonazepam (7-ACZP) is the major urinary metabolite (target

metabolite) of clonazepam. A microfluidic chip-based immunoassay with laser-

induced fluorescence (LIF) detection based on the water-soluble denatured bovine

serum albumin (dBSA)-coatedCdTe quantumdotswas prepared for the ultrasensitive

detection of 7-ACZP. The detection of 7-ACZP could be completedwithin 5min. This

method was compared with ELISA and showed a good correlation. The results were

confirmed by high-performance liquid chromatography and tandem mass spectrom-

etry (LC-MS/MS).36

The immuno- and affinity assays using mobile support allow the probes to move

freely through a liquid media. The mobile support usually consists of microspheres

made of latex or magnetic materials, or nanospheres such as QDs or gold nanopar-

ticles. When using microspheres for multiplex assays, antibodies with fluorescent

labels are attached to the microsphere or fluorescent microspheres are used for

identification.

Original bioassay for multiple types of antibodies (multiplex assay) was presented

byYoon and coworkers.37 They used an immunoassay (a type of immuno- and affinity

assays) with mobile support. The QDs were conjugated onto microspheres both to

enable multiplex assays and to enhance the limit of detection. This configuration was

called “nano-on-micro”or “NOM.”Upon radiationwithUV light (380 nm), a stronger

light scattering signal is observed with NOMs than QDs or microspheres alone. In

addition, NOMs are easier to handle than QDs. Since QDs also provide fluorescent

emission, they are able to utilize an increase in light scattering for detecting anti-

gen–antibody reaction and a decrease in QD emission to identify which antibody (or

antigen) is present. Two types of NOM combinations were used. One batch of

microspheres was coated with QDs emitting at 655 nm and mouse IgG (mIgG) and

theotherwithQDsemittingat605 nmandBSA.Amixtureof these twoNOMswasused

to identify either anti-mIgG or anti-BSA. NOM particles and target solutions were

mixed in amicrofluidic device (using highly carboxylatedmicrospheres as previously

demonstratedbythesamegroup)andon-chipdetectionwasperformedusingproximity

opticalfibers.Forwardlightscatteringat380 nmwascollected.Withthepositivetarget,

the scattering signal was increased. The LOD was as low as 50 ngml�1 (330 pM).

Fluorescent emission (655 or 605 nm)was simultaneously collected.With the positive

target, the emission signal was attenuated. Therefore, they were able to detect two

different antibodies simultaneously with two different detection protocols.

Chan’s group has created a diagnostic system capable of multiplexed, high-

throughput analysis of infectious agents in human serum samples. They have

demonstrated, as a proof-of-concept, the ability to detect serum biomarkers of the

most globally prevalent bloodborne infectious diseases (i.e., hepatitis B, hepatitis C,


and HIV) with low sample volume, rapidity, and 50 times greater sensitivity than that

of currently available FDA approved methods.16

3.2.4 DNA Bioassay

The field of molecular diagnostics has expanded rapidly over the past decade.

Applications include the detection of mutations responsible for human inherited

disorders, disease-causing and food-contaminating viruses, and research into bacteria

and forensics.

Detection of infectious species and genetic mutations at the molecular level opens

up the possibility of performing reliable diagnosis even before any symptom of a

disease appears. In addition, the development of novel therapeutics based on the

regulation of gene expression provides revolutionary new opportunities in the area of

pharmaceutical science.

To improve patient care, molecular diagnostics laboratories have been challenged

to develop new tests that are reliable, cost-effective, and accurate and to optimize

existing protocols by making them faster and more economical.

Conventional methods for the analysis of specific gene sequences are based on

either direct sequencing or DNA hybridization. Because of its simplicity, DNA

hybridization is more commonly used in the diagnostic laboratory than the direct

sequencingmethod. In DNA hybridization, the target gene sequence is identified by a

DNA probe that can form a double-stranded hybrid with its complementary nucleic

acidwith high efficiency and extremely high specificity in the presence of amixture of

many different, non-complementary nucleic acids. DNA probes are single-stranded

oligonucleotides, labeled with either radioactive or non-radioactive material, that

provide detectable signals indicating DNA hybridization. Radioactive labels are

extremely sensitive but have the obvious disadvantages of short shelf life, risks

associated with exposure of personnel to radiation, cost, storage, and disposal

problems. On the other hand, non-radioactive probes, such as enzymatic or lumines-

cence labels, are less sensitive and flexible in terms of design and application but are

clearly safer and more environmentally friendly. Over the past few years, advances in

robotics, microfluidics, electronics, and high-resolution optics have driven the

impressive development of both DNA microarrays and real-time PCR systems.

The majority of commercial microarrays (pioneered by Affymetrix Inc. with their

GeneChip�)38 and all real-time PCR instruments (e.g., ABI Prism� 7900HT

Sequence Detection System)39 rely upon the detection and quantitation of a fluores-

cent reporter, whose signal increases proportionally to the amount of hybridized target

or amplified PCRproduct.Detection thus requires imaging equipment or fluorescence

readers that are generally very expensive. In addition to these technologies already

impacting on themarket, new commercial research tools are expected to have amajor

influence in the coming years.

3.2.4.1 Nucleic Acids StructuresThe double helix structure of double-stranded DNA (dsDNA) is well known and is

easily recognized not just within the scientific community. Although common, the


double helix is not the only possible configuration for DNA to take, with some

configurations being radically different from theWatson andCrickproposed structure.

In physiological conditions, A-DNA, B-DNA (Watson and Crick structure), and Z-

DNA have all been observed. Figure 3.4a shows the crystal structures for two DNA

sequences, one is in the A conformation and the other is in the B. The conformation

assumedby a length ofDNAis dependent on avariety of factors including the base pair

sequence and the supporting environment. These three conformations consist of two

antiparallel strands bound together through hydrogen bonding. These single strands

(ssDNA) consist of twomain parts, namely, the phosphate–deoxyribose backbone that

forms the chain and the associated nucleobases (commonly referred to as bases). The

bases are carbon–nitrogen ring structures,which are of two types, the purines (adenine

(A) and guanine (G)) consisting of two fused rings and the pyrimidines (thymine (T)

and cytosine (C)) consisting of only one ring. These bases are joined to the phosphate

backbone and the order in which they occur provides the basic coding for genetic

material. The asymmetric ends of theDNAstrands are labeled as 50 that terminates in a

phosphate group and 30 that terminates with a hydroxyl group, and the nomenclature

refers to the position of the terminal carbon in the (deoxy)ribose ring. As a

FIGURE 3.4 (a) The crystal structures for two DNA sequences: one is in the A conforma-

tion and the other is in the B.48 (b) Chemical structures of the “Watson andCrick” paired bases;

adenine and thymine as well as guanine and cytosine.


consequence of both space constraints, and so as to maximize hydrogen bonding,

guanine preferentially binds to cytosine and adenine to thymine—this is known as

Watson and Crick base pairing, Figure 3.4b shows the chemical structures of these

nucleobases in their Watson and Crick pairs. Unlike the other two major conforma-

tions, Z-DNAdiffers in exhibiting a left-handed double helical structure (i.e., the helix

rotates around the axis in the opposite sense); this structure is found to often occurwith

G–C rich sequences in low salt conditions.With B-DNA the interwinding of the DNA

strands leads to a structure inwhich there are two distinct grooves spiraling around the

DNA duplex. The larger groove is called the “major groove” and the smaller one the

“minor groove”; these structural features are labeled in Figure 3.4a. The conformation

of DNAwhile attached to an electrode surface has been shown to depend upon the

strength of the interactions between the interface and the DNA strand. Through

heating dsDNA and breaking the weak hydrogen bonds, the two strands may be

separated. This process is known as “denaturation.” The melting temperature (Tm) at

which this occurs depends on both the length of the DNA chain and the constituent

bases. On cooling ssDNA, with a complementary sequence, the bases pair together to

form dsDNA. This is known as “hybridization.” It is possible for strands of DNA that

are not fully complementary to pair but the number of mismatches affects the stability

of the dsDNA and consequently lowers the melting temperature.

3.2.5 DNA Biosensors

DNA biosensors are analytical devices that result from the integration of a sequence-

specific probe (usually a short synthetic oligonucleotide) and a signal transducer. The

probe, immobilized onto the transducer surface, acts as the biorecognition molecule

and recognizes the targetDNA,while the transducer is the component that converts the

biorecognition event into a measurable signal. Assembly of numerous (up to a few

thousand) DNA biosensors onto the same detection platform results in DNA micro-

arrays (or DNA chips), devices that are increasingly used for large-scale transcrip-

tional profiling and single-nucleotide polymorphisms (SNPs) discovery. As clinical

diagnostics and other applications (e.g., environmental screening) do not generally

need themassive data accumulation typical of gene chips, alternative technologies are

in development whose promise is to provide flexible and economical alternatives for

applications that require relatively fewer measurements.

The DNA biosensor can be broadly divided into two main groups: label-free

systems and labeled. Those methods that use solution-phase reagents (e.g., metal

complexes or organic dyes) as markers of the hybridization process will be referred to

as label free. Label-free approaches typically rely on the measurement of changes in

the electrical and optical characteristics of sensing layer before and after the

hybridization reaction. By contrast, when organic and organometallic electroactive

compounds, nanoparticles, and catalytic and redox enzymes are permanently bound

(e.g., covalently or via (strept)avidin–biotin interactions) to one of the constituents of

the surface-tethered duplex, the method will be considered as label based. Sensitivity

and reliability of label-based approaches are often still unrivalled, as alsowitnessed by

the choice of these methodologies for the microarray platforms now on the market.


Tremendous research activities have been carried out to miniaturize the conven-

tional DNA analytical procedures in microchip platforms. These microdevices enjoy

the miniaturization advantages of small size, low sample, reagent and power con-

sumption, enhancedanalytical performance (e.g., shorter assay time), andhigh levelof

integration. An ideal microanalyzer should feature sample in result out kind of

automated operation, without any human intervention between individual assay steps.

There are three essential components in a complete DNA assay protocol that include

sample preparation, target amplification, and product detection.

The following sections aim to give an outline of the wide variety of methods that

have to date been employed to electrochemically and optically detect specific DNA

sequences. Typically, the design of an electrochemical DNA biosensor involves the

following steps:

1. Immobilization of the DNA probe

2. Hybridization with the target sequence

3. Labeling and electrochemical investigation of the surface

Optimization of each step is required to improve the overall performance of the

DNA sensing.

3.2.5.1 DNA ProbeAs the specificity of the hybridization reaction essentially depends on the biorecogni-

tion properties of the capture oligonucleotide, design of the capture probe is undoubt-

edly the most important preanalytical step. Thus, a number of probes, variable for

chemical composition and conformational arrangement, have been used to assemble

DNA biosensors.

Oligonucleotides (ODNs) are regularly used as the DNA probe in a biorecognition

layer.ODNsare short sequences ofDNAgenerally 20bases or less in length.Design of

linear probes takes now great advantage of decades of experience, which has led to

many commercially available types of software.

DNA probes are typically short (18–40-mer) oligonucleotides that are able to

hybridize with specific target sequences. While earlier work employed simple DNA

probe sequences as a model (e.g., oligo d(G)20), recent reports describe the use of

disease- or microorganism-related oligonucleotide sequences. Some alternative

methods use PNA and LNA such as DNA analogues. PNA possess an uncharged

pseudopeptide backbone (instead of the chargedphosphate–sugar backbone of natural

DNA).Because of their neutral backbone, PNAprobes offer greater affinity in binding

to complementary DNA and improved distinction between closely related sequences

(including single-base mismatches). Such mismatch discrimination has a particular

importance in the detection of disease-related mutations.

Locked nucleic acids (LNA�)40 are a class of nucleic acid analogues in which the

ribose ring is “locked” by a methylene bridge connecting the 20-O atom with the 40-Catom.DNAoligos incorporatingLNAnucleosides show increased thermal and further

improved discriminative power with respect to single-base mismatched targets.


The probe immobilization step plays the major role in determining the overall

performance of an electrochemical DNA biosensor. The achievement of high sensi-

tivity and selectivity requires maximization of the hybridization efficiency and

minimization of nonspecific adsorption, respectively.Control of the surface chemistry

and coverage is essential for assuring high reactivity, orientation, accessibility, and

stability of the surface-confined probe, as well as for minimizing nonspecific

adsorption events.

3.2.5.2 Hybridization ReactionThe kinetics and mechanism of the hybridization reaction in solution has been

widely studied. Hybridization involves a two-step process: nucleation and zippering.

Nucleation is the rate-limiting step. It is assumed that the nature of the hybridization

reaction at solid surfaces closely approximates that of the solution-phase reaction,

but its rate is about 10–100 times slower. Efficient hybridization of a target to

surface-bound probes can be impeded by several phenomena. For example, the

immobilized probe may be not accessible for hybridization because of some steric

hindrance. The rate of hybridization and the stability of the duplex depend on several

factors, such as salt concentration, temperature, use of accelerating agents, base

composition (G þ C content), and length of the probe sequence. The salt concen-

tration markedly affects the rate of hybridization reaction. Below 0.1M NaCl, a

twofold increase of salt concentration increases the hybridization rate by 5–10-fold

or even more. The rate levels off when the concentration exceeds 1.2M NaCl.

However, since this high salt concentration stabilizes mismatched duplexes, the use

of high ionic strength solutions is not recommended for single-base mutation

analysis.

The rate of hybridization strongly depends on the temperature. The maximum

rate is observed 20–25�C below Tm of the duplex. However, depending on salt

concentration, annealing may effectively occur at temperatures well below the

optimum value.

The overall sensitivity of a hybridization assay is strongly influenced by the

hybridization time. Moreover, the hybridization process can be facilitated using

appropriate reagents. The presence of guanidine HCl in the target solution was shown

to highly increase the rate of hybridization. The stringency of hybridization can be

additionally altered using formamide. Formamide decreases the Tm of nucleic acid

hybrids. Use of 30–50% formamide in the hybridization solutions allows the incuba-

tion temperature to be reduced to 30–42�C.The effect of sequence length on hybridization rate is well known. The lower

hybridization yield of assays in which long probes and/or targets are used is attributed

to steric hindrance and also to the slower mass transport rate of the target toward the

surface immobilized probe.

3.2.5.3 Hybridization DetectionOnce the target DNA has been captured onto the sensor surface, a range of different

approaches can be used for transducing the biorecognition event. The transducing

principles can be broadly divided into reagentless, label-free, and label-based


schemes. Labeling methods allow high sensitivity, and these approaches are devel-

oped to the point that they give reproducible results.

Among a range of options, current DNA hybridization detection methods have

mainly employed fluorescent labels, quantum dots, or heavy atom complex nanopar-

ticle labels.

3.2.5.4 Electrochemical DNA BiosensorThe recent appearance on the diagnosticmarket of electrochemical DNAmicroarrays

(e.g., Motorola eSensor�DNADetection System41 andXanthonXpressionAnalysis

System42) demonstrates the enormous potential of electrochemical DNA-based

biosensors. Electrochemical devices are highly sensitive, inexpensive, easy to use,

portable, and compatiblewithmicrofabrication technologies.Moreover, in contrast to

optical detection schemes, the electrical responses are independent of sample turbidi-

ty.TomakeDNAtestingmore convenient,more economically feasible, andultimately

morewidely used, the appealing promise of electrochemical detection technologies is

thus driving an intense research effort by hundreds of laboratories worldwide.

In the past few years, several excellent reviews have been published on both

electrochemical and DNA sensing.43–47

Moreover, the underlying physicochemical properties of DNA and its hybridiza-

tion as a basis for understanding howpresent electrochemicalmethodsmay enable the

detection of specific DNA sequences are reported by Compton and coworkers.48

In this section, only analytical procedures based on microfluidic platform coupled

to nanoparticles for hybridization electrochemical detection will be considered.

Liposomes encapsulating electroactive molecules have been used as labels for

DNA biosensor. A biosensor based on nucleic acid hybridization and liposome signal

amplification with an integrated microfluidic system and a minipotentiostat for the

quantification of denguevirusRNAwas reportedbyBaeumner and coworkers.49,50An

electrochemical microfluidic biosensor with an integrated minipotentiostat for the

quantification of RNA was developed based on nucleic acid hybridization and

liposome signal amplification. Specificity of the biosensor was ensured by short

DNAprobes that hybridizewith the targetRNAorDNAsequences. The reporter probe

was coupled to liposomes entrapping the electrochemically active redox couple

potassium ferri/ferrohexacyanide. The capture probes were coupled to superpara-

magnetic beads that were isolated on a magnet in the biosensor. Upon capture, the

liposomes were lysed to release the electrochemical markers that were detected on an

interdigitated ultramicroelectrode array in the biosensor just downstream of the

magnet. The current was measured, stored, and displayed byminiaturized instrumen-

tation (miniEC).

The same authors have presented an optimization of their studies.51 In the previous

study, IDUA were fabricated on Pyrex� 7740 as a substrate and overlaid PDMS

channels. Ingeneral, gold haspoor adhesionproperties tomost surfaces includingpoly

(methyl methacrylate). An intermediate adhesion layer of titanium is commonly used

between the substrate and the gold layer. However, for an electrochemical detection

system, a bimetallic system results in a galvanic cell with the less noble of the two

metals being solubilized. Since it cannot be guaranteed that the adhesion layer is not


coming in contact with the solution, it can result in limited lifetime of the electrode.

Alternatively, mercaptopropyltrimethoxysilane (MPTMS) has been used as an effec-

tive alternative to a metallic adhesion layer by providing the substrate surface with a

thiol monolayer. The gold electrodes are then adhered using gold–thiol interactions.

Although limitations were found for the usable applied potential, the electrodes were

more stable and had a longer lifetime compared to metallic adhesion systems. In the

recent study, cystamine was conjugated to the UV modified PMMA surface using

water-soluble carbodiimide chemistries, resulting in a thiolated surface. A liposomal

detection system was employed to aid in signal amplification. The liposome was

tagged with a DNA probe complementary to the target RNA. Superparamagnetic

beads tagged with a target complementary capture probewere used to immobilize the

target and the liposome complex over the IDUA (Figure 3.5a and b).

One of the major trend lines toward the research of novel diagnostic systems is the

concept of DNA chips (or DNA microarrays), usually associated to microfabrication

of diagnostic kits by screen printing techniques, inspired by planar, silicon-based

technologies.

The miniaturization of DNA analytical platforms has many advantages over the

conventional benchtop counterparts. These include low sample/reagent consumption

(volume of microliter down to picoliter) as well as short assay time (minutes rather

than days). Most important, they permit the integration of a number of functions

including sample preparation, target amplification, and product detection, thus

enabling a fully automated operation that can be used by untrained individuals.

Amplification of nucleic acids in biomedical and biochemical researches could be

used for diagnosing disease, sequencing, genotyping, and evolutionary studies. Such

applications of PCR require highly sensitive, fast, selective, and accurate detection

methods. Therefore, there has been recent interest in developing an accurate, sensitive,

selective, and fast detection method for PCR amplification.

Shiddiky et al.52 developed an electrochemical method for analyzing PCR ampli-

fication through the detection of inorganic phosphates (Pi). This method coupled a

microchip to a nanoparticle comprising poly-5,20-50,200-terthiophene-30-carboxylicacid (poly-TTCA)/pyruvate oxidase (PyO) modified microbiosensor. It detects Pi

produced from the pyrophosphate (PPi), which is released as a by-product of PCR.

After completion of PCR, PPi is hydrolyzed to Pi by inorganic pyrophosphatase. On

themicrobiosensor surface, pyruvatewas converted toH2O2 byPyO in the presence of

Pi and oxygen, and subsequently, the anodic current of enzymatically generatedH2O2

was detected at 10.5V versus Ag/AgCl. The CE-EC analysis was completed within

2min. An excellent operation stability of poly-TTCA/PyO was observed for a long

period of analysis.

Tin-doped indium oxide (ITO) is thematerial of choice for fabrication of a number

of optoelectronic devices. However, ITO electrodes have also found application for

DNA sequence-specific detection. An array of individually addressable ITO electro-

des was used as the transduction element in an integrated analytical device employed

for the multiplexed detection of E. coli and Bacillus subtilis cells. The choice of an

electrochemically driven immobilization strategy (electrochemical copolymerization

of pyrrole and pyrrole–oligonucleotide conjugates) allowed individual positioning of


the predefined probes at selected ITO surfaces. Notably, such immobilization chem-

istry provided the probes sufficient stability to tolerate the repeated thermal cycling of

in situ PCR amplification of the target DNA.53

3.2.5.5 Optical DNA BiosensorIn a recent review, Krull and coworkers have discussed the application of QDs, gold

nanoparticles, and molecular switches in optical nucleic acid diagnostics. The size-

FIGURE 3.5 (a) Firstly, sandwich hybridizations bind the liposomes to magnetic beads.51

Then, the bead/liposome complex is captured by a magnet placed over the channel. Finally, a

detergent is passed through the channel resulting in the lysis of the liposomes. The liposome

contents are then pumped over the IDUA for concentration determination. (b) (A) SEM of a

gold IDUA formed on a PMMA substrate.51 (B) A PMMA sheet containing a hot embossed

channel was then bonded to the PMMA containing the IDUA. The finished device contained a

500mm channel positioned along the IDUA. (C) The finished chip containing two channels.


dependent optical properties of nanoscalematerials, aswell as the ability to tailor both

material and surface composition of NPs, create exciting new possibilities in nucleic

acid analyses. Similarly,molecular assemblies of nucleic acids that generate “on–off”

responses at the single-molecule (or particle) level offer significant advantages in

diagnostics 54.

In nature, DNA and RNA are found at extremely low concentrations. As a result,

determining the concentration of DNA, especially without amplification, is a great

challenge.Manydifferent single-molecule opticalmethods have beenused to quantify

and characterize DNA and RNA.

Today, an ideal biosensor platform is required to be not onlyminiaturized and cost-

efficient but also capable of simultaneous detection of multiple analytes. The current

trend toward creating point-of-care molecular diagnostic biosensors and massively

parallel biorecognition arrays (microarrays) has introduced new technical challenges

for the probes, transducers, and their detection apparatus.

In the case of nucleic acid analyses, the idea is that each probe oligonucleotide has

an associated spectral code. The spectral code is created by assigning combinations of

different fluorophores to a particular probe oligonucleotide. The most common form

of barcodes areNPormicroparticle carriers that are dopedwith a specific combination

of fluorescent dyes or QDs. QDs are particularly well suited to barcoding since

multiple colors can be excited with a single wavelength and since each QD offers a

narrow, symmetric emission profile. Hybridization can be detected by labeling target

with a fluorophore of either shorter (e.g., fluorescein) or longer wavelength (e.g., Cy5)

than the encodingQDs.Hybridization assays have been carried out usingQD-barcode

technology.

Combining a biobarcode with microfluidic chip-based format, Mirkin and

coworkers have developed a new version of the biobarcode assay named as

Genomic Bio Bar Code Assay 55 (Figure 3.6). The assay utilizes oligonucleo-

tide-functionalizedmagneticmicroparticles to capture the target of interest from the

sample. A critical step in the new assay involves the use of blocking oligonucleo-

tides during heat denaturation of the double-stranded DNA. These blockers bind to

specific regions of the target DNA upon cooling and prevent the duplex DNA from

rehybridizing, which allows the particle probes to bind. Following target isolation

using the magnetic particles, oligonucleotide-functionalized gold nanoparticles act

as target recognition agents. The oligonucleotides on the nanoparticle (barcodes)

act as amplification surrogates. The barcodes are then detected using the scano-

metricmethod. The limit of detection for this assaywas determined to be 2.5 fM, and

this is the first demonstration of a barcode-type assay for the detection of double-

stranded genomic DNA. B. subtilis was chosen as a model system since it is a close

family member of the lethal bacterium B. anthracis, which in its spore form is the

biological weapon anthrax.

Thiswork paves theway for the transition of the biobarcode assay froma laboratory

technique to one that can be deployed in the field for the rapid and accurate detection of

biological terrorism agents.

The method proposed by Nie et al. used color-coded nanoparticles and dual-color

coincidence detection to quantify an array of biomolecules including DNA, RNA,


proteins, and viruses within a microfluidic channel. Red and green nanoparticles bind

to specific sites on the target molecule. Real-time coincidence measurements allow

discrimination of nanoparticle-bound target molecules and individual unbound

nanoparticles as they flow through the microfluidic channel. This method allows

the precise quantification of the biomolecules without the need for separation or

amplification.56

3.3 CONCLUSIONS AND FUTURE TRENDS

In this chapter, the new biosensor designs based on microfluidic and nanoparticles

have been presented. Compared to current state-of-the art DNA detection using

fluorophore labels and protein detection using ELISA, significant advances have been

made in terms of sensitivity, selectivity, and multiplexing capacity.

Nanotechnology is revolutionizing the development of biodevices, and it is

increasingly being used to design novel bioassays with high performance.

Nanotechnology-based biosensors have been integrated within tiny biochips with

on-board electronics, sample handling, and analysis. This greatly improves function-

ality by providing devices that are small, portable, easy to use, low in cost, disposable,

and highly versatile diagnostic instruments.

The combination of microfluidic technologies with nanoparticles is a very prom-

ising biosensor platform, and several examples have been presented.

FIGURE 3.6 Genomic Bio Bar Code Assay.56 The first step is to isolate the genomic DNA

from the bacterial cells and cut it with a restriction enzyme. This cut prevents the DNA from

supercoiling during heating and gives smaller target pieces. The next step is to introduce

blocking oligonucleotides designed to flank the probe binding sites and prevent strand

rehybridization after thermal denaturation. The blocking oligonucleotides are used in excess

to outcompete the native strand during hybridization. The target region is now “propped” open

and accessible for probe binding. Magnetic microparticles (oligo-MMPs) are used to capture

the targets from the sample and then washed. An excess of oligonucleotide modified gold

nanoparticle probes (oligo-AuNPs) is added to the assay solutions, which results in the

sandwiching of the target with the oligo-MMP. Unbound oligo-AuNPs are removed by

washing. The barcodes are chemically released for scanometric detection and quantification.


A special emphasis has placed on the challenges of integrating detection platforms

for encoded nanoparticles intomicrodevices for performingmultiplexed assay. This is

highly important in situations where the amount of sample is very limited, such as in

the analysis of blood from newborns, tumor tissue from biopsies, and so on. In

addition, multiplexing allows more efficient and therefore less expensive use of

reagents, and because the different targets are screened simultaneously, they experi-

ence equal conditions at each step of the assay procedure.

Overall, these microfluidic bioassays lead to inexpensive screening for multiple

diseases or biomolecule states in a simultaneous fashion,which should quickly change

theway inwhichmedicine is practiced, possibly leading to a preventionmindset rather

than a response after a diagnosis. Such applications will become more numerous. In

any case, the impact of nanoscale sensorswill have a profound effect onmedical, food,

and environmental testing.

REFERENCES

1. de la Escosura-Muniz A, Ambrosi A, Merkoci A. Electrochemical analysis with nanopar-

ticle-based biosystems. Trends Anal. Chem. 2008; 27(7); 568–584.

2. Gomez-Hens A, Fernandez-Romero JM, Aguilar-Caballos MP. Nanostructures as analyti-

cal tools in bioassays. Trends Anal. Chem. 2008; 27(5); 394–406.

3. Pingarron JM, Yanez-Sedeno P, Gonzalez-Cortes A. Gold nanoparticle-based electro-

chemical biosensors. Electrochim. Acta 2008; 53; 5848–5866.

4. Wang Z, Ma L. Gold nanoparticle probes. Coord. Chem. Rev. 2009; 253; 1607–1618.

5. Tamanaha CR, Mulvaney SP, Rife JC, Whitman LJ. Magnetic labeling, detection, and

system integration. Biosens. Bioelectron. 2008; 24; 1–13.

6. Pumera M, Merkoci A, Alegret S. New materials for electrochemical sensing VII.

Microfluidic chip platforms. Trends Anal. Chem. 2006; 25(3); 219–235.

7. Rıos A, Escarpa A, Gonzalez MC, Crevillen AG. Challenges of analytical microsystems.

Trends Anal. Chem. 2006; 25(5); 467–479.

8. Derveaux S, Stubbe BG, Braeckmans K, Roelant C, Sato K, Demeester J, De Smedt

SC. Synergism between particle-based multiplexing and microfluidics technologies

may bring diagnostics closer to the patient. Anal. Bioanal. Chem. 2008; 391;

2453–2467.

9. Choi J-W, Oh B-K, Kim Y-K, Min J. Nanotechnology in Biodevices. J. Microbiol.

Biotechnol. 2007; 17(1); 5–14.

10. Ligler FS. Perspective on optical biosensors and integrated sensor systems. Anal. Chem.

2009; 81(2); 519–526.

11. Lim CT, Zhang Y. Bead-based microfluidic immunoassays: the next generation. Biosens.

Bioelectron. 2007; 22; 1197–1204.

12. Liu G, Lin Y. Nanomaterial labels in electrochemical immunosensors and immunoassays.

Talanta 2007; 74; 308–317.

13. Tansil NC, Gao Z. Nanoparticles in biomolecular detection. Nanotoday 2006; 1(1);

28–37.

REFERENCES 121

14. Wang J, Bunimovich YL, Sui G, Savvas S, Wang J, Guo Y, Heath JR, Tseng H-R.

Electrochemical fabrication of conducting polymer nanowires in an integrated micro-

fluidic system. Chem. Commun. 2006; 3075–3077.

15. Edwards KA, Baeumner AJ. Liposomes in analyses. Talanta 2006; 68; 1421–1432.

16. Klostranec JM, Xiang Q, Farcas GA, Lee JA, Rhee A, Lafferty EI, Perrault SD, Kain KC,

Chan WCW. Convergence of quantum dot barcodes with microfluidics and signal

processing for multiplexed high-throughput infectious disease diagnostics. Nano Lett.

2007; 7(9); 2812–2818.

17. Marquette CA, BlumLJ. State of the art and recent advances in immunoanalytical systems.

Biosens. Bioelectron. 2006; 21; 1424–1433.

18. Bilitewski U. Protein-sensing assay formats and devices. Anal. Chim. Acta 2006; 568;

232–247.

19. www.diagnoswiss.com.

20. Rossier JS, Baranek S, Morier P, Vollet C, Vulliet F, De Chastonay Y, Reymond F. GRAVI:

Robotized microfluidics for Fast and Automated Immunoassays in Low Volume. JALA

2008; 13(6); 322–329.

21. Tothill IE. Biosensors for cancer markers diagnosis. Semin. Cell Dev. Biol. 2009; 20;

55–62.

22. Maeng JH, Lee BC, Ko YJ, Cho W, Ahn Y, Cho NG, Lee SH, Hwang SY. A novel

microfluidic biosensor based on an electrical detection system for alpha-fetoprotein.

Biosens. Bioelectron. 2008; 23; 1319–1325.

23. Goluch ED, Nam J-M, Georganopoulou DG, Chiesl TN, Shaikh KA, Ryu KS, Barron AE,

Mirkin CA, Liu C. A bio-barcode assay for on-chip attomolar-sensitivity protein detection.

Lab Chip 2006; 6(10); 1293–1299.

24. Goluch ED, Stoeva SI, Lee J-S, Shaikh KA, Mirkin CA, Liu C. A microfluidic detection

system based upon a surface immobilized biobarcode assay. Biosens. Bioelectron. 2009;

24; 2397–2403.

25. Wilson MS, Nie W. Multiplex measurement of seven tumor markers using an electro-

chemical protein chip. Anal. Chem. 2006; 78; 6476–6483.

26. Ko YJ, Maeng JH, Ahn Y, Hwang SY, Cho NG, Lee SH. Microchip-based multiplex

electro-immunosensing system for the detection of cancer biomarkers. Electrophoresis

2008; 29; 3466–3476.

27. Tang D, Yuan R, Chai Y. Magnetic control of an electrochemical microfluidic device with

an arrayed immunosensor for simultaneous multiple immunoassays.Clin. Chem. 2007; 53

(7); 1323–1329.

28. Ahmed MU, Hossain MM, Tamiya E. Electrochemical biosensors for medical and food

applications. Electroanalysis 2008; 20(6); 616–626.

29. Yang L, Bashir R. Electrical/electrochemical impedance for rapid detection of foodborne

pathogenic bacteria. Biotechnol. Adv. 2008; 26; 135–150.

30. VarshneyM, Li Y. Interdigitated array microelectrode based impedance biosensor coupled

with magnetic nanoparticle–antibody conjugates for detection of Escherichia coli O157:

H7 in food samples. Biosens. Bioelectron. 2007; 22; 2408–2414.

31. Varshney M, Li Y. Double interdigited array microelectrode-based impedance biosensor

for detection of viable Escherichia coli O15:H7 in growth medium. Talanta 2008; 74;

518–525.


32. Boehm DA, Gottlieb PA, Hua SZ. On-chip microfluidic biosensor for bacterial detection

and identification. Sens. Actuators B Chem. 2007; 126; 508–514.

33. BunyakulN, Edwards KA, Promptmas C, Baeumner AJ. Cholera toxin subunit B detection

in microfluidic devices. Anal. Bioanal. Chem. 2009; 393(1); 177–186.

34. Hiep HM, Yoshikawa H, Saito M, Tamiya E. An interference localized surface plasmon

resonance biosensor based on the photonic structure of Au nanoparticles and SiO2/Si

multilayers. ACS Nano 2009; 3(2); 446–452.

35. Jen C-P, Huang C-T, Lu Y-H. Simulation of biochemical binding kinetics on the micro-

fluidic biochip of fiber-optic localized plasma resonance (FO-LPR). Microelectron. Eng.

2009; 86; 1505–1510.

36. Chen W, Peng C, Jin Z, Qiao R, Wang W, Zhu S, Wang L, Jin Q, Xu C. Ultrasensitive

immunoassay of 7-aminoclonazepam in human urine based on CdTe nanoparticle bio-

conjugations by fabricated microfluidic chip. Biosens. Bioelectron. 2009; 24; 2051–2056.

37. Lucas LJ, Chesler JN, Yoon JY. Lab-on-a-chip immunoassay for multiple antibodies using

microsphere light scattering and quantum dot emission. Biosens. Bioelectron. 2007; 23;

675–681.

38. www.affymetrix.com.

39. www.appliedbiosystems.com.

40. www.exiqon.com.

41. www.motorola.com/lifesciences/esensor/.

42. www.xanthoninc.com.

43. Lucarelli F, Tombelli S, Minunni M, Marrazza G, Mascini M. Electrochemical and

piezoelectric DNA biosensors for hybridisation detection. Anal. Chim. Acta 2008; 609;

139–159.

44. Sadik OA, Aluoch AO, Zhou A. Status of biomolecular recognition using electrochemical

techniques. Biosens. Bioelectron. 2009; 24; 2749–2765.

45. Teles FRR, Fonseca LP. Trends in DNA biosensors. Talanta 2008; 77; 606–623.

46. Navani NK, Li Y. Nucleic acid aptamers and enzymes as sensors. Curr. Opin. Chem. Biol.

2006; 10; 272–281.

47. Lee TM, Hsing IM. DNA-based bioanalytical microsystems for handheld device applica-

tions. Anal. Chim. Acta 2006; 556; 26–37.

48. Batchelor-McAuley C, Wildgoose Gregory G, Compton RG. Biosens. Bioelectron. 2009;

doi: 10.1016/j.bios.2009.01.045.

49. Goral VN, Zaytseva NV, Baeumner AJ. Electrochemical microfluidic biosensor for the

detection of nucleic acid sequences. Lab Chip 2006; 6(3); 414–421.

50. Kwakye S, Goral VN, Baeumner AJ. Electrochemical microfluidic biosensor for nucleic

acid detection with integrated minipotentiostat. Biosens. Bioelectron. 2006; 21(12);

2217–2223.

51. Nugen SR, Asiello PJ, Connelly JT, Baeumner AJ. PMMAbiosensor for nucleic acids with

integrated mixer and electrochemical detection. Biosens. Bioelectron. 2009; 24;

2428–2433.

52. Shiddiky MJ, Rahman MA, Park JS, Shim Y-B. Analysis of polymerase chain reaction

amplifications through phosphate detection using an enzyme–based microbiosensor in a

microfluidic device. Electrophoresis 2006; 27; 2951–2959.

REFERENCES 123

53. Yeung S-W, Lee TM-H, Cai H, Hsing I-M. A DNA biochip for on-the-spot multiplexed

pathogen identification. Nucleic Acids Res. 2006; 34(18); e118.

54. Algar WR, Massey M, Krull UJ. The application of quantum dots, gold nanoparticles and

molecular switches to optical nucleic acid diagnostics. Trends Anal. Chem. 2009; 28(3);

292–306.

55. Hill HD, Vega RA, Mirkin CA. Non-enzymatic detection of bacterial genomic DNA using

the bio-barcode assay. Anal. Chem. 2007; 79; 9218–9223.

56. AgrawalA, ZhangC,Byasses T, TrippRA,Nie S. Counting single native biomolecules and

intact viruses with color-coded nanoparticles. Anal. Chem. 2006; 78; 1061–1070.


4MICROFLUIDIC ENZYMATICREACTORS USING NANOPARTICLES

CHUNHUI DENG AND YAN LI

Department of Chemistry, School of Pharmacy, Fudan University, Shanghai, China

4.1 INTRODUCTION

Enzymes are one of the catalysts that are useful for substance production in an

environment-friendly way and have high potential for analytical applications.1 The

use of enzymes for cleavage, synthesis, or chemical modification represents one of

the most common processes used in biochemical and molecular biology laboratories.

The continuing progress in medical research, genomics, proteomics, and related

emerging biotechnology fields has led to an exponential growth of applications of

enzymes and the development of modified or new enzymes with specific activities.2

In proteomic study, the rapid development of mass spectrometry and the associated

coupled technologies has provided an opportunity to perform protein and peptide

separation and identification in a highly automated and rapid fashion; hence, more

comprehensive protein mapping in cells or tissues can be obtained. Prior to protein

characterization by mass spectrometry analysis, it is necessary to perform the

controlled enzymatic degradation in a short time, resulting in well-defined and

reproducible peptide patterns. The ability of some proteases to cleave polypeptide

chains at specific cleavage sites makes them important tools of proteomics for

elucidation of the protein primary structure or post-translational modifications.3

However, the conventional techniques of in-solution digestion of proteins suffer

from limitations of long digestion time, chronic autodigestion of enzyme, and sample

loss, severely affecting the determination of comprehensive proteomic profiles.


125

On theother hand,microfluidic devices havebecomepowerful tools for performing

chemical or biological assays in recent years. Miniaturization to the microliter scale

reduces the requirements for reagents and solvents, providing flexibility in use and

faster reaction times, thereby supporting high-throughput experimentation.

Translation of processes from discovery to actual production could be accelerated

by using automated microscale processing methods that could lead to a reduction in

time and thus in cost for the product to reach the market. Microstructured reactors

minimally integrate the functionalmicrofluidic element in a suitable and appropriately

interfaced housing.4 Fluid flow as well as mass and heat transport are more easily

controlled on the microscale, often resulting in enhanced yield and improved

selectivity compared to conventional reactors.5–7

One of the promising solutions is the incorporation by patterning processes of

enzymes, such as proteases, within a microchannel to form a microfluidic enzymatic

reactor to carry out highly efficient and low-level protein digestion. Implementation of

enzymatic reactions inmicrochannels allows a decrease in the amount of consumables

and sample by several orders ofmagnitude.Detection sensitivity can also be improved

for a sample with small total volume since no dilution is necessary, and the smaller

scale increases the speed of diffusion-limited reactions allowing faster assays. Finally,

microfluidic enzymatic reactor may allow long-term higher automation and better

reproducibility.3

The aim of this chapter is to summarize recent advances in the field of immobilized

microfluidic enzymatic reactors (IMERs), which constitutes a new branch of nano-

technology.Without claiming tobe exhaustive, instead, this chapter focuseson IMERs

fabricatedwith enzyme-immobilized nanoparticles,while other formats of IMERs are

briefly discussed. Enzyme immobilization techniques and the main applications

of IMERs in the fields of peptide mapping, biosensing, and kinetic study are also

described.

4.2 ENZYME IMMOBILIZATION TECHNIQUES

Typically, enzymes have high efficiency under mild conditions and are highly

selective, but they are not stable for extended time in solution and during storage

their activity gradually decreases. Significant improvements in both the reaction rates

(much higher enzyme/substrate ratio can be achieved) and the storage stability can be

achieved with enzymes immobilized on the surface of a suitable carrier material.

Immobilized enzyme reactors are considerablymore stable and catalytically active for

a much longer duration than free enzymes.8,9 Immobilized molecules are more

resistant to the unfolding of their native structure that may be caused by heat and

pH changes. Besides, although solution-phase enzymatic reactors are simple, these

approaches canyield autodigestion, andmay require a delicate separation between the

enzyme and the reaction products. In contrast, the catalyst could be easily removed

from the reaction mixture when using immobilized enzyme, thus facilitating separa-

tion of product and recycled while its activity is preserved as well as avoiding

contamination of the digestion products by free enzyme molecules, which can be

126 MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

very detrimental to analysis.10 Also, continuous processes with the biocatalyst

immobilized on a solid support packed in a reactor could be designed. These

implementations were believed to significantly lower the cost of the biocatalyst

used per unit of product, which might represent significant savings in operations

requiring expensive enzymes.

Although immobilization may result in some changes in enzymatic activity,

optimum pH or affinity for the substrate, elimination of enzyme from the reaction

mixture, enhanced stability, reduced mixing and dilution-related problems, and the

possibility to reuse the reactors outweigh these changes.

Avariety ofmethods are nowavailable for immobilizationof enzymeon the surface

of the fused silica capillary or the channel of amicrofluidic chip or covalent binding to

the activated supports, physical adsorption of the enzyme on a solid matrix and

copolymerization of the enzyme with the polymers. The specific immobilization

chemistry depends on avariety of factors, character of the support, activationmethods,

and coupling procedure.2

4.2.1 Covalent Immobilization

The most intensively used immobilization technique is the formation of covalent

bonds between the protein and the support matrix,11 for example, immobilization in

the presence of carbodiimides, cross-linking by glutaraldehyde, or cyanogen bromide

activation of the support material. The main advantage of covalent binding to the

activated support is that it can prevent the desorption of enzyme from the support

matrix in thepresenceof substrates and solutions of high ionic strength andof reducing

spontaneous enzyme deactivation rates, as in proteases autodigestion. These benefits,

involving a longer IMER lifetime, are counterbalanced by the more easily altered

native tertiary structureof the enzymewith subsequent decrease in catalytic activity. In

addition, the use of a covalent binding mode involves a higher enzymatic thermal

stability since the strong interaction of enzyme to support causes rigidity of the protein

structure and consequently limits the thermal movement of the protein at a high

temperature. Therefore, the attached enzyme unfolds with difficulty, inactivation is

not so easily observed, and a higher reaction rate and fewer diffusional restrictions can

be achieved.12

Proteins usually have a number of potential immobilizing sites, corresponding to

particular functionalities on the molecules. The functional groups of the proteins

suitable for covalent binding include (i) the a-amino groups of the chain and the

e-amino groups of lysine and arginine, (ii) thea-carboxyl groups of the chain end andthe b- and g-carboxyl groups of aspartic and glutamic acids, respectively, (iii) the

phenol ring of tyrosine, (iv) the thiol group of cysteine, (v) the hydroxyl groups of

serine and threonine, (vi) the imidazole group of histidine, and (vii) the indole group

of tryptophan. The most common covalent immobilization procedures are summa-

rized in Figure 4.1.2

Themostwidely used, among all thesemethods, is based on the activation of amino

supports by using glutaraldehyde. The covalent Schiff’s base formation between the

aldehydic group and the e-amine group of lysine residue (Figure 4.1) is obtained under

ENZYME IMMOBILIZATION TECHNIQUES 127

mild reaction conditions (T, pH, and stirring) in accordancewith those required for the

optimal catalytic activity and enzymatic stability. However, the formation of Schiff’s

base is known to be reversible and could lead to gradual release of enzyme during

prolonged exposure to buffer solutions, particularly at elevated pH. Therefore,

reduction of Schiff’s base double bonds using a suitable reducing agent such as

NaCNBH3 has been proposed in order to produce a stable secondary amine that can

tolerate pH variations.13,14

Another preferablemethod for the covalent immobilizationof enzymes on supports

is via epoxide groups. A model study with beads having identical chemistry revealed

that e-amino group of lysine, indole group of tryptophan, and phenol group of tyrosine

residues were mostly involved in the reaction with epoxides forming a covalent bond

between the protein and the support.15While at neutral pH the reaction is slow and the

binding can take several days,16 it is much faster at pH above 9.11 Besides one-step

immobilization, enzyme can also be immobilized through a multistep binding

procedure. One of the most popular multistep immobilization techniques utilizing

the epoxide group involves themodification of epoxidegroupwith a diamine followed

by activation using a glutaraldehyde.17,18 The disadvantage of this immobilization

reaction is a potential for production of undesirable by-products, for example,

homoconjugates and various polymers.11 Another multistep binding procedure in-

volves hydrolysis of epoxide groups using hydrochloric acid or sulfuric acid.19 The

hydrolysis was followed by oxidation of hydroxide groups and reaction with TPCK-

trypsinmolecule.To suppress the reversibility of the formedSchiff’s base and stabilize

the bond with the enzyme, the immobilization was performed in the presence of a

reducing agent, sodiumcyanoborohydride.20Other functional groups that usually take

FIGURE 4.1 Schemes of immobilization procedures for covalent attachment of proteins:

(a) immobilization after support activation using cyanogen bromide; (b) immobilization

after support activation using trichlorotriazine; (c) immobilization of glycoprotein via their

carbohydrate moieties; (d) immobilization after support activation using glutaraldehyde;

(e) immobilization via epoxy group; (f) immobilization via azlactone group; and (g) immo-

bilization using carbodiimide as “zero linker.” Reprinted from Ref. 2, with permission.


part in covalent binding are carboxyl, diol, phenolic groups, and so on (as shown

in Figure 4.1).

To prevent modification of the enzymatic activity or complete inactivation of the

immobilized protein, it is important that the catalytic functional groups of the enzyme

are not involved in the covalent linkage to the support. Unfortunately, many of

the reactivegroups suitable for immobilization are often situated in the active center of

the enzyme. This problem can be sometimes eliminated by immobilization in the

presence of the substrate21 or competitive inhibitor22 of the enzyme. This also helps

stabilize the tertiary structureof theenzymeduring immobilization. Immobilizationof

small molecules on a substrate is typically easy; however, the active center of larger

proteins may no longer be accessible after immobilization. In these cases, improve-

ment can be achieved by introducing a spacer molecule.23 Good steric accessibility of

active sites can be obtained by oriented immobilization of glycoprotein enzymes

through their carbohydrate moieties.24

At the end of immobilization (alternatively during this process), unreacted active

groups of solid support must be blocked by reaction with inert moieties providing that

the same group was used for ligand immobilization. This blocking reaction is

necessary to prevent further nonspecific reactions between support and ligand that

could decrease its stability or specificity.13,25

4.2.2 Physical Adsorption

Physical adsorption of the enzyme onto a solid support is probably the simplest way of

preparing immobilized ligand molecule.26,27 The method is based on nonspecific

physical adsorption between the enzyme molecule and the surface of the

supports. Binding forces involve ionic interactions, hydrogen bonds, van der

Waals forces, hydrophobic interactions, and so on. Because no reactive species are

involved, the conformational changes that might result in change in the biological

activity are less significant.An advantage of adsorption is that usually no reagents, and

only a minimum of activation steps, are required. Unfortunately, the stability of the

adsorbed layer is typically much weaker than in the case of covalent bond, and

desorption of the ligand resulting from changes in temperature, pH, or ionic strength is

often observed. This situation can be partially overcome by a simple regeneration

achieved by the removal of the deactivated enzyme and by reloading with a fresh

active catalyst.28–30

One of the commonly used physical adsorption method is based on the enzyme

binding to the support media via the Lewis acid–base interaction through the divalent

cation chelators such as iminodiacetic acid (IDA), which is chemically bound to the

matrix. The conventional procedures for immobilization of chelator on silica matrix

would be (i) the derivatization of silica with the silane agent and (ii) the chemical

linkage of themetal chelator to the silane-modified silicamaterial. Unfortunately, this

immobilization approach results in low ligand densities because the secondary

reaction has to be performed below pH 8 for protecting the siloxane bonds between

silane and silica matrix, yet the optimum pH at 10–12.31 To solve this problem, the

chelator silane reagent was synthesized at a high pH (pH¼ 11) first, and then


immobilized them on the inner capillary wall to increase the density of immobilized

IDA.30,32 Themetal ion of copper and subsequently enzymewas specifically adsorbed

onto the capillary30,33 or microchip32 surface to form the IMER.

4.2.3 Layer-by-Layer Assembly

The technique of layer-by-layer (LBL) assembly has been developed as a versatile

method to functionalize surfaces.34,35 The process is based on the sequential deposi-

tion of interactive polymers from their solutions by electrostatic, van der Waals,

hydrogen bonding, and charge transfer interactions.36 Since the layer-assembled

microstructures with tailored composition and architecture can be used to incorporate

functional biomolecules, such as proteins, enzymes, and drug molecules,37,38 it is

attractive for applications in biocatalysis, immunosensing, and other biochemical

analysis. The process is superior to other techniques for preparing multilayer thin

films. The assembly is based on spontaneous adsorptions, no stoichiometric control is

necessary to maintain surface functionality, and the assembled films have a good

thermal and mechanical stability. The physicochemical properties of films could be

controlled by adjusting the deposition conditions or the outermost layer of the films.39

A pair of biomacromolecules, positively charged chitosan and negatively charged

hyaluronic acid, was assembled on the surface of a PETmicrofluidic chip using layer-

by-layer deposition for the formation of amicrostructured and biocompatible scaffold

to immobilize trypsin40 (as shown in Figure 4.2). The constructed microreactor

provides a large surface area to volume ratio and a confined microenvironment,

resulting in an increased reaction rate for the sensitive proteolysis of standard proteins

at lower detection limits and also of real biological samples.41

4.2.4 Biospecific (Affinity) Adsorption

Compared to nonspecific adsorption, much better results can be obtained by using

biospecific (affinity) adsorption, for example, the biotin–avidin or streptavidin

technique.23,42 The bonds between the water-soluble vitamin, biotin, and the egg

white protein, avidin, or its bacterial counterpart, streptavidin, are among the strongest

known and can be used for oriented immobilization. The strength of the binding

between biotin and avidin is so strong (Kd¼ 10�15M) that it can tolerate extreme

conditions of temperature, pH, and different solvent systems.42

The biotinylated polylysine was physically immobilized on a glass surface to

capture streptavidin-conjugated alkaline phosphatase.43 This microreactor was

FIGURE 4.2 Depiction of layer-by-layer process for enzyme immobilization. Reprinted

from Ref. 41, with permission.


applied for rapid determination of enzyme kinetics. Biotinylated lipid bilayer44 and

partial biotinylation by photopatterning on fibrinogen45 were also used for immobili-

zation. Themain advantage of the oriented immobilization is good steric accessibility

of the active binding site; however, theses methods are not suitable for long-term use

because of their instability. In addition, these applications are limited to streptavidin-

conjugated enzymes.

Anothermethod of immobilization using bioaffinity interaction is based on protein

A. Protein A, a coat protein extracted from the bacterium Staphylococcus aureus, has

the unique capacity to bind mammalian immunoglobulin G (IgG), especially Fc

constant region of IgG.One can also use other bacterial Fc binding proteins, instead of

protein A, such as streptococcal protein G, which has the advantage of binding to a

wider range of IgG species and subclasses.24 Pearson and coworkers developed a

new flow cytometry method allowing the rapid assessment of a large number of

particle-bound antibodies. Protein G-derivatized POROS beads were used to bind

affinity-purified antibodies specific for synthetic peptides designed from human

plasma proteins.46

4.2.5 Sol–Gel Encapsulation

Other approaches for immobilization may involve a physical step, for example, using

the low-temperature sol–gel technologies for protein encapsulation.47,48 The reaction

involves the hydrolysis and polycondensation of alkoxysilane monomers. During this

process, biomolecules are entrapped in the growing gel network rather than being

chemically or physically attached to the surfaces.49,50 There is no chemical bond

formation between protein and polymer matrix, but enzymes can be chemically

modified without loss of their activity by conjugation to one of the monomers of the

polymerization mixture prior to polymerization leading to covalent bond to the

matrix.51 Because encapsulation occurs under mild conditions, biomolecules retain

their structure and biological activity.With regard to stability, biomolecules entrapped

in sol–gel typically exhibit improved resistance to thermal and chemical denaturation

and increased storage and operational stability.49,52 Owing to the properties of sol–gel

such as large microstructured surface area, porous morphology, and hydrophilicity,

immobilization of enzymes in thematrix could result in high amount of loadingwith a

large extent of bioactivity remaining in the microfluidic device.53

So far, most sol–gel technologies for protein encapsulation describe the use of a

silica-based sol–gel. For example, Toyo’oka’s team developed a revised version of

tetramethoxysilane hydrogel for enzyme encapsulation.47 Trypsin encapsulation was

carried out in a single step under mild conditions within a capillary. In a similar

methodology by the same authors, the tetramethoxysilane hydrogel was applied to

fabricate a trypsin-encapsulated reactor within a sample reservoir of PMMA micro-

chip.48This encapsulated trypsin has proven tobe able to digestmodel peptides,which

were simply electrokinetically driven through the gel, or full proteins, which were

allowed to stay in the gel for 1 h. On-chip capillary electrophoresis (CE) with laser-

induced fluorescence (LIF) detection was then performed to analyze the digestion

products. Encapsulated trypsin was found to have a 19-fold higher activity than free


trypsin, and showed increased stability even after continuous use, compared to that in

free solution.

Despite the many advances that have been achieved in silica sol–gel processes for

bioencapsulation, the problem of miniaturization of the sol–gel structures on micro-

fluidic chips has not yet been resolved due to the fragility of the final gel structure,

manifested by shrinkage of thegel, pore collapse, and/or poor adhesion to the substrate

remains.54 Protein encapsulation techniques using themodified versions of titania and

alumina sol–gel matrixes were developed by Liu et al.55,56 They could get rid of the

fragility observed in silica gel structures and successfully perform the encapsulation of

enzyme to construct microfluidic enzymatic reactors for peptide mapping.57

Although sol–gel encapsulation possesses many advantages, there are still some

problems. The coating materials of sol–gel may hinder the conformational transition

of enzyme and the transport of substrate and product,58,59 resulting in low biocatalytic

activity. A promising solution is to fabricate a single-enzyme containing capsulate

with a thin, permeable coating.60,61 Kim and Grate have fabricated enzyme nano-

particles via a multistep procedure including surface modification, lyophilization,

polymerization in organic solvent, and shell condensation, and successfully obtained

enhanced enzyme stability at an insignificant increase in mass transfer resistance.62

Ouyang et al.63 presented a two-step procedure including surface acryloylation and in

situ aqueous polymerization to encapsulate a single enzyme in nanogel. Horseradish

peroxidase (HRP), an enzymewidely used in bioassay and biosynthesis but fragile to

phase transfer, was chosen as the model enzyme. Compared to the free HRP, the HRP

nanogel exhibited similar biocatalyticbehaviorwhile significantly improving stability

at high temperature and in the presence of polar organic solvent.

4.3 FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC

ENZYMATIC REACTORS

4.3.1 Immobilization of Enzyme on Microchannel Surface

So far, open-channel reactor designs that directly immobilize enzymes on a micro-

channel surface have been usedmost frequently for fabrication of IMERs, owing to the

relatively simple fabrication process. The continuing progress in microchip-based

bioanalysis will depend on the development of novel surface modification technolo-

gies in a simple and reliable fashion. The chemical patterning of a biocompatible

interface within microfluidic channels must be efficient to immobilize domains of

antibodies, enzymes, and other important biologically active compounds for a highly

sensitive detection. In particular, as protein analysis continues to push the limits, the

availability of new strategies will become more critical.41

Many of the reported microreactors are based on immobilization of enzyme

directly on the surface of a fused silica capillary. Amankwa andKuhr8,42 immobilized

the enzyme on the inner surface of a 50 mm ID aminoalkylsilane-treated fused silica

capillary via biotin–avidin–biotin coupling. Because the enzyme was coated on the

capillary wall, a very low flow rate (e.g., 40 nLmin�1) was needed to permit time for


diffusion of the protein sample to the immobilized enzyme. Kuhr’s group64,65 later on

found that the proteolysis reaction rate could be enhanced by applying low-power

acoustic vibration to the capillary, with digestion carried out in a batch-wise proce-

dure. Efficient tryptic digestion of large proteins has been carried out in as few as

30min.66 The capillary microreactors were used for protein characterization by

trypsin, pepsin, and carboxypeptidase Y digestion. The introduction of a functional

group on the microchannel surface was also used for covalent binding of enzyme.

Laurell et al. reported an enzyme reactor with high aspect ratio channels.67,68 The

glucose oxidase was immobilized on the whole of the sidewalls of the microchannels

by sequentially pumping the silane, glutaraldehyde, and glucose oxidase through the

microchannels. Trypsin was also covalently attached to the ultraviolet (UV)-modified

PMMA surface using coupling reagents N-(3-dimethylaminopropyl)-N0-ethylcarbo-diimide hydrochloride (EDC) and hydroxysulfosuccinimide (sulfo-NHS).69 The

bioreactor provided efficient digestion of a test protein, cytochrome c, at a flow

rate of 1mLmin�1, producing a reaction time of 24 s to give adequate sequence

coverage for protein identification.

Although the direct immobilization methods are easy, they are limited by both the

low surface area to volume (S/V) ratio of the capillary or microchannel and long

diffusion times. Because of relatively lowS/V ratio and long diffusion distances, avery

low flow rate was needed to guarantee sufficient time (half an hour or more) for

diffusion of the protein molecules to the immobilized enzyme. The S/V ratio can be

increased byusingverynarrowbore capillaries ormicrochannels, for example,with an

inner diameter of 10mm or less.70 Alternatively, problems related to long diffusion

times and low S/V ratio of the open tubular bioreactor can be avoided by capillary or

microchannel surface modification. The fused silica capillary was pretreated with

NH4HF2, in order to increase the surface area and wettability, permitting a homoge-

neous spreading of the hydrophilic carboxylsilyl layer.30 The silanol groups on the

roughed surface of capillary were then activated using HCl accompanied by removal

of themetal ions from the capillary inner surface. After that, GLYMO-IDA-silanewas

introduced to react with the capillary wall, and then Cu2þ solution and the enzyme

buffer solution were introduced to form an immobilized enzyme capillary micro-

reactor. The time necessary to complete the digestion of standard proteins in a 100 cm

long capillary IMER was about 15–30min.

Ekstr€om et al. developed a modified sol–gel technique to form nanostructures

on a silica microchannel surface71 that modifies the microchannel surface first with

silanizationusing apolymerized copolymer of (3-aminopropyl)triethoxysilane and/or

methylsilane, followed by glutaraldehyde activation, and final enzyme coupling.

Using this method, an increased surface area was obtained, and at least 10 times more

enzymes can be immobilized on these nanostructures by covalent cross-linking

through amide bond formation, disulfide orHis-tag, or by using amodifying succinate

spacer, compared to single-layer immobilization.72 Amicroreactor with immobilized

cucumisin on the nanostructured surface could process substrate 15 times faster than

the batch-wise reaction.72

The surface of polymeric microchannel is commonly hydrophobic that results in

poor wettability with aqueous solvents and promotes nonspecific protein adsorption.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC 133

It is also relatively inert to direct chemical modification. Therefore, surface modifi-

cation must be performed before enzyme immobilization in polymeric microchip.

Chemical modification could be performed by introducing carboxyl groups to poly

(dimethylsiloxane) (PDMS) surface based on ultraviolet graft polymerization of

acrylic acid.73 The covalent and physical immobilization of trypsin was carried

out using activation reagent 1-ethyl-3-(3-dimethyl aminopropyl)carbodiimide

(EDC)/N-hydroxysuccinimide (NHS) and a coupling reagent poly(diallyldimethy-

lammoniumchloride) (PDDA), respectively. The lab-made devices provided effective

digestion of several model proteins even at the fast flow rate of 3.5mLmin�1 for the

EDC/NHS-made device and 0.8mLmin�1 for the PDDA-made device, which af-

fordedvery short residence timesof 5 s and20 s.Avinyl group-containingPDMSplate

was fabricated with liquid silicon rubber containing pyrogenic silicic acid as a filler.4

Apart from its effect on elastomer properties, the silicic acid is expected to provide

additional silanol groups for surface chemistry. Bas-relief microstructures were

incorporated every 2.5mm along each microchannel of a PDMS plate, alternating

between the left and the right wall4 (see Figure 4.3a). They were included as flow

obstacles to improve mass transfer to and from the microchannel surface through a

passive mixing effect.6,74 Covalent protein attachment was then utilized, via cross-

linking with glutaraldehyde on the amino-silanized microstructured surface of the

reaction plate. Comparison of the activated and the untreated microchannels using

FIGURE 4.3 (a) Photograph of the microstructured PDMS multichannel plate. An SEM

picture of a passive mixing element is shown in the inset; electron micrographs of identical

sections of (b) microstructured plates without treatment; (c) after aminosilanization and

activation with glutaraldehyde; and (d) characteristically uneven distribution of immobilized

enzymes across the channels of the microstructured plate. Reprinted from Ref. 4, with

permission.


SEManalysis revealed surfacemodificationwith a layered structurewhose formation

is readily explained by the reaction of the surface silanol groups and most likely

requires the latter (see Figure 4.3b and c). SEM analysis was also used to characterize

b-glycoside hydrolase CelB immobilization with respect to spatial distribution of

enzyme along the microchannels. Protein binding was not uniform, and clusters of

aggregated protein were observed in the microchannels, often coating the passive

mixing elements (see Figure 4.3d)4.

Poly(methyl methacrylate) (PMMA) is another most commonly used polymer

substrate for IMER fabrication. Employing the strategy of silane-based chemistry to

introduce a variety of active groups, a craft copolymer has been designed and

synthesized, which afforded a firm, stable but easy-access silane-functionalized

chemical scaffold on a PMMA-based microchannel. These silane functional groups

readily reacted with the sol–gel to form a stable bonding through a silicon–oxygen–-

silicon bridge. Thus, anchorage of proteins could be realized on the hydrophobic

PMMA microchannels while preserving bioactivity.53

Surface modification methods incorporating the sol–gel technique were also

developed for IMERs fabricated with polymer substrates. A trypsin-encapsulated

titania (titanium dioxide) and alumina gel matrix was immobilized through the SiOH

group formed on a PDMS surface by plasma oxidation.57 These SiOH groups act as

anchors on themicrochannel wall linked covalently to the hydroxyl groups of trypsin-

encapsulated solmatrix.As a result, the trypsin-encapsulated gelmatrixwas anchored

to the wall of the microchannel, and the leakage of gel matrix from the microchannel

was effectively prevented. Using this device, digestion times were significantly

shortened (�2 s) and the application for high-throughput protein identification

was realized.

On the other hand, current chip fabrication protocols generally do not allow or will

destroy organic coatings when the two halves of the system are bonded. This means

that organic coatings must be applied to the fully fabricated channel system,

commonly by flowing reactants into the chip. The problem with this approach is

that reagents are dispersed throughout the channel system and all channels are coated.

This is undesirable in the case of some types of immobilized enzyme reactors. The

problem was addressed by using electroosmotic flow to direct reagents to specific

channels in a channel network.75 The route of transport, and thus the specificity of

channel coating, was controlled by the well to which negative potential was applied.

Flow in a multichannel network took the shortest route between the electrodes

delivering the motive potential. Different reagents in the reaction were delivered

from different wells and took different paths through the channel network. Only the

separation channel was in the flow path of all the reagents used in the coating process

and thus had channel-specific immobilization of the enzyme (Figure 4.4).

Rubloff et al. developedamethodology that enabled theprogrammableassemblyof

biomolecules on localized assembly sites inmicrochannels using electrodeposition of

the amine-rich polysaccharide chitosan to direct the assembly.76 They further dem-

onstrated that a metabolic pathway enzyme, S-adenosylhomocysteine nucleosidase

(Pfs), could be assembled in this way and that its catalytic action was retained in the

microfluidic environment, shownby conversion of substrate S-adenosylhomocysteine


(SAH) to products S-ribosylhomocysteine (SRH) and adenine (Figure 4.5a).77 While

promising as amethodology to replicatemetabolic pathways and search for inhibitors

as drug candidates, these investigations also revealed unintended (or parasitic) effects,

including products generated by the enzyme (1) in the homogeneous phase (in the

liquid) or (2) nonspecifically bound to microchannel surfaces. To reduce homoge-

neous reactions, a new packaging and assembly strategy was developed that

FIGURE 4.4 The flow direction of (a) the first-step chemical modification and (b) the

second-step chemical modification. The arrow indicates the direction of electroosmotic flow.

Reprinted from Ref. 75, with permission.

FIGURE 4.5 Minimize parasitic reactions by eliminating interconnect reservoirs and by

separating sequential flow directions in cross channels. To test the background signal

by parasitic reactions, Pfs enzyme solution was introduced without electroassembly followed

by buffer rinsing, then enzymatic substrate SAH was introduced, and products were collected

downstream to be analyzed by HPLC. (a) Single channel with interconnect reservoirs.

(b) Single channel without interconnect reservoirs. (c) Cross channel without interconnect

reservoirs. Reprinted from Ref. 78, with permission.


eliminated fluid reservoirs that were commonly used for fluidic interconnects with

external tubing (Figure 4.5b). To suppress reactions by nonspecifically bound enzyme

onmicrochannelwalls, a cross-flowmicrofluidic network designwas implemented so

that enzyme flow for assembly and substrate/product for reaction shared only the

region where the enzymewas immobilized at the intended reaction site, as illustrated

in Figure 4.5c.78

4.3.2 Packing Microchannels with Enzyme-Immobilized Micro/Nanoparticles

To overcome the limitations caused by low specific inner surface of microchannels,

besides surface modification, another alternative is to immobilize enzyme molecules

on the surface of different supports (micro/nanoparticles), followed by fixing them in

the microchannel by weirs, frits, or membranes.

Theprocedures for fabrication of IMERswith packingmicro/nanoparticles include

choosing suitable supports, immobilization of enzyme on the support surface, and

assembling the supports into the microchannel. Immobilization methods have been

described in Section 4.2. Immobilization of enzymes on supports can be obtained

namely “in situ” or “in batch.” When employing the in batch process, the enzyme is

first immobilized on the support and then packed into the microchannel using a slurry

packing technique,whereas in the in situ approach the enzyme is directly immobilized

on the prepacked column. Although the in-batch process takes the advantages of

reduced cost due to mass producibility, better reproducibility, and quality control, it

could result in a loss of catalytic activity.Massolini et al.28 have compared in batch and

in situ techniques of penicillinGacylase (PGA) immobilization onvarious derivatized

silica supports concluding that the in situ technique is the best way to obtain

satisfactory results in terms of bound amount of PGAand enzymatic activity retention.

Choosing the right supports is vital and getsmore andmore complicated because of

the increase in thenumberofnatural and synthetic supports available that greatlydiffer

in mechanical and physical properties. Furthermore, it is also necessary, at the same

time, to meet many requirements such as low cost, non toxicity, maximum activity,

high retention of catalytic activity over a long period, enzymatic stability, ease of

protein availability, and immobilization. The surface of the supports, on which the

enzyme is immobilized, has an important role to play in retaining the tertiary structure

of the enzyme that highly influences the thermal stability and catalytic activity of the

immobilized enzyme. Indeed an immobilized enzyme is known to acquire novel

kinetic properties that canmodify theMichaelis–Menten constant (Km) andmaximum

velocity (Vmax) and cause a shift of the pH and temperature-activity profile. Likewise,

groups involved in the attachment of proteins to the support must be different from the

active sites of enzymes. Therefore, the choice of both the support and the technique

depends on the nature of the enzyme, on the nature of the support, and on its ultimate

application. For this reason, it is not possible to recommend any universal immobili-

zation methods. The toxicity of immobilization reagents should also be considered

when final applications concern the food processing and pharmaceutical industries.12

An important factor in the preparation of a bioactive reactor is the structure of the

support since this determines accessibility of active sites to substrates. The ideal


support must be inert, stable, and resistant to mechanical strength. However, the other

physical properties, such as form, shape, porosity, pore size distribution, swelling

capability, and charges, are also very important because they influence the kinetic

process. Indeed, the reaction rates of the immobilized enzyme depend on the

enzymatic intrinsic activity, on the substrate accessibility to interact with the

active sites, on the amount of the loaded enzyme, and on the substrate concentration

and diffusivity.

The substrates must be able to diffuse from the bulk phase toward the surface and

product away (external diffusion) and within the pores of immobilized enzyme

particles (internal diffusion). These effects of diffusion, which are enhanced when

the enzyme is entrappedwithin amatrix, limit the reaction rate because they affect the

concentration of the substrate/cofactor in the vicinity of the enzyme. In this way, a

diffusional layer around the immobilized enzymes is formed and its thickness is

correlated with mass transfer effects. Thin diffusion layer, as opposed to a thick

layer, results in a low diffusional resistance. Methods to minimize the diffusional

effects could be the decreasing of enzyme loading and the increasing of substrate

concentration and diffusivity. The latter feature is strongly influenced by a hydrody-

namic parameter such as the flow rate that, as it increases causes a decrease in the

diffusional layer.

Care must be taken also in selecting the support materials because their character-

istics strongly influence, as previously mentioned, the accessibility of active sites to

substrates. For example, the more the pore diameter and size distribution increase the

more the surface area decreases. Therefore, it is generally preferable to choose pores

with a small diameter if the substrate has similar molecular dimensions. If substrates

with high molecular weight are implicated in the enzymatic reaction and their

diffusion in the active site is sterically hindered, a significant intraparticular mass

transfer resistance,which in turn significantly decreases the overall reaction rate,must

be evaluated.

Another feature influencing the enzymatic activity is the particle size, as is known,

the bigger the particle size, the greater the effect of diffusion control and less the

activity. To make the correct choice, it is also important to consider the relation of

particle size with pressure drop that are correlated in an inverse mode.

Therefore, the evolution of microchannel packing material is to minimize the

diffusional limitations by decreasing the size and optimizing the geometry of

immobilized biocatalyst particles, by decreasing the substrate concentration, by

enhancing the flow rate, by increasing the porosity, and by optimizing the biocatalyst

distribution in the beads.

4.3.2.1 Nonmagnetic Supports

Inorganic Supports The main supports used in enzyme immobilization are porous

inorganic solids such as the controlled pore glass (CPG) and silica. CPG presents a

higher thermal stability and a resistance to acids, whereas silica is characterized by a

larger specific surface area. Both supports must be derivatized with functional groups

that can interact covalently with enzymes. This feature may be obtained in the


laboratory or by using a commercial derivatized support.12 When uncoated CPG is

employed, the researchers use 3-aminopropyltriethoxysilane (APTES)79,80 to obtain

an aminopropyl-CPG that could be further activated with glutaraldehyde and then

reactedwith enzyme.81Microreactorswith enzymes immobilized on glass beads have

been prepared by filling the reaction chamber with beads; such a device was used for

the determination of xanthine using chemiluminescent detection82 and for online

protein digestion.81 In the case of uncoated silica, glicidoxypropyltrimethoxysilane83

is used to obtain epoxy-silica.

Organic Supports When the enzyme requires an alkaline medium, polymeric

supports are employed. Among these organic solids, poly(vinyl alcohol) (PVA)

activated with tresyl group is most used in in situ enzyme immobilization. The

performance of these phases depends on the kind of enzyme and on the number of

injections. In the case of leucine and 3-hydroxybutyrate dehydrogenase coupled to

NADHoxidase, a satisfactory reproducibility is obtainedwithin 10 days after 300 and

400 injections, respectively, of samples stored at 4�C when not in use.84 Crooks and

Seong developed advanced analytical microreactors using enzyme-immobilized

microbead mixing,85 and efficiently performed multistep enzyme reactions using

glucose oxidase and horseradish peroxidase immobilized on polystyrene.

Furthermore, the immobilization of enzymes on nickel–nitrilotriacetic acid

(Ni–NTA) agarose beads has also been reported and was applied to immobilize

bacterial P450;86 this immobilized enzyme was less denaturated because binding of

the enzyme was achieved using a histidine (His)-tag.

Manymethods have been reported for packing enzyme-immobilized supports into

microchannels. Microfabricated weirs87,88 or elevated structures (“bead stopper”)89

have been developed for keeping the supports packing in place. Andersson et al.

reported micromachined chambers surrounded by filter-like structures used for solid-

phase DNA sequencing.90 Sato et al. fabricated a barrier in a microchannel for

blocking derivatized beads used in chip-based immunoassays.91 A PDMS polymer

device for an on-chip, fritless, capillary electrochromatography has been realized,

characterized by a tapered column inwhich stationary phase particleswere retained.92

Fabrication using deep reactive ion etchingwas required to produce the siliconmaster.

Although successful to various extents, reported approaches often encountered

difficulties in the fabrication of on-chip packed column reactors, and procedures

commonly involved the use of sophisticated equipment, such as deep reactive ion

etching92 or microfabrication techniques involving multiple exposures to achieve

multiple-layered etching.89 Zhang et al. developed a novel fabrication approach that

enabled the enzyme-immobilizedmicrobeadswith controlled sizes to be placed at any

desired position on the microchip.93 The location of entrapped beads acting as the

support of enzymes can be easily controlled by spotting the slurry of beads at desired

position on the separation channel. The length of the immobilized beads region was

determined by the diameter of the spot, the low limitation of which is 0.5mm using a

10mL pipette tip. Different linear ranges of the biosensor can be obtained with varied

lengths. And the width of the region was controlled by the width of the separation

channel (200 mm). The concentration of microbeads slurry controlled the coverage


degree of microbeads immobilized on the surface of microchannels. The coverage

degree increased with the concentration of the slurry (Figure 4.6).

Even when successfully produced, the high backpressure generated even at low

flow rates often limits the use to relatively short column beds. To reduce the back-

pressure, functional groups could be directly attached to the walls of open-channel

reactors. However, such an approach significantly reduced the reactive capacity of

the reactor even after the surface modification. To decrease the flow resistance of

the nanoparticle-packed IMERs while maintaining reasonably high reactive area,

enzyme-immobilized nanoparticles could be immobilized on themicrochannel walls,

and left considerable space above the particles for low-resistance flow. CPG reactive

particles were immobilized on PDMSmicroreactor beds94 using a three-layer design

that was soft enough to allow implanting of the particles by applying gentle pressure

over the particles (Figure 4.7).

Following full polymerization, the particles were permanently immobilized. The

curing at a temperature of 37�C for 5 h produced no deleterious effects on the activity

of the immobilized glucose oxidase (GOD) particles. Silica nanoparticles were

immobilized on the surface using slow evaporation of the particle suspension in a

filled-in microchannel.95 The resultingmicrochannel was subjected to treatment with

3-aminopropyltriethoxysilane, and immobilization of enzyme was achieved by

covalent cross-linking through an amino group. Although physical stability needs

to be improved, a lipase-immobilized microreactor prepared by this method showed

1.5 times faster kinetics than those of a microreactor obtained by sol–gel surface

FIGURE 4.6 (a) Optical (left, A and C) and corresponding fluorescence images (right, B

and D) of the immobilization of microbeads with different concentrations. The concentrations

inA andCwere 1 and 30mgmL�1, respectively. Thewhite dots in optical images and the green

dots in fluorescence photos were the immobilized microbeads bonded with FITC-BSA.

(b) Optical image of microchannels with multireactors. The white part in the image represents

the enzyme reactors. Reprinted from Ref. 93, with permission.


modification.96 This result showed good correlation with the surface area: particle

arrangement has approximately 1.5 times larger surface area and could immobilize

more enzymes. Because of its good biocompatibility, charge-stabilized gold nano-

particles (AuNPs) provide a mild microenvironment similar to that of proteins in

native states and give protein molecules more freedom in orientation. They are

biocompatible and nontoxic and can offer large specific surface areas for ready

binding of a large range of biomolecules such as amino acids, proteins, and anti-

bodies.97 PDDA/AuNP multilayer films containing protease were assembled on the

surface of poly(ethylene terephthalate) (PET)microchannels to obtain a flow-through

protein digestion biochip. Sequential alternate adsorption of the cationic polyelectro-

lyte PDDA and the anion-coated AuNPs led to the formation of a biocompatible and

large specific surface to volume ratio network to immobilize the enzymes.98

Zeolite nanoparticles have been widely studied in the last decade and have drawn

much interest due to their large external surface area compared to conventional zeolite

crystals, high dispersibility in both aqueous and organic solutions, high thermal and

hydrothermal stabilities, and tunable surface properties such as adjustable surface

charge and hydrophilicity/hydrophobicity.99 The unique properties make nanozeo-

lites promising candidates for microfluidic surface modification and enzyme immo-

bilization. Silicalite-1 (S-1, all-silicaMFI-typezeolite nanoparticles)was selected and

successfully used to modify the PMMA surface. The silanol groups were introduced

and readily reacted with sol–gel to form stable microstructure matrices in micro-

channels (Figure 4.8). Trypsin was then stably immobilized within the PMMA

microchannel to fabricate an enzymatic on-chip microreactor.100

4.3.2.2 Magnetic SupportsAs mentioned above, difficulties were often encountered in packing of the enzyme-

immobilized nanoparticles in the microchannel. Besides the requirements of elaborate

in-capillary/microchannel chemistry, the reproducible filling of the microchannel still

remains a challenge. During the past decade, magnetic nanoparticles are gaining

increasing attention due to their ease of manipulation and recovery. On this basis,

magnetic nanoparticles have many uniquemagnetic properties such as superparamag-

netic, high coercivity, low Curie temperature, high magnetic susceptibility, and so on.

Therefore, they are of great interest for researchers from a broad range of disciplines,

FIGURE 4.7 CCD images of (a) the GOD-CPG particles (200–400 mesh) immobilized on

the PDMS surface of a section of the reactor and (b) a cross section of the immobilized particle

bed. Reprinted from Ref. 94, with permission.


including magnetic fluids, data storage, catalysis, and bioapplications.101–105 Thanks

to the unique magnetic properties, the magnetic nanoparticles offer the advantage

of straightforward and fast handling by using magnets or magnetic coils. Accordingly,

the use of magnetic nanoparticles, which does not require elaborate in-channel

chemistry or the use of sophisticated equipment for reversible packing of the supports

in the microchannel, is shown to have significant potentials in microfluidic

device.106,107

Generally, magnetic nanoparticles were prepared by encapsulating inorganic

magnetic particles (usually magnetite or maghemite) with organic materials or

inorganic materials, such as polymers, silica, metals, metal oxide, and so on.108,109

Synthesis and surface fictionalization strategies of magnetic nanoparticles were

discussed by Jiang et al.110

Polymer Encapsulation Among the different organic materials that can be used to

encapsulate themagnetic nanoparticles, polymers are of particular interest because of

their wide range of properties. Polymer coating will increase repulsive forces to

balance the magnetic and the van der Waals attractive forces acting on the magnetic

nanoparticles. In addition, polymer coating on the surface of magnetic nanoparticles

offer a high potential in the application of several fields. To use these materials for

FIGURE 4.8 SEM images of microchannel modified with silicalite-1. (a) Planform of the

channel at low magnification. (b) Planform of the channel at high magnification. (c) Cross

section of modified microchannel at high magnification. Reprinted from Ref. 100, with

permission.


fundamental or applied research, access to well-defined magnetic nanoparticle

samples whose properties can be “tuned” through chemical modification is necessary.

In a number of cases, it has now been shown that, through careful choice of the

passivating and activating polymers and/or reaction conditions, can producemagnetic

nanoparticles with tailored and desired properties. The use of polystyrene-encapsu-

lated superparamagnetic beads (2.8mmdiameter) has been exploredwith commercial

CE instrumentation for performing enzymatic and inhibition assays, as well as for

analysis of biological molecules such as antigens and substrates.107 Trypsin was

immobilized on nonporous COOH-functionalized polystyrenic magnetic particles

(�626 nm) andwas used for protein digestion.111Magnetic particles coated with poly

(N-isopropylacrylamide), polystyrene, poly(2-hydroxyethyl methacrylate-co-ethyl-

ene dimethacrylate), poly(glycidyl methacrylate), [(2-amino-ethyl)hydroxymethy-

len]biphosphonic acid or alginic acid were utilized for trypsin immobilization and

organized by an inhomogeneous external magnetic field in the microchannel.3

Silica Encapsulation Silica is also used for preparing the functionalized magnetic

nanoparticles. Easily replaceable and regenerable IMERs have been fabricated with

packing bed of magnetic silica nanoparticles that immobilized trypsin by metal-ion

chelated adsorption32,33 or covalent binding.13 Magnetic nanoparticles with small

size (�300 nm in diameter) and high magnetic responsivity to magnetic field

(68.2 emu g�1) were synthesized and modified with tetraethyl orthosilicate (TEOS)

(Figure 4.9a and b).

FIGURE 4.9 Schematic illustration of trypsin immobilization on magnetic silica nanopar-

ticles with (a) metal-ion chelated adsorption,32 and (b) covalent binding,13 and (c) amine-

functionalizedmagnetic nanoparticles.14 Reprinted fromRefs 32, 13, and 14, respectively, with

permission.


For metal-ion chelated adsorption of enzyme, the metal chelating agent of

iminodiacetic acid was reacted with glycidoxypropyltrimethoxysilane (GLYMO)

before its immobilization on the surface of magnetic silica nanoparticles. The metal

ion of copper and enzyme were subsequently adsorbed onto the surface

(Figure 4.9a).32,33 For covalent binding of enzyme, aminopropyltriethoxysilane

and glutaraldehyde (GA) were introduced to functionalize the magnetic silica

nanoparticles. Trypsin was then stably immobilized on the magnetic silica nanopar-

ticles through the reaction of primary amines of the proteins with aldehyde groups on

themagnetic silica nanoparticles (Figure4.9b).13However, despite popularity of silica

stationary phases in chromatography, application of silica-based supports for the

immobilization of proteins traditionally lags behind the use of organic polymers. The

reason may be the limited arsenal of reactive chemistries available for silica, the

danger of nonspecific interactions with surface silanols, and the limited hydrolytic

stability of the support.

The abovemethod proved to be effective and has been successfully used for protein

digestion. However, when using thesemethods,multiple steps of surfacemodification

on the magnetic microspheres were required prior to trypsin immobilization, which

resulted in a complicated and time-consuming procedure. Koh et al.112 modified

commercially available magnetic nanoparticles with APTES and then activated them

with glutaraldehyde prior to the immobilization of proteins on the nanoparticle

surface. Theirwork offered a simplerway of preparing protein-immobilizedmagnetic

nanoparticles; however, five reaction steps are still needed. The work done by

Nishimura et al.113 involves the in situ preparation ofmagnetic nanoparticles (mixture

of Fe3O4/g-Fe2O3) in the presence of trypsin at 4�C by chemical coprecipitation of

FeCl2 and FeCl3 using NH4OH as precipitator). This approach can directly lead to

trypsin-modifiedmagnetic nanoparticles; however, because the synthesis temperature

is too low, the obtained magnetic nanoparticles have poor crystallization as indicated

by theX-ray diffraction spectrum.As a result, themagnetic nanoparticles possess poor

magnetic response that may influence the practical application. In addition, because

the magnetic nanoparticles were synthesized in the trypsin aqueous solution, the

location of trypsin molecules in the trypsin-modified magnetic nanoparticles is ill

defined. Li et al. reported a novel and facile way of the preparation and application of

trypsin-immobilizedmagnetic nanoparticleswith superparamagetism.14First, amine-

functionalized magnetic nanoparticles were prepared through facile one-pot sol-

vothermal synthetic strategy. Then, magnetic nanoparticles were functionalized with

numerous aldehyde (–CHO) groups followed by immobilization of trypsin through

reaction of the aldehyde groups with amine groups of trypsin (Figure 4.9c). These

trypsin-immobilized magnetic nanoparticles were also successfully used for the

preparation of an easily replaceable on-chip enzymatic microreactor.114

One of the most interesting properties of a suspension of superparamagnetic

particles is its ability to self-organize in a magnetic field. When exposed to a uniform

external magnetic field, the magnetic particles acquire a magnetic moment. The

resulting dipole interactions cause an instantaneous self-organization of the suspen-

sion into a structure consisting of a columnar clustering in the direction of the field.

These columns are in turn organized in the direction perpendicular to the field, in


structures that depend on the container geometry, particle density, and magnetic field

history. The magnetic field can be provided by permanent magnets or magnetic coils.

The enzyme-immobilized magnetic nanoparticles were locally packed into the

microchannel by the application of one permanent magnet115 or two magnets. Two

permanentmagnets could be placed in a repulsive conformation, with the polarization

making a small angle (30�)with a straight channel that created amagnetic field parallel

to the flowwith a strong gradient pointing through the center of the chip channel. In the

beginning, particles self-organize in chain-like columns along the channel direction

(Figure 4.10, inset). When the concentration increases, the plug becomes opaque,

probably due to the formation of a “labyrinth like” structure116 made of tortuous and

ramified “walls”with onedirection collinear to thefield. Thedistance betweenwalls is

maintained by dipole–dipole repulsion, keeping in the bulk of the plug channels

collinear to the flow, with a thickness of a few micrometers.111

Thingsweredifferentwhen the twomagnetswereplacedwithopposite poles facing

each other perpendicular to the channel axis.3 When approaching the magnets, the

nanoparticles got magnetized and arranged in free-floating growing chains oriented

perpendicular to the flow field. In about 60min, these chains close to the center finally

staggered to a dense plug of particle clusters (Figure 4.11).3 The time needed for

formation of packing bed could be reduced within 1min by using magnetic nano-

particles with higher magnetic responsivity to magnetic field.13

4.3.3 Monoliths

In recent years,monolithic phaseshaveemergedas an attractiveand increasinglymore

popular alternative to packed columns due to simplicity of preparation and virtually

unlimited choice of chemistries they offer. In addition, there is no need for retaining

frits, and very fast separations can be achieved due to the typically lower flow

resistance even with smaller pore sizes. Perhaps the most appealing aspect of

FIGURE 4.10 Microreactor with the plug of magnetic beads maintained between the two

magnets; the inset is a 1006 microphotograph of the columns at the beginning of the formation

of the plug. Reprinted from Ref. 111, with permission.


monolithic materials is the ease of preparation. The simple polymerization process

starts from liquidprecursors (polymerizationmixture) and is performeddirectly inside

the capillary or a microfluidic chip. In contrast to packed beds, monolithic structures

exhibit excellent dimensional stability. The through-pores ofmonolithicmaterials can

be easily controlled allowing high-speed flows at low backpressures and the surface of

themonoliths can be easily chemicallymodified. Suchflexibility is ideal for designing

and developing the enzymatic reactor tailored for specific applications. Petro et al.17

describedcomparative studies inwhich trypsinwas immobilizedonbothmacroporous

poly(glycidyl methacrylate-co-ethylene dimethacrylate) (poly(GMA-co-EDMA))

beads and on chemically analogous monolith. Monolith and beads were modified

by a multistep process involving the modification of epoxide groups with ethylene-

diamine followed by activation using glutaraldehyde and final modification with

trypsin. Despite the relatively small size of the monodisperse beads used to minimize

the diffusional path length, the processivity of the enzyme immobilized on the

monolithic material was nearly two times higher compared to that of the bead-based

conjugates.17 Immobilized proteolytic enzyme reactors with monoliths as supports

have been summarized by Svec.117

The current monoliths can be subdivided in two categories (inorganic monoliths

and organic synthetic polymer-basedmonoliths) according to thematerial fromwhich

they are prepared. Themost frequently used inorganicmonoliths are silica monoliths,

while some other inorganicmonolithic materials have also been exploited for enzyme

immobilization. Yi et al. developed a novel immobilized trypsin reactor with titania

monolith as the carrier.118 The material was prepared from biocompatible precursors

FIGURE 4.11 Filling of microchannel by magnetic nanoparticles. (a–d) Obtained after 10,

30, 45, and 60min; flow rate ofmobile phase 1mLh�1. Reprinted fromRef. 3, with permission.


using aqueous processing conditions involving the formation of a glycerol–titania

composite sol and titania condensation.By adding poly(ethylene oxide),macroporous

titania monolith was obtained. g-Glutamyl transpeptidase, a clinically relevant

protease, was then entrapped in the monolithic network.

4.3.3.1 Silica MonolithThe conventional silica-based monolithic columns are usually prepared by sol–gel

approach, in which a porous silica rod could be formed by the hydrolytic polycon-

densation of alkoxysilane.Due to the existence ofmicrometer-size flow-throughpores

constituting amacroporousnetwork andnanometer-sizemesoporeson the skeleton, as

shown in Figure 4.12,119 silica-based monoliths have various merits, such as low

backpressure drop across the column, good permeability, and fast mass transfer

kinetics. So far, two different approaches have been used for the immobilization of

proteins on silica monoliths: activation of preformed silica monoliths followed by

enzyme immobilization or entrapment via sol/gel.

Activated Silica Column Calleri et al.120 used a thoroughly dried commercial

4.6mm ID monolithic silica column and activated its pore surface by reaction with

3-glycidoxypropyltrimethoxysilane in toluene. A combination of these two methods

enabled good characterization of the surface functionalities. b-Glucuronidase was

also immobilized on a silica monolith modified with 3-APTES and activated with N,

N0-disuccinimidylcarbonate,andwasusedfordeterminationofdextromethorphanand

dextrorphan in urine.121 Less typical is the immobilization of ascorbate oxidase by

physical immobilization on plain silica monolith prepared in situ in poly(ether ether

ketone) (PEEK) capillary and the use of the reactor for monitoring dopamine in the

presence of ascorbic acid.122 Presumably, the enzyme interacts with acidic silanol

functionalities.However, these coulombic forces arenotverystrongand the lifetimeof

the reactor can be significantly impaired.

Encapsulation in Sol–Gel The simplest, yet least used approach to immobilization

via sol/gel transition is encapsulation of the enzyme within the newly formed silica

matrix. Kawakami123 used immobilized protease P in a monolith formed in PEEK

capillary to afford a reactor for transesterification of vinyl butyrate. Kato and

FIGURE 4.12 SEM picture of the typical porous structure of (a) monolithic silica columns,

(b) the mesoporous structure of the silica skeleton, and (c) the macropores or through-pores.



coworkers developed a simple in situ encapsulation procedure to prepare the im-

mobilized trypsin reactor.124 After mixing with a fully or partially hydrolyzed silane,

trypsin could be well encapsulated in the hydrogel after several days. The enzymatic

activity of the resultantmonolithic reactor was about 700 times higher than that in free

solution. Itwas noteworthy that, by immobilizing trypsin in themonolith located at the

upstreamof a separation capillary, the authors enabled the enzymatic digestion andCE

separation in a single capillary. Thereafter, they improved this technique by coating

trypsin-containing gel on a porous silica monolith,125 which was subsequently fitted

into a 96-well plate for high-throughput proteome analysis. It was found that the

encapsulated trypsin within the gel matrix could offer high catalytic turnover rate

due to the large surface area of monoliths. All these experiments were carried out in

thebatchmode relayingon slowdiffusion to achieve themass transfer of both substrate

and products. As a result, the digestion is slow. Although interesting, this approach

does not fully utilize the potential of monolithic supports well demonstrated in flow-

through applications.

Photopolymerized Sol–Gels In contrast to the classical preparation of inorganic

monoliths using hydrolytically initiated polycondensation of alkoxysilanes, Zare’s

group126 introduced a novel approach to inorganic monolithic columns for CEC

combining photopolymerization and sol–gel transition. The monolith is obtained

in a single step during which both addition polymerization and polycondensation of

[3-(trimethoxysilyl)propyl] methacrylate simultaneously proceed in the presence of a

porogen. This photochemical route facilitates the exact placement of the monolith

within the device and the resulting material exhibits a high mechanical strength.

Perhaps themajor advantage of this approach is that there is no need for drying at high

temperature that may lead to cracking of the monolith. This technique was demon-

strated with a capillary integrating protein digestion into an immobilized enzyme

reactor with electrophoretic separation of the digest andmass spectrometric detection

of peptides.127Averyhigh percentage of protein in this solution (25%)was required to

achieve sufficient activity of the immobilized pepsin. Later on, they enhanced the

activity of enzymatic reactor by covalently bonding trypsin to such monolithic silica

via Schiff chemistry at room temperature, in which an alkoxysilane reagent with an

aldehyde functional group links to an inactive amine on trypsin to form an imine

bond.128 The results suggested that the proteolytic activity of such an immobilized

trypsin was increased by 2000-fold compared to that obtained in solution.

Another elegant approach includes the preparation of a 1 cm long monolithic plug

in a capillary via photopolymerization of amixture of condensed [3-(trimethoxysilyl)

propyl] methacrylate and PEG dimethacrylate followed by functionalization with a

toluene solution of (trimethoxysilyl) butyraldehyde passing through the monolith for

2 h.128 After a thorough wash with ethanol, the pores were filled with trypsin solution

and the immobilization reaction allowed to proceed for 19 h.

4.3.3.2 Organic Polymer MonolithsDue to a wide variety of chemistries and formats readily available, organic polymers

are very popular supports for enzyme immobilization.


Acrylamide Copolymers The use of acrylamide gels is almost synonymous with

enzyme immobilization via entrapment.129 Polyacrylamide is hydrophilic, that is,

protein-friendly; its cross-linking density is readily controlled through the percentage

of bisacrylamide cross-linker, and the redox-initiated polymerization proceeds at the

room temperature.Mersal and Bilitewski130 used polyacrylamide gel in both capillar-

ies and microfluidic chips for entrapment of glucose oxidase and determination of

glucose. They also admixed acrylic acid in the polymerization mixture to generate

EOF and drive the analytes through the device.

Palm and Novotny131 developed an interesting approach with single-step fabrica-

tion of immobilized trypsin reactor in a capillary using their acrylamide-based

monolith. The polymerization mixture consisting of monomers acrylamide, methy-

lenebisacrylamide, and N-acryloylsuccinimide dissolved in buffer solution together

with PEG and a redox initiator (TEMED and ammonium peroxodisulfate) was mixed

with trypsin. The polymerization reaction and enzyme immobilization could be

completed within 2 h for the high activity of NAS to enzyme. This kind of monolithic

reactor offered high flow permeability and biocompatibility. Using the same single-

step technique, Palm and Novotny132 also immobilized peptide-N-glycosidase and

used the conjugate for deglycosylation of proteins. However, the content of

monomers in the polymerization mixture was restricted by the poor solvability of

monomers in aqueous solution, resulting in a loose structure and low amount enzyme

immobilized.133

Obviously, two processes related to the enzyme can be envisioned to simulta-

neously proceed during the polymerization: (i) reaction of lateral nucleophilic

functionalities of trypsin with succinimide moieties in both monomeric and already

polymerized form and (ii) entrapment of the enzyme in the acrylamide matrix.

The extent of each of these processes is difficult to judge since no comparative

experiments without the succinimide monomer were performed.

Glycidyl Methacrylate Copolymers

DIRECT IMMOBILIZATION VIA EPOXY FUNCTIONALITIES Epoxide groups represent a

common tool in the field of enzyme immobilization. Using this reaction,

Ben�cina et al.134 immobilized protein A, deoxyribonuclease, and trypsin on 3mm

thick 12mm diameter poly(GMA-co-EDMA) disks using both static (disk immersed

in the protein solution) and dynamic (solution of protein pumped through the disk)

techniques. Comparative experiments clearly demonstrated significant effect of

benzamidine on the activity of the immobilized enzyme.

HYDROLYSIS FOLLOWED BY OXIDATION Since the direct reaction of epoxide function-

alities with proteins is slow, several alternative approaches have also been developed.

One of the oldest, which found its inspiration in the area of polysaccharide-based

supports,135 includes hydrolysis of the epoxide ring to a 1, 2-diol and its oxidation

using periodate. Detailed studies of both these reactions were carried out with poly

(GMA-co-EDMA) beads almost three decades ago.136,19 The protein predominantly

reacts with this support via its lateral primary amine groups of lysine residues. Since

the imine --C¼N-- double bond that forms is not very stable and easily undergoes


hydrolysis liberating again the immobilized protein, hydrogenation with sodium

cyanoborohydride is often used to convert this group to a secondary amine function-

ality --C--NH--. Luo et al.137 used this method to activate poly(GMA-co-EDMA)

monolith and immobilized papain on this support. The amount of papain immobilized

on the 50� 4mm ID monolith was 7.1mg g�1, but its efficiency was rather low. This

was ascribed to the steric hindrance resulting from the close vicinity of the enzyme to

the surface. Other explanation could be a multipoint immobilization that would

change the tertiary structure of the enzyme and deform the active site with the

concomitant decrease in activity. K�renkov�a et al. prepared a capillary enzymatic

reactor with covalent immobilization of trypsin on poly(GMA-co-EDMA) monolith

using this hydrolysis followed by oxidation method.20 For comparison, they also

immobilized trypsin on monolith by a single-step binding procedure. Since the

reaction of carboxylic functionalities with epoxides is not very efficient, they have

attempted to mask the epoxide groups with amino groups of aspartic acid to

increase the hydrophilic character of the pore surface and immobilize trypsin through

reactions of epoxide functionalities with amino groups of the protein molecule. No

significant difference in the enzymatic activity or stability was observed for the

reactors prepared in these two methods.20

Although inspired in recent years, the above methods still have some disadvan-

tages: the hydrophilicity of diol functionalities originating from the hydrolyzed poly

(GMA-co-EDMA) monolith is not sufficient to avoid adsorption of hydrophobic

albumin in a highly aqueous mobile phase that limits the application of the IMERs for

digestion of high molecular protein. To solve this problem, Svec’s group recently

demonstrated a novel approach to modification of the surface chemistry of the poly

(GMA-co-EDMA)monolith.138 Themonolith was first hydrophilized via photograft-

ing of poly(ethylene glycol) methacrylate followed by photografting of a 4-vinyl-2,2-

dimethylazlactone to provide the pore surface with reactive functionalities required

for immobilization. This new approach reduced the undesired nonspecific adsorption

of proteins and peptides and facilitated control of both the enzyme immobilization and

the protein digestion processes.

AMINOLYSIS FOLLOWED BY DIALDEHYDE ACTIVATION Another traditional and often

used path facilitating immobilization of glycidyl methacrylate-based supports com-

prises aminolysis of the epoxide ring using ammonia or a diamine, followed by

activation with dialdehyde most often glutaraldehyde.139 The aldehyde functionality

is then used for the reaction with an enzyme, and similarly to the previous technique,

the labile imine double bondmust be hydrogenated. Poly(GMA-co-EDMA)monolith

prepared using thermally initiated polymerization in a 50mm� 4mm ID columnwas

modified with 1,6-diaminohexane and glutaraldehyde.137 This reaction path was

believed to afford a spacer arm onwhich papain could reside thus decreasing the steric

constrains. Indeed, compared to the approach comprising hydrolysis followed by

oxidation to aldehyde described above, the immobilized papain exhibited twofold

higher effectiveness asmeasured by digestion of human IgG despite the lower amount

of attached protein. Yet, the enzymatic activity of the immobilized papain was only

17% of that observed with free enzyme. As expected, the immobilized enzyme was


significantlymore stable at a temperature of 75�C.Ye et al.18 developed amicroreactor

containing poly(GMA-co-EDMA) monolith prepared using thermally initiated po-

lymerization in a capillary. Aminolysis of the epoxide functionalities with 29%

aqueous ammonia solution afforded primary amine groups that were then activated

with glutaraldehyde. Finally, trypsin was immobilized in this monolith and stabilized

by reductionwith sodiumcyanoborohydride. The disadvantageof this immobilization

reaction is that it has a potential for producing undesirable by-products, for example,

homoconjugates and various polymers.

HYDROLYSIS FOLLOWED BY CARBONYLDIIMIDAZOLE ACTIVATION Carbonyldiimidazole

(CID) activation of hydroxyl groups containing supports for immobilization of

proteins was first suggested by Hearn et al.140 at the end of the 1970s, in response

to problems encountered at that time with widely used cyanobromide activation

method such as the formation of undesired ionized functionalities leading to nonspe-

cific interactions. Ben�cina et al.134 used this approach for immobilization of trypsin on

monolithic disks and achieved excellent activity. Once the immobilization was

complete, unreacted imidazole carbamate functionalities were quenched. The acti-

vated support contained 0.8mmolmL�1 imidazole carbamate functionalities. This is

said to be much less than the content of epoxide groups in the monolith. However, no

direct comparison between these two values can be made since a number of epoxide

functionalities are buried within the matrix. As the immobilization occurs only at the

pore surface, inaccessible groups cannot contribute to binding. In contrast, activation

with CID is likely to occur mostly on the pore surface where are located the most

accessible hydroxyl groups.

Vinylazlactone Copolymers Beads prepared by copolymerization of 1-vinyl-4, 4-

dimethylazlactone and methylenebisacrylamide using an inverse suspension process

were developedby3MCompanyat thebeginningof the 1990s.141 Their properties and

applications were summarized by Heilmann et al.142 in an excellent review. Due to

their enhanced reactivity to amine and thiol groups, they also found applications in

immobilization of various enzymes.143,144 An additional benefit of this chemistry is

the linkage of the protein through a dipeptide spacer that can contribute to enhance-

ment in activity.

Svec and coworkers introduced this chemistry in the field of monoliths and used it

for immobilization of trypsin.145 The initial monolith was prepared from amixture of

monomers that afforded reactivity (1-vinyl-4, 4-dimethylazlactone), hydrophilicity

(acrylamide), and cross-linking (ethylene dimethacrylate) dissolved in porogenic

solvent (tetradecanol). Azobisisobutyronitrile was used as the thermal initiator. The

porogen enabled formation ofmonolithswith a pore size of about 2.5mm.Trypsinwas

immobilized on this monolith located in a 20mm� 1mm PEEK tube by pumping its

solution through the device at a flow rate of 0.2mLmin�1 for 60min. The unreacted

azlactone functionalities were then quenched by reaction with 2-aminoethanol.

This protocol led to a conjugate with 38.8mgmL�1 or 90.7mg g�1 of immobilized

trypsin. They extended this azlactone chemistry to both capillary and microfluidic

formats.146–148


Since the ultimate goal is the fabrication of complex devices, the commonly

used thermally or redox-initiated free radical polymerization modes are not best

suited for the preparation of monoliths within a specific part of the capillary or a

microchip. To solve this problem, Svec further developed the photolithographic-like

technique involving photopolymerization through a mask.149 This simple approach

shown schematically in Figure 4.13 facilitates the formation of reactive porous

monoliths within a specific part of a microsystem. The as-prepared reactor afforded

suitable degrees of digestion of proteins even after very short residence time of less

than 1min.

4.3.4 Membrane

The incorporation of polymeric membranes into microfluidic networks has been

employed for enhancement in device functionality for years.150 Due to the porous

structures of membrane media, polymeric membranes exhibit a large surface to

volume ratio that serves to facilitate rapid solution exchange. Extremely large surface

area, at least 200 cm2 of internal surface per cm2 of frontal surface, is available for

protein adsorption and immobilization. Thus, the membranes containing adsorbed

proteins can be employed as miniaturized enzyme reactors.

FIGURE 4.13 Scheme of the preparation technique affording monolith with well-defined

size and location. (a) Empty capillary (or microfluidic chip); (b) capillary filled with the

polymerization mixture consisting of monomers, porogenic solvents, and a photoinitiator;

(c) capillary with attached photomask; (d) irradiation of the capillary contents through the

mask with UV light in the range of 220–330 nm for 10min to fabricate the monolith;

(e) removal of unreacted polymerization mixture from the dark parts and washing with a

solvent. Reprinted from Ref. 149, with permission.


Gao et al. reported a miniaturized membrane reactor by fabricating the PDMS

microfluidic channel on the PDMS substrate and coupling it to a poly(vinylidene-

fluoride) (PVDF)membrane providing large internal surface area for enzyme adsorp-

tion (Figure 4.14a).26 Despite the large S/V ratio of porous membrane media, this

reactor had a high total dead volume due to capillary connections with the micro-

channel. This problem was eliminated by placing a hydrophobic and porous PVDF

membrane around the end of a polymer sleeve (Figure 4.14b).27 The assembly of

capillary fitting, containing a length of fused silica capillary, provided the necessary

flow paths and the membrane media for performing rapid and effective proteolytic

digestion. This membrane-based proteolytic reactor can be directly coupled with

nano-ESI-MS for achieving speedy protein identification in seconds instead of hours.

In the case ofmembrane clogging, themembrane-based reactor can be regenerated by

simply replacing theoldmembranewith anewPVDFstrip, followedbyadsorptionof a

fresh enzyme solution.

Hisamoto et al. reported that a nylonmembrane could be formed at the interface of

two solutions formed in a microchannel (Figure 4.15a). Peroxidase was immobilized

on this membrane, which was used as a chemicofunctional membrane;151 however,

immobilization of the membrane is technically difficult, and application of this

method is limited because the nylon membrane is unstable in organic solvents.

Maeda and coworkers have developed a method for the preparation of an enzyme-

immobilized microreactor by simple loading of the enzyme solution and a mixture of

glutaraldehyde and paraformaldehyde into the microchannel forms a cross-linked

enzyme aggregate membrane on the microchannel wall (Figure 4.15b).152

4.3.5 Other Formats

Protein digestion can also be performed on a trypsin-immobilizedMALDI probewith

the advantage of obviating the need to handle samples before carrying out MALDI-

TOF-MS (matrix-assisted laser desorption ionization time-of-flight mass spectrome-

try) measurements. As described by Nelson et al.,153,154 enzyme was covalently

attached to theMALDI probe via a gold-coated stainless steel sample target. Proteins

were digested onenzyme-linked probesbydepositing the sample directly on the active

FIGURE 4.14 Schematic representation of the miniaturized trypsin membrane reactor

(a) by fabricating the PDMS microfluidic channel and coupling it to a PVDF membrane;

(b) by placing the PVDFmembrane inside the capillary fitting. Reprinted fromRefs 26 and 27,

respectively, with permission.


surface. Digestion was terminated by adding theMALDImatrix prior toMS analysis.

Picomoles of proteins could be efficiently digested in less than 30min. Houston and

Reilly155 used this technique for hemoglobin characterization. Recently, Lubman and

coworkers156 developed amethod combining capillarymonolithicRP-HPLCwith on-

plate enzymatic digestion for obtainingprotein identifications for humanbreast cancer

cells, which is a simple protocol with the advantage of effectively minimizing sample

loss. However, because the matrix used to enhance ionization was applied directly to

the probe, the enzyme-linked MALDI probe could not be reused for consecutive

digestion. To solve this problem, Li et al. introduced trypsin-linked magnetic

nanoparticles into the on-probe digestions (Figure 4.16).25 Due to their magnetic

property, the trypsin-linkednanoparticles couldbeeasily removed from theprobe after

digestion, which would benefit sample–matrix cocrystallization and avoid causing

possible contamination on the ion source chamber inMS.What ismore important is its

feasibility for reuse of MALDI probe for consecutive digestion.

More recently, Ota et al. developed one kind of elegant trypsin immobilized

monolithic silica with pipette-tip formula for high-throughput protein digestion.157

The silica-basedmonolith was first chemicallymodified by 3-aminopropyltrimethox-

ysilane, and then fixed into a 200mL pipette tip by supersonic adhesion. After the

carriers were activated by disuccinimidyl suberate (DSS), trypsin was finally im-

mobilized. The tip enabled the digestion of reduced and alkylated protein within 20

timesoperation, and the enzymatic activityof the immobilized trypsin tipwas about 50

times higher than that of the conventional in-solution format.

FIGURE 4.15 (a) Single and parallel dual nylon membranes; (b) cross-linked enzyme

aggregate membrane prepared inside the microchannel. Reprinted from Refs 151 and 152,

respectively, with permission.


Except for the above formats, Fan and Chen developed a fiber-packed channel

bioreactor for protein digestion by immobilizing trypsin on the fiberglass bundles

embedded in the substrate of a PMMA microchip (Figure 4.17a).158 A UV-sensitive

prepolymerized methyl methacrylate (MMA) molding solution containing a UV

initiator was sandwiched between a PMMA cover plate and a PMMA base plate

bearing glycerol-permeated fiberglass bundles and was exposed to UV light. During

polymerization, the fiberglass bundles were embedded in the PMMA substrate to

form fiberglass-packed microchannels. When the glycerol in the fiberglass bundles

sealed inside the PMMA substrate was flushed away with water, the obtained

porous fiberglass-packed microchannels could be employed as a support to immo-

bilize trypsin with the aid of chitosan (CTS) and glutaraldehyde. However, when

the enzyme activity decreased to some extent, the trypsin-immobilized layers

were permanently modified in the channel. Therefore, they further fabricated a

core-changeable needle enzymatic reactor by inserting a piece of trypsin-immobi-

lized glass fiber into the needle of a syringe (as shown in Figure 4.17b),159 which

could be regenerated by changing the core composed of a piece of glass fiber and

a layer of enzyme-entrapped polymer coating. The in-needle fiber bioreactor has

been coupled with MALDI-TOF MS for the digestion and peptide mapping of

model proteins.

FIGURE 4.16 (a) Trypsin-linkedmagnetic nanospheres were added to the protein solution.

(b) Nanospheres could be easily removed from the plate with a magnetized needle. (c) After

the removal of the nanospheres, the plate is ready for MALDI-MS analysis. Reprinted from

Ref. 25, with permission.

FIGURE 4.17 (a) SEM image of the cross section of a fiberglass-packed microchannel in

the PMMA substrate.158 (b) Schematic diagrams of the core-changeable needle bioreactor.159

Reprinted from Refs 158 and 159, respectively, with permission.


4.4 APPLICATION OF IMMOBILIZED MICROFLUIDIC

ENZYMATIC REACTORS

The fields of IMER application are becoming wider every year. A considerable

number of papers have been published reporting successful application of enzymatic

microreactors in chemistry and biochemistry. The principal field of application of

microreactors is tryptic digestion of proteins. Enzymatic microreactors also facilitate

characterization of enzyme activity as a function of substrate concentration, and

enable fast screening of new biocatalysts and their substrates. They may constitute

key parts of lab-on-a-chip and mTAS, assisting the analysis of biomolecules.160

Intentionally, in order to narrow the scope of this chapter, only applications in

peptide mapping, biosensing, and kinetic study are discussed in this section.

Readers interested in other application areas are advised to see comprehensive

reviews12,160–163 and recent publications.

4.4.1 Peptide Mapping

The ability to rapidly and efficiently digest and identify an unknown protein is of

great utility for proteome studies. Identification of proteins via peptide mapping is

generally accomplished through proteolytic digestion with enzymes such as trypsin.

Limitations of this approach consist in manual sample manipulation steps and

extended reaction times for proteolytic digestion. The use of immobilized trypsin

for cleavage of proteins is advantageous in comparison with application of its soluble

form. Thus, the greatest number of recent applications of IMERs refer to protein

analysis by peptide mapping, and for determination of post-translational modifica-

tions (PTMs) such as phosphorylation, glycosylation, and lipidylation, which are

essential in modulating biological functions of cells and can be associated with a

number of diseases.

Peptide mapping is typically performed using enzymatic cleavage of the protein

and the peptide fragments in the resulting mixture are identified using electrospray

ionizationmass spectrometry (ESI-MS) ormatrix-assisted laser desorption/ionization

mass spectrometry (MALDI-MS). In either case, separation of the peptidemixture, for

example, by micro-high-performance liquid chromatography (m-HPLC) or capillaryelectrophoresis, prior to mass spectrometric analysis, minimizes the ionization

suppression and improves the sequence coverage.164 One of the limiting steps in

peptide mapping is the manual sample manipulation and extended reaction times for

proteolytic digestion. Traditionally, enzymatic cleavage is performed in a homoge-

neous solution consisting of a mixture of the proteolytic enzyme and the protein.

K�renkov�aandForet summarized themost commonenzymesused for protein digestion

and their sites of cleavage in an excellent review (Table 4.1).2 Among these enzymes,

the most frequently used is trypsin, which catalyzes the process of protein digestion

through hydrolysis of peptide bonds at the C-end of the Arg and Lys residues and

typically provides peptides in a mass range suitable for high-resolution/high-sensi-

tivity mass mapping through mass spectrometry. A huge body of literature has been

reportedwith trypsin-immobilized IMERforpeptidemapping.Pepsin, an enzyme that


cleaves proteins at pH 2 at C-end of Phe and Leu residues, can be seen as a

complementary tool that may be used to both determine and confirm the respective

order of each peptide fragment obtained from the tryptic digestion. Since pepsin

digests proteins at acidic pH, it facilitates the direct coupling of the pepsin-immo-

bilized reactorwithESI-MS.The specificity of pepsin is lower than that of trypsin, and

some amino acid sequences are not cleaved despite the presence of Phe andLeu, while

other sequences can be cleaved even in their absence. Although this would be

detrimental for protein identification, this lack of specificity can be utilized for

quantitative protein analysis.165,166 Due to miscleavages, pepsin digestion generates

longer peptides that provide a signature of a specific protein. Thus, pepsin digestion

has been successfully applied to quantitative protein analysis affording the reproduc-

ible formation of specific peptide markers.166 Endoproteinases such as GluC, ArgC,

and LysC have been gaining more popularity due to their highly specific points of

cleavages in recent years. Theyoffer additional benefits such as digestion resulting in a

smaller number of larger, information-rich peptide fragments that simplifies their

separation and facilitates protein identification. These enzymes are also useful in

mapping of post-translational modifications such as methionine oxidation since the

large peptides allow lower quantization limits to be achieved using typical UV

detection. The larger peptides formed during LysC digestion are also more likely

to containmultiple sites allowing charging such as internal arginines, which form ions

better suited for MS/MS analysis. Endoproteinase LysC is also a more robust enzyme

that, compared to trypsin, maintains its activity even at relatively high concentrations

of denaturants.138

To achieve efficient and reproducible digestion results, maintaining optimum pH,

temperature, protein to enzyme ratio, and reaction time is critical. The variations in

TABLE 4.1 Some of the Most Common Enzymes Used for Protein Digestion

Enzymes

IUBMB Enzyme

Nomenclaturea Site of Cleavage

Trypsin EC 3.4.21.4 C-terminus of Arg and Lys

Chymotrypsin EC 3.4.21.1 C-terminus of Phe, Tyr, Trp, Leu,

and Met

Endoproteinase LysC EC 3.4.21.50 C-terminus of Lys

Endoproteinase AspN EC 3.4.24.33 N-terminus of Asp

Endoproteinase GluC

(S. aureus V8, pH 4)

EC 3.4.21.19 C-terminus of Glu

Endoproteinase GluC

(S. aureus V8, pH 8)

EC 3.4.22.19 C-terminus of Glu and Asp

Endoproteinase ArgC

(clostripain)

EC 3.4.21.8 C-terminus of Arg

Thermolysin EC 3.4.24.27 N-terminus of Leu, Ile, Val, Phe,

Met, and Ala

Pepsin EC 3.4.23.1 C-terminus of Phe, met, Leu, and Trp

Reprinted from Ref. 2, with permission.a IUBMB, International Union of Biochemistry and Molecular Biology.

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS 157

size, structure, type, and level of PTM make proteins significantly different in their

susceptibility to enzymatic digestion. Depending on the accessibility of the cleavage

sites, complete digestionmay require times ranging fromseveralminutes to overnight.

In principle, the time of digestion can be reduced using high concentration of the free

enzyme;167 however, such an approach has several disadvantages. Besides the added

cost, the enzymes often lose their activity and specificity, and the enzyme autodiges-

tion results in undesirable formation of additional peptides, which may lead to the

ionization suppression in MS analysis and complicate the interpretation of data.

Immobilizing the enzymeona solid support eliminatesunwanted autodigestion andan

extremely high local concentration of proteolytic enzyme providing rapid catalytic

turnover.Moreover, due to the relatively long incubation time the current workflow of

protein analysis includes protein digestion as an off-line step. With miniaturized

IMER, complete protein digestion can be performed in less than 1min and direct

online coupling with further peptide concentration and separation steps become

possible.

For high-throughput proteome analysis, it is important that separation and detec-

tion are coupled online to proteolysis. The online digestion of proteinswith IMER can

enable faster and more automated protein identification. Moreover, the use of IMER

avoids manual sample handling and the consequent possible contamination of the

sample, and the reactor can be coupled online to separation and detection with the

advantage of automation. Several possibilities can be envisioned for coupling

IMER with a separation and identification system and they are outlined in

Figure 4.18.161

4.4.1.1 IMER Coupled with MSEnzymes can be bound to the inner walls of the capillary (see Section 4.3.1). In this

case, there are no backpressure constraints during sample injection. Such micro-

reactors can be easily attached to an electrospray ionization (ESI) interfacewith mass

spectrometry (MS), providing an online system.147,168A typical experimental setup of

IMER hyphenation with ESI-MS is depicted in Figure 4.19.20 In this online system

developed by Foret’s group, the protein solution was pumped through the IMER at

selected flow rates (50–300 nLmin�1), and the digestion products were mixed in a T-

joint with a constant flow rate (1.2mLmin�1) of the spray solution (50% aqueous

ACN, 1% formic acid) supplied by the second syringe pump. The resulting streamwas

analyzedwith themicrospray interface suppliedwith theMS instrument.Recently, the

samegroupprepared IMERswith L-1-tosylamido-2-phenylethyl chloromethyl ketone

(TPCK)-trypsin and pepsin A covalently immobilized on thewall of a 10 mm ID fused

FIGURE 4.18 Modes of combining enzyme reactors with separation and identification

systems. Reprinted from Ref. 161, with permission.


silica capillary. The optimized reactorswere served as the nanospray needle in a liquid

junction interface for CE-ESI/TOF-MS analysis of protein mixtures. On-line diges-

tion of proteins on the capillary wall enabled faster and fully automated protein

identification using peptide mass fingerprinting.70

Enzymes can also be immobilized on a variety of nanoparticles and packed into the

microcolumn (seeSection4.3.2) and theproduct fractionsmaybe sampledprior to off-

line analysis by means of MALDI-TOF-MS.30,32,33,13 The primary advantage of the

MALDI approach compared to ESI, when coupling to microfluidic chips, is the

potential formultiplexing.Thedirectional control of theESI spray canbedifficultwith

a high-density array of spray tips. Furthermore, microfluidic chip interfaces with ESI

can suffer from stability problems when sprays are started or stopped. This limits the

speed of moving from one sample to another and therefore limits the throughput. In

most cases, digested peptides eluted from the IMERs were collected and deposited on

the MALDI probe followed by deposition of MALDI matrix. Alternatively, Lee et al.

developed an off-lineMALDI interface to combinematrix addition and deposition on

a MALDI target using a robotic plate spotter modified to accept the effluent from the

microfluidic chip automatically.69

Although less developed, online coupling with MALDI-MS can also be achieved.

For example, Ekstr€om et al.71 described a device that integrated them-chip IMERwith

a sample pretreatment robot and amicrofabricatedmicrodispenser to transfer digested

protein directly to a MALDI target plate for automated MS analysis (Figure 4.20).

Anodic etching in a hydrogen fluoride/ethanol solution was used to produce a porous

surface on the digestion chip. The use of porous silicon provided a 170-fold increase in

enzymatic activity compared to nonporous reactor.169,170 This increase in surface area

resulted in increased digestion efficiency and extremely fast digestions. The m-chipIMER allowed online enzymatic digestion of protein samples (1mL) within 1–3min,

FIGURE4.19 Schematic diagramof IMERhyphenated onlinewith ESI-MS.Reprinted from



about 200–1000 faster than digestion in solution. This integrated system provided a

throughput of 100 samples in 3.5 h.

The most common strategy for protein analysis is based on two-dimensional

polyacrylamide gel electrophoresis (2D PAGE). Typically, the identification of a

proteinmixture involves separation, extraction, proteolytic digestion, andMSanalysis

of each protein spot. New technology developed by Cooper and Lee171 involved an

online combination of electrophoretic protein transfer from a polyacrylamide gel with

proteolytic digestion in membrane-based IMER. After electrokinetic-based protein

extraction and stacking, real-time proteolytic cleavage of extracted protein, and direct

depositionofprotein digest onto theMALDI target, thepeptideswere identifiedbyMS

analysis (Figure 4.21). The sensitivity of the technologywas demonstrated both by the

detection of standard proteins from a gel protein loading as low as 1 ng and by the

identification of low-abundance proteins in complex yeast cell lysates.

4.4.1.2 IMER Coupled with Liquid Chromatography SystemLiquid chromatography in combination with tandem mass spectrometry (LC–MS/

MS) can overcome many of the limitations of 2D gel electrophoresis in proteomic

studies. The advantage of LC separation systems is that the automated operation is

much easier to perform. Therefore, it is not surprising that a lot of publications deal

with the development of IMERs compatible with liquid chromatographic systems.

Inmost cases, IMERsare coupledwithLCsystemswith severalvalves.Calleri et al.

reported the immobilization of trypsin on silica-based monoliths via epoxy groups.

With such an immobilized enzyme reactor, they achieved the hyphenation of online

digestion with HPLC via a switching valve. It was found that the cleavage efficiency

(aminoacidic recovery, %AA) achieved in 20min by the online protocol was at least

FIGURE 4.20 Microfluidic system for MALDI protein analysis. (a) Automated sample

pretreatment and injection; (b) m-chip IMER (the photo inset shows a SEM picture of the

lamella structure with the porous layer); (c) microdispenser used to deposit sample into m-vials; (d) shallow nanovials (300mm� 300mm� 20mm) on the MALDI target plate; and

(e) automated MALDI-TOF-MS analysis. Reprinted from Ref. 71, with permission.


comparable, or even better than the conventional off-line 4 h consuming method of

digestion. By using the online system, protein digestion and genetic variant identifi-

cation in serumsampleswere performedby the samegroup, andmutation sites in beta-

lactoglobulin A and B variants were successfully located.172 Markides’s group173

integrated IMER with a trapping column so that the peptides produced from the

enzymatic capillary reactor could be efficiently trapped and desalted that helped to

avoid detrimental signal suppression in the followingESI process. The peptides eluted

from the precolumnwere then separated on an analytical capillary column by a buffer

suitable for the ESI-MS process (Figure 4.22a). Different trapping columns were

tested in order to avoid losses of hydrophilic peptides during the elution from the

trypsin reactor on the separation column. The performance of the online system was

compared to that of the classical digestion in solution, with reference to peptide

FIGURE 4.21 Scheme of combined protein electronic transfer and miniaturized trypsin

membrane digestion for gel protein identification using MALDI-MS. (a) Electronic protein

transfer followed by (b) introduction of extracted proteins into a membrane reactor and

deposition of protein digest on a MALDI target. Reprinted from Ref. 171, with permission.

FIGURE 4.22 Schematic diagrams of IMERs coupled with LC. (a) Setup of online

proteolysis, peptide trapping/desalting, separation, and ESI-MS analysis. (b) Setup of online

system for peptide mapping of post-translational modified proteins. Reprinted from Refs 173

and 175, respectively, with permission.


sequence coverage and sensitivity. The peptide sequence coveragewas increased from

about 25% to 55%.

Hsieh et al.174 developed a fully automated five-column chromatography system

coupled to an ESI-MS/MS for the isolation, digestion, and characterization of human

hemoglobin. The employment of the trypsin reactor and the application of a multi-

valved apparatus allowed automated, direct transfer of analytes between the various

operation units. For analysis of post-translational modified proteins, Riggs175

described an automated multidimensional chromatographic system consisted of

different steps: (i) reduction and alkylation of the proteins were achieved in the

autosampler; (ii) proteolytic digestionwas carried out on trypsin column; (iii) specific

classes of peptides were selected by affinity chromatographic column; (iv) the

selected peptides were transferred to a reversed-phase chromatographic column

and further fractionated. The analytes were transferred between columns through

valves (Figure 4.22b). The chromatographic effluent was analyzed by ESI-MS for the

signature peptide approach. The model system chosen for this study was phosphory-

lated milk proteins, and the total analysis time in the tandem column mode of

operation was under 2 h.

Geiser et al. have developed an online system containing capillaries with two

different porous polymer monoliths for protein digestion with immobilized pepsin,

peptide preconcentration, and nanoliquid chromatography separation coupled to

electrospray ionization mass spectroscopy (nLC–ESI-MS).148 The first monolith

with well-defined porous properties was prepared by in situ copolymerization of

2-vinyl-4, 4-dimethylazlactone and ethylene dimethacrylate (poly(GMA-co-

EDMA)) and used as support for covalent immobilization of pepsin. The second,

poly(laurylmethacrylate-co-ethylene dimethacrylate) (poly(LMA-co-EDMA))

monolith with a different porous structure served for the preconcentration of peptides

from the digest and their separation in reversed-phase liquid chromatography mode

(Figure 4.23a).With no need for adding a trapping column to the system, the top of the

separation capillary was also served as a preconcentrator, thus enabling the digestion

ofverydilute solutions of proteins in the bioreactor and increasing the sensitivity of the

mass spectrometric detection of the peptides using a time-of-flight mass spectrometer

with electrospray ionization. The schematic diagram of the online system is provided

in Figure 4.23b. Myoglobin, albumin, and hemoglobin were digested to demonstrate

feasibility of the concept of using the two monoliths online. Successive protein

injections confirmed both the repeatability of the results and the ability to reuse the

bioreactor for at least 20 digestions.More recently, on the basis of the abovework, they

further modified the surface of the first monolithic poly(GMA-co-EDMA) via

multistep/multilayer photografting to obtain support for enzyme immobilization

with both reactive azlactone functionalities as well as largely eliminated nonspecific

adsorption of proteins and peptides, and extended the application of this online system

to analysis of high molecular weight human IgG.138

4.4.1.3 IMER Coupled with Capillary Electrophoresis SystemCapillary electrophoresis (CE) has become an important separation technique for

peptide mapping because of its simplicity, speed, high separation efficiency, and low


sample consumption. Different research groups have explored the possibility of

performing trypsin hydrolysis online with an electrophoretic separation. However,

online digestion of proteins before CE separation is difficult to accomplish due to the

challenges of manufacturing a bioreactor compatible with the small sample volumes

used inCE and because of the problems encountered in interfacing the reactorwith the

separation capillary. The two most common configurations are biocatalytic capillary

located upstreamof the separation capillary and trypsin immobilized in the first part of

the capillary devoted to separation.

IMER Coupled with CE Capillary In an on-capillary protein hydrolysis for peptide

mapping, a stop-flow incubation period within the capillary is required to allow

numerous hydrolysis products to accumulate before accomplishing their electropho-

retic separation. The resolving power of the separation suffers when diffusional

broadening of the reaction mixture zone becomes intolerable after lengthy incubation

FIGURE 4.23 (a) SEM micrographs of the cross section of 100 mm ID monoliths in

capillaries: (A) support for the enzymatic reactor; (B) HPLC column prepared. (b) Scheme

of the online system: (A) protein solution is injected into the immobilized pepsin reactor in 2%

aqueous formic acid, digested, and peptides are trapped on the top of HPLC column; (B) the

reactor is bypassed and peptides are separated in a gradient of acetonitrile. Reprinted from



periods, and increasing dilution of the reaction zone with incubation time results in a

concomitant reduction in the sensitivity of the on-capillary detection. Traditionally,

relatively slow reaction kinetics for protein hydrolysis generally require long incuba-

tion times, which can limit the usefulness of on-capillary proteolytic reactions.

Another critical consideration is the suitability of the separation buffer to the

enzyme-mediated chemistry. One must also have the capability to mix the enzyme

with the chosen substrate(s), maximize conversion of substrate to products, and

accomplish a separation of the resulting reaction mixture. These processes are

generally optimal under different solution conditions (pH, ionic strength, buffer

additives) in the capillary. The success of on-column peptide mapping is particularly

vulnerable when the optimal pH of the proteolysis reaction differs from the pH at

which optimal selectivity is obtained for the separation of the peptide.

Kuhr’s group demonstrated that trypsin can be readily immobilized on the surface

of a fused silica capillary via biotin–avidin–biotin technology42 and can be coupled to

the separation capillary to enable online digestion and separation through anopenfluid

junction (Figure 4.24a).8

The enzyme-modified fused silica microreactor was coupled through a 100 mmsolution gap to the separation capillary. Very little diffusional sample loss in the gap

FIGURE 4.24 Schematic presentation of online coupling of enzymatic reactor (a) coupled

with CE capillary via an open fluid junction8; (b) coupled with an SPE preconcentrator and

CE177; (c) coupled with transient isotachophoresis (CITP)/CZE-ESI-MS26; (d) CE-enzymat-

ic microreactor-CE-MS/MS.178 Reprinted from Refs 8,26,177, and 178, respectively, with

permission.


wasobserved at low ionic strength since thevoltagegradient across thegap is large and

the sample had little time for radial diffusion out of the junction. The utility of this

approach was that it not only offered an efficient mass transfer of the sample by

electromigration but also allowed subsequent separation of the transferred sample in

the second capillary without the need to move either capillary. This configuration

allowed independent optimization of digestion and separation conditions in an online

peptide mapping procedure. The enzyme-modified capillary was filled with the

protein solution and allowed to incubate at room temperature for approximately

2 h, during which proteolysis occurs. Subsequently, an aliquot of the digest was

injected into the separation capillary byapplying a potential across the two free ends of

both capillaries. The injected sample was then separated by applying the CZE

separation potential across the gap solution and the other end of the separation

capillary. Using this approach, they were able to perform online digestion and

separation of picomole quantities of protein by CZE in less than 3 h. The same

coupled capillary CZE instrument was used for further experiments and three

microreactors were prepared with two proteases (trypsin and pepsin) and a peptidase,

carboxypeptidase-Y.64 These proteolytic enzymes are distinguished by the large

differences in their specificities toward peptide bond cleavages and their pH of

greatest catalytic efficiency. Offline protein digestions in carboxypeptidase-Y and

trypsin-modified microreactors yielded C-terminal sequences and mass fingerprints,

by eitherMALDI-MSor plasma desorptionmass spectrometry,whichwere compared

with protein fragment databases for identification. On-line trypsin digestion followed

by CE/ESI-MS was also demonstrated to give the greatest efficiency in peptide

mapping analysis.65 Efficient digestion of the oxidized insulin b-chain occurred

within 40min, and the entire peptide mapping required about 1 h. The long digestion

time required in this system can be due to the fact that enzymes were directly

immobilized on the capillary inner wall with limited surface area that leads to a

low surface to volume ratio of the microreactor.

Bonneil et al.176 fabricated a microreactor with commercially available enzyme-

immobilized controlled pore glass beads packed in a capillary for peptidemapping by

capillary electrophoresis. The proteins were perfused through the microreactor for

about 2 h applying low pressure at the inlet. The digest was collected at the micro-

reactor outlet and the tryptic fragments were separated by CE and detected by UV

absorbance using a diode array detector. Subsequently, the microreactor was further

coupled online with an SPE preconcentrator and CE in an effort to improve mapping

sensitivity by minimizing sample handling that leads to peptide losses.177 The

proposed system depicted in (Figure 4.24b) allowed digestion of proteins and

preconcentration, separation, and detection of the peptides in 4 h. However, the

separation efficiency was poor due to the multiple-valve design of the system and

to the dispersion of the 60 nL desorption plug. Nevertheless, the maps were fairly

reproducible in terms of migration time.

On-line couplingof aminiaturizedmembrane reactor for proteolytic digestionwith

transient isotachophoresis (CITP)/CZE-ESI-MSwas proposed byGao et al.26 Trypsin

was adsorbed on a PVDF membrane. As reported in the schematic setup of

Figure 4.24c, proteins were forced by a syringe pump through pores into this


membrane, and the digested proteins were focused by transient isotachophoresis,

separated by CE, and detected by ESI. The microfluidic system enabled rapid

identification of proteins in minutes and consumed very little sample (nanogram).

The rather large dead volume of this device resulted in poor separation efficiency for

the digested peptides; theoretical plate counts were about 1500 for the peptides

resulting from digestion of cytochrome c. However, the abilities of the platform to

concentrate and resolve peptide mixtures enhanced the sensitivity and the dynamic

range of ESI-MS detection for the identification of minor proteins. In order to reduce

the dead volume, analysis time, and sample consumption, the same research group

employed a commonly used capillary fitting for directly housing a miniaturized

trypsin membrane reactor.27 The nanoscale trypsin reactor was integrated into a

platform for the concentration and separation of the proteolytic digest using chro-

matographic and electrophoretic methods prior to mass spectrometry analysis. The

application of sample stacking and separation techniques contributed to further

enhancement in the dynamic range and detection sensitivity for the analysis of

complex protein mixtures. The proposed nanoscale reaction system enabled rapid

proteolytic digestion in seconds instead of hours for a protein concentration of less

than 10�8M.

Dovichi and coworkers have described the production of a macroporous monolith

with immobilized trypsin that was coupled to capillary electrophoresis for peptide

mapping, which produced over 300,000 theoretical plates for peptide separation.18

That system employed postcolumn labeling with fluorescence detection but did not

separate proteins before digestion. Recently, they further developed a fully automated

bottom-up approach to protein characterization.178 Proteins were first separated by

capillary electrophoresis. A pepsin microreactor was incorporated into the distal end

of this capillary. The reason for choosing pepsin rather than trypsin as the digestion

enzymewas that pepsin allows to use a volatile acetic acid–ammonium acetate buffer

that is beneficial for combination with ESI-MS. Peptides formed in the reactor were

then transferred to a second capillary, where they were separated by capillary

electrophoresis and characterized by mass spectrometry. While peptides generated

from one digestion were being separated in the second capillary, the next protein

fraction underwent digestion in the microreactor (Figure 4.24d).

Enzyme Immobilized in the First Part of the CE Capillary Construction of an

IMER coupled with CE capillary is a quite complex operation; its reproducibility in

fabrication is low, and ensuring electrical and fluidic connections is not trivial either.

As an alternative, the enzyme can be directly immobilized on a portion of a capillary

column; the first part containing the immobilized enzyme then acts as a microreactor

and the second part of the column is devoted to the separation of the peptides. This

approach can in principle reduce systematic errors associatedwithmoving the sample

from the microreactor into the separation zone. One of the most serious issues that

restricted the application of this approach is the difficulty in controlling the position

and the sizeof thepatchof immobilized enzymes.Another critical consideration is that

since the enzymatic reaction and the separation were performed in the same capillary

column, the buffers are exactly the same for both procedures. It is known that the


optimal pH of the enzymatic reaction probably differs from the pH at which optimal

selectivity is obtained for the separation of the peptide. Therefore, the buffer

conditions should be carefully selected considering from both standpoints.

Trypsin was encapsulated in tetramethoxysilane-based hydrogel in a single step

under mild conditions within a capillary, and a 1.5-cm length of gel was formed at the

inlet of the capillary.47 The average amount of trypsin in the capillary was about

0.90mg cm�1 gel. The resultant monolithic reactor showed enzymatic activity ap-

proximately 700 times higher than that in free solution, without stopping the flow. The

substrates were introduced electrokinetically from the inlet of the capillary and then

cleaved into products while they flowed through the trypsin-encapsulated gel by

electrophoresis and EOF. Unreacted substrates and products were separated by

electrophoresis (Figure 4.25). The buffer conditionwas considered for both enzymatic

reaction and CE separation and 50mM Tris–HCl (pH 7.5) was finally selected. The

system was used successfully for hydrolyzing a-N-benzoyl-L-arginine ethyl ester

(BAEE) and separating its product without stopping the flow. However, the method

was tested only on peptides, as proteins larger than trypsin could not pass through the

hydrogel and were not digested, severely limiting the application of this system.

Enzyme can also be immobilized on a portion of silica capillary through a

photocoupling reaction. Bossi et al. constructed a CE-microreactor for peptide

mapping using this technology.179 The bioreactor was characterized by being a single

piece, thus ensuring no fluidic or electrical leakage typical of the reactors constructed

asmultiassemblies.The immobilization procedurewasoptimized, and the activity and

stability of the reactor were tested with proteins of different dimensions (cytochrome

c, hemoglobin, and carbonic anhydrase). Mapping online in the CE-microreactor was

quite competitive in terms of time (completed map within 15min) and exhaustive for

FIGURE 4.25 Schematic illustration of an online enzyme reactor integrated into CE.



the mapping of small proteins. However, the CE-microreactor was not coupled toMS

for complete identification of the peptides mapped.

4.4.2 Immobilized Enzyme Biosensors

Immobilized enzymes have been used as component of detection device in various

analytical and biomedical applications. Some of these devices have been associated

with miniaturized electrochemical detectors and fluidic systems. Since numerous

oxidase and dehydrogenase enzymes generate oxidizable products (hydrogen perox-

ide and NADH, respectively), enzyme-based biochip assays were integrated with

various amperometric detectors byWang et al.180–182 Such electrochemical detectors

offer additional advantages for CE microchips, including compatibility with micro-

machining technologies, miniaturization of both the detector and control instrumen-

tation, and high sensitivity and selectivity. On-line precolumn and on-column

reactions of glucose oxidase and alcohol dehydrogenase have been employed both

for selective measurements of glucose (in the presence of ascorbic acid and uric

acid)180 and for the simultaneousmeasurements of glucose and ethanol (in connection

with electrophoretic separation of the peroxide and NADH products).181 On-chip

assay of amino acids based on their electrophoretic separation, postcolumn reaction

with amino-acid oxidase and amperometric detection of the hydrogen peroxide

product was also developed by the same group.182

L’Hostis et al.183 tested the enzymatic microreactor for glucose detection via

glucose oxidase immobilized to glass beads. Glucose detection was carried out by

electrochemical measurements of hydrogen peroxide generated enzymatically with a

platinum electrode. Glucosewas alsomeasuredwithmicrofluidic biosensors based on

immobilizing glucose oxidase in poly(dimethylsiloxane) electrophoretic micro-

chips.93 An immobilized enzymatic fluorescence capillary biosensor was developed

for determination of sulfated bile acid in urine.184 Amicroelectromechanical systems

(MEMS) device consisting of two identical freestanding polymer diaphragms, resis-

tive heaters, and a thermopile between the diaphragms was fabricated.185 Enzymes

specific to ametabolic analyte systemwere immobilized onmicrobeads packed in the

chambers. When a sample solution containing the analyte was introduced to the

device, the heat released from the enzymatic reactions of the analyte was detected by

the thermopile. The device had been tested with glucose solutions at physiologically

relevant concentrations and shown its potentiality for continuous monitoring of

glucose and other metabolites.

De Boer et al. designed and implemented a continuous-flowmicrofluidic assay for

screening (complex) mixtures for bioactive compounds.186 The microfluidic chip

featured two microreactors (1.6 and 2.4mL) in which an enzyme inhibition and a

substrate conversion reaction were performed, respectively. Enzyme inhibition was

detected by continuously monitoring the products formed in the enzyme–substrate

reaction by electrospray ionizationmass spectrometry. In order to enable the screening

of mixtures of compounds, the chip-based assay was coupled online to capillary

reversed-phase high-performance liquid chromatography with the HPLC column

being operated either in isocratic or in gradient elution mode. In order to improve the


detection limits of the current method, sample preconcentration based on a micro

online solid-phase extraction column was employed (Figure 4.26).

IMERs are also utilized for screening inhibitors as potential drugs.187,188 Brennan

and coworkers recently prepared protein-doped monolithic silica columns for im-

mobilized enzyme reactors, which allowed the screening of enzyme inhibitors with

MS as the detector.189 Kang and coworkers created an immobilized capillary

acetylcholinesterase (AChE) reactor based on a layer-by-layer assembly for inhibitor

screening. A 0.5 cm long plug of solution of the cationic polyelectrolyte polydial-

lyldimethylammoniumwas injected into the capillary to produce a positively charged

coating on the surface of the capillary; subsequently, the enzyme solution with the

same plug length was injected into the capillary and incubated for 10min to

immobilize the enzyme on the capillary wall via electrostatic interaction; third,

PDDA solution with the same plug length was injected again into the capillary to

cover the immobilized enzyme by forming a PDDA–AChE–PDDA sandwich-like

structure. The substrate solution was injected and incubated for a short time, followed

by applying a voltage to separate the product from the unreacted substrate

FIGURE 4.26 (a) Schematic overview of the online continuous-flow system: 1, pump and

autosampler; 2, trapping column; 3, switching valve; 4, analytical column; 5, syringe pumps; 6,

microfluidic chip; 7, mass spectrometer. (b) The microfluidic chip as used for bioactivity

screening: 1, substrate solution; 2, LCeffluent; 3, enzyme solution; 4, open tubularmicroreactor

with a volume of 1.6mL; 5, open tubular microreactor with a volume of 2.4mL; 6, flow toward

mass spectrometer. Reprinted from Ref. 186, with permission.


(Figure 4.27). Screening a small compound library containing 4 known AChE

inhibitors and 42 natural extracts was demonstrated, and species with inhibition

activity can be straightforwardly identified with the system.188

TwocytochromeP450 (CYP)-based immobilized enzyme reactorswere developed

to perform automated online phase I drug metabolism studies.190 For this purpose,

biotinylated recombinant CYP2D6 or CYP3A4 reconstituted systems were anchored

to the surface of two monolithic minicolumns (2mm� 6mm ID), which had been

covalently grafted with NeutrAvidin. After optimization of immobilization condi-

tions, the obtained IMERs were integrated online into a LC hyphenated to an

FIGURE 4.27 Schematic representation of the immobilized capillary enzyme reactor with

CE separation for inhibitor screening. Reprinted from Ref. 188, with permission.


electrospray ionization MS/MS system. Studies with probe substrates and a known

competitive inhibitor were performed, showing the potential of CYP-based IMERs in

drug metabolism.

4.4.3 Kinetic Studies

Microreactors offer significant advantages for online monitoring of biocatalysis and

characterization of kinetics of supported enzymes. First, application of IMER offers a

great advantage by shortening the analysis time. Inmany cases, an enzymatic reaction

is very fast and can reach equilibrium within a single passage of substrate stream

through the microreaction channel. However, several biotransformations, for exam-

ple, those catalyzed by lipases, are slower. Scaling down the dimensions of the

microreactor, and immobilizing the enzyme (lipase) inside a fused silica capillary,

leads to very short times for the hydrolysis,191 while in batch reactions, completion of

enzyme-catalyzed transesterification may take days for some supported lipases.192

Second, application of IMER greatly decreases the amount of biocatalyst used. In

comparisonwith standard assays, the amount of enzymeusedwas very small: Seong et

al. estimated that only 200 pmol (3� 109 molecules of enzyme) was required for the

analysis.193 Moore et al. presented an assay for 500 lipase molecules that could be

applied to single cells.194

The kinetics model described by Lilly et al.195 is appropriate for systems with

continuous flow of the substrate and under steady-state conditions, and can be

summarized by the following equation:

PS0 ¼ K 0m lnð1�PÞþC=Q

where P is the fraction of substrate reacted in the column, S0 is the substrate

concentration at the beginning, K 0m denotes the apparent Michaelis constant, C is

the reaction capacity of the reactor, andQ is the flow rate of the substrate. This formula

allows determination of the apparentMichaelis constant of the catalytic process when

all other parameters are known. The Michaelis constant is typically measured with a

series of experiments at different substrate concentrations in a well-mixed container.

Seong et al. showed that theMichaelis constant determinedwith amicrofluidic device

with immobilized horseradish peroxidase was similar to the value obtained during

homogeneous catalysis in batchmode.193 If anymass transfer effects contribute to the

dynamics, an extrapolation to zero flow rate is required to obtain the value of the

Michaelis constant for comparisonwith that of free enzyme.193An interestingmethod

for determiningKm and nmax was presented by Jiang et al., who applied online frontal

analysis of peptides originating from the digestion by trypsin immobilized on glycidyl

methacrylate-modified cellulose.196 The Lineweaver–Burke diagrams were easily

constructed based on the effects of injection of different concentrations and variation

of flow rate of the substrate solution. Ristenpart et al. demonstrated a microfluidic

technique for measuring Michaelis–Menten rate constants with only a single experi-

ment.197 Enzyme and substratewere brought together in a coflowmicrofluidic device,

and they established analytically and numerically that the initial concentration of


product scales with the distance x along the channel as x5/2. Measurements of the

initial rate of product formation, combined with the quasi-steady rate of product

formation further downstream, yielded the rate constants. They corroborated the x5/2scaling result experimentally using the bioluminescent reaction between ATP and

luciferase/luciferin as a model system.

Mass transfer is always an important issuewhen considering enzymes entrapped in

supports, and the ideal situation iswhen diffusion of substrate and product into and out

of the bulk solution is not the rate limiting process. Koh and Pishko determined

Michaelis constants of enzymes entrapped in hydrogel micropatches in microfluidic

channels using Lineweaver–Burke graphs.198 Values were found to be lower, by

approximately an order ofmagnitude, than those obtained from experiments using the

homogeneous enzymes. The influence of entrapment in the hydrogel nanostructure on

the kinetic properties of the enzymes was discussed.

4.5 SUMMARY AND FUTURE PERSPECTIVE

In this chapter, we have focused on the devolvement of immobilized microfluidic

enzymatic reactors using nanoparticles. Avariety ofmethods are now available for the

immobilization of enzymes on nanoparticles. Immobilization techniques, such as

physical adsorption, covalent binding, and copolymerization of enzyme with the

polymers are discussed in this chapter.The specific immobilization chemistrydepends

on a variety of factors, characters of the supports, activation methods, and coupling

procedure. The ideal supports for immobilization process should have the following

characteristics: (i) large surface area, (ii) permeability, (iii) hydrophilic character, (iv)

insolubility, (v) chemical, mechanical, and thermal stability, (vi) high rigidity, (vii)

chemical reactivity for coupling of the ligands, and (viii) resistance to microbial and

enzymatic attack.2 With the rapid acceleration of research in the area of nanotech-

nology, new nanoparticles (nonmagnetic or magnetic) that can serve as enzyme

immobilization supports emerge in increasing numbers with lower cost, nontoxicity,

higher activity, and stability, aswell as easier enzyme availability and immobilization.

Intentionally, several types of supports are now commercially available for immobi-

lization processes. Poros� particles (poly(styrene–divinylbenzene)) possessing large

through-pores that allow analyte molecules to “perfuse” rapidly through the interior

of the particles, as well as very short “diffusive” pores, have been introduced by

Applied Biosystems (Foster City, CA, USA) into the market as Poroszyme� Enzyme

trypsin bulk media and cartridges (30mm� 2.1mm ID) to perform rapid online

tryptic digestion of protein.173–175 CIM� Disk Monolithic Columns developed by

BIA Separations (Ljubjana, Slovenia) are also utilized for trypsin immobilization and

online protein digestion.199,200 In the light of current developments in nanotechnolo-

gy, it is to be hoped that the availability of new support materials will lead to faster,

more effective, and more stable IMERs.

Due to the flow-through format of IMERs, they can be easily combined with other

flow-through techniques such as HPLC or CE separation withMS detection, allowing

their broad applications in tremendous areas. However, some drawbacks such as band


broadening due to nonspecific binding to the enzyme and the influence of the flow rate

on the substrate conversion efficiency in the case of specificmass transfer regimen can

limit the application of IMERs in an online system. Now with the introduction of

monolithic supports, it is possible to overcome all diffusion limitations so that the

enzymatic kinetic parameters are flow-unaffected.

IMERs are commonly used in medical diagnostics and therapy, enzyme-based

electrodes, organic synthesis, kinetic study, biosensors, and many other applications

such as removal of waste metabolites, blood detoxification, and/or corrections of

inborn metabolic deficiency. Among these applications, most of the IMER applica-

tions are aimed at protein analysis by peptide mapping. High immobilized enzyme

concentrations, in combination with the superior solute transport properties of

perfusivemedia, allow achievement of complete digestion inminutes or even seconds

rather than in hours when compared to solution-phase digest. Unlike solution-phase

methods, the degree of protein digestion can be controlled through the flow rate in the

IMER. The highly stable enzyme immobilization chemistry enables the reuse of the

same IMERs for times of assays,with reproducible results. The integration of different

steps (digestion, separation, and identification) in a single system is particularly

attractive for the analysis of complex protein mixtures. However, it is noticed that,

although numerous publications related to application of IMER contained online

system in protein analysis have been reported so far, most of these studies concerned

digestion of standard proteins, such as myoglobin, cytochrome c, R-lactalbumin,

bovine serumalbumin, andholo-transferrin. Incontrast, fewworks concernedanalysis

of proteins in real samples (tissue, cell, etc.).

Further technical improvements are needed to enable the IMERs contained online

system for the detection and characterization of low-abundance analytes, and to

increase the throughput of the methodologies. The final goal is to find robust,

automated, and sensitive high-throughput analytical tools in bioanalytical science

and in applied biotechnology.

REFERENCES

1. Miyazaki, M.; Maeda, H. Microchannel enzyme reactors and their applications for

processing. Trends Biotechnol. 2006, 24, 463–470.

2. K�rrenkov�a, J.; Foret, F. Immobilized microfluidic enzymatic reactors. Electrophoresis

2004, 25, 3550–3563.

3. B�ılkov�a, Z.; Slov�akov�a, M.;Minc, N.; F€utterer, C.; Cecal, R.; Hor�ak, D.; Benes, M.; Potier,

I.; K�renkov�a, J.; Przybylski, M.; Viovy, J.-L. Functionalized magnetic micro- and

nanoparticles: optimization and application to m-chip tryptic digestion. Electrophoresis

2006, 27, 1811–1824.

4. Thomsen, M.S.; Nidetzky, B. Microfluidic reactor for continuous flow biotransformations

with immobilized enzymes: the example of lactose hydrolysis by a hyperthermophilic b-glycoside hydrolase. Eng. Life Sci. 2008, 8, 40–48.

5. J€ahnisch, K.; Hessel, V.; L€owe, H.; Baerns, M. Chemistry in microstructured reactors.

Angew. Chem., Int. Ed. 2004, 43, 406–446.

REFERENCES 173

6. DeMello, A.J. Control and detection of chemical reactions inmicrofluidic systems.Nature

2006, 442, 394–402.

7. Watts, P.; Wiles, C. Recent advances in synthetic micro reaction technology. Chem.

Commun. 2007, 5, 443–67.

8. Amankwa, L.N.; Kuhr, W.G. Online peptide mapping by capillary zone electrophoresis.

Anal. Chem. 1993, 65, 2693–2697.

9. Cobb, K.A.; Novotny, M. High-sensitivity peptide mapping by capillary zone electropho-

resis and microcolumn liquid chromatography using immobilized trypsin for protein

digestion. Anal. Chem. 1989, 61, 2226–2231.

10. Davis, M.T.; Lee, T.D.; Ronk, M.; Hefta, S.A. Microscale immobilized protease reactor

columns for peptide mapping by liquid chromatography/mass spectral analyses. Anal.

Biochem. 1995, 224, 235–244.

11. Hermanson, G.T. Bioconjugate Techniques; Academic Press: San Diego, CA, 1996.

12. Girelli, A.M.; Mattei, E. Application of immobilized enzyme reactor in on-line high

performance liquid chromatography: a review. J. Chromatogr. B 2005, 819, 3–16.

13. Li, Y.; Yan, B.; Deng, C.-H.; Yu,W.-J.; Xu, X.-Q.; Yang, P.-Y.; Zhang, X.-M. Efficient on-

chip proteolysis system based on functionalized magnetic silica microspheres.

Proteomics 2007, 7, 2330–2339.

14. Li, Y.; Xu, X.-Q.; Deng, C.-H.; Yang, P.-Y.; Zhang, X.-M. Immobilization of trypsin on

superparamagnetic nanoparticles for rapid and effective proteolysis. J. Proteome Res.

2007, 6, 3849–3855.

15. Drobn�ık, J.; Vlas�ak, J.; Pila�r, J.; Svec, F.; K�alal, J. Synthetic polymers in the study of

protein immobilization on glycidylmethacrylate carriers.EnzymeMicrob. Technol. 1979,

1, 107–112.

16. Vodopivec, M.; Berovic, M.; Jancar, J.; Podgornik, A.; Strancar, A. Application of

convective interaction media disks with immobilised glucose oxidase for on-line glucose

measurements. Anal. Chim. Acta 2000, 407, 105–110.

17. Petro, M.; Svec, F.; Fr�echet, J.M.J. Immobilization of trypsin onto “molded” macro-

porous poly(glycidyl methacrylate-co-ethylene dimethacrylate) rods and use of the

conjugates as bioreactors and for affinity chromatography. Biotechnol. Bioeng. 1996,

49, 355–363.

18. Ye, M.; Hu, S.; Schoenherr, R.M.; Dovichi, N.J. On-line protein digestion and peptide

mapping by capillary electrophoresis with post-column labeling for laser-induced

fluorescence detection. Electrophoresis 2004, 25, 1319–1326.

19. Svec, F.; Hrudkova, H.; K�alal, J. Reactive polymers. 20. Preparation of a carrier with

aldehyde groups by oxidation of a macroporous glycidyl methacrylate-ethylenedi-

methacrylate copolymer with periodic acid. Angew. Makromol. Chem. 1978, 70,

101–108.

20. K�renkov�a, J.; Bilkov�a, Z.; Foret, F. Characterization of a monolithic immobilized trypsin

microreactor with on-line coupling to ESI-MS. J. Sep. Sci. 2005, 28, 1675–1684.

21. Jiang, H.-H.; Zou, H.-F.; Wang, H.-L.; Ni, J.-Y.; Zhang, Q.; Zhang, Y.-K. On-line

characterization of the activity and reaction kinetics of immobilized enzyme by high-

performance frontal analysis. J. Chromatogr. A 2000, 903, 77–84.

22. Xie, S.; Svec, F.; Fr�echet, J.M.J. Design of reactive porous polymer supports for high

throughput bioreactors: poly(2-vinyl-4,4-dimethylazlactone-co-acrylamide-co-ethyl

dimethacrylate) monoliths. Biotechnol. Bioeng. 1999, 62, 30–35.


23. Nouaimi, M.; M€oschel, K.; Bisswanger, H. Immobilization of trypsin on polyester fleece

via different spacers. Enzyme Microb. Technol. 2001, 29, 567–574.

24. Turkov�a, J. Oriented immobilization of biologically active proteins as a tool for revealing

protein interactions and function. J. Chromatogr. B 1999, 722, 11–31.

25. Li, Y.; Yan, B.; Deng, C.-H.; Tang, J.; Liu, J.-Y.; Zhang, X.-M. On-plate digestion of

proteins using novel trypsin-immobilized magnetic nanospheres for MALDI-TOF-MS

analysis. Proteomics 2007, 7, 3661–3671.

26. Gao, J.; Xu, J.; Locascio, L.E.; Lee, C.S. Integrated microfludic system enabling protein

digestion, peptide separation, and protein identification. Anal. Chem. 2001, 73,

2648–2655.

27. Cooper, J.W.; Chen, J.; Li, Y.; Lee, C.S. Membrane-based nanoscale proteolytic reactor

enabling protein digestion, peptide separation, and protein identification using mass

spectrometry. Anal. Chem. 2003, 75, 1067–1074.

28. Massolini, G.; Calleri, E.; DeLorenzi, E.; Pregnolato,M.; Terreni,M.; Felix, G.; Gandini,

C. Immobilized penicillin G acylase as reactor and chiral selector in liquid chromatog-

raphy. J. Chromatogr. A 2001, 921, 147–160.

29. Calleri, E.; Temporini, C.; Furlanetto, S.; Loiodice, F.; Fracchiolla, G.; Massolini, G.

Lipases for biocatalysis: development of a chromatographic bioreactor. J. Pharm.

Biomed. Anal. 2003, 32, 715–724.

30. Guo, Z.; Xu, S.-Y.; Lei, Z.-D.; Zou, H.-F.; Guo, B.-C. Immobilized metal-ion chelating

capillary microreactor for peptide mapping analysis of proteins by matrix assisted laser

desorption/ionization-time of flight-mass spectrometry. Electrophoresis 2003, 24,

3633–3639.

31. Anspach, F.B. Silica-based metal chelate affinity sorbents. I. Preparation and characteri-

zation of iminodiacetic acid affinity sorbents prepared via different immobilization

techniques. J. Chromatogr. A 1994, 672, 35–49.

32. Li, Y.; Xu, X.-Q.; Yang, B.; Deng, C.-H.; Yu,W.-J.; Yang, P.-Y.; Zhang, X.-M.Microchip

reactor packed with metal-ion chelated magnetic silica microspheres for high efficient

proteolysis. J. Proteome Res. 2007, 6, 2367–2375.

33. Li, Y.; Yan, B.; Xu, X.-Q.; Deng, C.-H.; Yang, P.-Y.; Shen, X.-Z.; Zhang, X.-M. On-

column tryptic mapping of proteins using metal-ion-chelated magnetic silica micro-

spheres by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry.

Rapid Commun. Mass Spectrom. 2007, 21, 2263–2268.

34. Lvov, Y.; Decher, G.; Mohwald, H. Assembly, structural characterization, and thermal

behavior of layer-by-layer deposited ultrathin films of poly(vinyl sulfate) and poly

(allylamine). Langmuir 1993, 9, 481–486.

35. Decher, G. Fuzzy nanoassemblies: toward layered polymeric multicomposites. Science

1997, 277, 1232–1237.

36. Serizawa, T.; Yamaguchi, M.; Akashi, M. Enzymatic hydrolysis of a layer-by-layer

assembly prepared from chitosan and dextran sulfate. Macromolecules 2002, 35,

8656–8658.

37. Ichinose, I.; Takaki, R.; Kuroiwa, K.; Kunitake, T. Electrostatic desorption of cytochrome

c on ultrathin ZrO2-gel layers and preparation of alternate multilayers. Langmuir 2003,

19, 3883–3888.

38. Salloum, D.S.; Schlenoff, J.B. Protein adsorption modalities on polyelectrolyte multi-

layers. Biomacromolecules 2004, 5, 1089–1096.

REFERENCES 175

39. Zhou, X.-C.; Wu, L.-Y.; Zhou, J.-Z. Fabrication of DNAmicroarrays on nanoengineered

polymeric ultrathin film prepared by self-assembly of polyelectrolyte multilayers.

Langmuir 2004, 20, 8877–8885.

40. Liu, Y.; Lu, H.-J.; Zhong,W.; Song, P.-Y.; Kong, J.-L.; Yang, P.-Y.; Girault, H.H.; Liu, B.-

H. Multilayer-assembled microchip for enzyme immobilization as reactor toward low-

level protein identification. Anal. Chem. 2006, 78, 801–808.

41. Liu, Y.; Zhong,W.;Meng, S.; Kong, J.-L.; Lu, H.-J.; Yang, P.-Y.; Girault, H.H.; Liu, B.-H.

Assembly-controlled biocompatible interface on a microchip: strategy to highly efficient

proteolysis. Chem. Eur. J. 2006, 12, 6585–6591.

42. Amankwa, L.N.; Kuhr, W.G. Trypsin-modified fused-silica capillary microreactor

for peptide mapping by capillary zone electrophoresis. Anal. Chem. 1992, 64,

1610–1613.

43. Gleason, N.J.; Carbeck, J.D. Measurement of enzyme kinetics using microscale steady-

state kinetic analysis. Langmuir 2004, 20, 6374–6381.

44. Mao, H.-B.; Yang, T.-L.; Cremer, P.S. Design and characterization of immobilized

enzymes in microfluidic systems. Anal. Chem. 2002, 74, 379–385.

45. Holden, M.A.; Jung, S.-Y.; Cremer, P.S. Patterning enzymes inside microfluidic channels

via photoattachment chemistry. Anal. Chem. 2004, 76, 1838–1843.

46. Anderson, N.L.;Haines, L.R.; Pearson, T.W.An effective and rapidmethod for functional

characterization of immunoadsorbents using POROS beads and flow cytometry. J.

Proteome Res. 2004, 3, 228–234.

47. Sakai-Kato, K.; Kato, M.; Toyo’oka, T. On-line trypsin-encapsulated enzyme reactors by

the sol–gel method integrated into capillary electrophoresis. Anal. Chem. 2002, 74,

2943–2949.

48. Sakai-Kato, K.; Kato, M.; Toyo’oka, T. Creation of on-chip enzyme reactors by

encapsulating trypsin in sol–gel on a plastic microchip. Anal. Chem. 2003, 75, 388–393.

49. Dave, B.C.; Dunn, B.; Valentine, J.S.; Zink, J.I. Sol–gel encapsulation methods for

biosensors. Anal. Chem. 1994, 66, 1120–1127.

50. Gill, I.; Ballesteros, A. Bioencapsulation within synthetic polymers. Part 1. Sol–gel

encapsulated biologicals. Trends Biotechnol. 2000, 18, 282–296.

51. Hahn, R.; Podgornik, A.; Merhar, M.; Schallaun, E.; Jungbauer, A. Affinity monoliths

generated by in situ polymerization of the ligand. Anal. Chem. 2001, 73, 5126–5132.

52. Kim, Y.-D.; Park, C.B.; Clark, D.S. Stable sol–gel microstructured and microfluidic

networks for protein patterning. Biotechnol. Bioeng. 2001, 73, 331–337.

53. Qu,H.-Y.;Wang,H.-T.; Huang,Y.; Zhong,W.; Lu,H.-J.; Kong, J.-L.; Yang, P.-Y.; Liu, B.-

H. Stable microstructured network for protein patterning on a plastic microfluidic

channel: strategy and characterization of on-chip enzyme microreactors. Anal. Chem.

2004, 76, 6426–6433.

54. Park, C.B.; Clark, D.S. Sol–gel encapsulated enzyme arrays for high-throughput screen-

ing of biocatalytic activity. Biotechnol. Bioeng. 2002, 78, 229–235.

55. Liu, Z.-J.; Liu, B.-H.; Kong, J.-L.; Deng, J.-Q. Probe trace phenols based onmediator-free

alumina sol–gel derived tyrosinase biosensor. Anal. Chem. 2000, 72, 4707–4712.

56. Jiang, D.-C.; Tang, J.; Liu, B.-H.; Yang, P.-Y.; Kong, J.-L. Ultrathin alumina sol–gel-

derived films: allowing direct detection of the liver fibrosis markers by capacitance

measurement. Anal. Chem. 2003, 75, 4578–4584.


57. Wu, H.-L.; Tian, Y.-P.; Liu, B.-H.; Lu, H.-J.; Wang, X.-Y.; Zhai, J.-J.; Jin, H.; Yang, P.-Y.;

Xu, Y.-M.; Wang, H.-H. Titania and alumina sol–gel-derived microfluidic enzymatic-

reactors for peptide mapping: design, characterization, and performance. J. Proteome

Res. 2004, 3, 1201–1209.

58. Frenkel-Mullerad, H.; Avnir, D. Sol–gel materials as efficient enzyme protectors:

preserving the activity of phosphatases under extreme pH conditions. J. Am. Chem.

Soc. 2005, 127, 8077–8081.

59. Itoh, T.; Ishii, R.; Ebina, T.; Hanaoka, T.; Fukushima, Y.; Mizukami, F. Encapsulation of

myoglobin with a mesoporous silicate results in new capabilities. Bioconj. Chem. 2006,

17, 236–240.

60. Kumar, R.; Maitra, A.N.; Patanjali, P.K.; Sharma, P. Hollow gold nanoparticles encap-

sulating horseradish peroxidase. Biomaterials 2005, 26, 6743–6753.

61. Ma, D.; Li, M.; Patil, A.J.; Mann, S. Fabrication of protein/silica core–shell nanoparticles

by microemulsion-based molecular wrapping. Adv. Mater. 2004, 16, 1838–1841.

62. Kim, J.; Grate, J.W. Single-enzyme nanoparticles armored by a nanometer-scale organic/

inorganic network. Nano Lett. 2003, 3, 1219–1222.

63. Yan, M.; Ge, J.; Liu, Z.; Ouyang, P.-K. Encapsulation of single enzyme in nanogel

with enhanced biocatalytic activity and stability. J. Am. Chem. Soc. 2006, 128,

11008–11009.

64. Licklider, L.; Kuhr, W.G. Optimization of online peptide mapping by capillary zone

electrophoresis. Anal. Chem. 1994, 66, 4400–4407.

65. Licklider, L.; Kuhr, W.G.; Lacey, M.P.; Keough, T.; Purdon, M.P.; Takigiku, R. Online

microreactors/capillary electrophoresis/mass spectrometry for the analysis of proteins

and peptides. Anal. Chem. 1995, 67, 4170–4177.

66. Licklider, L.; Kuhr, W.G. Characterization of reaction dynamics in a trypsin-modified

capillary microreactor. Anal. Chem. 1998, 70, 1902–1908.

67. Laurell, T.; Rosengren, L. A micromachined enzyme reactor in (110)-oriented silicon.

Sens. Actuators B 1994, 19, 614–617.

68. Laurell, T.; Drott, J.; Rosengren, L. Silicon-wafer integrated enzyme reactors. Biosens.

Bioelectron. 1995, 10, 289–299.

69. Lee, J.; Musyimi, H.K.; Soper, S.A.; Murray, K.K. Development of an automated

digestion and droplet deposition microfluidic chip for MALDI-TOF MS. J. Am. Soc.

Mass Spectrom. 2008, 19, 964–972.

70. K�renkov�a, J.; Klep�arn�ık, K.; Foret, F. Capillary electrophoresis mass spectrometry

coupling with immobilized enzyme electrospray capillaries. J. Chromatogr. A 2007,

1159, 110–118.

71. Ekstr€om, S.; Onnerfjord, P.; Nilsson, J.; Bengtsson, M.; Laurell, T.; Marko-Varga, G.

Integrated microanalytical technology enabling rapid and automated protein identifica-

tion. Anal. Chem. 2000, 72, 286–293.

72. Miyazaki, M.; Kaneno, J.; Uehara, M.; Fujii, M.; Shimizu, H.; Maeda, H. Simple method

for preparation of nanostructure on microchannel surface and its usage for enzyme

immobilization. Chem. Commun. 2003, 5, 648–649.

73. Wu, H.-L.; Zhai, J.-J.; Tian, Y.-P.; Lu, H.-J.;Wang, X.-Y.; Jia,W.-T.; Liu, B.-H.; Yang, P.-

Y.; Xu, Y.-M.; Wang, H.-H. Microfluidic enzymatic-reactors for peptide mapping:

strategy, characterization, and performance. Lab Chip 2004, 4, 588–597.

REFERENCES 177

74. Hessel, V.; L€owe, H.; Sch€onfeld, F. Micromixers: a review on passive and active mixing

principles. Chem. Eng. Sci. 2005, 60, 2479–2501.

75. Xiong, L.; Regnier, F.E. Channel-specific coatings on microfabricated chips. J.

Chromatogr. A 2001, 924, 165–176.

76. Park, J.J.; Luo,X.-L.;Yi,H.-M.;Valentine, T.M.; Payne,G.F.; Bentley,W.E.;Ghodssi, R.;

Rubloff, G.W. Chitosan-mediated in situ biomolecule assembly in completely packaged

microfluidic devices. Lab Chip 2006, 6, 1315–1321.

77. Luo, X.; Lewandowski, A.T.; Yi, H.; Payne, G.F.; Ghodssi, R.; Bentley, W.E.

Programmable assembly of a metabolic pathway enzyme in a pre-packaged reusable

bioMEMS device. Lab Chip 2008, 8, 420–430.

78. Luo, X.-L.; Berlin, D.L.; Buckhout-White, S.; Bentley, W.E.; Payne, G.F.; Ghodssi, R.;

Rubloff, G.W. Design optimization for bioMEMS studies of enzyme-controlled meta-

bolic pathways. Biomed. Microdevices 2008, 10, 899–908.

79. Tomer, S.; Dorsey, J.G.; Berthod, A.Nonionicmicellar liquid chromatography coupled to

immobilized enzyme reactors. J. Chromatogr. A 2001, 923, 7–16.

80. Shu, H.-C.; Wu, N.-P. A chemically modified carbon paste electrode with D-lactate

dehydrogenase and alanine aminotranferase enzyme sequences for D-lactic acid analysis.

Talanta 2001, 54, 361–368.

81. Lim, L.W.; Tomatsu, M.; Takeuchi, T. Development of an on-line immobilized-enzyme

reversed-phase HPLC method for protein digestion and peptide separation. Anal.

Bioanal. Chem. 2006, 386, 614–620.

82. Richter, T.; Shultz-Lockyear, L.L.; Oleschuk, R.D.; Bilitewski, U.; Harrison, D.J. Bi-

enzymatic and capillary electrophoretic analysis of non-fluorescent compounds in

microfluidic devices determination of xanthine. Sens. Actuators B 2002, 81, 369–376.

83. Bertucci, C.; Petri, A.; Felix, G.; Perini, B.; Salvatori, P. Lipase-based HPLC stationary

phase: enantioselective synthesis of 2-substituted 1,3-propanediol monoacetates.

Tetrahedron: Asymmetry 1999, 10, 4455–4462.

84. Kiba, N.; Saegusa, K.; Furosawa, M. Post-column enzyme reactors for chemilumino-

metric detection of glucose 1,5-anhydroglucitol and 3-hydroxybutyrate in an anion-

exchange chromatographic system. J. Chromatogr. B 1997, 689, 393–398.

85. Seong, G.H.; Crooks, R.M. Efficient mixing and reactions within microfluidic channels

using microbead-supported catalysts. J. Am. Chem. Soc. 2002, 124, 13360–13361.

86. Srinivasan, A.; Bach, H.; Sherman, D.H.; Dordick, J.S. Bacterial P450-catalyzed

polyketide hydroxylation on a microfluidic platform. Biotechnol. Bioeng. 2004, 88,

528–535.

87. Oleschuk, R.D.; Shultz-Lockyear, L.L.; Ning, Y.-B.; Harrison, D.J. Trapping of bead-

based reagents within microfluidic systems: on-chip solid-phase extraction and electro-

chromatography. Anal. Chem. 2000, 72, 585–590.

88. Kerby, M.B.; Legge, R.S.; Tripathi, A. Measurements of kinetic parameters in a

microfluidic reactor. Anal. Chem. 2006, 78, 8273–8280.

89. Nakamura, H.; Murakami, Y.; Yokoyama, K.; Tamiya, E.; Karube, I. A compactly

integrated flow cell with a chemiluminescent FIA system for determining lactate

concentration in serum. Anal. Chem. 2001, 73, 373–378.

90. Andersson, H.; van der Wijngaart, W.; Enoksson, P.; Stemme, G. Micromachined flow-

through filter-chamber for chemical reactions on beads. Sens. Actuators B 2000, 67,

203–208.


91. Sato, K.; Tokeshi,M.; Odake, T.; Kimura, H.; Ooi, T.; Nakao,M.; Kitamori, T. Integration

of an immunosorbent assay system: analysis of secretory human immunoglobulin A on

polystyrene beads in a microchip. Anal. Chem. 2000, 72, 1144–1147.

92. Ceriotti, L.; de Rooij, N.F.; Verpoorte, E. An integrated fritless column for on-chip

capillary chromatography with conventional stationary phases. Anal. Chem. 2002, 74,

639–647.

93. Zhang, Q.; Xu, J.-J.; Chen, H.-Y. Glucose microfluidic biosensors based on immobilizing

glucose oxidase in poly(dimethylsiloxane) electrophoretic microchips. J. Chromatogr. A

2006, 1135, 122–126.

94. Xu, Z.-R.; Fang, Z.-L. Composite poly(dimethylsiloxane)/glass microfluidic systemwith

an immobilized enzymatic particle-bed reactor and sequential sample injection for

chemiluminescence determinations. Anal. Chim. Acta 2004, 507, 129–135.

95. Miyazaki, M.; Kaneno, J.; Kohama, R.; Uehara, M.; Kanno, K.; Fujii, M.; Shimizu, H.;

Maeda, H. Preparation of functionalized nanostructures on microchannel surface and

their use for enzyme microreactors. Chem. Eng. J. 2004, 101, 277–284.

96. Miyazaki, M.; Kaneno, J.; Yamaori, S.; Honda, T.; Briones, M.P.P.; Uehara, M.;

Arima, K.; Kanno, K.; Yamashita, K.; Yamaguchi, Y.; Nakamura, H.; Yonezawa, H.;

Fujii, M.; Maeda, H. Efficient immobilization of enzymes on microchannel surface

through His-tag and application for microreactor. Protein Pept. Lett. 2005, 12,

207–210.

97. Merkoci, A.; Marta, A.; Tarraso’n, G.; Eritja, R.; Alegret, S. Toward an ICPMS-linked

DNA assay based on gold nanoparticles immunoconnected through peptide sequences.

Anal. Chem. 2005, 77, 6500–6503.

98. Liu, Y.; Xue, Y.; Ji, J.; Chen, X.; Kong, J.-L.; Yang, P.-Y.; Girault, H.H.; Liu, B.-H. Gold

nanoparticles assembly microfluidic reactor for efficient on-line proteolysis. Mol. Cell.

Proteomics 2007, 6, 1428–1436.

99. Tosheva, L.; Valtchev, V.P. Nanozeolites: synthesis, crystallization mechanism, and

applications. Chem. Mater. 2005, 17, 2494–2513.

100. Huang, Y.; Shan,W.; Liu, B.-H.; Liu,Y.; Zhang,Y.-H.; Zhao,Y.; Lu,H.-J.; Tang,Y.;Yang,

P.-Y. Zeolite nanoparticle modified microchip reactor for efficient protein digestion. Lab

Chip 2006, 6, 534–539.

101. Patel, D.; Moon, J.Y.; Chang, Y.; Kim, T.J.; Lee, G.H. Poly(D,L-lactide-co-glycolide)

coated superparamagnetic iron oxide nanoparticles: synthesis, characterization and in

vivo study as MRI contrast agent. Colloid Surf. A 2008, 313, 91–94.

102. Zhao, M.; Josephson, L.; Tang, Y.; Weissleder, R. Magnetic sensors for protease assays.

Angew. Chem., Int. Ed. 2003, 42, 1375–1378.

103. Mornet, S.; Vasseur, S.; Grasset, F.; Veverka, P.; Goglio, G.; Demourgues, A.; Portier, J.;

Pollert, E.; Duguet, E.Magnetic nanoparticle design formedical applications.Prog. Solid

State Chem. 2006, 34, 237–247.

104. Stevens, P.D.; Fan, J.; Gardimalla, H.M.R.; Yen, M.; Gao, Y. Superparamagnetic

nanoparticle-supported catalysis of Suzuki cross-coupling reactions. Org. Lett. 2005,

7, 2085–2088.

105. Jun, Y.-W.; Choi J.-S.; Cheon, J. Heterostructured magnetic nanoparticles: their versatil-

ity and high performance capabilities. Chem. Commun. 2007, 1203–1214.

106. Choi, W.; Oh, K.W.; Thomas, J.H.; Heineman, W.R.; Halsall, H.B.; Nevin, J.H.;

Helmicki, A.J.; Henderson, H.T.; Ahn, C.H. An integrated microfluidic biochemical

REFERENCES 179

detection system for protein analysis with magnetic bead-based sampling capabilities.

Lab Chip 2002, 2, 27–30.

107. Rashkovetsky, L.G.; Lyubarskaya, Y.V.; Foret, F.; Hughes, D.E.; Karger, B.L. Automated

microanalysis using magnetic beads with commercial capillary electrophoretic instru-

mentation. J. Chromatogr. A 1997, 781, 197–204.

108. Xu, X.; Friedman, G.; Humfeld, K.D.; Majetich, S.A.; Asher, S.A. Superparamagnetic

photonic crystals. Adv. Mater. 2001, 13, 1681–1684.

109. Lu, Y.; Yin, Y.; Mayers, B.T.; Xia, Y. Modifying the surface properties of superpar-

amagnetic iron oxide nanoparticles through a sol–gel approach. Nano Lett. 2002, 2,

183–186.

110. Wu,W.; He, Q.-G.; Jiang, C.-Z. Magnetic iron oxide nanoparticles: synthesis and surface

fictionalization strategies. Nanaoscale Res. Lett. 2008, 3, 397–415.

111. Slov�akov�a, M.; Minc, N.; B�ılkov�a, Z.; Smadja, C.; Faigle, W.; F€utterer, C.; Taverna, M.;

Viovy, J.-L. Use of self assembledmagnetic beads for on-chip protein digestion.LabChip

2005, 9, 935–942.

112. Koh, I.; Wang, X.; Varughese, B.; Isaacs, L.; Ehrman, S.E.; English, D.S. Magnetic iron

oxide nanoparticles for biorecognition: evaluation of surface coverage and activity. J.

Phys. Chem. B 2006, 110, 1553–1558.

113. Nishimura, K.; Hasegawa, M.; Ogura, Y.;et al. 4�C preparation of ferrite nanoparticles

having protein molecules immobilized on their surfaces. J. Appl. Phys. 2002, 91,

8555–8556.

114. Liu, J.-Y.; Lin, S.; Qi, D.-W.; Deng, C.-H.; Yang, P.-Y.; Zhang, X.-M. On-chip enzymatic

microreactor using trypsin-immobilized superparamagnetic nanoparticles for highly

efficient proteolysis. J. Chromatogr. A 2007, 1176, 169–177.

115. Hayes, M.A.; Polson, N.A.; Phayre, A.N.; Garcia, A.A. Flow-based microimmunoassay.

Anal. Chem. 2001, 73, 5896–5902.

116. Lawrence, E.M.; Ivey, M.L.; Flores, G.A.; Lui, J.; Bibette, J.; Richard, J. Field-induced

structure of confined ferrofluid emulsion. Int. J. Mod. Phys. B 1994, 8, 2765–2777.

117. Svec, F. Less common applications of monoliths. I. Microscale protein mapping with

proteolytic enzymes immobilized on monolithic supports. Electrophoresis 2006, 27,

947–961.

118. Yi, Y.Y.; Chen, Y.; Brook, M.A.; Brennan, J.D. Development of macroporous titania

monoliths by a biocompatible method. Part 2. Enzyme entrapment studies.Chem.Mater.

2006, 18, 5336–5342.

119. Cabrera, K. Applications of silica-basedmonolithic HPLC columns. J. Sep. Sci. 2004, 27,

843–852.

120. Calleri, E.; Massolini, G.; Lubda, D.; Temporini, C.; Loiodice, F.; Caccialanza, G.

Evaluation of amonolithic epoxy silica support for penicillinGacylase immobilization. J.

Chromatogr. 1031, 2004, 93–100.

121. Calleri, E.;Marrubini,G.;Massolini,G.; Lubda,D.; de Fazio, S.S.; Furlanetto, S.;Wainer, I.

W.; Manzo, L.; Caccialanza, G. Development of a chromatographic bioreactor based on

immobilized beta-glucuronidase onmonolithic support for the determination of dextrome-

thorphan and dextrorphan in human urine. J. Pharm. Biomed. Anal. 2004, 35, 1179–1189.

122. Chen, Z.; Hayashi, K.; Iwasaki, Y.; Kurita, R.; Niwa, O.; Sunaawa, K. On-line monolithic

enzyme reactor fabricated by sol–gel process for elimination of ascorbic acid while

monitoring dopamine. Electroanalysis 2005, 17, 231–238.


123. Kawakami,K.; Sera, Y.; Sakai, S.; Ono, T.; Ijima,H.Development and characterization of

a silica monolith immobilized enzyme micro-bioreactor. Ind. Eng. Chem. Res. 2005, 44,

236–240.

124. Sakai-Kato,K.;Kato,M.;Toyo’oka,T.On-line trypsin-encapsulatedenzyme reactor by the

sol–gelmethod integrated intocapillaryelectrophoresis.Anal.Chem.2002,74, 2943–2949.

125. Kato, M.; Inuzuka, K.; Sakai-Kato, K.; Toyo’oka, T. Monolithic bioreactor immobilizing

trypsin for high-throughput analysis. Anal. Chem. 2005, 77, 1813–1818.

126. Dulay, M.T.; Quirino, J.P.; Bennett, D.C.; Kato, M.; Zare, R.N. Photopolymerized

sol–gel monoliths for capillary electrochromatography. Anal. Chem. 2001, 73,

3921–3926.

127. Kato, M.; Sakai-Kato, K.; Jin, H.M.; Kubota, K.; Miyano, H.; Toyo’oko, T.; Dulay, M.T.;

Zare, R.N. Integration of on-line protein digestion, peptide separation, and protein

identification using pepsin-coated photopolymerized sol–gel columns and capillary

electrophoresis/mass spectrometry. Anal. Chem. 2004, 76, 1896–1902.

128. Dulay, M.T.; Baca, Q.J.; Zare, R.N. Enhanced proteolytic activity of covalently bound

enzymes in photopolymerized sol gel. Anal. Chem. 2005, 77, 4604–4610.

129. Kostner, A.; Mandel, M. In Immobilized Enzymes; Mosbach, K., Ed.; Academic Press:

New York, 1976; pp 191–195.

130. Mersal, G.A.M.; Bilitewski, U. Development of monolithic enzymatic reactors in glass

microchips for the quantitative determination of enzyme substrates using the example of

glucose determination via immobilized glucose oxidase. Electrophoresis 2005, 26,

2303–2312.

131. Palm, A.; Novotny, M.V. Analytical characterization of a facile porous polymer mono-

lithic trypsin microreactor enabling peptide mass mapping using mass spectrometry.

Rapid Commun. Mass Spectrom. 2004, 18, 1374–1382.

132. Palm, A.; Novotny, M.V. A monolithic PNGase F enzyme microreactor enabling glycan

mass mapping of glycoproteins by mass spectrometry. Rapid Commun. Mass Spectrom.

2005, 19, 1730–1738.

133. Hilder, E.F.; Svec, F.; Fr�echet, J.M.J. Polymericmonolithic stationary phases for capillary

electrochromatography. Electrophoresis 2002, 23, 3934–3953.

134. Ben�cina, K.; Podgornik, A.; Strancar, A.; Ben�cina,M. Enzyme immobilization on epoxy-

and 1,10-carbonyldiimidazole-activated methacrylate-based monoliths. J. Sep. Sci. 2004,

27, 811–818.

135. Hermanson, G.T.; Mallia, A.K.; Smith, P.K. Immobilized Affinity Ligand Techniques;

Academic Press: New York, 1992.

136. Hrudkov�a, H.; Svec, F.; K�alal, J. Reactive polymers. 14. Hydrolysis of epoxide groups of

copolymer glycidylmethacrylate-ethylenedimethacrylate.Br. Polym. J. 1977, 9, 238–240.

137. Luo, Q.-Z.; Mao, X.-Q.; Huang, X.; Zou, H.-F. High-performance affinity chromatogra-

phy for characterization of human immunoglobulin G digestion with papain. J.

Chromatogr. B 2002, 776, 139–147.

138. K�renkov�a, J.; Lacher, N.A.; Svec, F. Highly efficient enzyme reactors containing trypsin

and endoproteinase LysC immobilized on porous polymer monolith coupled to MS

suitable for analysis of antibodies. Anal. Chem. 2009, 81, 2004–2012.

139. Valentov�a, O.; Marek, M.; Svec, F.; Stamberg, J.; Vodr�a�zka, Z. Comparison of different

methods of glucose oxidase immobilization. Biotechnol. Bioeng. 1981, 23, 2093–2104.

REFERENCES 181

140. Hearn, W.; Bethell, G.S.; Ayers, J.S.; Hancock, W.S. Application of 1,10-carbonyldii-midazone-activated agarose for the purification of proteins. II. The use of an activated

matrix devoid of additional charged groups for the purification of thyroid proteins. J.

Chromatogr. 1979, 185, 463–470.

141. Coleman, P.L.; Walker, M.M.; Milbrath, D.S.; Stauffer, D.M.; Rasmussen, J.K.; Krepski,

L.R.; Heilmann, S.M. Immobilization of proteinA at high density on azlactone-functional

polymeric beads and their use in affinity chromatography. J. Chromatogr. A 1990, 512,

345–363.

142. Heilmann, S.M.; Rasmussen, J.K.; Krepski, L.R. Chemistry and technology of 2-alkenyl

azlactones. J. Polym. Sci., Polym. Chem. 2001, 39, 3655–3677.

143. Drtina, G.J.; Haddad, L.C.; Rasmussen, J.K.; Gaddam, B.N.;et al. Azlactone-reactive

polymer supports for immobilizing synthetically useful enzymes. II. Important prelimi-

nary hydrogen bonding effects in the covalent coupling of penicillin G acylase. React.

Funct. Polym. 2005, 64, 13–24.

144. Heilmann, S.M.; Drtina, G.J.; Haddad, C.; Rasmussen, J.K.;et al. Azlactone-reactive

polymer supports for immobilizing synthetically useful enzymes. Part I. Pig liver esterase

on dispersion polymer supports. J. Mol. Catal. B 2004, 30, 33–42.

145. Xie, S.; Svec, F.; Fr�echet, J.M.J. Design of reactive porous polymer supports for high

throughput bioreactors: poly(2-vinyl-4,4-dimethylazlactone-co-acrylamide-co-ethyle-

nedimethacrylate) monoliths. Biotechnol. Bioeng. 1999, 62, 30–35.

146. Peterson, D.S.; Rohr, T.; Svec, F.; Fr�echet, J.M.J. Enzymatic microreactor-on-a-chip:

protein mapping using trypsin immobilized on porous polymer monoliths molded in

channels of microfluidic devices. Anal. Chem. 2002, 74, 4081–4088.

147. Peterson, D.S.; Rohr, T.; Svec, F.; Fr�echet, J.M.J. High-throughput peptidemass mapping

using a microdevice containing trypsin immobilized on a porous polymer monolith

coupled to MALDI TOF and ESI TOF mass spectrometers. J. Proteome Res. 2002, 1,

563–568.

148. Geiser, L.; Eeltink, S.; Svec, F.; Fr�echet, J.M.J. In-line system containing porous polymer

monoliths for protein digestion with immobilized pepsin, peptide preconcentration and

nano-liquid chromatography separation coupled to electrospray ionization mass spec-

troscopy. J. Chromatogr. A 2008, 1188, 88–96.

149. Svec, F. High-throughput peptide mass mapping using an integrated capillary device

coupled to a mass spectrometer. LC-GC Eur. 2005, 18, 17–20.

150. Wang, P.-C.; DeVoe, D.L.; Lee, C.S. Integration of polymeric membranes with micro-

fluidic networks for bioanalytical application. Electrophoresis 2001, 22, 3857–3867.

151. Hisamoto, H.; Shimizu, Y.; Uchiyama, K.; Tokeshi, M.; Kikutani, Y.; Hibara, A.;

Kitamori, T. Chemicofunctional membrane for integrated chemical processes on a

microchip. Anal. Chem. 2003, 75, 350–354.

152. Honda, T.; Miyazaki, M.; Nakamura, H.; Maeda, H. Immobilization of enzyme on a

microchannel surface through cross-linking polymerization. Chem. Commun. 2005,

5062–5064.

153. Dogruel, D.; Williams, P.; Nelson, R.W. Rapid tryptic mapping using enzymatically

active mass-spectrometer probe tips. Anal. Chem. 1995, 67, 4343–4348.

154. Nelson, R.W.; Dogruel, D.; Krone, J.R.; Williams, P. Peptide characterization using

bioreactive mass-spectrometer probe tips. Rapid Commun. Mass Spectrom. 1995, 9,

1380–1385.


155. Houston, C.T.; Reilly, J.P. Toward a simple, expedient, and complete analysis of human

hemoglobin by MALDI-TOFMS. Anal. Chem. 1999, 71, 3397–3404.

156. Zheng, S.P.; Yoo, C.; Delmotte, N.; Miller, F.R.; Huber, C.G.; Lubman, D.M. Monolithic

column HPLC separation of intact proteins analyzed by LC-MALDI using on-plate

digestion: an approach to integrate protein separation and identification. Anal. Chem.

2006, 78, 5198–5204.

157. Ota, S.; Miyazaki, S.; Matsuoka, H.; Morisato, K.; Shintani, Y.; Nakanishi, K. High-

throughput protein digestion by trypsin-immobilized monolithic silica with pipette-tip

formula. J. Biochem. Biophys. Methods 2007, 70, 57–62.

158. Fan, H.-Z.; Chen, G. Fiber-packed channel bioreactor for microfluidic protein digestion.

Proteomics 2007, 7, 3445–3449.

159. Wang, S.; Chen, Z.; Yang, P.-Y.; Chen, G. Trypsin-immobilized fiber core in syringe

needle for highly efficient proteolysis. Proteomics 2008, 8, 1785–1788.

160. Urban, P.L.; Goodall, D.M.; Bruce, N.C. Enzymatic microreactors in chemical analysis

and kinetic studies. Biotechnol. Adv. 2006, 24, 42–57.

161. Massolini, G.; Calleri, E. Immobilized trypsin systems coupled on-line to separation

methods: recent developments and analytical applications. J. Sep. Sci. 2005, 28, 7–21.

162. Laurent, N.; Haddoub, R.; Flitsch, S.L. Enzyme catalysis on solid surfaces. Trends

Biotechnol. 2008, 26, 328–337.

163. Ma, J.-F.; Zhang, L.-H.; Liang, Z.; Zhang, W.-B.; Zhang, Y.-K. Recent advances in

immobilized enzymatic reactors and their applications in proteome analysis. Anal. Chim.

Acta 2009, 632, 1–8.

164. Foret, F.; Preisler, J. Liquid phase interfacing and miniaturization in matrix-assisted laser

desorption/ionization mass spectrometry. Proteomics 2002, 2, 360–372.

165. Julka, S.; Regnier, F.Quantification in proteomics through stable isotope coding: a review.

J. Proteome Res. 2004, 3, 350–363.

166. Bruyneel, B.; Hoos, J.S.; Smoluch, M.T.; Lingeman, H.; Niessen, W.M.A.; Irth, H. Trace

analysis of proteins using postseparation solution-phase digestion and electrospray mass

spectrometric detection of marker peptides. Anal. Chem. 2007, 79, 1591–1598.

167. Lazar, I.M.; Ramsey, R.S.; Ramsey, J.M.On-chip proteolytic digestion and analysis using

“wrong-way-round” electrospray time-of-flight mass spectrometry. Anal. Chem. 2001,

73, 1733–1739.

168. Peterson, D.S.; Rohr, T.; Svec, F.; Fr�echet, J.M.J. Dual-function microanalytical device by

in situ photolithographic grafting of porous polymer monolith: integrating solid-phase

extraction and enzymatic digestion for peptide mass mapping. Anal. Chem. 2003, 75,

5328–5335.

169. Drott, J.; Rosengren, L.; Lindstr€om, K.; Laurell, T. Pore morphology influence on

catalytic turn-over for enzyme activated porous silicon matrices. Thin Solid Films

1998, 330, 161–166.

170. Drott, J.; Rosengren, L.; Lindstr€om, K.; Laurell, T. Porous silicon carrier matrices in

micro enzyme reactors-influence of matrix depth.Mikrochim. Acta 1999, 131, 115–120.

171. Cooper, J.W.; Lee, C.S. Integrated and ultrasensitive gel protein identification. Anal.

Chem. 2004, 76, 2196–2202.

172. Calleri, E.; Temporini, C.; Perani, E.; Palma, A.D.; Lubda, D.; Mellerio, G.; Sala, A.;

Galliano, M.; Caccialanza, G.; Massolini, G. Trypsin-based monolithic bioreactor

REFERENCES 183

coupled on-linewith LC/MS/MS system for protein digestion and variant identification in

standard solutions and serum samples. J. Proteome Res. 2005, 4, 481.

173. Samskog, J.; Bylund, D.; Jacobsson, S.P.; Markides, K.E. Miniaturized on-line proteo-

lysis–capillary liquid chromatography–mass spectrometry for peptide mapping of lactate

dehydrogenase. J. Chromatogr. A 2003, 998, 83–91.

174. Hsieh,Y.L.F.;Wang,H.;Elicone,C.;Mark, J.;Martin,S.A.;Regnier,F.Automatedanalytical

system for the examination of protein primary structure. Anal. Chem. 1996, 68, 455–462.

175. Riggs, L.; Sioma,C.; Regnier, F.E. Automated signature peptide approach for proteomics.

J. Chromatogr. A 2001, 924, 359–368.

176. Bonneil, E.; Mercier, M.; Waldron, K.C. Reproducibility of a solid-phase trypsin

microreactor for peptide mapping by capillary electrophoresis. Anal. Chim. Acta

2000, 404, 29–45.

177. Bonneil, E.; Waldron, K.C. On-line system for peptide mapping by capillary electropho-

resis at sub-micromolar concentrations. Talanta 2000, 53, 687–699.

178. Schoenherr, R.M.; Ye, M.-L.; Vannatta, M.; Dovichi, N.J. CE-microreactor-CE-MS/MS

for protein analysis. Anal. Chem. 2007, 79, 2230–2238.

179. Bossi, A.; Guizzardi, L.; D’Acuto, M.R.; Righetti, P.G. Controlled enzyme-immobilisa-

tion on capillaries for microreactors for peptide mapping. Anal. Bioanal. Chem. 2004,

378, 1722–1728.

180. Wang, J.; Chatrathi, M.P.; Tian, B.; Polsky, R. Microfabricated electrophoresis chips for

simultaneous bioassays of glucose, uric acid, ascorbic acid, and acetaminophen. Anal.

Chem. 2000, 72, 2514–2518.

181. Wang, J.; Chatrathi, M.P.; Tian, B. Microseparation chips for performing multienzymatic

dehydrogenase/oxidase assays: simultaneous electrochemical measurement of ethanol

and glucose. Anal. Chem. 2001, 73, 1296–1300.

182. Wang, J.; Chatrathi, M.P.; Ib�anez, A.; Escarpa, A. Micromachined separation chips with

post-column enzymatic reactions of “class” enzymes and end-column electrochemical

detection: assays of amino acids. Electroanalysis 2002, 14, 400–404.

183. L’Hostis, E.; Michel, P.E.; Fiaccabrino, G.C.; Strike, D.J.; de Rooij, N.F.; Koudelka-Hep,

M.Microreactor and electrochemical detectors fabricated using Si and EPONSU-8. Sens.

Actuators B 2000, 64, 156–162.

184. Li, Y.-S.; Liu, W.-P.; Gao, X.-F.; Chen, D.-D.; Li, W.G. Immobilized enzymatic fluores-

cence capillary biosensor for determination of sulfated bile acid in urine. Biosens.

Bioelectron. 2008, 24, 538–544.

185. Wang, L.; Sipe, D.M.; Xu, Y.; Lin, Q. A MEMS thermal biosensor for metabolic

monitoring applications. J. Microelectromech. Syst. 2008, 17, 318–327.

186. De Boer, A.R.; Bruyneel, B.; Krabbe, J.G.; Lingeman, H.; Niessen, W.M.A.; Irth,

H. A microfluidic-based enzymatic assay for bioactivity screening combined with

capillary liquid chromatography and mass spectrometry. Lap Chip 2005, 5,

1286–1292.

187. Tang, Z.-M.; Wang, Z.-Y.; Kang, J.-W. Screening of acetylcholinesterase inhibitors in

natural extracts by CE with electrophoretically mediated microanalysis technique.


188. Tang, Z.-M.; Wang, T.-D.; Kang, J.-W. Immobilized capillary enzyme reactor based on

layer-by-layer assembling acetylcholinesterase for inhibitor screening by CE.



189. Besanger, T.R.; Hodgson, R.J.; Green, J.R.A.; Brennan, J.D. Immobilized enzyme reactor

chromatography: optimization of protein retention and enzyme activity in monolithic

silica stationary phases. Anal. Chim. Acta 2006, 564, 106–115.

190. Nicoli, R.; Bartolini, M.; Rudaz, S.; Andrisano, V.; Veuthey, J.-L. Development of

immobilized enzyme reactors based on human recombinant cytochrome P450 enzymes

for phase I drug metabolism studies. J. Chromatogr. A 2008, 1206, 2–10.

191. Kaneno, J.; Kohama, R.; Miyazaki, M.; Uehara, M.; Kanno, K.; Fujii, M.; et al.

Development of surface modification method and its application for preparation of

enzyme-immobilized microreactor. Kagaku Kogaku Ronbunshu 2004, 30, 154–158.

192. Kamal, A.; Sandbhor, M.; Ramana, K.V. One-pot lipase-catalyzed synthesis of enantio-

pure secondary alcohols from carbonyl compounds: a new protocol for lipase-mediated

resolution. Tetrahedron: Asymmetry 2002, 13, 815–820.

193. Seong, G.H.; Heo, J.; Crooks, R.M.Measurement of enzyme kinetics using a continuous-

flow microfluidic system. Anal. Chem. 2003, 75, 3161–3167.

194. Moore, B.D.; Stevenson, L.; Watt, A.; Flitsch, S.; Turner, N.J.; Cassidy, C.; et al. Rapid

and ultra-sensitive determination of enzyme activities using surface-enhanced resonance

Raman scattering. Nat. Biotechnol. 2004, 22, 1133–1138.

195. Lilly, M.D.; Hornby, W.E.; Crook, E.M. Kinetics of carboxymethylcellulose ficin in

packed beds. Biochem. J. 1966, 100, 718–723.

196. Jiang, H.; Zou, H.; Wang, H.; Ni, J.; Zhang, Q.; Zhang, Y. On-line characterization of the

activity and reaction kinetics of immobilized enzyme by high-performance frontal

analysis. J. Chromatogr. A 2000a, 903, 77–84.

197. Ristenpart, W.D.; Wan, J.; Stone, H.A. Enzymatic reactions in microfluidic devices:

Michaelis–Menten kinetics. Anal. Chem. 2008, 80, 3270–3276.

198. Koh, W.-G.; Pishko, M. Immobilization of multi-enzyme microreactors inside micro-

fluidic devices. Sens. Actuators. B 2005, 106, 335–342.

199. Nicoli, R.; Gaud, N.; Stella, C.; Rudaz, S.; Veuthey, J.-L. Trypsin immobilization on three

monolithic disks for on-line protein digestion. J. Pharm. Biomed. Anal. 2008, 48,

398–407.

200. Nicoli, R.; Rudaz, S.; Stella, C.; Veuthey, J.-L. Trypsin immobilization on an ethylene-

diamine-based monolithic minidisk for rapid on-line peptide mass fingerprinting studies.

J. Chromatogr. A 2009, 1216, 2695–2699.

REFERENCES 185

5MICROFLUIDIC DEVICES FORNANODRUG DELIVERY

CLEMENT KLEINSTREUER

Department of Mechanical and Aerospace Engineering and Department of Biomedical

Engineering, North Carolina State University, Raleigh, NC, USA

JIE LI

Department of Mechanical and Aerospace Engineering, North Carolina State University,

Raleigh, NC, USA

5.1 INTRODUCTION

Nanodrug delivery, employing microscale devices, is a broad and complex research

topicwith severalmajor application areas. For example, cost-effective drug discovery,

development, and testing are of great concern to the pharmaceutical industry, while

clinical diagnostics and drug delivery, ideally in combined form, are of interest to

healthcare providers. The associated microfluidic devices include lab-on-a-chip

(LOC) systems for drug discovery/development and bio-MEMS (biological/biomed-

icalmicroelectromechanical system) for controlled biological processing and optimal

(nano-) drug delivery. Powered by microfluidics, the use of LOC devices can be a

robust and fast method to discover, refine, and test a drug. This is important in light

of the fact that presently only one-tenth of the drug compounds that enter the clinical

trial phase succeed in becoming commercially available (see 03/31/07 Report at

BioMarket Research.com).

Bio-MEMSs are being used for controlled biological processes, such as cell sorting

and multinodal bioimaging/identification, as well as for targeted drug delivery. The


187

latter entails biochemical or mechanical methodologies, that is, either a passivemode

or different active delivery modes.1

1. Passive multifunctional nanoparticle systems (MFNPSs) include injected

porous (micrometer) particles carrying nanodrugs and releasing them near/at

the desired site.

2. Active nanodrug carriers (NDCs) of engineered size, shape, and surface

characteristics circulate in the bloodstream and may actively attach to diseased

cells/tissues.

3. Active (mechanical) drug delivery systems (DDSs) concentrate on 100%

targeting methodologies, nanofluid flow in microchannels, nanodrug mixing,

and microdevice optimization.

This chapter focuses on both biochemical and mechanical drug delivery systems

with an emphasis on experimental/computational simulation aspects of bio-MEMSs,

where some function as implanted microfluidic devices for controlled nanodrug

release. In any case, the overriding optimization objectives include biocompatibility,

controlled nanodrug release, performance accuracy and reliability, minimization of

side effects, size reduction, and cost-effectiveness.

Previous reviews concentrated on particular topics. For example, Suh et al.2 discus-

sed biological MFNPSs in nanotechnology, stressing nanotoxicity concerns and

nanodrug applications to neuroscience, while Emerich and Thanos3 outlined the

potential of nanomedicine enabling targeted delivery of diagnostic and therapeutic

agents, and Kim et al.4 provided a past-to-future overview on nanotechnology in drug

delivery. Riehemann et al.5 outlined recent developments and applications in nanome-

dicine. Parallel reviews on nanodrug (and gene) delivery systems are treatment specific

toward particular diseases or organs. For example, Kasuya and Kuroda6 summarized

nanomedicine for the human liver and Subramani7 considered nanodrug treatment

applications for cancer and diabetes, while Kleinstreuer et al.8 reviewed targeted

delivery of inhaled drug aerosols to predetermined sites to combat lung tumors or

evensystemicdiseases,outliningtheunderlyingmethodologyofasmart inhalersystem.

5.2 MICROFLUIDIC DEVICES

To appreciate the mechanics of microfluidic devices as well as ongoing modeling and

simulation aspects, this section starts out reviewing a few basic elements of micro-

fluidics andmicrosystems as well as their modeling assumptions. This brief discourse

is especially useful for readers interested in a state-of-the-art sample application given

in Section 5.4.

The main focus of microscale research and development is on device fabrication

and expansion of microsystem application areas, which implies innovative advances

in the material sciences, manufacturing technology, as well as supportive design

software creation. Electromechanical components of consumer goods, vehicles, and

machinery, as well as entire devices, especially medical implants and laboratory test

188 MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

equipment, are being built on a microscale. Examples include MEMS, microheat

sinks, iPods, and appliance control parts, aswell as sensors anddrug-release patches in

medicine, or lab-on-a-chip units and reactors in biomedical and chemical engineering.

Clearly, it is the low production cost, compactnesswith a very high surface-to-volume

ratio, rapid throughput with very small sample volumes, and integrated multifunc-

tionality, for example, nanodrug mixing or particle separation or stream positioning,

that make microscale fluid devices attractive alternatives to conventional flow

systems.9

5.2.1 Microfluidics and Microsystems

Microfluidics is the study of transport processes in microchannels. Of interest are

methods and devices for controlling and manipulating fluid flow, finite liquid volume

delivery, and particle transport on a nano- and microscale. Although microfluidics

deals with fluid behavior in systems with “small” length scales, conventional (i.e.,

macroscale) flow theory is typically applied, at least for liquid flows inmicrochannels

with Dhydraulic� 10 mm and standard gas flows when Dh� 100 mm. However, for

microchannelgasflows in the slip regime, that is, 0:001 � Kn ¼ l=L � 0:1 (where theKnudsen number is the ratio of the molecular mean free path over a system length

scale), modification to the velocity and temperature boundary conditions has to be

made. Clearly, when the Knudsen number is above 0.1, alternative system equations

and numerical solution techniques have to be considered.

Microfluidic devices (or microsystems, or bio-MEMS) typically consist of re-

servoirs, channels, actuators, pumps, valves, mixers, sensors, controllers, filters,

and/or heat exchangers. Associated with microfluidic devices are the following

R&D areas:

. Microfabrication of components or entire devices, using silicon, glass, polymer,

or steel

. Microfluidic transport phenomena, includingmechanicalmicropumps aswell as

nonmechanical surface effects

. Task-specific devices, such as micrototal analysis systems (mTAS), LOC, orDDSs

. Reliable detection and measuring systems

. Power systems and microdevice packaging

. Data communication, including telemetry for monitoring system performance

. Biocompatibility and adherence to regulations

Bio-MEMSs for drug delivery are of interest in this chapter, where we focus on

nanodrug transport phenomena in microchannels. Such devices offer a number of

advantages, such as controlled drug release, reliable accurate dosing, targeted

treatment, and automated feedback control, all resulting in small size and operational

convenience, efficacy, and cost-effectiveness. Basic background information, includ-

ing microscale device manufacturing methods, may be found in the books by

Tabeling,10Nguyen andWereley,11Saliterman,12 andTesar.13Reviewsof engineering

MICROFLUIDIC DEVICES 189

flows in small devices have been provided by Stone et al.,14 Hilt and Peppas,15

Whitesides,16 Hu and Zengerle,17 and Geipel et al.,18 to name a few.

5.2.2 Microsystem Modeling Assumptions

One of the key elements of all microsystems is the microchannel (soon becoming a

nanochannel) with hydraulic diameters, that is, circular-tube-equivalent diameters,

typically ranging from 10 to 500 mm. This is rather small in light of the fact that the

diameter of the human hair is about 80 mm. When considering fluid flow in such tiny

conduits, we should recall that the underlying macroscale modeling assumptions

are valid only when

1. the fluid is “infinitely divisible,” that is, the fluid forms a continuum, and hence

we can use the conservation laws in terms of the continuity, momentum, and

heat/mass transfer equations, often summarized for all practical purposes as the

Navier–Stokes equations;

2. all flowquantities are in local thermodynamic equilibrium, that is, novelocity or

temperature jumps at fluid–wall interfaces.

Concerning the continuum assumption (1), the two main classes of fluids, that is,

gases (in case of nanospray delivery) and liquids (primarily nanodrugs in aqueous

solutions) differ primarily by their densities and by the degrees of interaction between

the constituent molecules. Focusing on aqueous solutions, water density is typically

rliquid � 103 kg m�3 with an intermolecular distance lIM ¼ 0:3 nm. Now, if the key

macroscopic length scale, for example, microchannel effective diameter (or height or

width), is of the order of 10 mm or more, fluids with those characteristics appear

continuous and hence the Navier–Stokes equations hold. The local thermodynamic

equilibrium condition (2) implies that all macroscopic quantities within the fluid have

sufficient time to adjust to their surroundings. That process depends on the time

betweenmolecular collisions and hence themagnitude of themean free path traveled.

Clearly, a rarefied gas in a small microchannel does not form a continuum and hence

would exhibit velocity and temperature jumps at the channel walls, requiring more

exotic solution methods, for example, the lattice Boltzmann method (LBM), direct

simulation Monte Carlo (DSMC), or molecular dynamics simulation (MDS).

The conventional driving force for flow in microchannels is still the net pressure

force, using micropumps, when substantial flow rates, that is, Re ¼ vh=u > 1:0, aredesired, as for rapid nanodrug mixing and delivery. However, certain microfluidic

devices for biomedical, chemical, and pharmaceutical applications employ more

esoteric driving forces, such as surface tension (i.e., capillary or Marangoni effects)

and electrokinetic phenomena (i.e., electrophoresis or electroosmosis). In general, the

surface-to-volume ratio varies as the inverse of the system�s length scale, that is, 1/L,and hence microsystems with relatively large surface areas may cause significant

viscous resistance. In turn, it would require relatively powerful actuators, including

pumps, valves, and so on, to operate a microfluidic device. In order to have such

pumps/actuators/valves as integral parts of the microfluidic device, new principles


had to be employed. Thus, complementary to mechanical actuators with moving

parts, microscale phenomena were used when the inlet Reynolds number was low

(Re�O(1)), such as electrokinetic pumping (e.g., electroosmosis) and capillary

surface tension effects, electromagnetic force fields, and acoustic streaming.

Another contrasting macroscale versus submicrometer scale consideration is that

conventional fluid flow is described by velocity and pressure fields and by its

properties. Hence, they are characterized as interacting groups, such as kinematic

(i.e., velocity and strain rate), thermodynamic (i.e., pressure and temperature),

transport (i.e., viscosity, conductivity, and diffusivity), and miscellaneous parameters

(i.e., surface tension, vapor pressure, etc.). However, on the submicrometer scale,

matter, that is, solid, liquid, or gaseous, is more realistically described in terms of

interacting molecules. For example, molecules in a solid are densely packed and

arranged in a lattice, where each molecule is held in place by large repulsive forces

according to the Lennard-Jones (L-J) potential.19 Nevertheless, when solving pro-

blems of fluid flow in microchannels, the continuum mechanics assumption is

preferred over any molecular approach. For the latter approach, the state of each

molecule in terms of position and velocity has to be known, and then one has to

evolve/simulate that state forward in time for each molecule. That implies the

solution of Newton�s second law of motion with the L-J force (i.e., the spatial

derivative of the L-J potential) for billions of molecules. In contrast, when continuous

fluid flow behavior can be assumed, that is, system length scales Lgas> 100 mm (or

Kn� 0.1) andLliquid> 10 mm,we just numerically solve the conservation laws subject

to key assumptions and appropriate boundary conditions, as exemplified in

Section 5.4.1.

In summary, it is not surprising that fluid flow in microchannels may differ from

macrochannel flow behavior in terms of entrance, wall, and thermal flow effects.20

Specifically, because of the typically short microchannel length, entrance effects

(i.e., developing 2Dor 3Dflows)may be dominant. At themicrochannelwall, the “no-

slip” conditionsmaynot hold for hydrophobic liquids, electrokinetic forcesmay come

into play, and surface roughness effects may be substantial. Early onset of laminar-

to-turbulent flow transition may occur and viscous dissipation of heavy liquids in

high shear rate fields may increase the fluid temperature measurably.

5.2.3 Categories of Microfluidic Devices

The development and use of microfluidic devices, including bio-MEMS (see

Figure 5.1), are naturally being driven by application areas, that is, drug delivery

routes and targets for specific clinical treatment needs, and ultimately by business

interests. For successful treatment, rapid administration of the right dosage of

medication is important: too low a dosage may be ineffective and too high may be

harmful. Furthermore, dose frequency and duration, drug toxicity and interaction, as

well as allergies must all be considered on a patient- and disease-specific basis.

A smart drugdelivery system (SDDS) connects a patient, that is, the specificdisease

site, with an appropriate drug. An SDDS is a formulation (or device) with which

nanomedicine is introduced into the body, released at a controlled rate, and


subsensitively transported across cell membranes for therapeutic action with minimal

side effects.

The two most common delivery routes are oral, that is, drugs taken by mouth and

swallowed and hence a device is not needed, and via injection (i.e., parenteral

administration) directly into the bloodstream or affected area. More modern routes

include targeted drug aerosol inhalation for lung, sinus, and systemic diseases

employing smart inhaler systems, transdermal delivery via microneedles, and body

implanted microfluidic devices with controlled nanomedicine release. The oral route

is the typical way to deliver drugs into the body because it is cheap and most

convenient. However, it may not be very efficient because of drug absorption and/

or degradation before it reaches the bloodstream and ultimately the affected area or

organ. While injection is effective for relatively high quantities of large-molecule

drugs, it is also inconvenient, somewhat expensive, and not easy to control. Hence,

with the advent of nanodrugs and gene therapy, new delivery devices, including bio-

MEMS,had tobedeveloped.Keycomponents of such systems includemicrochannels,

micropumps (i.e., mechanical and electrokinetic), microvalves, microreservoirs, and

micromixers.21–25 Application-driven drug delivery devices can be categorized

into several groups—for example, microneedles providing active medicine infusion

FIGURE 5.1 Bio-MEMS components and flow chart.


through transdermal patches; drug-eluting stents maintaining patency of, say, coro-

nary arteries; smart inhalers, with which drug aerosol streams (nasal sprays) are

controlled, to improve targeted particle deposition; and self-contained external or

implantable bio-MEMS.

Possibly, the largest group directly benefitting from bio-MEMS is diabetic patients

ensuring exact glucose control via embeddedmonitors and insulin dispensers. Patients

with pacemakers or defibrillators may receive needed medication from an implanted

bio-MEMS during an arrhythmia event. Severe asthma patients typically requiring

several drugs, such as bronchodilators and anti-inflammatory medicine, may rely on

real-time disease analysis and subsequently controlled, targeted drug release. Pain

management can be handled by a programmable pump for, say, small-dose intraspinal

morphine administration toblockneurotransmitters fromreaching thebrain.Clearly, a

lot of R&Dwork and clinical testing have to be accomplished before most conceived

bio-MEMSgains public acceptance. The book edited byDesai and Bhatia26 discusses

bio-MEMS for drug delivery in several chapters, while the review by Elman et al.27

briefly summarizes next-generation bio-MEMS, which the authors classified as

passive or active delivery devices.

Microneedles,made out of silicon, polymer, steel, or metal oxides, are only a few

hundred micrometers long; that is, they generate microconduits past the outer skin

permeation barrier without encountering a nerve. In array formation, connected to a

liquid drug reservoir, they allow for dispersion and systemic uptake of macromolecu-

lar drugs, possibly replacing hypodermic injections, say, for vaccinations and insulin

delivery.28–30

Drug-eluting stents are slow-release nanodrug implants that mainly function as

scaffolds to keep arteries open after coronary angioplasty, reduce the likelihood of

restenosis, and reject foreign object. After initially very positive responseworldwide,

they have recently encountered mixed reviews because of postoperative complica-

tions, such as late stent thrombosis in some patients.31–35

Micropumps are vital for direct drug delivery when connected to a microreservoir,

or for nanodrug mixing in microchannels (see Section 5.4.1). Other micropump

applications include movement of nano- to microliter solutions in LOC and mTASdevices, molecular particle sorting with microfilters or via hydrodynamic focusing,

and flowmeasurements with microsensors. However, one should note that a majority

of hydrodynamic microscale and certainly most nanoscale processes are driven by

electrokinetic flow or surface-mediated transport.36 Thus, due to the very high

frictional resistance, pressure-driven flow in a nanofluidic device is inappropriate.

The reason is that Dp � L=D4h, where Dp is the pressure drop across the conduit of

length L and hydraulic diameterDh¼ 4A/P, with A being the cross-sectional area and

P the wetted perimeter. For example, Zahn37 reviewed the physics and fabrication

of mechanical and nonmechanical micropumps.

Hundreds ofmicroreservoirs can be embedded into a single silicon microchip that

is covered by a thinmetal or polymermembrane. Themicroreservoirsmay contain any

combination of drugs, chemicals, and/or biosensors, where the membrane seal can be

activated for controlled drug release, using preprogrammedmicroprocessors,wireless

telemetry, or biosensor feedback. Clearly, these microchips can store and release


nanodrugs in a controlled fashion, and they are advantageous because of their small

size, low power consumption, and absence of moving parts.

5.3 NANODRUG DELIVERY

It is apparent from the previous sections that drug delivery systems based on bio-

MEMSare just beginning to reach themarket. Self-regulated insulin delivery systems,

drug-eluting stents, and microneedle arrays with reservoirs on a chip are some of the

moremature examples. In this section (Figure 5.2), nanodrug carriers for biochemical

drug/gene delivery systems as well as associated clinical application areas and

mechanical nanodrug delivery methodologies are discussed.

5.3.1 Nanodrugs

Nanodrugs (or genetic material) embedded in nano/microspheres are promising

candidates for treatment of various diseases, such as cancer, infections, metabolic

and autoimmune diseases, and diseases related to the brain (http://nano.cancer.gov).

FIGURE 5.2 Strategies of targeted nanodrug delivery systems.


Such nanomedicine carriers can be microparticles made of soluble, insoluble, or

naturally biodegradable polymers, microcapsules, porous particles, cells, liposomes,

and so on. Emerich and Thanos3 provided an overview of typical nanoparticles used

in drug and gene delivery, while Kasuya and Kuroda6 summarized the desirable

properties and characteristics of nanomedicine carriers. Their modified and updated

lists are given below.

5.3.1.1 Solid NanoparticlesCeramic nanoparticles aremade from inorganic nonmetallic compounds with porous

characteristics such as oxides, that is, silica (SiO2), alumina (Al2O3), hydroxyapatite

(HA), and zirconia (ZrO2). They are stable in the typical range of temperatures and

pH encountered in the body and can be used to deliver proteins and genes. However,

their lack of biodegradation and slowdissolution raises safety questions. For example,

it was found that those made of silica can efficiently transport therapeutic genes to

the spleen and trigger a potent immune response capable of attacking tumors.38 The

results released in 2008 in Chemical & Engineering News (http://pubs.acs.org/

isubscribe/journals/cen/86/i35/html/8635scic.html#6) showed that iron oxide nano-

particles caused little DNAdamage andwere nontoxic, zinc oxide nanoparticles were

slightly worse, and titanium dioxide caused only DNA damage.

Carbon nanotubes (CNTs) are extremely small tubes that can be categorized as

single-walled nanotubes (SWNTs) and multiwalled nanotubes (MWNTs). These

compounds havebecome increasingly popular invarious fields simplybecause of their

small size and amazing optical, electric, and magnetic properties when used alone

or with other materials, such as drugs. Carbon nanotubes have potential therapeutic

applications in the field of drug delivery, diagnostics, and biosensing. For example,

SWNTs have been shown to shuttle various cargos across cellular membrane without

cytotoxicity. SWNTs can be used as a platform for investigating surface–protein and

protein–protein binding. These nanotubes can act as highly specific electronic sensors

for detecting clinically important biomolecules such as antibodies associated with

human autoimmunediseases. Functionalized carbonnanotubes can also act as vaccine

delivery systems. The basic concept is to link the antigen to carbon nanotubes while

retaining its conformation, thereby inducing antibody response with the right speci-

ficity. Overall, the future use of carbon nanotubes in drug delivery systems may

enhance detection sensitivity in medical imaging, improve therapeutic effectiveness,

and decrease side effects.

Nanocrystals are aggregates of molecules that can be combined in a crystalline

formof the drug surroundedby a thin surfactant coating.Highdosages can be achieved

andpoorly soluble drugs canbe formulated for improvedbioavailability.Both oral and

parenteral delivery systems are possible and the limited carrier in the formulation

reduces potential toxicity. Limitations include poor drug stability.

Polymers such as albumin, chitosan, and heparin occur naturally and have been

a material of choice for the delivery of oligonucleotides, DNA, protein, and drugs.

The drug is physically entrapped in the polymer capsule. The characteristics can be

summarized as follows: (i) water soluble, nontoxic, and biodegradable; (ii) surface

modification (pegylation); (iii) selective accumulation and retention in tumor tissue;

NANODRUG DELIVERY 195

and (iv) specific targeting of cancer cells while sparing receptor-mediated targeting

of normal cells with a ligand. Polymer nanostructured fibers, core–shell fibers, hollow

fibers, and nanorods and nanotubes provide a platform for a broad range of applica-

tions. For example, biological objects, including drugs, of different complexities

carrying specific functions can be incorporated into such nanostructured polymer

systems. Biosensors, tissue engineering, drug delivery, and enzymatic catalysis are

just a few applications. Another example is superparamagnetic particles, known to

display strong interactions with external magnetic fields leading to large saturation

magnetization. By using periodically varying magnetic fields, the nanoparticles can

be heated to provide a trigger for drug release.

Solid lipid nanoparticles are lipid-based submicron colloidal carriers. They have a

solid hydrophobic core surrounded by a monolayer of phospholipid. The system is

stabilized by the inclusion of fairly high levels of surfactants. They are less toxic than

polymer nanoparticles and can be used to deliver drugs orally, topically, or via the

pulmonary route. While stability is a concern, it is better than that observed with

liposomes.

5.3.1.2 Colloidal Soft MatterDendrimers are artificial, polymer-basedmolecules formed frommonomers such that

each layer of branching units doubles or triples the number of peripheral groups (i.e.,

they look like a foam ball). The void area within a dendrimer, its ease of modification/

preparation, and size control offer great potential for targeted gene and drug delivery.

Improvements in cytotoxicity profiles, biocompatibility, and biodistribution are

needed. Dendrimers are repeatedly branched molecules. They are emerging as a

rather new class of polymeric nanosystems with applications in drug delivery. The

properties of dendrimers are dominated by the functional groups on the molecular

surface.Dendritic encapsulation of functionalmolecules allows for the isolation of the

active site, a structure that mimics the structure of active sites in biomaterials because

dendritic scaffolds separate internal and external functions. For example, a dendrimer

can bewater soluble when its end group is a hydrophilic group, like a carboxyl group.

It is theoretically possible to design a water-soluble dendrimer with internal hydro-

phobicity, which would allow it to carry a hydrophobic drug in its interior. Another

property is that the volume of a dendrimer increases when it has a positive charge.

If this property can be applied, dendrimers can be used for drug delivery systems that

can give medication to the affected part inside a patient�s body directly.

Hydrogels (also called aquagels) are a network of polymer chains that are water

insoluble, and sometimes found as a colloidal gel in which water is the dispersion

medium. Hydrogels are superabsorbent (they can contain over 99% water) natural

or synthetic polymers. Hydrogels also possess a degree of flexibility very similar to

natural tissue due to their significant water content. Hydrogels are responsive to

specific molecules, such as glucose or antigens, that can be used as biosensors as well

as in drug delivery system.

Liposomes were tiny bubbles (vesicles) made out of the same material as a cell

membrane. Liposomes are small spherical systems that are synthesized from


cholesterol and nontoxic phospholipids. Liposomes can be filled with drugs and used

to deliver drugs for cancer and other diseases. Membranes are usually made of

phospholipids, which are molecules having a head group and a tail group. The head is

attracted towater and the tail,which ismadeof a longhydrocarbonchain, is repelledby

water. Because they are natural materials, liposomes are considered attractive,

harmless drug delivery carriers that can circulate in the bloodstream for a long

time.Another interesting property of liposomes is their natural ability to target cancer.

The endothelial wall of all healthy humanblood vessels is encapsulated by endothelial

cells that are bound together by tight junctions. These tight junctions stop any large

particle in the blood from leaking out of the vessel. Tumor vessels do not contain the

same level of seal between cells and are diagnostically leaky. This ability is known as

the enhanced permeability and retention effect. Liposomes of certain sizes, typically

less than 400 nm, can rapidly enter tumor sites from the blood but are kept in the

bloodstream by the endothelial wall in healthy tissue vasculature. Anticancer drugs

such as doxorubicin (Doxil), camptothecin, and daunorubicin (DaunoXome) are

currently being marketed in liposome delivery systems. Despite a relatively long

history of investigation, liposomes exhibit limited stability and have not made

significant medical impact.

Micelles provide considerable advantages among drug carrier systems for their

solubilization to contribute the increasing bioavailability of poorly soluble drugs and

the characteristics to stay in the blood long enough to afford a gradual accumulation in

aparticular area.Amicelle is an aggregate of surfactantmolecules dispersed in a liquid

colloid. A typical micelle in aqueous solution forms an aggregatewith the hydrophilic

“head” regions in contact with surrounding solvent, sequestering the hydrophobic tail

regions in the micelle center. This type of micelle is known as a normal phase micelle

(oil inwater micelle). Inversemicelles have the head groups at the center with the tails

extending out (water in oil micelle). Micelles are approximately spherical in shape.

Other phases including shapes such as ellipsoids, cylinders, and bilayers are also

possible. Micelle formation is essential for the absorption of fat-soluble vitamins and

complicated lipids within the human body. When membrane phospholipids are

disrupted, they can reassemble themselves into tiny spheres, smaller than a normal

cell, either as bilayers or as monolayers. The bilayer structures are liposomes and the

monolayer structures are called micelles.

Microemulsions have great thermodynamic stability, which allows for self-emul-

sification at a wide range of temperatures and affords easy preparation. The structural

variability of microemulsions together with its composition and the pH of the

environment can obviously influence the drug release rate. Other positive character-

istics include the low viscosity of the majority of the system.

Organogels are systems resembling the structure of microemulsions but are

semisolid. Most organogels utilized in pharmaceuticals are lecithin, gelatin, or

sorbitan ester-based systems in biocompatible solvents. Complex aqueous phases

such as vesicle suspensions entrapping drugs can be eventually incorporated into the

organogel systems. They can entrap hydrophilic or hydrophobic drugs and antigens;

it is possible to achieve controlled release systems.


5.3.1.3 Desirable Characteristics of Nanodrug CarriersWith this large family of nanodrugs available (see Sections 5.3.1.1 and 5.3.1.2), it is

important to consider suitable (or even optimal) attributes of nanodrugs and their

carriers. Specifically, the following properties should be optimized for the in vivo use

of conventional nanoparticle carriers for drug or gene delivery.

1. Acceptability to Versatile Payloads: The nanoparticle carrier (NPC) should

allow payloads of various therapeuticmaterials. This facilitates the development

of a general purpose carrier and concurrent administration of different drugs.

2. Low or No Toxicity: The NPC should not be made of or contain potentially

dangerous materials.

3. Active Targeting: The NPC should recognize and attach target cells and tissues

by a sensormolecule displayed on its surface. NPC of about 100 nm are likely to

accumulate spontaneously in tumors after systemic injection, but this passive

targetingmechanism based on the enhanced permeability and retention effect39

is excluded from these criteria.

4. Appropriate Size: The size of NPC should be about 40–150 nm in diameter. Too

small nanoparticles (<40 nm) and too large nanoparticles (>150 nm) are

nonspecifically removed from blood circulation by the function of kidney

and reticuloendothelial system in the liver, respectively.40 NPC of about

40–150 nm could be used to target both tumors utilizing the enhanced perme-

ability and retention effect and hepatocytes by passing through the fenestrae in

liver endothelial cell.41

5. Appropriate Surface Charge: The surface charge of NPC is known to affect

severely the stability and biodistribution of systemically administrated nano-

particle carriers. For ideal delivery, the surface of nanoparticle carriers should

be optimized so as not to be entrapped by unexpected tissues. For example, one

positively charged nanoparticle (i.e., polyplex of polyethyleneimine and DNA)

was efficiently accumulated in the lung after systemic administration.42

6. Efficient Cell-Penetrating Activity: The nanoparticle carriers should possess

cell-penetrating activity for active and rapid intrusion across the plasma mem-

brane or the endosomal membrane of target cells and tissues because many

therapeutic materials (particularly, genes and siRNAs) function intracellularly.

7. Mechanism of Intracellular Targeting: The nanoparticle carriers in target cells

should bring and release payloads to the intracellular destination precisely (e.g.,

genes and siRNAs should be released in the nucleus and cytoplasm, respec-

tively, not in endosomes).

5.3.2 Nanodrug Targeting

Asmentioned, targeting is the ability to direct in controlled fashion the drug particles,

or a drug-loaded system, to the predetermined site of interest. There are mechanical

and biochemical methodologies of drug targeting (Figures 5.2 and 5.3a and b). As

Kaparissides et al.43 mentioned, in biochemical targeting two approaches can be


distinguished, that is, passive and active targeting. Theygive as an example for passive

targeting the observed preferential accumulation of chemotherapeutic agents in

“solid” tumors as a result of the enhanced vascular permeability of tumor tissues

compared to healthy tissue (Figure 5.3a). As a variation to nanodrugs cruising in the

bloodstream and hopefully reaching the right site, drug carriers with surface func-

tionalities, for example, ligands interacting with tumor cell receptors, can seek out,

bind to, andpenetrate target cells (Figure 5.3b). In contrast,mechanical drug targeting

is the delivery of a controlled fluid–particle stream from an optimal release (or arterial

injection) point to a predetermined deposition site.44,45 Specifically, Kleinstreuer45

FIGURE 5.3 (a) Passive drug targeting by nanodrug carriers in blood vessels and (b) active

nanodrug targeting with conjugated antibodies.


discussed several mechanical drug delivery applications within the framework of

fluid–structure interaction simulations. For example, targeted drug aerosol delivery

can be potentially accomplished with a smart inhaler system.46 Another application is

the targeting of liver tumors with radioactive microspheres,47 where the particles are

released, via a microcatheter, from an optimal inlet position of the hepatic artery.

Most of the drug targeting literature deals with biochemical targeting, especially

“colloidal soft matter” (see Section 5.3.1.2) as reviewed by Bonacucina et al.48 Such

“soft matter,” for example, microemulsions and organogels (as passive drug carriers)

as well as liposomes, micelles, and dendrimers (as active drug carriers), can increase

drug solubility and bioavailability and can attach/penetrate tumor cells, especially

wheremicelles are used. Clearly, all present biochemical targeting efforts are disease/

treatment specific.1,49–54

Drug targetingviamechanical deliverydevices include reservoirs, pumps, inhalers,

catheters, needles, drug-eluting stents, and implants releasing drugs (Section 5.2.3).

The next two sections provide some physical insight into the microfluidics of bio-

MEMSs, focusing on nanodrug mixing, microchannel flow, and device optimization.

5.4 Bio-MEMS APPLICATIONS

To illustrate some aspects of bio-MEMS research and development, nanodrugmixing

and microchannel designs are discussed in the next two sections.

5.4.1 Nanofluid Flow Simulations

Microfluidics deals with methods and devices for manipulating and controlling

fluid–particle flow in microchannels.55 A recent application area is nanomedicine

with the goal of controlled nanodrug delivery to specified target areas.12,56 A key

aspect of this goal is the development of integrated drug delivery systems to monitor

and control target cell responses to pharmaceutical stimuli, to understand biological

cell activities, or to facilitate drug development processes. An important part of such

drug delivery systems, belonging to the family of bio-MEMS, is active or passive

micromixers25,57 to assure near-uniform nanodrug concentrations. Static micromix-

ers, not requiring any external energy source, rely on chaotic advection and/or

enhanced diffusion, typically to mix two fluids.58,66 For example, Hardt et al.61

reviewed recent developments in micromixing technology, focusing on liquid mixing

with passive micromixers. Four kinds of mixers that employed different hydrody-

namic principles are discussed: hydrodynamic focusing, flow separation, chaotic

advection, and split and recombine flows. Diffusive mixing can be improved by

increasing the interfacial contact area between the different fluids and reducing the

diffusion length scale. Thus, selecting the right type of micromixer for a specific

application is very important.

Li and Kleinstreuer67 analyzed rapid nanoparticle mixing in a carrier fluid,

employing low-cost micromixers and heat transfer to achieve two system design

goals, that is, uniform exit particle concentration and minimum required channel

length. Specifically, a microfluidics device for controlled nanofluid flow in


microchannels (Figure 5.4) is investigated for basic nanomedicine applications.

Presently planned for laboratory-scale testing, uniform,predetermined concentrations

of a stimulus (e.g., cocaine particles) should be delivered via multiple microchannels

to an array of wells containing brain cells to measure cell responses (e.g., dopamine

production levels). Their study focuses on device miniaturization in light of the

ultimate goal of bio-MEMS implantation into the diseased brain region of, say,

Parkinson�s patients. Most important, the impact of two types of static micromixers

(Figure 5.4c) is analyzed to achieve uniform nanoparticle concentrations at the exit of

a representative microchannel of minimum length.

Figure 5.5 shows Lmin(Pe) for the different scenarios. Clearly, any micromixer

module reduces Lmin significantly for all P�eclet numbers.While an increase in slotted

baffle plates reduces Lmin, the simple three-sided injection unit performs best.

FIGURE 5.4 Microfluidics system: (a) laboratory-scale nanodrug supply device, (b) rep-

resentative microchannels, and (c) static mixer inserts.

Bio-MEMS APPLICATIONS 201

Alternative to the P�eclet number, which is based on the average velocity of

nanofluid plus carrier fluid, the Reynolds number ratio of nanofluid to carrier fluid

is a suitable operational parameter. The associated Reynolds numbers are

Rei ¼ ðuDhÞini

ð5:1Þ

with i¼ 1 indicating the carrier fluid in the solution channel and i¼ 2 denoting the

nanofluid channel (Figure 5.4b). For the given system, Figure 5.6 indicates that the

main microchannel length can be below 4mm when employing an injection micro-

mixer, that is, a 70% reduction in channel length.

The addition of nanoparticles and certainly the installment of micromixers

increases the pressure drop in both channels and hence the pumping power require-

ments. Pumpingpower is defined as the product of the pressure drop across the channel

(Dp) and the volumetric flow rate (Q), that is,

P ¼ Dp �Q ð5:2ÞThe pressure differences occur between the nanofluid inlet or carrier fluid inlet and

the system outlet, that is, requiredminimal length. The volumetric flow rate is the sum

of the volumetric flow rates of both nanofluid and carrier fluid.

Figure 5.7a and b depicts the relationship of pressure drop and pumping power for

the two cases. Clearly, the addedmicromixer increases the local pressure drop, but the

decreased system length may reduce any negative effect caused by the micromixer.

As shown in Figure 5.8a and b, the power requirement even decreases in some cases,

FIGURE 5.5 Micromixer influence on minimal uniformity length/system dimension.67


that is, when employing the injection micromixer. The employment of baffle-slit

micromixers slightly increases the pressure drop; however, when the pumping power/

volumetric flow rate gets larger and larger, the negative effect appears to be less and

less. For example, for the two- or four-baffle micromixer, the pressure drop is even

smaller than that without anymicromixer when the nanofluid supply rate is increased

to 8mm s�1. In summary, employing an appropriatemicromixer decreases the system

dimension and the associated power requirement.

A heat flux was used to ensure that mixture delivery to the living cells occurs at a

required temperature of 37 �C. The change of fluid properties and nanoparticle

diffusivity, caused by the added heat flux, also benefits system miniaturization. As

shown in Figure 5.6, the added heat flux greatly decreases the system dimension; that

is, an average 35% reduction is observed.

5.4.2 Device Optimization

It is obvious that the better an engineering device performs, the lower the (irreversible)

losses, that is, the closer it operates at an isentropic efficiency.This directly implies that_Sgen should be minimized as part of any device/process design or improvement.

For example, considering simple heat transfer from an ambient reservoir at T0 ¼ ¢,

the entropy balance equation can be written as

_Sgen ¼ @S

@t�X _Q

T0�X

_mSjin þX

_mSjout > 0 ð5:3Þ

FIGURE 5.6 Minimal uniformity length versus Reynolds number ratio.67


FIGURE 5.7 Pressure drop versus pumping power: (a) carrier fluid inlet and (b) nanofluid

inlet.67


Effectively, _Sgen � _W loss; that is, in general, power loss is due to system, heat

transfer, and fluid flow irreversibilities:

Ploss � _W loss ¼ T0 _Sgen ¼ T0@S

@t�

_Q

T0�X

_mSjin�X

_mSjout� �

ð5:4Þ

Focusing on friction (or viscous effects) as the main cause of irreversibilities and

hence entropy generation, the rate of irreversible conversion from flow energy to heat

can be expressed as

tij@vi@xj

� mF ¼ Ploss

8 ¼ T0_Sgen8 ð5:5Þ

Clearly, to minimize entropy production in pressure-driven microchannels, we

have to reduce the viscous dissipation function F, that is, achieve minimization of

_Sgen8 ¼ m

TF ¼ m

T

@u

@yþ @v

@x

� �2

þ 2@u

@x

� �2

þ 2@v

@y

� �2" #

ð5:6Þ

In general, and exclusively for fully developed flow, the term ðm=TÞð@u=@yÞ2 ismost important. If wall slip is significant, then u(y) velocity profile is greatly affected

and hence the channel pressure drop.

FIGURE 5.8 Heat flux influence on minimal uniformity length.67


5.4.2.1 Liquid Flow in a MicrochannelMinimizationof entropygeneration as a design tool to determinebest devicegeometry

and operation, especially for heat exchangers, has been established for macroscale

configurations.68–75 However, fluid flow inmicrochannels exhibits dominant features

oftennonexistingor less influential inmacrochannels, for example,wall slip velocities

for somegases, entrance effects because of the short conduit length, significant surface

roughness in relation to microchannel height (or hydraulic diameter), and so on.11

Thus, application of entropy generation minimization principles may assist in the

optimal design of microchannel heat sinks and bio-MEMS in light of geometric and

operational conditions.67,76–79 Classical methods for enhanced heat transfer, for

example, an increase of heat transfer area and/or inlet Reynolds number, are limited

options for microchannel flow. Thus, the use of nanofluids as coolants, for example,

CuO or Al2O3 nanospheres with diameters in the range of 5 nm< dp< 150 nm in

water, oil, or ethylene glycol, is a third option. In case of nanomedicine delivery with

bio-MEMS, nanofluid flow analysis is important andmeasurable reduction of entropy

generation is desirable.

In this section, entropy generation is minimized for steady laminar pure water and

nanofluid flows in a representative trapezoidal microchannel in terms ofmost suitable

channel aspect ratio and Reynolds number range.

One effective operational parameter is the inlet Reynolds number; Figure 5.9

indicates a desirable range of 425�Re� 1100 for all fluids and aspect ratios

considered, ignoring “slit flow” for AR¼ 0.9337. Due to slightly enhanced frictional

FIGURE 5.9 System entropy generation versus Reynolds number.80


effects (due to the increase of viscosity), Snanofluid

G;total > Swater

G;total. The overall flow

field entropy is generated for different scenarios in terms of the following integral

form:

SG;total ¼ 1

_mcp

ðððV

Sgen

8 dV ð5:7Þ

An important geometric design parameter is the aspect ratio. Figure 5.10 shows

SG;totalðARÞ for three fluids and different inlet Reynolds numbers. Specifically, for

Uin¼ 4m s�1 (implying Re¼ 425, 437, and 466 for water, 1% CuO nanofluids, and

4% CuO nanofluids, respectively), the larger aspect ratio generates smaller SG;totalvalues, while for Uin¼ 10m s�1 (implying Re¼ 1063, 1092, and 1165), the lower

aspect ratio generates smaller SG;total values. Clearly, the 1%CuO–water pairingyields

more favorable results than the nanofluid with 4% CuO particles.

Figure 5.10 also reveals that there are significant trend changes of SG;total(AR, Re)

for all three fluids occurring at critical values, that is, AR� 0.55 and Re� 700. The

main reasons are that with elevated Reynolds numbers, temperature gradients are

reduced and the frictional effects become dominant, even more pronounced as

AR ! 1.0. The additional test run for pure water at Re¼ 638 confirms that

Re� 700 is critical, while Figure 5.9 has these trend changes in SG;total embedded

as well.

FIGURE 5.10 System entropy generation as a function of aspect ratio and Reynolds

number.80


5.5 CONCLUSIONS AND FUTURE PERSPECTIVE

The development of nanodrugs and their carriers and the design ofmicroscale devices

for optimal drug targeting are presently very important research topics. Particularly,

lab-on-a-chip systems for drug discovery, development and testing, active nanodrug

carriers that attach to tumor cells, and mechanical drug delivery systems (also called

bio-MEMSs) have attractedmuch attention and enjoy increasing use in the healthcare

industry.

This chapter summarizes some fundamentals and applications of microfluidics

and bio-MEMS. Then, it discusses types of nanodrugs and characteristics of their

carriers, aswell as biochemical andmechanical nanodrug targeting. To illustrate some

aspects of bio-MEMS components and applications, computational results for nano-

drug mixing and nanofluid flow in microchannels and optimal microchannel design/

operation are presented.

Future work will concentrate on the improvements in smart drug delivery systems

on themarket, ranging fromnanodrug carriers tomicrofluidic devices for optimal drug

targeting. One of the more challenging tasks will be to combine biochemically active

NDC with bio-MEMSs. They could deliver up to 100% of the NDCs to the pre-

determined, disease-related targetwhere theNDCs attach to the tumor cells, penetrate

them,and release thenanodrugs.Theunderlyingprincipleofparticle targeting isbased

on the backtracking methodology45 that can be applied to smart delivery of inhaled

drug aerosols and injected radioactive microspheres, NDCs, and so on.

REFERENCES

1. Jain, K.K. Drug Delivery Systems. Methods in Molecular Biology, Humana Press: 2008;

Vol. 437.

2. Suh, W.H.; Suslick, K.S.; Stucky, G.D.; Suh, Y.-H. Nanotechnology, nanotoxicology,

neuroscience. Prog. Neurobiol. 2009, 87, 133–170.

3. Emerich, D.F.; Thanos, C.G. Targeted nanoparticle-based drug delivery and diagnosis.

J. Drug Target. 2007, 15(3), 163–183.

4. Kim, S.; Kwon, I.I.K.; Kwon, I.C.; Park, K. Nanotechnology in drug delivery: past,

present, and future. In Nanotechnology in Drug Delivery; de Villiers, M.M.; Aramwit, P.;

Kwon, G.S., Eds.; Springer/AAPS Press: 2009.

5. Riehemann, K.; Schneider, S.W.; Luger, T.A.; Godin, B.; Ferrari, M.; Fuchs, H.

Nanomedicine: challenge and perspectives. Angew. Chem. Int. Ed. 2009, 48, 872–897.

6. Kasuya, T.; Kuroda, S. Nanoparticles for human liver-specific drug and gene delivery

systems: in vitro and in vivo advances. Expert Opin. Drug Deliv. 2009, 6(1), 39–52.

7. Subramani, K. Applications of nanotechnology in drug delivery systems for the treatment

of cancer and diabetes. Int. J. Nanotechnol. 2006, 3(4), 557–580.

8. Kleinstreuer, C.; Zhang, Z.; Donohue, J.F. Targeted drug-aerosol delivery in the human

respiratory system. Annu. Rev. Biomed. Eng. 2008, 10, 195–220.

9. Bayraktar, T.; Pidugu, S.B. Characterization of liquid flows in microfluidic systems. Int. J.

Heat Mass Transfer 2006, 49, 815–824.


10. Tabeling, P. Introduction to Microfluidics, Oxford University Press: Oxford, UK, 2005.

11. Nguyen, N.-T.;Wereley, S.T. Fundamentals, Applications of Microfluidics, Artech House:

Boston, MA, 2006.

12. Saliterman, S.S. Fundamentals of bioMEMS and Medical Microdevices, Wiley–

Interscience/SPIE Press: Bellingham, WA, 2006.

13. Tesar, V. Pressure-Driven Microfluidics, Artech House: Norwood, MA, 2007.

14. Stone, H.A.; Stroock, A.D.; Ajdari, A. Engineering flows in small devices: microfluidics

toward a lab-on-a-chip. Annu. Rev. Fluid Mech. 2004, 36, 381–411.

15. Hilt, J.Z.; Peppas, N.A. Microfabricated drug delivery devices. Int. J. Pharm. 2005, 306,

15–23.

16. Whitesides, G.M. The origins and the future of microfluidics: insight overview. Nature

2006, 442, 368–373.

17. Hu, M.; Zengerle, R. Microfabricated devices for controlled biochemical release. In 10th

International Conference on New Actuators, Bremen, Germany, 2006, pp 263–271.

18. Geipel, A.; Goldschmidtboeing, F.; Jantscheff, P.; Esser, N.; Massing, U.;Woias, P. Design

of an implantable active microport system for patient specific drug release. Biomed.

Microdevices 2008, 10, 469–478.

19. Bird, G.A. Molecular Gas Dynamics and Direct Simulation of Gas Flow, Oxford

University Press: Oxford, UK, 1994.

20. Koo, J.; Kleinstreuer, C. Liquid flow in microchannels: experimental observations and

computational analysis of microfluidics effect. J. Micromech. Microeng. 2003, 13,

568–579.

21. Dutta, D.; Ramachandran, A.; Leighton, D.T. Jr Effect of channel geometry on solute

dispersion in pressure-driven microfluidic systems. Microfluid. Nanofluid 2006, 2,

275–290.

22. Zhang, C.; Xing, D.; Li, Y. Micropumps, microvalves, and micromixers within PCR

microfluidic chips: advances and trends. Biotechnol. Adv. 2007, 25, 483–514.

23. Oh, K.W.; Ahn, C.H. A review of microvalves. J. Micromech. Microeng. 2006, 16,

R13–R19.

24. Tsai, N.-C.; Sue, C.-Y. Review of MEMS-based drug delivery and dosing systems. Sens.

Actuators 2007, 134, 555–564.

25. Nguyen, N.-T. Micromixers: Fundamentals, Design and Fabrication, William Andrew:

Norwich, NY, 2008.

26. Desai, T.; Bhatia, S. Therapeutic Micro/Nano Technology, BioMEMS and Biomedical

Nanotechnology; Springer: 2006.

27. Elman,N.M.; Patta,Y.; Scott, A.W.;Masi, B.;Duc,H.L.H.; Cima,M.J. The next generation

of drug delivery microdevices. Clin. Pharmacol. Ther. 2009, 85, 544–547.

28. Kim, K.; Lee, J.-B. MEMS for drug delivery. In Bio-MEMS: Technologies and

Applications; Wang, W.; Soper, S.A. Eds.; CRC Press/Taylor & Francis: Boca Raton,

FL, 2007.

29. Roxhed, N.; Samel, B.; Nordquist, L.; Griss, P.; Stemme, G. Painless drug delivery through

microneedle-based transdermal patches featuring active infusion. IEEE Trans. Biomed.

Eng. 2008, 55(3), 1063–1071.

30. Wu, Y.; Qiu, Y.; Zhang, S.; Qin, G.; Gao, Y. Microneedle-based drug delivery: studies on

delivery parameters and biocompatibility. Biomed. Microdevices 2008, 10, 601–610.

REFERENCES 209

31. Nakazawa, G.; Finn, A.V.; Kolodgie, F.D.; Virmani, R. A review of current devices and a

look at new technology: drug-eluting stents. Expert Rev. Med. Devices 2009, 6(1), 33–42.

32. Melikian, N.; Wijns, W. Drug-eluting stents: a critique. Heart 2008, 94, 145–152.

33. Wykrzykowska, J.J.; Onuma, Y.; Serruys, P.W. Advances in stent drug delivery: the future

is in bioabsorbable stents. Expert Opin. Drug Deliv. 2009, 6(2), 113–126.

34. Kukreja, N.; Onuma, Y.; Daemen, J.; Serruys, P.W. The future of drug-eluting stents.

Pharmacol. Res. 2008, 57, 171–180.

35. Jamshidi, P.; Mahmoody, K.; Erne, P. Covered stents: a review. Int. J. Cardiol. 2008, 130,

310–318.

36. Karniadakis, G.; Beskok, A.; Aluru, N. Microflows and Nanoflows: Fundamentals and

Simulation, Springer: 2005.

37. Zahn, J.D. Micropump applications in bio-MEMS. In Bio-MEMS: Technologies and

Applications; Wang, W.; Soper, S.A.Eds.; CRC Press/Taylor & Francis: Boca Raton,

FL, 2007.

38. Tan, K.; Cheang, O.; Ho, I.A.; Lam, P.Y.; Hui, K.M. Nanosized bioceramic particles could

function as efficient gene delivery vehicleswith target specificity for the spleen.Gene Ther.

2007, 14, 828–835.

39. Maeda, H.; Wu, J.; Sawa, T.; Matsumura, Y.; Hori, K. Tumor vascular permeability and

the EPR effect in macromolecular therapeutics: a review. J. Control. Release 2000, 65,

271–284.

40. Drummond, C.D.; Meyer, O.; Hong, K.; Kirpotin, D.B.; Papahadjopoulos, D. Optimizing

liposomes for delivery of chemotherapeutic agents to solid tumors. Pharmacol. Rev. 1999,

51(4), 691–744.

41. Snoeys, J.; Lievens, J.; Jacobs, F.; Duimel, H.; Collen, D.; Frederik, P.; Geest, B.D. Species

differences in transgeneDNAuptake in hepatocytes after adenoviral transfer correlatewith

the size of endothelial fenestrae. Gene Ther. 2007, 14, 604–612.

42. Bragonzi, A.; Diana, G.; Villa, A.; Galori, G.; Biffi, A.; Bordignon, C.; Assael, B.M.;

Conese, M. Biodistribution and transgene expression with nonviral cationic vector/DNA

complexes in the lungs. Gene Ther. 2000, 7(20), 1753–1760.

43. Kaparissides, C.; Alexandridou, S.; Kotti, K.; Chaitidou, S. Recent advances in novel drug

delivery systems. J. Nanotechnol. 2006, 2, 1–11.

44. Kleinstreuer, C. Biofluid Dynamics: Principles and Selected Applications, Taylor &

Francis: Boca Raton, FL, 2006.

45. Kleinstreuer, C. Optimal drug delivery based on fluid-particle dynamic simulations.

ANSYS Advantage Vol. 3 (4), 2009.

46. Kleinstreuer, C.; Zhang, Z.; Li, Z.; Roberts, W.L.; Rojas, C. A new methodology for

targeting drug-aerosols in the human respiratory system. Int. J. Heat Mass Transfer 2008,

51, 5578–5589.

47. Kennedy, A.; Kleinstreuer, C.; Basciano, C.A.; Dezarn, A. Computer Modeling of 90Y

Microsphere Transport in the Hepatic Arterial Tree to Improve Clinical Outcomes. Int. J.

Radiol. Oncol. Biol. Phys. (Red Journal), 2010, 76(2), 631–637.

48. Bonacucina, G.; Cespi, M.; Misici-Falzi, M.; Palmieri, G. Review: colloidal soft matter as

drug delivery system. J. Pharm. Sci. 2009, 98(1), 1–42.

49. Tomioka, H.; Tatano, Y.; Yasumota, K.; Shimizu, T. Recent advances in anti-tuberculosis

drug development and novel drug targets. Expert Rev. Respir. Med. 2008, 2(4), 455–471.


50. Muthu,M.S.; Singh, S. Targeted nanomedicines: effective treatment modalities for cancer,

AIDS and brain disorders. Nanomedicine 2009, 4(1), 105–118.

51. Lammers, T.; Hennink, W.E.; Storm, G. Tumor-targeted nanomedicines: principles and

practice. Br. J. Cancer 2008, 99, 392–397.

52. Tosi, G.; Bostantino, L.; Ruozi, B.; Forni, F.; Vandelli, M.A. Polymeric nanoparticles for

the drug delivery to the central nervous system. Expert Opin. Drug Deliv. 2008, 5(2),

155–174.

53. Polyak, B.; Friedman, G. Magnetic targeting for site-specific drug delivery: applications

and clinical potential. Expert Opin. Drug Deliv. 2009, 6(1), 53–70.

54. Bohmer, M.R.; Klibanov, A.L.; Tiemann, K.; Hall, C.S.; Gruell, H.; Steinbach, O.C.

Ultrasound triggered image-guided drug delivery. Eur. J. Radiol. 2009, 70(2), 242–253.

55. Kleinstreuer, C.; Li, J.; Koo, J. Microfluidics of nanodrug delivery. Int. J. Heat Mass

Transfer 2008, 51, 5590–5597.

56. Labhasetwar, V.; Leslie-Pelecky, D.L. Biomedical Applications of Nanotechnology,

Wiley–Interscience: Hoboken, NJ, 2007.

57. Nguyen,N.-T.;Wu, Z.Micromixers: a review. J.Micromech.Microeng. 2005,15, R1–R16.

58. Stroock, A.D.; et al. Chaotic mixer for microchannels. Science 2002, 295, 647–651.

59. Kim, D.S. A barrier embedded chaotic micromixer. J. Micromech. Microeng. 2004, 14,

798–805.

60. Munson, M.S.; Yager, P. Simple quantitative optical method for monitoring the extent of

mixing applied to a novel microfluidic mixer. Anal. Chim. Acta 2004, 507, 63–71.

61. Hardt, S.; Drese, K.S.; Hessel, V.; Schonfeld, F. Passivemicromixers for applications in the

microreactor and uTAS fields. Microfluid. Nanofluid. 2005, 1, 108–118.

62. Floyd-Smith, T.M.; Golden, J.P.; Howell, P.B.; Ligler, F.S. Characterization of passive

microfluidic mixers using soft lithography. Microfluid. Nanofluid. 2006, 2, 180–183.

63. Chang, C.-C.; Yang, R.-J. Electrokinetic mixing in microfluidic systems. Microfluid.

Nanofluid. 2007, 3, 501–525.

64. Chung, C.K.; Shih, T.R. Effect of geometry on fluid mixing of rhombic micromixers.

Microfluid. Nanofluid. 2008, 4, 419–425.

65. Kang, T.G.; Singh, M.K.; Kwon, T.H.; Anderson, P.D. Chaotic mixing using periodic and

aperiodic sequences of mixing protocols in a micromixer.Microfluid. Nanofluid. 2008, 4,

589–599.

66. Brotherton, C.M.; Sun, A.C.; Davis, R.H. Computational modeling and comparison of

three co-laminar microfluidic mixing techniques. Microfluid. Nanofluid. 2008, 5, 43–53.

67. Li, J.; Kleinstreuer, C. Microfluidics analysis of nanoparticle mixing in a microchannel

system. Microfluid. Nanofluid. 2009, 6, 661–668.

68. Bejan, A. Entropy Generation Minimization: The Method of Thermodynamic

Optimization of Finite-Size System and Finite-Time Processes, CRC Press: Boca

Raton, FL, 1996.

69. Bejan, A. Fundamentals of exergy analysis, entropy generation minimization, and genera-

tion of flow architecture. Int. J. Energy Res. 2002, 26, 545–565.

70. Sahin, A.Z. A second law comparison for optimum shape of duct subjected to constant wall

temperature and laminar flow. Heat Mass Transfer 1998, 33, 425–430.

71. Sahin, A.Z. Entropy generation in a turbulent liquid flow through a smooth duct subjected

to constant wall temperature. Int. J. Heat Mass Transfer 2000, 43, 1469–1478.

REFERENCES 211

72. Mahmud, S.; Fraser, R.A. The second law analysis in fundamental convective heat transfer

problems. Int. J. Thermal Sci. 2003, 42, 177–186.

73. Mansour, R.B.; Galanis, N.; Nguyen, C.T. Dissipation and entropy generation in fully

developed forced and mixed laminar convection. Int. J. Thermal Sci. 2006, 45, 998–1007.

74. Khan, W.A.; Culham, J.R.; Yovanovich, M.M. Optimal design of tube banks in crossflow

using entropy generation minimization method. J. Thermophys. Heat Transfer 2007,

21(2), 372–378.

75. Ko, T.H.; Wu, C.P. A numerical study on entropy generation induced by turbulent forced

convection in curved rectangular ductswith various aspect ratios. Int. Commun.HeatMass

Transfer 2009, 36, 25–31.

76. Chein, R.; Chuang, J. Analysis of microchannel heat sink performance using nanofluids.

Appl. Thermal Eng. 2005, 25, 3104–3114.

77. Heris, S.Z.; Etemad, S.G.; Esfahany, M.N. Experimental investigation of oxide nanofluids

laminar flow convective heat transfer. Int. Commun. Heat Mass Transfer 2006, 33,

529–535.

78. Jang, S.P.; Choi, S.U.S. Cooling performance of a microchannel heat sink with nanofluids.

Appl. Thermal Eng. 2006, 26, 2457–2463.

79. Kleinstreuer, C.; Li, J. Microscale Cooling Devices, Encyclopedia of Micro and

Nanofluidics, Li, D.; Ed.; Springer-Verlag: Heidelberg, 2008.

80. Li, J.; Kleinstreuer, C. Entropy generation analysis for nanofluid flow in microchannels.

ASME J. Heat Transfer, 2010, in press.


6MICROCHIP AND CAPILLARYELECTROPHORESIS USINGNANOPARTICLES

MUHAMMAD J. A. SHIDDIKY AND YOON-BO SHIMDepartment of Chemistry and Institute of Biophysio Sensor

Technology, Pusan National University, Busan, South Korea

6.1 INTRODUCTION

Nanoparticles (NPs), generally defined as materials with a particle size of less than

100 nm in at least one dimension, are of interest because their chemical and physical

behavior is unprecedented and remarkably different from those in bulk form. They

have great potential for applications in electronic, chemical, ormechanical industries,

aswell as in technologies, including superconductors, catalysts, drug carriers, sensors,

magnetic materials, pigments, separation science, and in structural and electronic

materials. In separation science alone, significant advances have been made in

electrophoresis and microchip separations employing nanoparticles.1,2 The small

size of nanoparticles is responsible for their novel, unrevealed properties (electrical,

magnetic, chemical, optical, and mechanical).3 In addition, some nanoparticles

possess advantages such as a large surface area, good chemical and thermal stability,

significantmechanical strength, ease ofmodification, andbiomolecular compatibility.

Owing to these unique properties, nanoparticles have attracted immense attention in

separation science.

Chemical separations are inevitable for analysis of complex samples, in particular

biological samples. In this vein, capillary electrophoresis (CE) and microchip

electrophoresis (MCE) are, among all, two promising separation techniques.4–6


213

Since 1989, when nanoparticles were first applied as a pseudostationary phase in CE

by Wallingford and Ewing,1 new separation media containing a large variety of

nanoparticles with a wide range of chemically useful characteristics were developed

for CE and MCE analyses. A pseudostationary phase with a large surface area, in

combination with an electroosmotic flow (EOF)-driven system, has great potential in

a highly efficient separation technique. For example, commercially available silica

nanoparticles have been used as separation buffer additives in CE by Fujimoto and

Muranaka.7 In the past decade, extensive research on nanomaterial applications in CE

and MCE has evolved and different types of nanoparticles have been tested, such as

gold nanoparticles (AuNPs), silica nanoparticles, magnetic nanoparticles, polymer

nanoparticles, metal oxide nanoparticles, and carbon nanotubes. Until recently,

nanoparticle-mediated CE and MCE have been studied in the analysis of proteins,

DNA separation and sequencing, separation of drugs and drug delivery analysis in

biomedical sciences, separation of environmental pollutants (organic and inorganic),

and many other small molecule analyses. Much of the earlier works regarding use of

nanoparticles in CE andMCE separation has been summarized in several reviews.8–13

This chapter gives an overview of the new developments and innovative applications

of nanoparticles, including metal and metal oxide nanoparticles, carbon nanotubes,

silica nanoparticles, and polymeric nanoparticles as stationary and/or pseudosta-

tionary phases in CE and MCE. The use of nanoparticles in the pseudostationary

phases of CE and MCE is discussed in detail.

6.2 MICROCHIP ELECTROPHORESIS

6.2.1 Microfluidic Devices

Since the conception of the microfluidic bioanalytical device over a decade ago by

Widmer’s group,14 it has become a very efficient platform for the analysis of

biologically, clinically, and environmentally important analytes. The microfluidic

device frequently referred to either as lab-on-a-chip or micrototal analysis systems

(m-TAS), typically consists of microchannel networks, miniaturized chambers/reac-

tors, microseparation/detection units, and combinations thereof to accommodate

the sensitive and large-scale analysis of analytes. Among many other components

in a microfluidic device, valves and pumps, mixers, injectors, microreactors, pre-

concentrators and microseparators, filters, detectors, and temperature measurement

units are the major components.15

6.2.2 Advantages and Applications of Microfluidic Devices

Themicrofluidic devices have numerous advantages over traditional systems based on

robotics and conventional analysis instrumentation. First, mature microfabrication

techniques adapted from the semiconductor industry allowmass production of lab-on-

a-chipdevices, thereby reducingcost per device andallowingmassiveparallel systems

to be constructed easily. Other advantages include high speed, high efficiency, high

214 MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

throughput, low sample/reagent consumption, low waste generation, portability, and

disposability.16 Possible applications include clinical instrumentation for point-of-

care testing, real-timemonitoring for biodefense,methods for glucosemonitoring and

the subsequent release of insulin, forensic applications including DNA analysis at

crime scenes, and environmental applications such as on-site testing of explosives or

phenolic pollutants in groundwater.16–20

6.2.3 Separation Techniques in Microfluidic Devices

Separation in a microfluidic device is based on a variety of principles, from simple

microchip zone electrophoresis (MZE) to more complex methods, such as microchip

gel electrophoresis (MGE), micellar electrokinetic chromatography (MEKC), iso-

tachophoresis, microchip electrochromatography, isoelectric focusing, and micro-

emulsion electrokinetic chromatography.

6.2.3.1 Microchip Zone ElectrophoresisZone electrophoresis is the simplest form of microchip electrophoresis. In MZE, the

species to be separated are dissolved in buffer only, so that the ions can freely move

along the solution by diffusion and/or under the influence of an electric field. Every

charged molecule can be characterized by its electrophoretic mobility (mep), which is

largely determined by its size radius (r) and charge (q):

mep ¼ q=6prh ð6:1Þ

whereh is the viscosity of the buffer system.When there is no electroosmotic flow, the

final velocity (u) of a charged molecule is derived as

u ¼ uep ¼ mep �E ð6:2Þ

whereE is the electric field strength. However,most often, electroosmotic flow occurs

as well, and the final velocity is found as the vector sum of the electrophoretic velocity

(uep) and the electroosmotic velocity (ueo). Because the electroosmotic flow, and

hence the velocity associated with it, is a bulk property, it uniformly affects all

molecules. In addition, the absolutemagnitude of the electroosmotic velocity is larger

than anyof the electrophoretic velocities, and therefore all species, positively charged,

uncharged, or negatively charged, move in the same direction, namely, the direction

defined by the electroosmotic flow (Figure 6.1).

A sample containing both cations and anions can be injected at one end of

a capillary or microchannel and then, after separation has occurred, the individual

bands of ions can be detected at the other end. This makes the necessary experimental

set up forMZEmuch simpler. Separation in such systems is governed by the charge to

size ratio of the molecules. Therefore, the small, highly charged cations are first to

migrate through the channel and arrive at the detector, followed by larger, less charged

cations, then all unchargedmolecules, followedby larger, less chargedanions,with the

highly charged small anions arriving last at the detector (Figure 6.1). Due to the high

MICROCHIP ELECTROPHORESIS 215

separation power of MZE, even small differences in the charge to mass ratio can be

sufficient to achieve separation. In MZE, due to the absence of the other interaction

equilibrium, and ignoring contributions from the injection plug width and detector

geometry, the only source of dispersion is diffusion, the effects of which can be

minimized by decreasing the time from injection to detection.A large number ofMZE

separations have been performed mainly due to the relative ease of implementation

and good separation efficiency.22

6.2.3.2 Separation of DNA by Microchip Gel ElectrophoresisThe mobility of DNA depends on its charge to mass ratio. In a free solution, in tris-

acetate or tris-borate buffer (pH8.4), the difference of its charge tomass ratio is almost

FIGURE 6.1 (a) An open-ended microchannel extends between two reservoirs, across which

a high voltage is applied. This voltage causes analytes to migrate from the site of sample

application at the cathode buffer through a detector to the anode. The EOF that results fromwall

pumping drives the separation of analytes. Reproduced from Ref. 21, with permission. (b)

Vector addition of electrophoretic mobilities of individual ions and the electroosmotic mobility

(meo) to yield total mobility (mtot).


identical for fragments from approximately 400 to 48,500 bp. DNA is a biopolymer

consisting ofmany repeating units called nucleotides, and the addition and subtraction

of such units change only the charge and absolutemass or size but not the ratio of these

two parameters. Because of this property, the separation of DNA fragments from

approximately 400 to 48,500 bp is not possible in a free solution.23 Thus, to separate

DNAmolecules, channels are typically filled with a polymer gel that acts as a sieving

matrix. This sieving matrix includes noncross-linked polymers, such as linear

polyacrylamide, polyethylene glycol, and cellulose derivatives, as well as cross-

linked polymers or gels, such as polyacrylamide and agarose. The entangled polymer

network inside the microchannel serves as a molecular sieve in which smaller DNA

and protein molecules migrate faster than larger ones.

Electrophoresis ofDNA in a gelmatrix can be thought of as a type of “gel filtration”

where a mixture of DNA molecules of different sizes is forced to move through the

pores of the matrix under the influence of an electric field. Any DNA fragment to be

fractionated encounters the gel network of polymer threads or pores that increases the

effective friction and consequently lowers the velocity of the molecules. Small

molecules move more rapidly through the gel, while large molecules move relatively

slow. The existence of the gel medium significantly contributes to the observed

electrophoretic mobility of the DNAmolecules. The pore size in the gel matrix plays

a critical role in determining the relative electrophoretic mobility and separation

efficiency of the DNA fragments.

Unlike MZE, electroosmotic flow is mostly undesired in MGE, particularly for

separation of negatively charged DNA fragments. In bare silica channels, a large

magnitudeof theEOFdrivesDNAmolecules toward the cathodeand thereforedistorts

the separation of the DNA. In some cases, electroosmotic forces within the channel

may cause the gel to migrate out of the channel. In addition, analyte–wall interactions

due to ionic interactions and hydrogen bonding significantly interfere with the

separation as well. Because of these issues, it is desirable to suppress EOF within

the channel for separation of DNA by MGE. EOF can be controlled by applying

a proper coating, either covalent or dynamic, at the inner surface of the channel.24 The

coatings can increase, decrease, reverse, or eliminate the EOF depending on the

presence or absence of certain functional groups contained within the coating

materials. Coated microchannels also prevent sample interaction with the channel

and feature good separation efficiencies and excellent run-to-run reproducibility in

migration times. Most DNA separations by MGE are conducted in coated channels

where the EOF is completely eliminated or significantly reduced to achieve high-

resolution separations.

6.2.3.3 Micellar Electrokinetic ChromatographyMEKC is particularly useful for separating small and neutral molecules, which has

been impossible by gel or zone electrophoresis.25 The separation mechanism is based

on partitioning of the analyte between themicelle and the surrounding aqueous phase.

MEKC can be performed by dissolving an ionic surfactant in the CE running solution

at a concentration higher than the critical micelle concentration (cmc), with no

MICROCHIP ELECTROPHORESIS 217

instrumental modification. In general, neutral or alkaline buffer solutions are used to

create conditions for a strong EOF that moves the entire liquid stream in the capillary

toward the cathode (Figure 6.2). Therefore, even anionic micelles such as sodium

dodecyl sulfate (SDS) migrate toward the cathode. The neutral analyte is not

solubilized by or is free from the micelle and thus migrates at the same velocity as

that of the EOF; the analyte that is totally incorporated into themicelle migrates at the

samevelocity as that of themicelle. Other neutral analytes are detected between t0 and

tmc, migration time of the EOF marker and the micelle, respectively. The interval

between t0 and tmc is called the migration time window. The wider the window, the

larger the peak capacity, which is the number of peaks that can be separated during

a run.Migration time can bemeasured by usingmarkers such asmethanol forEOFand

Sudan III for the micelle. The retention factor k can be defined as

k ¼ nmc=naq ð6:3Þ

in which nmc and naq are the number of moles of the analyte in the micelle and

surrounding aqueous phase, respectively; k can be measured by

k ¼ ðtR�t0Þ=ft0ð1�tR=tmcÞg ð6:4Þ

inwhich tR is themigration time of the analyte.26 The difference between this equation

and the conventional one used in chromatography is the limited migration time

window in MEKC. Although the micelle is not fixed inside the capillary, it plays the

same role as the stationary phase in chromatography and is therefore called the

pseudostationary phase. The MEKC resolution Rs, equation is

Rs ¼ pN=4½ða�1Þ=a� � ½k2=ð1þ k2Þ� � ð½ð1�t0=tmc�=½1þðt0=tmcÞk1�Þ ð6:5Þ

FIGURE 6.2 Separation principle of MEKC. Reproduced from Ref. 25, with permission.


in which N is the plate number and a the selectivity factor equal to k2/k1.27

Equation 6.5 is similar to the one used in conventional chromatography, except for

the addition of the last term on the right-hand side. This variable comes from the

migration of the micelle or pseudostationary phase inside the capillary, thus, the

migrationof thepseudostationaryphase causes reductionof the column length.28 If the

micellemigration is completely suppressed or tmc is infinity, the resolution equation is

the same as the conventional one.

The third termon the right-hand side of equation (6.5) is the retention factor term. It

is not independent of the other variable because it includes k. The optimum kopt value

to maximize resolution is easily determined by25

kopt ¼ p ðtmc=t0Þ ð6:6Þ

which differentiates the product of the last two terms in equation (6.5). Under neutral

or acidic conditions, tmc/t0 is 3–4 and kopt is 1.7–2.0.25 To adjust k in MEKC, the

concentration of the surfactant can be increased or decreased since k can be expressed

as

k ¼ KVmc=Vaq ¼ KvðCsf � cmcÞ ð6:7Þ

in which K is the distribution coefficient of the analyte between the micelle and

the aqueous phase, Vmc and Vaq are the volume of the micelle and the aqueous

phase, respectively, v is the partial specific volume of the micelle, and Csf is the

surfactant concentration.27 As shown by equation (6.7), k is linearly proportional

to the surfactant concentration, an advantage of MEKC because the Csf needed to

obtain a given k can be calculated provided the cmc and k at a certain Csf are

known.

MEKC rarely separates extremely hydrophobic analytes with high k-values.

However, several strategies are possible. Adding an organic solvent significantly

reduces k and gives better resolution of extremely hydrophobic analytes. The organic

aqueous solution has a higher viscosity, and the migration times will be long. Adding

too much organic solvent may destroy the micellar structure and/or decrease the

migration timewindowbecause the electrophoreticmobility of themicelle is reduced,

a likely result of the reduced charge on the micelle or the increased size of the micelle

due to swelling caused by the organic solvent.25

6.3 APPLICATION OF NANOPARTICLES IN CE AND MCE

6.3.1 Why Nanoparticles

The emergence of nanotechnology is opening up new horizons for the application of

nanoparticles’ physics, chemistry, biology, medicine, materials science, and interdis-

ciplinary fields. In particular, nanoparticles are of considerable interest in separation

sciences due to their attractive physical and chemical properties. The unique

APPLICATION OF NANOPARTICLES IN CE AND MCE 219

properties of nanoparticles offer excellent prospects for the development of separation

media because of the following reasons:

1. Nanoparticles exhibit higher ratios of surface area to volume than their bulk

counterparts for organofunctional groups that can interact with the channel/

capillary surface, target analytes, or both leading to enhanced separation

selectivity and efficiency.

2. Nanoparticles offer additional interaction sites where the solute can interact

with the running buffer additives and therefore lead to a higher selectivity for

target molecules.

3. Some nanoparticles, particularly metal nanoparticles, can be conjugated with

polymeric materials (e.g., poly(diallyldimethylammonium chloride)) and also

act as an excellent pseudostationary phase that offers excellent separation

performances with advanced functional properties for constructing separation

media for DNA and proteins.

4. The use of the nanoparticles in detection is compatible with the miniaturization

and integration of the analytical instruments, alongwith the inherent simplicity,

speed, high selectivity, and excellent catalytic activity.

The application of nanoparticles inCE andMCEhas been studiedmainly due to the

enhanced separation performances (selectivity and efficiency) owing to these impor-

tant features.

6.3.2 Nanoparticle-Mediated Capillary Electrophoresis

Using nanoparticles as run buffer additives is similar to using micelle additives in

micellar electrokinetic chromatography.25–27,29 In both cases, the purpose of using the

additive is to provide additional interaction sites with which the solutes can interact.

Both MEKC and nanoparticle-mediated electrophoresis (NME) offer some definite

advantages. (i) The presence of nanoparticles andmicelles in the run buffer avoids the

need to pack the capillary with a stationary phase. (ii) Since there is no conventional

stationary phase, the need for frits and other retaining techniques is eliminated. (iii)

Micelles and nanoparticles in the run buffer move through the capillary with an

apparent mobility that takes into account the effects of both the electroosmotic flow

and their own native electrophoretic mobility. As a result, there is a constant turnover

in the interacting media. The solutes are moving under the influence of the electric

field and are separated on the basis of their different effective charges and by

differential partitioning between the aqueous buffer and the nanoparticle or micellar

pseudostationary phase. This additional partitioning effect increases the separation

degrees of freedom and allows separation not only of the charged solutes but also of

the neutral ones. Thus, the user can control the nature of interactions with additives

and tailor the CE system to specific analytes. Although the use of nanoparticles and

micelles in CE and MCE has many similarities, there are also some important

differences between these two additives. One main difference between MEKC and


NME is that in the former approach the solutes can penetrate the core of the micelles,

whereas in NME, the interactions with the solutes occur always at the outer surface of

the nanoparticles. These interactions can be either with the surface itself or with the

organic moieties attached to the surface of the nanoparticle. In addition, NME extends

the operable range to much higher fields. The operable field in CE is limited to a large

extent by thermal effects induced by the applied voltage and the generated current.

Therefore, it is advantageous, whenever possible, to operate CE at a low ionic strength.

In the case of MEKC, it is necessary to use high surfactant concentrations that exceed

the critical micelle concentration. Stabilized nanoparticles do not suffer from this

drawback and stable colloidal solutions are attained even without dissolved species.

6.3.2.1 Gold NanoparticlesThe presence of gold nanoparticles in the run buffer in CE has the ability to

significantly affect the apparent mobility (map) of a certain solute and change the

electroosmotic mobility (meo) of the run buffer. Ovadia et al.29 used citrate-stabilized

AuNPs in conjunction with channels treated with poly(diallyldimethylammonium

chloride) (PDDC) tomanipulate the selectivity between solutes inCE.Modification of

the channel wall with only PDDC covers the silanol groups with positive quaternary

ammonium groups that adsorb the negatively charged gold nanoparticles. The

adsorption of PDDC on the capillary and the subsequent introduction of the cit-

rate-stabilized gold nanoparticles all change the apparent mobilities. Both the

apparent selectivities and the observed selectivities, which take into account the

effect of the electroosmotic flow, are affected. The extent of the change depends on

the charge of the solutes and their functional groups. With the negatively charged

solutes, the introduction of PDDC decreases the absolute value of the apparent

mobilities. The introduction of the citrate-stabilized gold nanoparticles to the run

buffer resulted in a further decrease in the apparent mobility of the solutes. Increasing

theconcentrationof thenanoparticles in thebuffer caused littlechange to themapvalues.

The electrophoretic selectivity (aef) is defined here as the ratio of the electropho-

retic mobilities (mef) of two neighboring solutes in an electropherogram:

aef ¼ mef;2 =mef;1 ð6:8ÞSimilarly, the apparent selectivity (aap) is the ratio of the two neighboring apparent

mobilities:

aap ¼ map;2 =map;1 ð6:9ÞIn the presence of the electroosmotic flow (meo), the observed selectivity (aobs) can

be substantially different from the electrophoretic or apparent selectivity. The EOF

vector can either enhance or diminish the electrophoretic or apparent selectivity. The

observed selectivity is described by

aobs ¼ tm;2=tm;1 ¼ mobs;2=mobs;1 ¼ map;2=map;1 þmeo=meo ð6:10Þ

where tm represents migration times and indexes 1 and 2 denote the solutes. For

the positively charged solutes, aobs decreases slightly with increasing meo. For the


negatively charged solutes, since |meo|< |map|,1 the selectivities increasewith increas-

ingmeo. The presence of the PDDC-gold coatingmodifies the electroosmoticmobility

and the observed mobility of the solutes. These changes in the mobilities are

manifested in selective alterations and allow obtaining separations that cannot be

achieved without nanoparticles.

AuNPs have been employed as pseudostationary phases in AuNP-coated capillar-

ies for CE separation of acidic and basic proteins at low pH (3), achieving high plate

numbers and high run-to-run reproducibility.31 The AuNPs were first protected by

a bilayer of didodecyl dimethylammonium bromide (DDAB). The inner wall of the

capillary was then coated with these nanoparticles to prevent protein adsorption.

Another effect of the adsorbed, positively charged nanoparticle coating was a reversal

in EOF. In addition, the DDAB-capped gold nanoparticles were noncovalently

coated with polyethylene oxide (PEO) and dispersed in the electrolyte and used as

a pseudostationary phase to enhance interactions with proteins. PEO was attached

through the hydrophobic regions with the hydrophilic groups able to interact with

proteins. At neutral pH, plate numbers of 23,000 were achieved for bovine serum

albumin, which rose to 62,000 at pH 3.5. For lysozymes, respective plate numbers of

81,000 and 450,000were achieved. Introducing PEO also raised the theoretical plates

to over 1000,000 for lysozymes.

Li and coworkers reported the synthesis of bifunctional Au-Fe3O4 nanoparticles

that possess high catalytic activity for separation and detection of proteins.32 Au-

Fe3O4 nanoparticles combined the merits of both gold and Fe3O4 nanoparticles and

were formed by chemical bond linkage. Owing to the introduction of AuNPs, the

bifunctional Au-Fe3O4 nanoparticles can be easily modified with other functional

molecules to realize various nanobiotechnological separations and detections. For

example,Au-Fe3O4 nanoparticles can bemodifiedwith nitrilotriacetic acidmolecules

through Au–S interaction and used to separate proteins simply with the assistance of

a magnet. Bradford protein assay and sodium dodecyl sulfate–polyacrylamide gel

electrophoresis were performed to examine the validity of the separation procedure,

and the phosphate determination method suggested that the as-separated protein

maintained catalytic activity. This result showed the efficiency of such a material in

protein separation and suggested that its use could be extended tomagnetic separation

of other biosubstances. The author concluded that the developed synthetic strategy

could be useful for facile preparation of diverse bifunctional and evenmultifunctional

nanomaterials.

A method for enrichment and separation of acidic and basic proteins using

centrifugal ultrafiltration, followed by nanoparticle-filled capillary electrophoresis,

has been described by Tseng and coworkers.33 To improve stacking and separation

efficiencies of proteins, a separation buffer containing 1.6% PDDAwas added to the

AuNPs, PEO, cetyltrimethylammonium bromide (CTAB), and poly(vinyl alcohol).

The use of AuNP as additives exhibited better efficiency in separation, stacking,

and analysis time. The separation efficiencies of acidic and basic proteins remained

greater than 10,000 and 100,000 platesm�1, respectively. To further enhance detec-

tion sensitivity, protein samples were enriched using centrifugal ultrafiltration,

followed by our proposed stacking method. The detection sensitivity was improved


up to 314-fold compared to normal hydrodynamic injection. In addition, the limits of

detection for most proteins were down to a nanomolar range. The proposed method

was also successfully applied to the analysis of egg-white proteins. More recently,

Tween 20-capped AuNPs were used as selective probes for extraction and preconcen-

tration of aminothiols from an aqueous solution (Figure 6.3).34 Tween 20 molecules

were noncovalently attached to the surface of AuNPs to form Tween 20–AuNPs.

Thesemodified-AuNPswere then used for selective extraction of aminothiols through

FIGURE 6.3 Online concentration and separation of urine samples by PDDA-filled CE (a)

before and (b and c) after extraction of 5.0mL urine using 100� Tween 20–AuNPs (200mL).

Urine samples (b) without and (c) with spiking of 5mM GSH, Cys, and HCys were hydrody-

namically injected by raising the capillary inlet 20 cm high for 180 s. Analytes attached to the

surface of AuNPs are released upon the addition of 500mM DTT (20mL). A 50 cm capillary

(20 cm to detector) is filled with 1.6% v/v PDDAC solution, which is prepared in 20mM

phosphate solution at pH 2.0. The applied voltage is �8.5 kV whereas the electric current is

80mA. The detection wavelength is set at 200 nm. Reproduced from Ref. 34, with permission.


the formation of Au–S bonds. After extraction and centrifugation, aminothiols were

detached from the surface of AuNPs by addingDTT in high concentration. This probe

was then used in combination with CE and UV absorption detection. On-line

concentration and separation of the released aminothiols were performed using

1.6% v/v PDDA as an additive in CE. Under optimal extraction and stacking

conditions, the detection limit for glutathione (GSH), cysteine (Cys), and homocys-

teine (HCys) were 28, 554, and 456 nM, respectively. In comparison to the normal

injection without the extraction procedure, approximately 2280-, 998-, and 904-fold

improvement in sensitivity was observed for GSH, Cys, and HCys, respectively. It is

believed that this approach had a significant potential to be extended to clinical

diagnosis.

Spherical AuNPs with sizes ranging from 2.0 to 3.0 nmwere used for open-tubular

CE for separation of thiourea, naphthalene, biphenyl, and four polycyclic aromatic

compounds with high efficiency.35 The capillary was first etched with ammonium

hydrogen difluoride, followed by prederivatization with (3-mercaptopropyl)-tri-

methoxysilane and immobilization of the dodecanethiol-capped AuNPs. An etching

process was used to increase the surface area of the capillary inner wall, by a factor

of up to 1000. The methodology in the article presented high separation efficiencies

(up to 200,000 platesm�1), although the etching process is time consuming and

complicated.

For the development of a low-viscosity polymer solutionwith high sieving and self-

coating abilities in DNA sequencing by CE, Wang and coworkers prepared AuNPs

with particle sizes of approximately 20, 40, and 60 nm and added them to a quasi-

interpenetrating network (quasi-IPN) composed of linear polyacrylamide (LPA) with

different viscosity-averagemolecular masses of 1.5, 3.3, and 6.5MDa, and poly-N,N-

dimethylacrylamide (PDMA) to form polymer/metal composite matrices, respective-

ly.36 These matrices improve ssDNA sequencing performances due to interactions

betweenAuNPs andpolymer chains and the formationof physical cross-linkingpoints

as demonstrated by intrinsic viscosities and glass transition temperatures. Addition of

AuNPs to quasi-IPN containing 1.5MDa LPA could significantly improve sieving

ability for both small and large DNA fragments. The sieving ability of quasi-IPN/

AuNPs with lower molecular mass LPA and AuNPs approximated those of quasi-IPN

with higher molecular mass LPAwithout AuNPs. Thus, the use of quasi-IPN/AuNPs

with lower molecular mass LPA could avoid problems in relation to LPAwith higher

molecularmass, such as difficult preparation andveryhighviscosity, and thus promote

full automation. As a result of the interactions of AuNPs with polymer chains,

more dilute solutions with lower viscosities possessed much improved sieving

performances in terms of resolution and migration time than relatively concentrated

quasi-IPN, without AuNPs, containing the same LPA.

Understanding nanoparticle behavior in the presence of an electric field will have

a significant impact on separation science where nanoparticles can serve to improve

either the mobility or detection sensitivity of target molecules. In this effort, Haes and

coworkers exploited nanoparticle behavior in the presence of an electric field and

investigated their impacts on separation science where nanoparticles are employed

as pseudostationary phase comprising only 2% of the total capillary volume.37 The


optical properties of covalently functionalized gold nanoparticles have been used to

investigate both the stability of nanoparticles and the mobility of dopamine, epineph-

rine, and pyrocatechol in capillary electrophoresis, with either primarily covalently

functionalized carboxylic acid (Au-COOH) or amine (Au-NH2) surface groups being

characterized using extinction spectroscopy, transmission electron microscopy, and

zeta potential measurements and utilized in separating neurotransmitters. The authors

anticipated at least three nanoparticle-specific mechanisms, those effecting the

separations. First, the degree of nanoparticle–nanoparticle interactions is quantified

using a new parameter termed the critical nanoparticle concentration (CNC). CNC is

defined as the lowest concentration of nanoparticles that induces predominant

nanoparticle aggregation under specific buffer conditions and is determined using

dual-wavelength photodiode array detection. Once the CNC has been exceeded,

reproducible separations are no longer observed. Second, nanoparticle–analyte

interactions are dictated by electrostatic interactions that depend on the pKa of the

analyte and surface charge of the nanoparticle. Finally, nanoparticle–capillary inter-

actions occur depending on surface chemistry. As a result of these three nanoparticle-

specific interactions, the mobility of neurotransmitters increases in the presence

of amine-terminated nanoparticles but decreases slightly with carboxyl-termi-

nated nanoparticles. Furthermore, these interactions also directly influence the

mobility of nanoparticles. They also found that run buffer viscosity was influ-

enced by the formation of a nanoparticle steady-state pseudostationary phase

along the capillary wall. Despite differences in buffer viscosity leading to

changes in neurotransmitter mobilities, no significant changes in electroosmotic

flow were observed.

6.3.2.2 Silica NanoparticlesChang and coworkers described two methods for the analysis of biologically active

amines using silica nanoparticles as a pseudostationary phase modifier in conjunction

with laser-induced native fluorescence detection.38,39 The methods were applied for

the analysis of urine samples. In the first method, the CE capillary wall was

dynamically coatedwith poly(vinylpyrrolidone) and poly(ethylene oxide) to suppress

EOF and minimize interactions between the capillary wall and analytes. Addition of

silica nanoparticles to the electrolyte caused an increase in EOF, thereby enhancing

separation speed, due to adsorption of silica nanoparticles onto the capillary wall. In

addition, the adsorption of the silica nanoparticles onto the capillary wall suppressed

nonspecific interactions of analytes with the inner wall of the capillary, enabling

sharper peaks and thus achieving higher plate numbers. In the second method,

capillaries were dynamically coated with silica nanoparticles and poly(L-lysine).

The EOF direction was controlled by varying the outermost layer of the capillaries

with poly(L-lysine) and SiO2 NPs. Over the pH range 3.0–5.0, the (poly(L-

lysine)–SiO2NP)n–poly(L-lysine) capillaries had an EOF toward the anodic end

and were more suitable for the separation of acids with respect to speed, while the

(poly(L-lysine)–SiO2NP)n capillaries had an EOF toward the cathodic end and were

more suitable for the separation of biogenic amines regarding speed and sensitivity.

The separation of standard solutions containing five amines and two acids by CE


with LIF detection using (poly(L-lysine)–SiO2NP)2–poly(L-lysine) and (poly(L-

lysine)–SiO2NP)3 capillaries was accomplished within 10 and 7min, providing plate

numbers of 38,000 and 50,000 platesm�1 for 5-hydroxytryptamine, respectively.

A CE separation method employing aminopropyl-modified nanoparticles as

a pseudostationary phase was evaluated for improving separation efficiency using

a mixture of six aromatic acids.40 An average theoretical plate value of 10,000 plates

m�1 was achieved. Excellent reproducibility of the migration time, peak area, and

peak height were obtained. The ability to separate six aromatic acids with a plate

number greater than 50,000 platesm�1 and 100-fold enhanced sample capacity and

sensitivity suggested their method may hold great promise to be incorporated into

multidimensional separation approaches.

6.3.2.3 TiO2

For separation of oligopeptides and proteins in open-tubular CE,41 TiO2-based

nanoparticles were attached to the inner wall of the capillary by a condensation

reaction. Nanoparticles with a size of 10 nmwere stabilized by polyethylene glycol. It

was suggested that the main separation mechanismwas based on the ligand exchange

of the analytes with the phosphate ion groups adsorbed onto the TiO2-based nano-

particles. The CE system was used for separation of the angiotensin-type oligopep-

tides in phosphate buffer (pH 8.0), with an average separation efficiency of

31,000 platesm�1. In anotherwork, the systemwasused for separationof conalbumin,

apo-transferrin, ovalbumin, and bovine serum albumin.42 Egg-white proteins (lyso-

zyme, conalbumin, and ovalbumin a and b) were separated with an efficiency of

10,000 platesm�1 for ovalbumin, with five different glycol isoforms of ovalbumin

resolved. This was an interesting approach for protein separation at neutral pH and

moderate ionic strength with reasonable selectivity. However, plate numbers were

low and separation time was long. In addition, the capillary coating procedure is time

consuming and complicated.

6.3.2.4 Carbon NanotubesCarbon nanotubes (CNTs) have been widely used in a variety of areas, including

biosensors, solar cells, field emission devices, and molecular electronics.43 In

electrochemical assays, CNTs play a dual role in both the recognition and the

transduction events, namely, as a building block material for biomolecule attachment

via covalent bond formation between the carboxylic acid group of the CNTs and the

amine groups of the biomolecules, and as molecular wires to allow electrical

communication between the underlying electrode and the enzyme labels attached

to the ends of the CNTs.44 CNTs also possess a large surface area, good chemical

stability, and significant mechanical strength. These unique properties of CNTs make

them extremely attractive for CE and MCE separation.

The first use of CNTs as a pseudostationary phase in CE separation was described

byWang et al.45 who used carboxylic single-wall nanotubes (SWNTs) for separation

of caffeine and theobromine with improved resolution. Xu and Li described the use

ofmultiple-wall carbon nanotubes (MWCNs) for the separation ofDNA fragments by

a CE-contactless conductivity detection method with improved resolution.46 Na et al.


compared the use of four different kinds of b-cyclodextrin (CD)-modified nanopar-

ticles, MWNTs, polystyrene nanoparticles, TiO2-based nanoparticles, and Al2O3-

based nanoparticle, as pseudostationary phases for enantioseparation of theb-blockerclenbuterol.47 The use of surfactants was necessary to form stable nanoparticle

suspensions. The use ofb-CD-modified nanoparticles improved resolution compared

to the use of the same amount of freeb-CDdue to the orientation of theb-CDadsorbed

onto the nanoparticles, which allowed a better contact with analytes. MWNTs were

alsoused for separationof sevenpurine andpyrimidinebases inCE.48 Themethodwas

also applied to determine purine and pyrimidine bases in yeast RNA. Separation of the

adenine and thymine bases was accomplished in an electrolyte containing carboxyl-

ated MWNTs, previously impossible in the absence of MWNTs. MWCNTs interact

with molecules via electrostatic interaction and hydrogen bonding. A network of

MWNTs with molecular sieving properties was proposed to be responsible for

separation. Itwas suggested that the presence ofMWNTs decreased analyte–capillary

wall interactions and stacked the analytes into narrower zones due to their large surface

area and abundance of functional groups on the surface. The use of such systems

could be favorable for protein separationwhere analyte-capillarywall interaction is an

issue.

6.3.2.5 Lipid-Based Liquid Crystalline NanoparticlesNanoparticle was utilized by Nilsson et al.49 for separation of proteins with similar

mass to charge ratio at neutral pH without organic modifier using a hydrophobic

interaction chromatography-based mechanism. Lipid-based liquid crystalline nano-

particles were prepared and used as pseudostationary phase. These nanoparticles have

potential benefits including high biocompatibility, ease of preparation, and suspension

stability at high concentrations.Using laser-induced fluorescence enabled detection at

high nanoparticle concentrations. Green fluorescent protein (GFP) and mutants of

GFP harboring single or double amino acid substitutions with the same charge were

separated in the described system but not in conventional capillary electrophoresis

(Figure 6.4). Separation was achieved by increasing the salt concentration to promote

hydrophobic interactions by shielding the repulsive electrostatic interactions. In

addition, the method was adapted to a capillary with an effective length of 6.7 cm,

enabling fast separations and future applications on chip.

Charged characteristics of biodegradable nanoparticles would influence not only

the particulate flocculation during preservation but also the colloidal drug-releasing

behavior and their interaction with biological cells. Hence, the electrophoretic

mobility and the zeta potential of bovine knee chondrocytes (BKCs) and the three

synthetic biocompatible NPs were evaluated under various ionic strengths and ionic

species. Kuo and Lin50 have studied the electrophoretic mobility, zeta potential, and

fixed charge density ofBKCs,methylmethacrylate-sulfopropylmethacrylate (MMA-

SPM) nanoparticles, polybutylcyanoacrylate (PBCA) NPs, and solid lipid nanopar-

ticles (SLNs) under the influences of Naþ , Kþ , and Ca2þ with various ionic

strengths. Results revealed that, for a specific cationic species, the absolute values

of the electrophoretic mobility, the zeta potential, and the fixed charge density

decreased with an increase in ionic strength. For a constant ionic strength, the effect


of ionic species on the reduction in the absolute values of the electrophoretic mobility,

the zeta potential, and the fixed charge density followed the orderNaþ >Kþ >Ca2þ

for the negatively charged BKCs, MMA-SPM NPs, and SLNs. The reverse order is

true for the positively charged PBCA NPs.

6.3.3 Nanoparticle-Mediated Microchip Electrophoresis

Microchip capillary electrophoresis emerged in the early 1990s as a novel approach to

the high-speed separation of biological compounds, including DNA and proteins.

Since the early development in this area, growth in the research field has exploded and

now includes chemists and engineers focused on developing new and better micro-

chips, as well as biologists and biochemists who have begun to apply this exciting and

FIGURE6.4 (a) 3D structure of native green fluorescent protein (PDB ID: 1QYO) visualized

by PyMOL molecular graphics system (DeLano Scientific, Palo Alto, CA). (b) Separation of

GFP, (þ )GFP, (�)GFP, and (H)GFP on a nondenaturing 5% polyacrylamide gel in 0.75M

Tris–HCl, pH8.8. Electrophoresiswas conducted in 192mMglycine and 8.3mMTris–HCl, pH

8.3, for 30min at 150V. The fluorescence was visualized on a UV table set to an excitation

wavelength of 365 nm. (c) Electropherogram showing separation of GFP, (þ )GFP, and (�)

GFP by capillary electrophoresis. (d) Electropherograms showing separation of GFP and (H)

GFPby capillary electrophoresis. (c and d)Capillary: 17 cmeffective length, 24 cm total length,

50mm ID and 375mmOD; electrolyte: 100–250mM tricine, pH 7.5; separation voltage: 10 kV

(normal polarity); detection: laser-induced fluorescence, excitation at 488 nm and emission

at 520 nm; sample: GFP, (þ )GFP and (�)GFP (c) and GFP and (H)GFP (d), 0.01mgmL�1 of

each protein in 50mM tricine, pH 7.5; injection: 3 kV, 3 s. Reproduced from Ref. 49, with

permission.


still relatively new methodology to real-world problems. A significant advance has

been achieved in microchip electrophoresis using nanoparticles.

6.3.3.1 Gold NanoparticlesWang and coworkers showed the use of AuNPs in conjunction with microchip

electrophoresis to improve the selectivities between solutes and to increase the

efficiency of separation.51 They coated the channel wall with a layer of PDDC,

followed by coating with citrate-stabilized AuNPs. The resolution and plate numbers

of the solutes were almost double in the presence of AuNPs. Such selectivity

improvements reflected changes in the observed mobility accrued from interaction

of solutes with the particle surfaces. They discussed that the addition of the AuNPs in

the separation buffer in MCE adds another separation vector to the orthogonal

electrophoretic vector. The coexistence of these two vectors resulted in resolution

and plate number enhancement of the solutes.

In this vein, Chang and coworkers tested DNA separation in AuNP-filled micro-

fluidic channels by MCE. To avoid the aggregation of the AuNPs and allow strong

interactions with the DNA molecules, AuNPs were modified with a polymer matrix,

poly(ethylene oxide) (PEO), via noncovalent bonding.52 Coating the separation

channels on a poly(methyl methacrylate) plate with PEO/AuNP composites was

effective in improving reproducibility and selectivity (Figure 6.5). To obtain more

hydrophilic and stable channels with high EOFs, Chen and coworkers described

a layer-by-layer assembly technique to coat the PDMSmicrochip channel with PDDA

and silica nanoparticles.53 The cationic polymer PDDAwas strongly attached to the

surface of the PDMS channel; negatively charged silica nanoparticles were then

immobilized by electrostatic interactions. In a similar manner, they also modified the

PDMS surfaces with AuNPs and polyethyleneimine.54 Both channels displayed

a long-time stability and good reproducibility and selectivity. Dopamine and

epinephrine were separated with good selectivity, employing the coated channels.

An EOF-switchable PDMS microfluidic channel modified with cysteine has been

developed by Zhang and coworkers, PDMS channels were coated with PDDA and

gold nanoparticles with a layer-by-layer technique to immobilize cysteine.55 The

electroosmotic mobility, and hence the observed mobility inside the PDDA-gold-

coated channels, can be reversibly switched between the cathodic (high pH) and

anodic direction (low pH) by varying the pH of the running buffer. This pH-dependent

selectivity can be divided into three surface charge states depending upon the

protonated degree of the amino and carboxyl groups of the cysteine: positively

charged, neutral, and negatively charged. At pH 5.0, near the isoelectric point of

the chemisorbed cysteine, the surfaces of the channels are neutral. When the pH is

above 6.0 or below 4.0, themagnitude of the observedmobility varies within a narrow

range. Separation of dopamine and epinephrine, as well as arginine and histidine, was

performed employing these AuNP-modified channels.

Liu et al. described a microchip reactor coated with an AuNP network entrapping

trypsin.56The reactor is designed for the proteolysis of low-level proteins and complex

extracts originating from mouse macrophages. The nanostructured surface coating

was assembled via a layer-by-layer electrostatic binding of PDDC and AuNPs. The


PDDC/AuNP multilayer assembly constructed on the surface of a PET microfluidic

chip offered a biocompatible interface with a large surface area, desirable for the

controlled adsorption of trypsin. Due to a high concentration of trypsin confined to

the microchannel, low levels of the standard protein samples are rapidly digested

on the microchip reactor within a few seconds. Furthermore, the digestion of real

protein mixtures isolated frommouse macrophages illustrated the performance of the

online microchip bioreactor. The protein mixtures extracted from the mouse

FIGURE 6.5 Separations of a mixture of equal volume of 10mgmL�1 DNA markers Vand

VI under different conditions using a three-layer coated PMMA microdevice. PEO(GNPs)

(1.5%) solutions containing 0.5mgml�1 EtBr were prepared in 100mM glycine buffer, pH 9.1

in (a); 100mM glycine buffer, pH 9.1, containing 1M urea in (b); and 100mM glycine–citrate

buffer, pH 9.2 in (c). Current: 20mA in (a) and (b), 30mA in (c). The separations were

conducted at 2400V. Hydrodynamic injections were conducted by dipping the DNA sample

with a 30 cm� 350mm ID capillary. The separations were conducted at 2800V. Reproduced

from Ref. 52, with permission.


macrophages were efficiently identified by online digestion and LC–ESI-MS/MS

analysis. This method is feasible for the characterization of various real protein

extracts.

Microfluidic-based bioreactors have become the focus of research interest for

immunoassays and biomarker diagnostics. Lin et al.57 and Ahn et al.58 developed

electroimmunosensing microchip bioreactors based on AuNPs for real-time mea-

surement of antigen–antibody reactions. Electrochemical impedance spectroscopy

was used by Lin et al. to detect the antigen–antibody interaction, where a 10-fold

detection sensitivity was observed employing AuNPs. In Ahn’s approach, using a

microfilter and microbeds, target proteins were immobilized in the detection zone of

the microreactor where the microelectrodes were located. The immunoreaction was

detected by measuring the electrical resistance between the microelectrodes using

AuNPs with silver enhancement. The electrical resistance varied according to the

concentration of the target antibody; the detection limit of the method was

10 ngmL�1. In another electroimmunosensing microchip bioreactor, Ahn et al.

utilized pillar-type microfilters within a reaction chamber and immunogold

silver staining to amplify the electrical signal that corresponded to the immune

complex.59 To demonstrate this approach, they simultaneously assayed three cancer

biomarkers, namely, alpha-fetoprotein, carcinoembryonic antigen, and prostate-

specific antigen (PSA), on the microchip. The electrical signal generated as a result

of immunoreaction was measured and monitored within 55min and the

working range of the proposed microchip was 10�3–10�1 mgmL�1 of the target

antigen.

A nanomosaic network of AuNPs for the detection of ultralow concentrations of

proteins was reported by Girault and coworkers using two planar microelectrodes

embedded in amicrochip (Figure 6.6) that permitted generation of capacitive coupling

to the nanomosaic system without the need for direct electrical contact with the

channel.60 By tailoring the microchannel surface using a sandwich configuration of

polyethylene terephthalate/AuNPs/poly(L-lysine), the surface charge was modified

following the biomolecule interactions and monitored using a noncontact admittance

technique. The main process governing the electrical properties of the sensor was the

adsorption of charged protein onto a highly sensitive polyethylene terephthalate/

AuNPs/poly(L-lysine) sandwich network. Under experimental conditions, the

b-lactoglobulin is negatively charged and its adsorption on the surface poly(L-lysine)layer caused a change in the charged region of the Gouy–Chapman layer. The surface

charge is modified during adsorption of macromolecules and monitored using

capacitive admittance tomoscopy. This nanodevice system behaves like a tunable

capacitor and can be employed for the detection of any kind of molecule. The

femtomolar detection of an anionic protein, such as b-lactoglobulin in phosphate

buffered saline medium, was taken as an example. AuNPs with fully matched DNA

duplexes on their surfaces aggregate together without molecular cross-linking at high

salt concentrations.

The mechanism of this noncross-linking interaction between the duplex with

different DNA sequence-modified AuNPs and a duplex-modified flat gold surface

has been explainedbyHosokawaand coworkers in a recent study.61They immobilized


15-base single-stranded DNA (ssDNA) on the surfaces of AuNPs with a diameter of

40 nm and on a flat gold substrate. AuNPs were hybridized with 15-base ssDNA at

a low salt concentration. A microfluidic device was used for simultaneous delivery of

the following three components onto the gold substrate: the duplex-modified AuNPs,

15-base ssDNA to be hybridized onto the substrate, and NaCl at a high concentration.

Adsorption of the AuNPs onto the substrate was monitored using surface plasmon

resonance imaging.When the AuNPs and the substrate had an identical sequence, the

adsorption behaviorwas analogous to the aggregation behavior ofAuNPs in test tubes.

Furthermore, they also investigated cases in which the AuNPs and the substrate had

completely different sequences, and obtained results suggesting the noncross-linking

attraction force primarily depends on the terminal base pairs of the duplexes. They

claimed that the main mechanism of the noncross-linking interaction is likely to be

interduplex base stacking rather than formation of Holliday junctions.62

6.3.3.2 Magnetic NanoparticlesZhang and coworkers described a microchip enzymatic microreactor based on the

glass microchip with trypsin-immobilized superparamagnetic nanoparticles.63

Magnetic nanoparticles with a small size (d¼ 50 nm) and strong magnetism were

synthesized. At first, amine-functionalized magnetic nanoparticles are prepared

FIGURE 6.6 SEM images of (a) the PET photoablated microchannel with a cross section of

45mm� 100mmand a length of 1.4 cm; (b) trapezoidal section of themicrochannel; and (c) the

planarmicroelectrodes. Vertically, the two band electrodes beneath the horizontal flow channel.

The detection zone represents the area where capacitive coupling takes place. (d) Side view of

the contactless zone, which is about 5mm and represents the thickness separation between the

flowmicrochannel and the planar microelectrodes. Reproduced fromRef. 60, with permission.


applying a facile one-pot strategy. Magnetic nanoparticles are then functionalized

with numerous aldehyde groups by treating the amine-functionalized magnetic

nanoparticles with glutaraldehyde. Finally, immobilization of trypsin on aldehyde-

functionalized magnetic nanoparticles is achieved through reaction of the aldehyde

groups with the amine groups of the trypsin. The magnetic nanoparticles are then

locally packed in the glass microchip by application of a strong magnetic field using

a magnet to form an on-chip magnetic nanoparticle packing bed. Capability of the

proteolyticmicroreactor is demonstrated by cytochrome c, bovine serumalbumin, and

myoglobin as model proteins. Complete protein digestion is achieved in a short time

(10 s) under a flow rate of 5.0mLmin�1. These results are expected to open up new

possibilities for proteolysis analysis as well as a new application of magnetic

nanoparticles. Replacement of nanoparticles and new microreactor construction is

facile, and the packing bed can be used at least five times without any treatment. Since

the preparation and surface functionality of magnetic nanoparticles is low-cost and

reproducible, the preparation method and application approach of magnetic nano-

particles may find much potential in proteome research.

Chen and coworkers described an integrated microfluidic sorting device that

utilized sugar-encapsulated magnetic nanoparticles to separate a specific strain of

bacteria from a mixture solution.64 In this system, a microfluidic device consisting

of two inlets and an electromagnet or permanent magnet is constructed by a soft

lithography process. The magnetic field generated by either the electromagnet or the

permanent magnet is strong enough to attract the bacteria bound to the magnetic

nanoparticles to cross the stream boundary of the laminar flow. The sorting efficiency

is found to depend on both flow rate and strength of the magnetic field. Themaximum

sorting efficiency was measured to be higher than 90% with selectivity near 100%.

The reactor was able to separate 1000 bacterial cells within 1minwithmore than 70%

sorting efficiency. However, the reactor has a drawback. The formation of large

aggregates at high bacterial and nanoparticle concentrations could block the

microchannels.

A stimuli-responsive magnetic nanoparticle-based microsystem for diagnostic

target capture and concentration has been developed for microfluidic lab card settings

by Stayton and coworkers.65 Telechelic poly(N-isopropylacrylamide) (PNIPAAm)

polymer chains were synthesized with dodecyl tails at one end and a reactive

carboxylate at the opposite end by the reversible addition fragmentation transfer

technique.ThesePNIPAAmchains self-associate intonanoscalemicelles andare used

as dimensional confinements to synthesize the magnetic nanoparticles. The resulting

superparamagnetic nanoparticles exhibited a Fe2O3 core (�5 nm) with a layer of

carboxylate-terminated PNIPAAm chains as a corona on the surface. The carboxylate

group was used to functionalize the magnetic nanoparticles with biotin and subse-

quently with streptavidin. The functionalized magnetic nanoparticles can be revers-

ibly aggregated in solution as the temperature is cycled through the PNIPAAm lower

critical solution temperature (LCST). While the magnetophoretic mobility of the

individual nanoparticles below the LCST is negligible, the aggregates formed

above the LCST are large enough to respond to an applied magnetic field.

Magnetic nanoparticles can associatewith biotinylated targets as individual particles,


and then subsequent application of a combined temperature increase and magnetic

field can be used to magnetically separate the aggregated particles onto the poly

(ethylene glycol)-modified polydimethylsiloxane channel walls of a microfluidic

device. When the magnetic field is turned off and the temperature is reversed, the

captured aggregates redisperse into the channel flow stream for further downstream

processing (Figure 6.7). The dual magnetic- and temperature-responsive nanoparti-

cles can thus be used as soluble reagents to capture diagnostic targets at a controlled

time point and channel position. They can then be isolated and released after the

nanoparticles have captured the target molecules, overcoming the problem of low

magnetophoreticmobility of the individual particleswhile retaining the advantages of

FIGURE 6.7 Particle capture and release scheme (a) and the correspondingmicrographs (b).

The PNIPAAm mNP capture/release was demonstrated in PEGylated PDMS microfluidic

channels whose channel width was 500mm. The magnetic field was introduced by embedding

a magnet at the lower side of the channel. The mNP solution (4mgmL�1) was injected into the

channels with a constant flow (�1mLmin�1) during the entire experiment. mNPs are soluble

and free flowing in the PEGylated channels when temperature is below the LCSTof PNIPAAm.

As they flow into the heated region, the temperature is above the LCST of the PNIPAAm, and

the mNPs aggregate but do not stick to the nonfouling, PEGylated channel walls in the absence

of an applied magnetic field. mNPs are captured onto the PEGylated channel walls only when

the temperature is raised above the LCSTand the magnetic field is applied. The reversal of the

temperature and applied magnetic field results in the redissolution of the aggregated magnetic

nanoparticles and their diffusive re-entry into the flow stream. Reproduced from Ref. 65, with

permission.


a high surface to volume ratio and faster diffusive properties during target capture.

These dual magnetic- and temperature-responsive magnetic nanoparticles are de-

signed to facilitate diagnostic target isolations and assays in point-of-caremicrofluidic

diagnostic devices.

A faster method of preparing an easily replaceable protease microreactor for

a microchip application is described.66 Magnetic particles coated with poly(N-

isopropylacrylamide), polystyrene, poly(2-hydroxyethyl methacrylate-co-ethylene

dimethacrylate), poly(glycidyl methacrylate), [(2-amino-ethyl)hydroxymethylen]bi-

phosphonic acid, or alginic acid, with immobilized trypsin, were utilized for hetero-

geneous digestion. To obtain the highest digestion efficiency, submicrometer spheres

were organized by an inhomogeneous external magnetic field perpendicular to the

direction of the channel. Kinetic parameters of the microchip-immobilized magnetic

enzyme reactor were determined. The capability of the proteolytic reactor was tested

with five model glycoproteins, ranging in molecular mass from 4.3 to 150 kDa.

Digestion efficiency of proteins invarious conformationswas investigated usingSDS-

PAGE, HPCE, RP-HPLC, and MS. The compatibility of the microchip-immobilized

magnetic enzyme reactor system with total and limited proteolysis of high molecular

weight glycoproteins was discussed, and subsequently paves the way for automated,

high-throughput proteomic microchip applications.

6.3.3.3 Zeolite NanoparticlesZeolite nanoparticles have drawn much interest in microchip-based applications due

to their excellent properties including large external surface areas compared to

conventional zeolite crystals, high dispersibility in both aqueous and organic solu-

tions, high thermal andhydrothermal stabilities, and tunable surface properties such as

adjustable surface charge and hydrophilicity/hydrophobicity. Yang and coworkers

fabricated an enzymatic microreactor based on the poly(methyl methacrylate)

(PMMA) microchip surface modified with zeolite nanoparticles.67 The hydropho-

bic–hydrophobic bonding interaction is used to immobilize zeolite nanoparticles on

the PMMA surface due to the surface hydrophobic property of silicalite-1, with an

aluminum-free framework. Immobilization of proteins on the silicalite-1/PMMA

surface occurred via a silicon–oxygen–silicon bridge. Using MALDI-TOF mass

spectrometry, the silicalite-1/PMMA microreactor provided an efficient digestion

of cytochrome c and bovine serum albumin at a flow rate of 4.0mLmin�1, affording

a reaction time of less than 5 s.

An on-chip microreactor has been developed for the acceleration of protein

digestion through the construction of a nanozeolite-assembled network.68 The

nanozeolite microstructure was assembled using a layer-by-layer technique based

on PDDA and zeolite nanocrystals. Adsorption of trypsin in the nanozeolite network

was theoretically studied based on the Langmuir adsorption isotherm model. It was

found that the controlled trypsin-containing nanozeolite networks assembled within

a microchannel could act as a stationary phase with a large surface to volume ratio for

the highly efficient proteolysis of both proteins at low levels and with complex

extracts. The maximum proteolytic rate of the adsorbed trypsin was measured to be

350mMmin�1mg�1, much faster than that in solution. Moreover, due the large


surface to volume ratio and biocompatible microenvironment provided by the

nanozeolite-assembled films, as well as the microfluidic confinement effect, low-

level proteins down to 16 fmol per analysis were confidently identified using the

as-prepared microreactor within a very short residence time, coupled to matrix-

assisted laser desorption time-of-flight mass spectrometry. The approach was further

demonstrated in the identification of complex extracts from mouse macrophages

integrated with two-dimensional liquid chromatography–electrospray ionization-

tandem mass spectrometry. This microchip reactor is promising for the development

of a facile means of protein identification.

The electrophoretic mobility of AuNPs with different sizes was studied using

platinum-coated alumina membranes in a microfluidic device.69 The electrophoretic

mobility of gold nanoparticles depends on the nature of the mobile phase and

interfacial properties of the alumina channels. The transport performance of nano-

particles are improved with the addition of SDS to the mobile phase because SDS not

only decreases the physical adsorption of gold nanoparticles onto the nanochannel

wall of the alumina membrane but also reduces the thickness of the electric double

layer (decreasing apparent particle size). When the alumina membranes were modi-

fied with 6-aminohexanoic acid, it was further confirmed that the physical adsorption

played a key role in the electrophoretic mobility of AuNPs.

6.3.3.4 TiO2 NanoparticlesCremer and coworkers developed a facile and simple method for patterning metal

nanoparticle films of arbitrary geometry inside sealed PDMS/TiO2/glass microfluidic

devices and illustrated the ability to biofunctionalize these films with ligands for

protein capture (Figure 6.8).70 A 6.0 nm TiO2 film is first deposited onto a planar

Pyrex or a silica substrate subsequently bonded to a PDMS mold. UV light is then

exposed through the device to reduce the metal ions in an aqueous solution to create

a monolayer-thick film of metal nanoparticles of varying sizes by independently

controlling the solution conditions in each microchannel where the film is formed. In

terms of simplicity and design flexibility, this method has advantages overmultiphase

laminar flow-based assays. In addition, functionalizing nanoparticle films inside

microfluidic channels may afford new opportunities for biosensors or screening

assays. The ability to address individual ligands atop nanoparticle films inside

microfluidic devices could be combined with such technologies as transmission

surface plasmon resonance spectroscopy or surface-enhanced fluorescence, allowing

the development of powerful lab-on-a-chip devices with label-free detection or

fluorescence detection with enhanced sensitivity.

6.3.3.5 Polymer NanoparticlesLanger and coworkers demonstrated the interaction of PEGylated poly(lactic acid)

nanoparticles/microparticles and similar particles conjugated to aptamers that recog-

nize the transmembrane prostate-specific membrane antigen (PSMA), with cells

seeded in microchannels.71 Binding of particles to the cells that expressed or did not

express the PSMA (LNCaP or PC3) was evaluated with respect to changes in fluid

shear stress, PSMAexpressionon target cells, andparticle size.At static and lowshear,


FIGURE 6.8 Schematic diagram for the deposition of a silver nanoparticle film. First,

a AgNO3 solution is introduced into the microchannel. Next, UV radiation is passed through

a photomask onto the backside of the TiO2 thin film. Agþ ions adsorbed at the interface are

selectively reduced by photoelectrons, which grow into nanoparticle films. This process can be

used in combination with thiol chemistry inside sealed microfluidic channels to address

surface chemistries in almost any desired location or pattern. Reproduced from Ref. 70,

with permission.


nanoparticle–aptamer bioconjugates selectively adhered toLNCaP, and not PC3 cells,

but not under higher shear conditions. Control nanoparticles and microparticles

lacking aptamers and microparticle–aptamer bioconjugates did not adhere to

LNCaP cells, even under very low shear conditions. This model can be used to

determine the ideal particle size and ligand density on the particle surface for binding

to target cells under fluid flow conditions. The author claims that similar microfluidic

models can be designed to simultaneously test multiple parameters and therefore

maximally optimize the physical and chemical properties of therapeutic and diag-

nostic particles prior to their in vivo evaluation.

Ishihara and coworkers have reported amethod of preparing polymer nanoparticles

for the selective capture of a specific protein from amixturewith high effectiveness.72

The nanoparticle surfacewas coveredwith hydrophilic phosphorylcholine groups and

active ester groups for easy immobilization of antibodies. Phospholipid polymers

(PMBN) composed of 2-methacryloyloxyethyl phosphorylcholine, n-butyl methac-

rylate, and p-nitrophenyloxycarbonyl polyethyleneglycol methacrylate were synthe-

sized for the surfacemodificationofpoly(L-lactic acid) nanoparticles. Surface analysis

of the nanoparticles using laser-Doppler electrophoresis and X-ray photoelectron

spectroscopy revealed that the surface of nanoparticles was covered with PMBN.

Protein adsorption was evaluated with regard to nonspecific adsorption onto the

nanoparticles that was effectively suppressed by phosphorylcholine groups. The

immobilization of antibodies on nanoparticles was carried out under such physiologi-

cal conditions as to ensure specific binding of antigens. The antibody immobilized on

the nanoparticles exhibited high activity and strong affinity for the antigen similar to

that exhibitedbyanantibody in a solution.The selectivebindingof a specificprotein as

an antigen from a protein mixture was relatively high compared to that observed with

conventional antibody-immobilized polymer nanoparticles. The authors found that

the nanoparticles with both phosphorylcholine and active ester groups for antibody

immobilization have strong potential for use in highly selective separation based on

biological affinities between biomolecules.

6.3.3.6 Carbon NanotubesIn recent years, CNTs have attracted much attention as a novel monolithic stationary

phase for high-performance liquid chromatography and capillary electrochromato-

graphy.73 Both the retention and the separation efficiencies were enhanced by

incorporation of CNTs into the stationary phase. Carboxylic CNTs were used as

a pseudostationary phase in CE, as described above, where electrodiffusion and

adsorptionwere greatly suppressed between the capillarywall and solutes and thus led

to better peak shapes of isomers. Nevertheless, CNTs have attracted much interest for

their biocompatibility, especially for conjugation with proteins. Kong and coworkers

tested a PMMA microfluidic chip for enantioseparation of tryptophan enantiomers

using bovine serum albumin-conjugated CNTs as stationary phase.74 Successful

separation of tryptophan enantiomers was achieved within 70 s with a resolution

factor of 1.35, utilizing a separation length of 32mm. The theoretical plate number

was 24,000 platesm�1 for D-tryptophan and 7700 platesm�1 for L-tryptophan,

respectively.


6.3.3.7 AuNP-Mediated On-Chip PreconcentrationThe use of the MCE method in trace analysis of analytes are somewhat

restricted owing to the small sample injection volume and short path length available

for optical measurements.75,76 To overcome this limitation, one promising alter-

native is to preconcentrate the sample prior to detection. In order to further enhance

the separation selectivity and detection sensitivity of the preconcentration method,

one can modify the separation and preconcentration buffers with charge-stabilized

metal nanoparticles. A detailed description of the experimental procedures

and principles associated with this method are included in a paper by Shiddiky

and Shim.77 They demonstrated an on-chip preconcentration method for DNA

preconcentration, separation, and EC detection employing AuNP-modified buffers

and electrodes (Figure 6.9). The device consisted of three parallel channels: the first

two were the field-amplified sample stacking (FASS) and subsequent field-amplified

sample injection (FASI) steps, the third was for microchip gel electrophoresis with

electrochemical detection step. The stacking and separation buffers containing the

hydroxypropyl cellulose (HPC) matrix were modified with AuNPs. The conducting

polymer/AuNP-modified electrode was used to detect amplified DNA based on their

direct oxidation in a solution phase.

When both the FASS and the FASI methods were sequentially applied, total

sensitivity was improved by approximately 25,000-fold compared to a conventional

MGE-ED analysis. In the FASS step, DNA stacking occurs at the boundary between

the low- and high-conductivity buffers due to the sharp decreases in the velocity of

DNAmigrating from the higher field strength (in low-conductivity sample buffer) to a

relatively lower (in high-conductivity stacking buffer) zone. In the FASI step, stacking

occurs in low-pH/high-conductivity stacking buffers containing HPC/AuNP matrix.

DNAs were stacked first at the boundary between the water and the low-pH/high-

conductivity buffer, possibly for the same reason as mentioned above for the FASS

step. The initially stacked DNA samples were subjected to a second stacking upon

interaction with the HPC/AuNP matrix present at the concentration boundary. The

second stacking occurred mainly because of retardation of the DNAmovement by the

HPC/AuNP matrix. In addition, the presence of trisodium citrate ions and citrate-

stabilized AuNPs in the stacking buffer solutionmight have affected the enhancement

of stacking efficiency due to three possibilities, the increase in relative conductivity of

the stacking buffer; adsorption of DNA onto HPC/AuNP surfaces through the

protruding part of the HPC matrix that adsorbed onto the AuNP surfaces; or

minimization of the DNA adsorption on the wall of the channel in the presence of

strongly adsorbed AuNPs.

6.3.3.8 Colloidal Au Self-Assembly in MCE SeparationSelf-assembled colloidal crystals have attracted interest in the fields of separation and

polymer dynamics owing to the relatively simple fabrication and uniquely ordered

porous structure. The microfluidic-based colloidal self-assembly technique dramati-

cally reduces preparation time and avoids formation of cracks caused by drying.Using

this technique, robust colloidal lattices of various pore sizes and materials can be

readily incorporated into microfluidic devices for rapid separation of biomolecules


FIGURE6.9 Schematic illustration of the preconcentration, separation, and electrochemical

detection of DNA. (1) Sample loading: a voltage of þ 100V cm�1 was applied to R1 while R2

was grounded and R3 was left floating. (2) FASS step: (i) a potential of þ 200V cm�1 was

applied to R2 for 130 s while R3 was grounded and R1 and R4 were left floating. (ii) Water/

stacking buffer injection: during FASS step, a water plug was injected hydrodynamically from

R5 to R4 at a flow rate of approximately 0.1mLmin�1 for 110 s. Thereafter, the stacking buffer

was injected for 60 s at the same flow rate. (3) FASI step: initially preconcentrated sample was

then injected into channel 2 by applying a voltage of approximately 150V cm�1 to R3 with R5

grounded, leaving all other reservoirs floating. (4) Sample loading and injection: a voltage of

þ 100V cm�1 was applied to R7 for 40 s, while R4 was grounded and R3, R5, R6, and R8

remained floating. Injection was effected by applying an injection voltage of þ 200V cm�1 for

5 s to R4. (5) Separation and detection: MGE-ED was performed by applying separation field

strength of�340V to R6 and þ 1500V to R8with R3, R4, R5, and R7 floating. Amperometric

detection potential: þ 0.8V versus Ag/AgCl. Reproduced from Ref. 77, with permission.


with a wide size distribution. Recently, Zeng and Harrison developed a facile,

microfluidic, colloidal self-assembly strategy to create ordered, robust, three-dimen-

sional nanofluidic sieves within microfluidic devices, with which a fast separation of

DNA and proteins of a wide size range was achieved.78 Fabrication and characteriza-

tion of self-assembled colloidal arrays within microfluidic systems is schematically

illustrated in Figure 6.10. This approach offers a significantly greater assembly speed

over conventional colloidal deposition approaches,which usually take tens of hours to

days to construct colloidal arrays within microdevices. The high assembly speed can

be attributed to the large surface area of the protruding evaporation interface.

Compared to the design of microfabricated barrier structures with a gap size smaller

than the particles to be trapped, this approach uses the air–liquid interface as a virtual

frit to retain the microspheres, offering advantages in terms of a much larger

FIGURE 6.10 Fabrication and characterization of self-assembled colloidal arrays within

microfluidic systems. (a) Schematic illustration of microfluidic colloidal self-assembly in

a one-dimensional separation microchip (PDMS chip layout: (1) buffer, (2) sample, (3) sample

waste, (4) buffer waste). (b) Optical micrograph of a translucent 0.9 mm silica sphere array

growing inside a microchannel (as indicated by the left arrow), showing a convex evaporation

interface at the channel opening. (c) Digital images of a PDMS chip packed with 0.9mm silica

spheres before drying. When illuminated by white light vertically from the bottom, the array

exhibited monochromic transmitted light at various angles due to Bragg diffraction. (d and e)

SEM images of a matrix of 330 nm silica spheres at different magnifications. (f) SEM image of

a hexagonally closed packed 2mm PS colloidal array fabricated within a microchannel. The

arrows indicate lattice defects. The scale bars are 200, 2, and 10 mm in (d–f), respectively.

Reproduced from Ref. 78, with permission.


evaporation area and simplified device fabrication. The high-speed assembly allows

the packing of large particles of various materials.

Separation of Proteins and DNA Nanometer-sized interstices in the close-packed

sphere lattice create a nanofluidic sieve that consists of voids interconnected by

narrower pores, with an equivalent diameter approximately 15% of the sphere size.

Molecules experience a loss in entropy by steric constriction while traveling through

a constraining pore.79 Thus, the porous structure imposes periodically modulated,

free-energy barriers to molecular transport that are presumably responsible for size-

dependent separation of molecules with dimensions comparable to the narrower pore

sizes. Harrison’s approach thus provided separation of biomolecules with a wide size

distribution, ranging from proteins (20–200 kDa) to dsDNA (0.05–50 kbp). For

example, sieving-based protein separation in these porous beds was an obvious

objective,made challenging by thevery small pore size required. Figure 6.11a exhibits

separation of SDS-denatured protein markers using differently sized silica particles.

Four proteins of 20–205 kDa were separated in a matrix of 330 nm silica spheres, as

shown in Figure 6.11a. The resolution was 2.64, between the 20.1 and the 116 kDa

proteins, and 3.92, between the 116 and the 205 kDa proteins, respectively, which

indicate a better size selectivity for larger proteins than smaller ones. Four consecutive

runs of a lowDNAmass ladder containing six equimolar dsDNA fragments from 100

to 2000 bpwere performed in a 0.9mmsilica particle array; the electropherograms are

superimposed in Figure 6.11b. In this case, the detector was located 5.0mm from the

injection point. The device was operated for approximately 5 h before this reproduc-

ibility test. The standard deviation of the migration time between these runs was less

than2%, indicating the stability of the self-assembled colloidal sieveunder the applied

electric field. The bed stability is likely a result of van derWaals forces between silica

particles and the contribution from hydrogen bonds between the surfaces of close-

packed silica particles.

FIGURE6.11 (a) Separation of four proteins using 330 nm silica particles (E¼ 19.1V cm�1,

L¼ 8mm): (1) trypsin inhibitor, 20.1 kDa; (2) BSA, 66 kDa; (3) b-galactosidase, 116 kDa; and(4) myosin, 205 kDa. (b) Four consecutive runs of a low DNA mass ladder obtained using

0.9mm silica beads in a device that had been operated for approximately 5 h (E¼ 19.2V cm�1;

L¼ 5mm, and the microchip has approximately100 mm wide and approximately 20 mm deep

microchannels with a cross-injection design). Reproduced from Ref. 78, with permission.


6.3.3.9 Surface Displacement Reactions on Colloidal Gold in MicrofluidicChannelsSpatial imaging of the fluorescence from derivatized organomercaptans in a flowing

microfluidic channel has been developed by Bohn and coworkers to monitor the

surface displacement kinetics of organomercaptans to the surface ofAuNPs.80 In these

experiments, surface displacement of the tagged reagent to the surface of the colloid is

accompanied by a fluorescence decrease associated with quenching of fluorescence

from the adsorbed fluorophore by coupling to the surface plasmon of Au. Several

characteristic features of the hybrid nanofluidic–microfluidic devices have been

exploited while making these measurements. The ability to electrokinetically inject

fluid across a nanocapillary array membrane separating two microfluidic channels

provides a convenient means of reaction initiation, and once begun, the constant

velocity of the microfluidic flow allows a distance–time conversion so that spatial

images canbe translateddirectly into temporal plots of reactant concentrations.As test

cases, the surface displacement kinetics of two fluorescently tagged species, a small

organomercaptan (SAMSA), and an octapeptide (CDWAK�WAD) have been mea-

sured. Excellent fits of the kinetic data toLangmuir kineticmodelswere obtained in all

cases, obviating the need to invoke more complicated kinetic models. However, the

surface displacement rate constants determined for thiol surface displacement on Au

nanoparticles were roughly one order of magnitude larger than those measured for

similar thiols on planarAu surfaces, indicating faster kinetics in the colloid–adsorbate

system. These results highlight the utility of colloidal Au nanoparticles as molecular

carriers for the sequestration of analytes, allowing the manipulation of mass-limited

samples and ultimately the capture and delivery of selected analytes from a micro-

fabricated device to an off-line detector.

6.3.3.10 Microchip-Based Bio-Barcode AssayA single disposable chip has been developed for carrying out a multistep process that

employed nanoparticles for single protein marker detection.81 To illustrate the

capability of this bio-bar code assay (BCA), Liu et al tested the presence of PSA in

buffer solution and goat serum. The proposed BCA protocol can be divided into two

stages—targetproteinseparation (stage1)andbarcodeDNAdetection (stage2).Major

steps are illustrated in Figure 6.12. In stage 1, magnetic microparticles are introduced

into a microfluidic channel reactor. The sample fluid is then flowed into the channel

along with functionalized AuNP probes. Hybridized magnetic microparticle–

protein–AuNPs sandwiches formwhen target proteins are present in the sample fluid.

Themagneticmicroparticles andmagnetic microparticle–protein–AuNPs conjugates

are then immobilized to thechannelwallwithamagnetwhile the supernatant iswashed

away with several column volumes of buffer. Subsequently, barcode DNA strands are

released from the AuNP probes by applying deionized water that dissociates the

barcode DNA from the AuNPs. For stage 2, the released barcode DNA strands are

transferred to a detection channel. The bottom surface of the detection channel is

functionalizedwithcapture strands thatarehalfcomplementary to thebarcodeDNA.A

second set of AuNP probes, functionalized with the remaining complimentary

sequence, is then introduced. It should be noted that the AuNP probes used here are


not the same as the AuNP probes employed in the previous stage. The barcode DNA

molecules allow the functionalized NP probes to be hybridized to the surface. Chip-

immobilizedAuNP probes thus signify the presence of the barcodeDNA. The amount

of chip-immobilized AuNP probes can be detected with a number of methods. One

method involves detecting light scattering off the surface-bound AuNP probes.

Exposing the captured goldAuNP probes to a silver staining solution further enhances

the detectable optical signal.82 Gold serves as a catalyst for silver staining, hence

enlarging the size of thegold nanoparticles. Enhancedparticles arevisible to the naked

eye and can be detected using commercially available scanners. Detection was

accomplished at PSA concentrations as low as 500 aM. This corresponds to only

300 copies of protein analytes using 1mL total sample volume.

6.3.3.11 Microfluidic Fabrication of Biopolymer Micro- and NanoparticlesDue to the advantages of the formation of polymer-based microparticles using

microdevices over conventional emulsification techniques,with respect to size control

and polydispersity of the final sample, fabrication of simple, reliable, and robust

methods in microfluidic devices has gained interest. Kumacheva and colleagues

FIGURE 6.12 Implementation of the bio-barcode assay within a microfluidic device. First,

magnetic particles functionalized with monoclonal PSA antibodies are introduced into the

separation area of the chip. The particles are then immobilized by placing a permanent magnet

under the chip, followed by introduction of the sample and AuNPs that are decorated with both

polyclonal antibodies and barcode DNA. The barcode DNA is then released from the AuNPs

and is transported to the detection area of the chip. The detection area of the chip is patterned

with capture DNA. Salt and a second set of AuNPs functionalizedwith complementary barcode

DNA sequences are introduced into the detection area to allow hybridization. Finally, the signal

from the AuNPs is amplified using silver stain. Reproduced from Ref. 81, with permission.


described two methods for the formation of monodispersed, polymer-based micro-

particles inmicrofluidicdevices frombiologicallyderivedandsyntheticpolymers.83,84

In the former method, they reported an approach generating capsules of biopolymer

hydrogels. Droplets of an aqueous solution of a biopolymer were emulsified in an

organic phase comprising a cross-linking agent. Polymer gelation was achieved in

a microfluidic chip by diffusion-controlled, ionic cross-linking of the biopolymer,

following the transfer of the cross-linking agent from the continuous phase to the

droplets. Collecting particles in a large pool of cross-linking, agent-free liquid

quenchedgelation.Microgel structure, fromcapsules togradientmicrogels toparticles

withauniformstructure,wascontrolledbyvaryingresidence timeof thedropletson the

microfluidic chip and concentration of the cross-linking agent in the continuous phase.

The described approach was applied to prepare capsules of several polysaccharides

such as alginate, carrageenan, and carboxymethylcellulose. For continuous and

scalableproductionofcoreshelldropletsandpolymercapsules inmicrofluidicdevices,

they employed a capillary instability-driven breakup of a liquid jet formed by two

immisciblefluidstoachieveprecisecontroloveremulsificationofeachliquid,allowing

the production of highlymonodispersed core shell droplets with a predetermined core

diameter and shell thickness. Polymer particleswithvarious shapes andmorphologies,

including spheres, truncated spheres and hemispheres, and single and multicore

capsules were obtained via fast photopolymerization of monomeric shells.

Lin and coworkers described the manipulation of Ca-alginate microspheres, using

a microfluidic chip, for the encapsulation of gold nanoparticles, based on hydrody-

namic focusing on the formation of a series of self-assembling sphere structures, the

so-called water-in-oil (w/o) emulsions, in the cross-junction microchannel.85 These

fine emulsions, consisting of aqueous Na-alginates, are then dripped into a solution of

20% calcium salt to accomplish Ca-alginate microspheres in an efficient manner.

Experimental data show that microspheres with diameters ranging from 50 to

2000mm, with a variation less than 5%, were precisely generated. The size and

gap of the droplets are tunable by adjusting the relative sheath/sample flow rate ratio.

Furthermore, the application of these particles toward encapsulated AuNPs was

successfully performed. Developed microfluidic chip fabrication and setup are

proficient and easily programmed to generate a large set of ordered Ca-alginate

microspheres.

Uniform, spherical, and molecularly imprinted polymer beads were prepared via

controlled suspension polymerization in a spiral-shaped microchannel using mineral

oil and a perfluorocarbon liquid as continuous phases.86 Monodisperse droplets

containing the monomers, template, initiator, and porogenic solvent were introduced

into the microchannel; particles of uniform size were produced by subsequent UV

polymerization, quickly and without wasting polymer materials. Change in flow

conditionswithin themicrofluidic device allowed for variation in droplet/particle size.

The diameter of the resulting products typically had a coefficient of variation below

2%. The specific binding sites created during the imprinting process were analyzed by

radioligand binding analysis. The molecularly imprinted microspheres produced in

the liquid perfluorocarbon continuous phase had a higher binding capacity compared

to the particles produced in themineral oil continuous phase, though it should be noted


that the aim of this studywas not to optimize ormaximize imprinting performance but

rather to demonstrate broad applicability and compatibility with knownMIP produc-

tion methods. The successful imprinting against a model compound using two very

different continuous phases (one requiring a surfactant to stabilize the droplets, the

other not) demonstrates the wide-ranging of this approach.

Rondeau and Cooper-White described a microfluidic-based methodology for the

synthesis of monodispersed biopolymer (alginate) particles utilizing a multiphase

microfluidic device that relied upon a novel pseudoequilibrium solvent diffusion

process during laminar flow in PDMS-based microdevices.87 The two alternative

routes have been explored in detail in order to understand the parameters influencing

the resultant product, solvent diffusion, and competing condensation and/or cross-

linking reactions (Figure 6.13). A detailed description of the experimental procedures

and principles associated with this method, including the dependence of droplet

generation, solvent diffusion, and cross-linking kinetics on changes in device and flow

parameters and on polymer solution properties, is available in Ref. 87. These

microdevice formats were used to make microparticles of controlled size and

polydispersity and to reproducibly produce nanoparticles from alginate. The results

indicate the process will allow production of particles from any synthetic or biologi-

cally derived polymer that can be solvated within a liquid partially miscible with

FIGURE 6.13 Diagram showing the two processes used for producing alginate micropar-

ticles and nanoparticles. (a) In process 1, the alginate solution and the CaCl2 solution are

injected into the channel. The two solutions mix together before the droplet formation with

DMC as the continuous phase. As the droplets flow downstream, two processes are competing,

the water diffuses out of the drops that harden due to the alginate cross-linking reaction. The

particles are collected in a CaCl2 solution to achieve full extension of cross-liking and to store

the particles. (b) In process 2, the same alginate solution is injected via the two Y-shaped inlet

channels that join together to form the main flow channel. DMC is injected further downstream

via the secondary Y-shaped channels to generate alginate droplets at their junction with the

main flow channel. As the droplets flow downstream, the water is diffusing from the drops into

the DMC. The condensed alginate particles are collected in a highly concentrated CaCl2solution. Reproduced from Ref. 87, with permission.


another liquid. The size of particles can be explicitly controlled through varying the

relative solvency of the two liquids, physical properties of the polymer solution,

including polymer concentration, configuration, ionic state, and solvent conditions,

and processing parameters, including flow rate, temperature, and device design.

A sensor for characterization of nanoparticle colloidal suspensions employing

a diffraction grating under total internal reflection for investigation of nanodisperse

fluids, passing through an integrated microfluidic channel, has been developed.88

Dispersions containing polymeric, metallic, and ferromagnetic nanoparticles were

studied and accurately determined in real time with the specific refractive index

for nanoparticle suspension and nanoparticle concentration. Nanoparticle concentra-

tions were calculated at a resolution of 0.3–0.5wt % for polymeric nanoparticles,

0.03–0.05wt % for metallic nanoparticles, and 0.05–0.1wt % for ferromagnetic

nanoparticles.This translatedtoeffective refractive indicesdeterminedwithanaccuracy

of 7� 10�4 for the polymeric and 2� 10�4 for the metallic and ferromagnetic

dispersions.

6.3.3.12 In-Channel Modification of Au MicroelectrodesEnhancing the coulometric efficiency (Ceff) of amperometric detection interfacedwith

capillary electrophoresis is of considerable importance. The Ceff is defined as the

percentage of the injected analyte detected at the amperometric detector, governed by

the following equation:

Ceff ¼ ðNd=NiÞ � 100% ð6:11ÞwhereNd andNi correspond to thenumberofmoles detected and injected, respectively.

Nd can be calculated based on Faraday’s equation:

QC ¼ nFNd ð6:12Þwhere n is the number of electrons, F is Faraday’s constant, and Qc the total charge

(Qc¼ i� t).Qc can be determined by integrating the area of the corresponding peak in

the electropherogram.However, a smooth surface restrains the interactionbetween the

analytes and the electrode, which results in lower sensitivity. Jankowiak and cow-

orkers described amethod for in-channel electrochemical deposition of gold particles

on a PDMS/glassmicrofluidic device in order to vertically increase the surface area of

the Au-sensing microelectrode.89 The modified electrodes provided well-resolved

separation of dopamine and catechol in less than 60 s with enhanced Ceff; no peak

broadening was observed, which can be attributed to vertical enlargement of the

electrode’s size. The electroplated Au microelectrode offered a stable detection

background current (<2 nA) with minimal noise level (<15 pA), which lowered

the LOD by a factor of 2 for dopamine and a factor of 3 for catechol.

6.4 CONCLUSIONS

Nanoparticles offer unique opportunities for designing excellent separationmedia for

analyte separation inCEandMCE.The studiesdescribed abovedemonstrate thebroad

CONCLUSIONS 247

potential of nanoparticles for new separation media with excellent selectivity and

efficiency.Although significant advances have beenmade in the area of nanoparticles,

many theoretical and technical problems still need to be solved. The successful

realization of these new systems requires proper attention to aggregation of nano-

particles that commonly controls the performances of the method. The remarkable

selectivity of the new nanoparticle-based CE and MCE protocols opens up the

possibility for detecting proteins, DNA, peptides, amino acids, or numerous target

agents that cannot be separated by conventional methods. All these approaches with

various nanoparticles have been designed with nanometer dimensions and may be

fully implemented in ultrasensitive detection and monitoring of biological events,

where the role of mechanics in these events will be fully appreciated.

REFERENCES

1. Wallingford, R. A.; Ewing, A. G. Capillary electrophoresis. Adv. Chromatogr. 1989, 29,

1–76.

2. Hayat, M. A.,Ed. Colloidal Gold: Principles. In Methods and Applications; Academic

Press: 1991.

3. Schmid, G. G., Ed. Clusters and Colloids; VCH: Weinheim, Germany, 1994.

4. Shiddiky, M. J. A.; Kim, R.-E.; Shim, Y.-B. Microchip capillary electrophoresis with

a cellulose-DNA-modified screen-printed electrode for the analysis of neurotransmitters.


5. Shiddiky, M. J. A.; Won, M.-S.; Shim, Y.-B. Simultaneous analysis of nitrate and nitrite in

a microfluidic device with a Cu-complex-modified electrode. Electrophoresis 2006, 27,

4545–4554.

6. Lee, K.-S.; Shiddiky, M. J. A.; Park, S.-H.; Park, D.-S.; Shim, Y.-B. Electrophoretic

analysis of food dyes using a miniaturized microfluidic system. Electrophoresis 2008, 29,

1910–1917.

7. Fujimoto, C.; Muranaka, Y. Electrokinetic chromatography using nanometer-sized silica

particles as the dynamic stationary phase. J. High Resolut. Chromatogr. 1997, 20,

400–402.

8. Wang, Y.; Ouyang, J.; Baeyens,W. R. G.; Delanghe, J. R. Use of nanomaterials in capillary

and microchip electrophoresis. Expert Rev. Proteomics 2007, 4, 287–298.

9. Nilsson, C.; Birnbaum, S.; Nilsson, S. Use of nanoparticles in capillary and microchip

electrochromatography. J. Chromatogr. A 2007, 1168, 212–224.

10. Nilsson, C.; Nilsson, S. Nanoparticle-based pseudostationary phases in capillary electro-

chromatography. Electrophoresis 2006, 27, 76–83.

11. Guihen, E.; Glennon, J. D. Nanoparticles in separation science—recent developments.

Anal. Lett. 2003, 36, 3309–3336.

12. Lin, Y.-W.; Huang, M.-F.; Chang, H.-T. Nanomaterials and chip-based nanostructures for

capillary electrophoretic separations of DNA. Electrophoresis 2005, 26, 320–330.

13. Zhang, Z.; Wang, Z.; Liao, Y.; Liu, H. Applications of nanomaterials in liquid chromatog-

raphy: opportunities for separation with high efficiency and selectivity. J. Sep. Sci. 2006,

29, 1872–1878.


14. Manz,A.;Graber, N.;Widmer, H.M.Miniaturized total chemical analysis systems: a novel

concept for chemical sensing. Sens. Actuators B 1990, 1, 244–248.

15. Kutter, J. P.; Geschke, O. InMicrosystemEngineering of Lab-on-a-ChipDevices; Geschke,

O.; Klank, H.; Tellemann, P., Eds.; Wiley-VCH: Weinheim, Germany, 2004.

16. West, J.; Becker, M.; Tombrink, S.; Manz, A. Micro total analysis systems: latest

achievements. Anal. Chem. 2008, 80, 4403–4419.

17. Dittrich, P. S.; Tachikawa, K.;Manz, A.Micro total analysis systems. Latest advancements

and trends. Anal. Chem. 2006, 78, 3887–3908.

18. Wang, J. Real-time electrochemical monitoring: toward green analytical chemistry. Acc.

Chem. Res. 2002, 35, 811–816.

19. Shiddiky, M. J. A.; Park, D.-S.; Shim, Y.-B. Detection of polymerase chain reaction

fragments using a conducting polymer-modified screen-printed electrode in a microfluidic

device. Electrophoresis 2005, 26, 4656–4663.

20. Shiddiky, M. J. A.; Rahman, M. A.; Park, J.-S.; Shim, Y.-B. Analysis of polymerase chain

reaction amplifications through phosphate detection using an enzyme-based microbio-

sensor in a microfluidic device. Electrophoresis 2006, 27, 2951–2959.

21. Linhardt, R. J.; Toida, T. Ultra-high resolution separation comes of age. Science 2002, 298,

1441–1442.

22. Dolnik, V.; Liu, S.; Jovanovich, S. Capillary electrophoresis onmicrochip.Electrophoresis

2000, 21, 41–54.

23. Stellwagen, N. C.; Gelfi, C.; Pier Giorgio, R. The free solution mobility of DNA.

Biopolymers 1997, 42, 687–703.

24. Kato, M.; Gyoten, Y.; Sakai-Kato, K.; Nakajima, T.; Toyo’oka, T. Cationic starch

derivatives as dynamic coating additives for analysis of amino acids and peptides

using poly(methyl methacrylate) microfluidic devices. Anal. Chem. 2004, 76,

6792–6796.

25. Terabe, S. Micellar electrokinetic chromatography. Anal. Chem. 2004, 76, 240A–246A.

26. Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Electrokinetic separa-

tions with micellar solutions and open-tubular capillaries. Anal. Chem. 1984, 56,

111–113.

27. Terabe, S.; Otsuka, K.; Ando, T. Electrokinetic chromatographywithmicellar solution and

open-tubular capillary. Anal. Chem. 1985, 57, 834–841.

28. Zhang, C.-X.; Sun, Z.-P.; Lin, D.-K. Micellar electrokinetic capillary chromatography

theory based on conventional chromatography. J. Chromatogr. A 1993, 655, 309–316.

29. Neiman, B.; Grushka, E.; Lev, O. Use of gold nanoparticles to enhance capillary

electrophoresis. Anal. Chem. 2001, 73, 5220–5227.

30. Kenndler, E. InHigh Performance Capillary Electrophoresis; Khaledi, M. G., Ed.; Wiley:

New York, 1998; pp 25–76.

31. Yu, C. J.; Su, C. L.; Tseng, W. L. Separation of acidic and basic proteins by nanoparticle-

filled capillary electrophoresis. Anal. Chem. 2006, 78, 8004–8010.

32. Bao, J.; Chen, W.; Liu, T.; Zhu, Y.; Jin, P.; Wang, L.; Liu, J.; Wei, Y.; Li, Y. Bifunctional

Au-Fe3O4 nanoparticles for protein separation. ACS Nano 2007, 1, 293–298.

33. Lin, C.-Y.; Liu, C.-H.; Chang,H.-C.; Tseng,W.-L. Enrichment and separation of acidic and

basic proteins using the centrifugal ultrafiltration followed by nanoparticle-filled capillary

electrophoresis. Electrophoresis 2008, 29, 3024–3031.

REFERENCES 249

34. Li, M.-D.; Cheng, T.-L.; Tseng, W.-L. Nonionic surfactant-capped gold nanoparticles

for selective enrichment of aminothiols prior to CE with UV absorption detection.


35. Yang, L.; Guihen, E.; Holmes, J. D.; Loughran, M.; O’Sullivan, G. P.; Glennon, J. D. Gold

nanoparticle-modified etched capillaries for open-tubular capillary electrochromatogra-

phy. Anal. Chem. 2005, 77, 1840–1846.

36. Zhou, D.; Wang, Y.; Yang, R.; Zhang, W.; Shi, R. Effects of novel quasi-interpenetrating

network/gold nanoparticles composite matrices on DNA sequencing performances by CE.


37. Ivanov, M. R.; Bednar, H. R.; Haes, A. J. Investigations of the mechanism of gold

nanoparticle stability and surface functionalization in capillary electrophoresis. ACS

Nano 2009, 3, 386–394.

38. Kuo, I. T.; Huang, Y. F.; Chang, H. T. Silica nanoparticles for separation of biologically

active amines by capillary electrophoresis with laser-induced native fluorescence detec-

tion. Electrophoresis 2005, 26, 2643–2651.

39. Huang, Y.-F.; Chiang, C.-K.; Lin, Y.-W.; Liu, K.; Hu, C.-C.; Bair, M.-J.; Chang, H.-T.

Capillary electrophoretic separation of biologically active amines and acids using nano-

particle-coated capillaries. Electrophoresis 2008, 29, 1942–1951.

40. Liu, D. N.; Wang, J.; Guo, Y. G.; Yuan, R. J.; Wang, H.; Bao, J. J. Separation of aromatic

acids by wide-bore electrophoresis with nanoparticles prepared by electrospray as

pseudostationary phase. Electrophoresis 2008, 29, 863–870.

41. Hsieh, Y. L.; Chen, T. H.; Liu, C. P.; Liu, C. Y. Titanium dioxide nanoparticles-coated

column for capillary electrochromatographic separation of oligopeptides. Electrophoresis

2005, 26, 4089–4097.

42. Hsieh, Y. L.; Chen, T. H.; Liu, C. Y. Capillary electrochromatographic separation of

proteins on a coated titanium dioxide nanoparticles.Electrophoresis 2007, 27, 4288–4294.

43. Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chemistry of carbon nanotubes. Chem.

Rev. 2006, 106, 1105–1136.

44. Gooding, J. J.;Wibowo, R.; Liu, J.; Yang,W.; Losic, D.; Orbons, S.; Mearns, F. J.; Shapter,

J. G.; Hibbert, D. B. Protein electrochemistry using aligned carbon nanotube arrays. J. Am.

Chem. Soc. 2003, 125, 9006–9007.

45. Wang, Z.; Luo, G.; Chen, J.; Xiao, S.; Wang, Y. Carbon nanotubes as separation carrier in

capillary electrophoresis. Electrophoresis 2003, 24, 4181–4188.

46. Xu, Y.; Li, S. F. Y. Carbon nanotube-enhanced separation of DNA fragments by a portable

capillary electrophoresis system with contactless conductivity detection. Electrophoresis

2006, 27, 4025–4028.

47. Na, N.; Hu, Y. P.; Jin, O. Y.; Baeyens, W. R. G.; Delanghe, J. R.; Taes, Y. E. C.; Xie, M. X.;

Chen, H. Y.; Yang, Y. P. On the use of dispersed nanoparticles modified with single layer

b-cyclodextrin as chiral selector to enhance enantioseparation of clenbuterolwith capillaryelectrophoresis. Talanta 2006, 69, 866–872.

48. Xiong, X.; Ouyang, J.; Baeyens, W. R. G.; Delanghe, J. R.; Shen, X. M.; Yang, Y. P.

Enhanced separation of purine and pyrimidine bases using carboxylic multiwalled carbon

nanotubes as additive in capillary zone electrophoresis. Electrophoresis 2006, 27,

3243–3253.

49. Nilsson, C.; Becker, K.; Harwigsson, I.; B€ulow, L.; Birnbaum, S.; Nilsson, S.

Hydrophobic interaction capillary electrochromatography of protein mutants. Use of


lipid-based liquid crystalline nanoparticles as pseudostationary phase. Anal. Chem.

2009, 81, 315–321.

50. Kuo, Y.-C.; Lin, T.-W. Electrophoretic mobility, zeta potential, and fixed charge

density of bovine knee chondrocytes, methyl methacrylate-sulfopropyl methacrylate,

polybutylcyanoacrylate, and solid lipid nanoparticles. J. Phys. Chem. B 2006, 110,

2202–2208.

51. Pumera, M.; Wang, J.; Grushka, E.; Polsky, R. Gold nanoparticle-enhanced microchip

capillary electrophoresis. Anal. Chem. 2001, 73, 5625–5628.

52. Lin, Y.-W.; Huang, M.-J; Chang, H.-T. Analysis of double-stranded DNA by microchip

capillary electrophoresis using polymer solutions containing gold nanoparticles.

J. Chromatogr. A 1014, 2003, 47–55.

53. Wang, W.; Zhao, L.; Zhang, J.-R.; Wang, X.-M.; Zhu, J.-J.; Chen H.-Y. Modification of

poly(dimethylsiloxane) microfluidic channels with silica nanoparticles based on layer-by-

layer assembly technique. J. Chromatogr. A 2006, 1136, 111–117.

54. Wang, A.-J.; Xu, J.-J.; Zhang, Q.; Chen, H.-Y. The use of poly(dimethylsiloxane) surface

modification with gold nanoparticles for the microchip electrophoresis. Talanta 2006, 69,

210–215.

55. Wang,W.; Zhao, L.; Zhou, F.; Zhu, J.-J.; Zhang, J.-R. Electroosmotic flow-switchable poly

(dimethylsiloxane) microfluidic channel modified with cysteine based on gold nanopar-

ticles. Talanta 2007, 73, 534–539.

56. Liu,Y.;Xue,Y.; Ji, J.; Chen,X.; Kong, J.; Yang, P.; Girault, H.H.; Liu, B.Gold nanoparticle

assemblymicrofluidic reactor for efficient on-line proteolysis.Mol. Cell. Proteomics 2007,

6, 1428–1436.

57. Li, M.; Lin, Y.-C.; Su, K.-C.; Wang, Y.-T.; Chang, T. C.; Lin, H. P. A novel real-time

immunoassay utilizing an electro-immunosensing microchip and gold nanoparticles for

signal enhancement. Sens. Actuators B 2006, 117, 451–456.

58. Ko, Y. J.; Maeng, J. H.; Ahn, Y.; Hwang, S. Y.; Cho, N. G.; Lee, S.-H. Real-time

immunoassay with a PDMS-glass hybrid microfilter electro-immunosensing chip using

nanogold particles and silver enhancement. Sens. Actuators B 2008, 132, 327–333.

59. Ko, Y.-J.; Maeng, J.-H.; Ahn, Y.; Hwang, S. Y.; Cho, N.; Lee, S.-H. Microchip-based

multiplex electro-immunosensing system for the detection of cancer biomarkers.


60. Gamby, J.; Abid, J.-P.; Tribollet, B.;Girault, H.H.Nanomosaic network for the detection of

proteins without direct electrical contact. Small 2008, 4, 802–809.

61. Sato, Y.; Hosokawa, K.; Maeda, M. Detection of non-cross-linking interaction between

DNA-modified gold nanoparticles and a DNA-modified flat gold surface using surface

plasmon resonance imaging on a microchip. Colloids Surf. B: Biointerfaces 2008, 62,

71–76.

62. Ortiz-Lombardia, M.; Gonzalez, A.; Eritja, R.; Aymami, J.; Azorin, F.; Coll, M. Crystal

structure of a DNA Holliday junction. Nat. Struct. Biol. 1999, 6, 913–917.

63. Liu, J.; Lin, S.; Qi, D.; Deng, C.; Yang, P.; Zhang,X.On-chip enzymaticmicroreactor using

trypsin-immobilized superparamagnetic nanoparticles for highly efficient proteolysis.

J. Chromatogr. A 2007, 1176, 169–177.

64. Shih, P.-H.; Shiu, J.-Y.; Lin, P.-C.; Lin,C.-C.;Veres, T.; Chen, P.On chip sorting of bacterial

cells using sugar-encapsulated magnetic nanoparticles. J. Appl. Phys. 2008, 103, 07A316/

311–307A316/313.

REFERENCES 251

65. Lai, J. J.; Hoffman, J. M.; Ebara, M.; Hoffman, A. S.; Estournes, C.;Wattiaux, A.; Stayton,

P. S. Dual magnetic-/temperature-responsive nanoparticles for microfluidic separations

and assays. Langmuir 2007, 23, 7385–7391.

66. ZuzanaB�ılkov�a, Z.; Slov�akov�a,M.;Minc, N.; F€utterer, C.; Cecal, R.; Hor�ak, D.; Benes,M.;

Potier, I. I.; K�renkov�a, J.; Przybylski, M.; Viovy, J.-L. Functionalized magnetic micro- and

nanoparticles: optimization and application to m-chip tryptic digestion. Electrophoresis

2006, 27, 1811–1824.

67. Huang, Y.; Shan,W.; Liu, B.; Liu, Y.; Zhang, Y.; Zhao, Y.; Lu, H.; Tang, Y.; Yang, P. Zeolite

nanoparticle modified microchip reactor for efficient protein digestion. Lab Chip 2006, 6,

534–539.

68. Ji, J.; Zhang, Y.; Zhou, X.; Kong, J.; Tang, Y.; Liu, B. Enhanced protein digestion through

the confinement of nanozeolite-assembled microchip reactors. Anal. Chem. 2008, 80,

2457–2463.

69. Yuan, H.; Cheow, P.-S.; Ong, J.; Toh, C.-S. Characterization of the electrophoretic mobility

of gold nanoparticles with different sizes using the nanoporous aluminamembrane system.

Sens. Actuators B 2008, 134, 127–132.

70. Castellana, E. T.; Kataoka, S.; Albertorio, F.; Cremer, P. S. Direct writing of metal

nanoparticle films inside sealed microfluidic channels. Anal. Chem. 2006, 78, 107–112.

71. Farokhzad, O. C.; Khademhosseini, A.; Jon, S.; Hermmann, A.; Cheng, J.; Chin, C.;

Kiselyuk, A.; Teply, B.; Eng, G.; Langer, R. Microfluidic system for studying the

interaction of nanoparticles and microparticles with cells. Anal. Chem. 2005, 77,

5453–5459.

72. Goto, Y.; Matsuno, R.; Konno, T.; Takai, M.; Ishihara, K. Polymer nanoparticles covered

with phosphorylcholine groups and immobilizedwith antibody for high-affinity separation

of proteins. Biomacromolecules 2008, 9, 828–833.

73. Li, Y.; Chen, Y.; Xiang, R.; Ciuparu, D.; Pfefferle, L. D.; Horv�ath, C.; Wilkins, J. A.

Incorporationof single-wall carbonnanotubes intoanorganicpolymermonolithic stationary

phase form-HPLCand capillary electrochromatography.Anal.Chem. 2005, 77, 1398–1406.

74. Weng, X.; Bi, H.; Liu, B.; Kong, J. On-chip chiral separation based on bovine serum

albumin-conjugated carbon nanotubes as stationary phase in a microchannel.


75. Gong, M.; Wehmeyer, K. R.; Limbach, P. A.; Arias, F.; Heineman, W. R. On-line sample

preconcentration using field-amplified stacking injection in microchip capillary electro-

phoresis. Anal. Chem. 2006, 78, 3730.

76. Shiddiky, M. J. A.; Park, H.; Shim, Y.-B. Direct analysis of trace phenolics with

a microchip: in-channel sample preconcentration, separation, and electrochemical detec-

tion. Anal. Chem. 2006, 78, 6809–6817.

77. Shiddiky, M. J. A.; Shim, Y.-B. Trace analysis of DNA: preconcentration, separation, and

electrochemical detection in microchip electrophoresis using Au nanoparticles. Anal.

Chem. 2007, 79, 3724–3733.

78. Zeng, Y.; Harrison, D. J. Self-assembled colloidal arrays as three-dimensional nanofluidic

sieves for separation of biomolecules on microchips. Anal. Chem. 2007, 79, 2289–2295.

79. Giddings, J. C.; Unified Separation Science; Wiley: New York, 1991.

80. Kirk, J. S.; Sweedler, J. V.; Bohn, P. W. Nanofluidic injection and heterogeneous kinetics

of organomercaptan surface displacement reactions on colloidal gold in a microfluidic

stream. Anal. Chem. 2006, 78, 2335–2341.


81. Goluch, E.D.;Nam, J.-M.;Georganopoulou,D.G.;Chiesl, T.N.; Shaikh,K.A.;Ryu,K. S.;

Barron, A. E.; Mirkin, C. A.; Liu, C. A bio-barcode assay for on-chip attomolar-sensitivity

protein detection. Lab Chip 2006, 6, 1293–1299.

82. Taton, T. A.; Mirkin, C. A.; Letsinger, L. R.; Scanometric, DNA array detection with

nanoparticle probes. Science 2000, 289, 1757–1760.

83. Nie, Z.; Xu, S.; Seo,M.; Lewis, P. C.; Kumacheva, E. Polymer particleswith various shapes

and morphologies produced in continuous microfluidic reactors. J. Am. Chem. Soc. 2005,

127, 8058–8063.

84. Zhang, H.; Tumarkin, E.; Peerani, R.; Nie, Z.; Sullan, R.M.A.;Walker, G. C.; Kumacheva,

E. Microfluidic production of biopolymer microcapsules with controlled morphology.

J. Am. Chem. Soc. 2006, 128, 12205–12210.

85. Huang, K.-S.; Lai, T.-H.; Lin, T.-C. Manipulating the generation of Ca-alginate micro-

spheres using microfluidic channels as a carrier of gold nanoparticles. Lab Chip 2006, 6,

954–957.

86. Zourob, M.; Mohr, S.; Mayes, A. G.; Macaskill, A.; P�erez-Moral, N.; Fielden, P. R.;

Goddard, N. J. A micro-reactor for preparing uniform molecularly imprinted polymer

beads. Lab Chip 2006, 6, 296–301.

87. Rondeau, E.; Cooper-White, J. J. Biopolymer microparticle and nanoparticle formation

within a microfluidic device. Langmuir 2008, 24, 6937–6945.

88. Sarov,Y. E.; Capek, I.; Ivanov, T. B.; Ivanova, K. Z.; Sarova, V.A.; Rangelow, I.W.On total

internal reflection investigation of nanoparticles by integrated microfluidic system. Nano

Lett. 2008, 8, 375–38.

89. Dawoud, A. A.; Kawaguchi, T.; Jankowiak, R. In-channel modification of electrochemical

detector for the detection of bio-targets on microchip. Electrochem. Commun. 2007, 9,

1536–1541.

REFERENCES 253

7PILLARS AND PILLAR ARRAYSINTEGRATED IN MICROFLUIDICCHANNELS: FABRICATIONMETHODS AND APPLICATIONS INMOLECULAR AND CELL BIOLOGY

JIAN SHIECOLE NORMALE SUP�eRIEURE, PARIS, FRANCE

YONG CHEN

Institute for Integrated Cell-Material Sciences, Kyoto University, Kyoto, Japan

7.1 INTRODUCTION

Nanofabrication and microdevice technologies are now becoming more and more

important for the continuous improvement of our basic understanding in cell biology

and our ability in diagnosis, drug discovery, environment control, and clinical

treatment.1–3 Among many others, nanofabrication can be used to pattern functional

surfaces to control cell adhesion, migration, differentiation, and so on, whereas

microdevices particularly designed for the manipulation of small volume of liquid

samples, that is, microfluidic chips, can be used for highly efficient cell handling,

biomolecule separation, high-throughput screening, and so on. The reasons are

obvious: the sizes of the fabricated elements (patterned features and channels) not

only are comparable to the size of cells and large biomolecules but can also be well

designed for highly integrated microsystems.4–9 In addition, it will be possible to


255

produce such systems with high throughput and low cost, leading to a similar impact

generated by CMOS-based semiconductor industry during the last few decades.

Pillars are certainly the simplest and universal pattern elements. Nature uses

pillars for different purposes. At micrometer scale, for example, one can observe

dense pillar arrays on lotus leaves for self-cleaning, whereas more complicated

pillar structures can be found on the butterfly wings showing splendid colors for

seduction and/protection10,11 (Figure 7.1). With the rapid progress in nanofabrica-

tion tools, one can now easily produce high-density pillars of different sizes

and different materials, also showing biomimetic or designed functionalities. In

particular, patterned pillars can be used as artificial gels for biomolecule separation

or preconcentration. They can also be used as artificial extracellular matrix (ECM)

for controlling and/or monitoring cell growth. On the other hand, microfluidic chips

are ideal systems for the integration of pillar arrays, although only a few investiga-

tions have been reported previously.

Microfluidics itself is an emerging field with great promises.12–14 Indeed, the

alreadydemonstrated applicationpotential has attracted a continuousgrowing interest

FIGURE 7.1 Photograph of lotus (a) and butterfly (c), and scanning electronic micrograph

(SEM) of pillars on the lotus leave (b) and wings of the butterfly (d).10,11

256 PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

during the past 15 years. A large variety of functionalities can now be integrated for

molecular and cell biology studies. One can, for example, use integrated microelec-

trodes to trap, separate, or fuse cells.15–17 One can also use integrated fiber or

gratings for highly sensitive detection.18,19 The integrated pillars are often used as

barriers to trap cells or as filters to separate biomolecules. It should also be possible to

explore more functionalities based on their unique electric, magnetic, or mechanic

properties.

In this chapter, first we discuss pillars and pillar arrays integrated into microfluidic

chips. Then, fabricationmethods are reviewed and application examples are described

to illustrate different aspects of their utilities.

7.2 PATTERNING TECHNIQUES

Both top-down and bottom-up techniques can be used for the fabrication of high-

density pillars. The so-called “top-down” techniques have been issued from semicon-

ductor industry, where lithography and pattern transfer techniques play the most

essential role in the manufacturing of integrated circuits. By using standard photoli-

thography, for example, the designed features are transferred from a mask to a resist

layer coated on a substrate. Then, the features are produced in substrate by etching,

regrowth, doping, and liftoff to obtain desired functionalities. In contrast, the “bottom-

up” techniques are based on self-assembling at nanoscale or molecular levels.

Obviously, the two approaches can be combined and used for the fabrication of

highly integrated microdevices with enhanced functionalities.

7.2.1 Lithography and Related Techniques

Conventional lithography methods refer to optical lithography, electron beam lithog-

raphy, focused ion beam lithography, X-ray lithography, and extremeUV lithography

that are commonly used or studied by semiconductor industry and research laborato-

ries for different purposes. During the past decades, efforts have been extensively

devoted to the improvement of these methods in terms of resolution and throughput.

However, there is still no high-throughput method for themanufacturing of integrated

features of sizes down to 50 nm. Moreover, even for the fabrication of large features,

the ownership cost is excessively high and the required environment is not suitable for

different applications other than semiconductor industry. Electron beam lithography,

for instance, is a high-resolution technique based on sequential writing. Typically,

a high-energy (1–200 keV) electron beam is formed and projected on the samplewith

an electron beam optical column. The resolution of electron beam lithography is

limited by scattering of second electrons in the resist layer and substrate, providing a

feature size down to 10 nm.Although electron beam lithography is flexible and of high

resolution, it is not suitable formass production because of the limitation of its writing

speed. Similarly, focused ion beam lithography provides excellent resolution but low

processing speed. As high-energy ions penetrate thematerial, they lose their energy at

a rate several orders of magnitude higher than that of electrons because of their large

PATTERNING TECHNIQUES 257

mass values. For this reason, it is generally used for micromachining or localized

implantation.

Optical lithography is the commonmethod for pattern reproduction. The resolution

of projection optical lithography is defined by the Rayleigh criterion R¼ k1l/NA,wherel is thewavelength of the light,NA the numerical aperture of the optical system,

and k1 an empirical factor depending on details of experimental conditions. To

improve the resolution, sophisticated mask design, off-axial illumination, and top

surface image techniques are used, but most important is to utilize short-wavelength

exposure sources. The early projection systems worked with mercury lamp for

a wavelength range between 350 and 400 nm. Now, advanced production systems

use 248 or 193 nmwavelength radiation providedbykryptonfluoride or argonfluoride

excimer laser. To further improve the resolution, liquid is injected into thegap between

the focusing lens and the substrate to obtain a larger NA (immersion lithography).

Finally, extreme UV lithography has been studied for many years as candidate of the

so-called next-generation lithography because of its much reduced exposure wave-

length (11–13 nm). This technique is extremely complex and expensive, and it is not

clear yet when this technique will be applied for the mass production.

Unlike conventional methods, more recent techniques including soft lithography

and nanoimprint lithography are not based on the use of photon, electron, and ion

beams. Accordingly, these nonconventional methods are flexible, low cost, and

promising for a large variety of applications. Soft lithography has been proposed

in 1998 byWhitesides,20 which is based on the use of PDMS (polydimethylsiloxane)

as lithography template. Typically, it is used for microcontact printing where

molecules are released from template and then self-assembled in the contact area.

The printed features can be used for different purposes, but theymight be resistant for

wet etching. The resolution of microcontact printing is limited by the mechanical

stability of the template and the diffusion of inkingmolecules around the contact area.

In practice, features of sizes down to 200 nm can be obtained and it is possible to

achieve a higher resolution by using PDMS templates of improved stiffness and

molecules of smaller diffusion length. Actually, soft lithography refers to a large class

of patterning and device technologies that use PDMS features obtained by casting.

Nanoimprint lithography was proposed by Chou in 199521 and soon it became

a widespread method for both academic and industrial applications. Nanoimprint

lithography is based on pressure-induced deformationwith a high-resolutionmold. In

most cases, a resist layer deposited on substrate is patterned using amold fabricated by

electron beam lithography. After separation of themold, the resist pattern is treated by

reactive ion etching to remove the residual part of the imprinted area before further

processing. Nanoimprint lithography is intrinsically of high resolution, high through-

put, and low cost. No apparent limiting factor does exist, such as diffraction in optical

lithography or scattering in electron beam lithography. Nanoimprint tools can be

relatively simple since neither optical lens system for projection nor high-energy

column for beammanipulation is required. Finally, once amold is fabricated, it can be

used many times, which makes this technique attractive for low cost and mass

production. Currently, two types of nanoimprint methods coexist, that is, thermal

nanoimprint lithography initially proposed by Chou and UV nanoimprint lithography


(UV-NIL) that uses photocurable resin and UV light for solidification.22,23 UV

imprinting works at room temperature with a low pressure, which should be more

suitable for fine alignment compared to the thermal imprinting technique. For both

thermal andUV-based processes, themold fabrication is a critical issue because of the

high cost. Pattern replication over large area is also difficult because of the nonflatness

of wafers. Although large pressure can be applied, it is not reasonable to work with

fragile sample and fine aligning. One solution is to use a step-and-repeat strategy but

this may require a much more sophisticated imprinting tool. For these reasons, it is

interesting to use soft molds for a more conformable imprinting over large wafer

area.24,25 Soft molds such as PDMS can be easily produced by casting so that many

copies can be obtained at low cost. Furthermore, PDMS has low surface energy that

facilitates the mold separation after imprint. PDMS has, however, low Young�smodule that prevents the replication of both high resolution (<50 nm) and very large

features (>100 mm) because of the pressure-induced mechanic deformation.

Fortunately, pillar arrays of a typical feature size in the range between 50 nm and

100mm are useful for the most recent applications. Anyway, large features can be

easily replicated by photolithography and small features can be produced by electron

beam lithography. Then, mix-and-match techniques can be applied to cover all

feature sizes in a more cost-effectiveway.26 To illustrate the replication performance,

we show in Figure 7.2 a few examples of the fabrication ability of both thermal and

soft UV-NIL techniques.

FIGURE 7.2 SEM images of fabricated high-density dot and pillar arrays. (a and c) Nickel

dots (period 60 nm) and SiO2 pillars (period 300 nm) obtained by thermal nanoimprint

lithography; (b and d) PMMA pillars on silicon (diameter 120 nm) and in plexiglas (height

3mm) obtained by soft UV nanoimprint lithography.24–27

PATTERNING TECHNIQUES 259

7.2.2 Auto Assemblage

Nanopillars can also be produced by nonlithographic methods. One solution is to use

the so-called bottom-up approach where a functional material can be constructed by

self-assembly of elementary building blocks into ordered arrays. These building

blocks can be typically metallic or polymer nanoparticles. After assembling, the

pattern of the ordered arrays can be transferred to substrate by etch or other pattern

transfer techniques. Similarly, copolymer demixing can also be used for the fabrica-

tion of nanopillar arrays.

Monodispersed nanospheres made of silica, metal, and polymers are now com-

mercially available and a large number of investigations have already been reported on

their organization ontoflat or structured surfaces.Most of themethods used so farwere

basedoncapillary force assisted autoorganization. In the simplest casewhereadroplet

of the colloid suspension is dried slowly on an unpatterned polar surface, the particles

aggregate at the rim of the droplet because of the attractive capillary forces between

particles. Several methods including controlled deposition, dipping, and microfluidic

flowing have been proposed. By controlling more accurately the capillary force

and the evaporation, nanoparticles could be assembled on patterned substrates with

various geometries.27

Once highly ordered and large area particles are assembled, they can be used as

mask for etching or liftoff. When reactive ion etching technique is applied, the space

between the pillars can be controlled by adjusting the etch parameters. Otherwise,

pillar arrays of different periods can be obtained by using nanospheres of different

sizes.When liftoff technique is applied, a thinmetal layer is deposited and the resulted

hexagonal triangle pattern serves as etchmask for the substrate. By changing the angle

of metal deposition (shadow masking), other forms of patterned arrays could also be

obtained.28 Although both fabrication processes are simple, which can be used for a

number of applications, it is in general difficult to produce perfect and large surface

pillar arrays without defect.

Block copolymer demixing has been proposed to achieve ultrahigh-density

nanostructure patterning. A block copolymer consists of two chemically distinct

polymer chains covalently bound at one end. Because of the tendency of phase

separation for unlike chains and the constraint imposed by chain connectivity, the

block copolymer can be self-organized to form periodic domains on a molecular

length. In general, the competition between the interfacial and chain stretching

energies governs the bulk equilibrium phase behavior, and the relative volume

fractions of the blocks control the curvature, size, and the periodicity of the

nanodomains. In most cases, functionalized block copolymers can be used for

spontaneous organization on a flat substrate. However, the nanostructure formation

can be improved by using a patterned substrate. Other kinds of physical or

chemical means such as mechanic sliding and electric field induction can also

be used for the improvement of the pattern formation, resulting in regular line and

dot arrays with a feature size down to a few nanometers. Finally, the self-organized

copolymers can be used for selective ion etching or liftoff for further device

processing.


7.2.3 Growth

Chemical vapor deposition (CVD), physical vapor deposition (PVD), andmany other

derived deposition methods can be used for the fabrication of small features on

a substrate. Indeed, with or without mask, a large variety of materials could be

deposited as monocrystalline, polycrystalline, or amorphous, including silicon,

silicon–germanium, silicon dioxide, silicon carbide, silicon nitride and silicon

oxynitride, carbon fiber and carbon nanotubes, tungsten, titanium nitride, and various

high-k dielectrics. Electroplating using either electrolytic or electroless deposition

techniques can be applied to obtain high-density metallic pillars. With electrolytic

deposition, the surface on which the metal is to be deposited is used as cathode and

deposition occurs when themetal ions travel under the influence of the electric field in

the aqueous solution ofmetal salts. ADC voltage is applied between the substrate and

the metal to be deposited. With a patterned substrate, it is important to accurately

control the current density that strongly influences the deposition rate, plating

adherence, and plating quality. It is also important to note that in the case of high-

density nanopillar growth, the current density cannot be homogeneous over the plating

surface. In general, the current density can be significantly larger at the edge than in the

center of a patterned area.

7.3 OTHER FABRICATION ASPECTS

Depending on the application domain and the manufacturing cost, materials of

different types can be used. Among them, silicon-based materials and polymers

arewidely used because of their well-known physical and chemical properties and the

availability of their processing parameters. In many cases, surface fuctionalization is

needed to enhance the functionality of the fabricated pillar arrays.

7.3.1 Material Choice

Silicon and silicon-based materials are mainly used for the production of integrated

circuits by semiconductor industry. They are also commonly used for themanufactur-

ing of microelectromechanical systems (MEMS). More recently, silicon-based

materials are studied for the fabrication of integrated optoelectronic devices.

Silicon nanopillars can be easily obtained by lithography and reactive ion-etching

techniques, which explains why most nanopillars fabricated so far were based on

silicon processing. A thin layer of silicon dioxide can be easily obtained on the surface

of silicon by native or controlled oxidation, which is thermally stable and chemically

resistant to organic solvents and acids. Then, surface fuctionalization can be easily

done for different molecules. Compared to the polymers, silicon-based materials are,

however, more expensive and their batch processing sequences are in general more

complicated than the one-step replicationmethods. Finally, silicon is not desirable for

in vivo applications and it is not suited for applications such as electrophoresis with

a high electric field, observation with an inverted light microscope, and so on.

OTHER FABRICATION ASPECTS 261

The fabrication of polymer pillars can be simple and cost effective using

nonconventional techniques. Another advantage of using polymers is the availabil-

ity of a broad range of material choices, including PMMA, PS, PC, cyclic olefin,

elastomer, and others. In practice, the material choice depends on many factors such

as the processing ability, cost, optical transparency, mechanic stiffness, and

dielectric properties. When applied to microfluidics, appropriate sealing methods,

such as lamination, ultrasonic welding, thermal bonding, have to be carefully

studied to achieve the desired performance. Due to the limitation of the polymers�thermal stability and the stability against organic solvents, the fabricated pillars can

only be operated at temperature below the glass transition temperature and they

have to avoid many types of organic solvents also. Finally, a surface treatment is

generally required for microfluidic application since most of the polymers are not

hydrophilic.

Materials of other types are also studied to demonstrate more specific functionali-

ties. For example, metallic pillar structures have shown unique electrical, magnetic,

and optical properties. Gold pillar arrays can be used, in particular, for label-free

detection because of much enhanced surface plasmon effect. Although magnetic

nanopillars are generally considered as media for ultrahigh-density data recording,

they can also be used for bioprocessing or bioanalyses. Semiconductor nanopillars of

ZnOhave shownnot only interesting piezoelectric properties but also the usefulness in

the studyofUV lasers, dye-sensitized solar cells, antireflection coating, and soon.29–31

Therefore, functional materials should be more attractive than silicon and polymers,

and fortunately they can also be patterned and integrated into microdevices made of

silicon or polymers.

7.3.2 Surface Fuctionalization

To endow the nanopillars with more specific applications, one can apply different

physical or chemical methods to treat the surface of pillars made of silicon, silicon

dioxide, or polymers. In general, a treated surface has to be stable in time and the

appliedmethod has to be compatiblewith thewhole process of device fabrication. It is

also desirable to have a high resolution when a local surface modification is applied.

Typically, a hydrophilic surface treatment improves the device aqueous solubility and

the device biocompatibility. When appropriate biomolecules are bound to the pillars�surfaces, the fabricateddevices canbeused forbiosensing, sorting, or purification.The

surface treatment of pillars can be performedusing avariety of techniques for different

purposes. Physical modification can be done by CVD, PVD, spin coating, solution

cast, and plasma processes, whereas chemical modification can be obtained by

grafting and self-assembly. Both techniques can be applied for the improvement of

device biocompatibility and the reduction or elimination of solute interactions with

device surfaces. They can also be used for the modification of electroosmotic flow

and surface immobilization of reactive biomolecules such as enzymes, antibodies,

proteins, DNA, and so on. Finally, they can be used as sieving matrices in separation

devices.


7.3.3 Integration into Microdevices

Integration of pillar arrays is a critical issue for the functional device fabrication for

which different methods are studied. Conventional lithography methods such as

electron beam lithography and focused ion beam lithography can be used for device

prototyping, but they are not cost-effective and cannot be used for large-scale

manufacturing. Nanoimprint lithography and self-assembling are able to pattern

nanostructures cost-effectively, but it is more difficult to use them for the fabrication

of complex andmultilevel devices.One solution is to combine bothmethods, that is, to

produce nanopillar arrays by nanoimprint lithography or self-organized assembling

and other large features (cavities, channels, posts, etc.) by photolithography and soft

lithography.High-density nanopillars could be integrated intomicrofluidic devices by

in situ polymerization cast molding.32 The more general approach can be based on

“mix-and-match” lithography methods.26 Indeed, we obtained high-density nano-

pillars by soft UV nanoimprint lithography and microfluidic cross-channels and

reservoirs by standard optical lithography.As shown in Figure 7.3, softUV-NIL is first

applied. After the first liftoff with a thin film of nickel, larger features such as

microfluidic channels are defined by photolithography and the second liftoff of a thin

layer of nickel. Then, the substrate is etched by reactive ion etching and high-density

and high aspect ratio nanopillars are obtained simultaneously with microfluidic

channels. Finally, the device is sealed by a PDMS cover plate with connection holes.

Clearly, thismix-and-match approach has the following advantages: (1) It is easy to

control the nanoimprint process so that high-quality pattern definition can be achieved

without considering particular mold and process design. (2) It is easy to align

microfluidic channel pattern with the nanopillar areas, since the excess areas of

the nanopillar arrays will be simply masked by Ni layer of the second liftoff. (3) It is

FIGURE 7.3 Mix-and-match fabrication process for the integration of pillar arrays into a

microchannel.

OTHER FABRICATION ASPECTS 263

easy to cover the patterned structures with a flat PDMS plate, and there will be no

leakage problem since the nanopillar areas are completely integrated into the channel

with the same etched height. In a particular case, we demonstrated large DNA

separation by using this kind of microfluidic device. Obviously, the mix-and-match

approach is applicable for the fabrication of other types of microdevices with

integrated nanopatterns.

7.4 APPLICATION EXAMPLES

Pillar arrays can be used for different purposes in cell biology. Although their whole

application potential cannot be predicted at the present stage, we select the following

examples to illustrate the current efforts on several aspects without paying particular

attention on the pillars� geometric design or their device configuration. Indeed, all

fabricated patterns should be scalable and all pillar arrays could be integrated into

microfluidic devices.

7.4.1 Pillars for Genomics and Proteomics

Genomic and proteomic information is of great importance in cell biology, cancer

diagnostics, and drug discovery. Although the human genomics project was finished

earlier than expected, there are still remaining challenges for both fundamental

research and clinic uses. One of the challenges is the fabrication of miniaturized

biomedical devices for high-quality and high-throughput analyses. In such a context,

nanopillar arrays integrated into microfluidic devices hold a great promise for

improved genomic and proteomic investigation.

7.4.1.1 Biomolecule PreconcentrationOne critical issue in genomics and proteomics is the sample preparation. It is known

that more than 10,000 different biomolecule species with concentrations varying over

nine orders of magnitude exist in a typical blood sample. Such diversities of

molecules, as well as their huge differences in concentration, cause a big problem

for the sample preparation. In general, extracted and purified genomic materials are

limited in both concentration and volume, which makes it difficult for both analysis

processing and detection. For proteomics, advanced detection methods such as laser-

induced fluorescent and mass spectroscopy are often required. This issue is not

exceptional for the microfluidic analysis systems, although it is able to handle and

manipulate the liquid sample in pL–nL scale with high efficiency. Besides, in

proteomics, this problem is exacerbated by the fact that information-rich signaling

molecules are present only in trace concentrations (nM–pM range). Furthermore,

there is no signal amplification technique, such as polymerase chain reaction (PCR),

for proteins and peptides. Efficient tools for sample amplification or preconcentration

are required for the manipulation and detection of highly diluted analytes in extreme

small volumes. For this particular application, the integrated nanopillar arrays may

have some advantages.


Silicon microstructures with high surface to volume ratios have been used for

capturing, washing, and eluting short- (500 bp) and medium-size (48,000 bp) DNA.33

In this study, chaotropic (GuHCl) salt solutions were used as binding agents, whereas

ethanol-based solutions andwater were used as wash and elution agents, respectively.

DNA quantities approaching 40 ng cm�2 of binding area were captured from input

solutions in the 100–1000 ngmL�1 concentration range. For dilute samples of interest

for pathogen detection, PCR and gel electrophoresis were used to demonstrate

extraction efficiencies of about 50%, and concentration factors of about 10� using

bacteriophage l DNA as the target.

Alternatively, the ability to purify single-stranded DNA (ssDNA) sequencing

ladders has been demonstrated by using microfluidics-based solid-phase and revers-

ible immobilization technology.34 Here, hot embossing has been used for the

fabrication of polycarbonate (PC) containing microposts with large surface areas.

An immobilization bedwas prepared by exposing PC surfaces to UV radiation, which

resulted in the formation of surface carboxylate groups through a photooxidation

reaction. Such functionalities can serve as a capture medium for a wide range of

molecular weight-sized DNAs for the isolation, preconcentration, or purification of

DNAembedded in complex samplematrices. As observed byUVspectroscopy, a load

of 7.6� 1.6mgmL�1 of gDNAwas immobilized onto the PPC bed. The recovery of

DNA following purification was estimated to be 85� 5%.

Considerable efforts have also been made to the problem of microfluidics-based

protein preconcentration. In particular, porous silica membranes were used to pre-

concentrate protein samples prior to electrophoretic separations.35 The preconcen-

tration rate in such devices, however, could not be easily predicted because of the

problems of membrane clogging and aging.Most of the recent works have focused on

nanochannel-based preconcentration. It is known that co-ions of the surface charges

can be excluded from the nanochannel while counterions are enriched. If an electric

field is applied, the enhanced counterionic transport can be used for electroprecon-

centration of charged biomolecules. Four preconcentration regimes can then be

distinguished, and results suggested that both the mobility and the valence of the

species are important parameters in the determination of the preconcentration rates.36

Alternatively, protein samples can be prepared by isoelectrical focusing (IEF) with

patterned substrates.37 By using PDMS pillar arrays fabricated by soft lithography,

a proteinmixturewith pI ranging from4.7 to 10.6 has been successfully separatedwith

good resolution. Compared to the classical gel-based system, this method consider-

ably reduces the separation time from several hours to 10min. The pillar chips can be

reused several times while classical gels are disposable.

7.4.1.2 DNA Stretching and SeparationSeparation of DNA fragments by size is at the heart of genome mapping and

sequencing. It can substantially enhance the capabilities of diagnosis, pharmacoge-

netics, and forensic tests. Compared to the commonly used gel-based electrophoresis,

microfluidic channel electrophoresis provides clear advantages of improved separa-

tion speed, reduced consumption of reagents, and ease of automation.38–40However, it

is difficult to introduce high-viscous polymer solution intomicroscale channel anduse

APPLICATION EXAMPLES 265

such simple device for long DNA separation. Microfluidic devices with integrated

pillar arrays can then be used. On the one hand, using nanopillars obviates the

introduction of gel matrix. On the other hand, during electrophoresis, DNAmolecules

will not be trapped by nanopillar arrays as that happened in gel matrix. With such

a device, new separation mechanisms can be studied in correlation with theoretical

modeling.

The conformation change of long DNA molecules plays an important role during

electrophoretic separation. Normally, a long DNA molecule in its relax state has

a spherical shape. When migrating under electric field in nanopillars� sieving matrix,

the long DNA molecules are stretched. This deformation is not entropically favored

and the stretchedDNAmolecules try to escape from the trap. Indeed, real-time images

of a single T4 DNAmolecule in the nanopillar region under an electric field show the

step-by-step changes in conformation of theDNAmolecule from spherical to linear.41

In the same nanopillar arrays, themigrationmobility of different longDNAmolecules

is different. By varying the size and geometry of the nanopillar arrays, one can easily

adjust the migration speed of long DNAmolecules and hence separate them in a short

time limit.

By using electron beam lithography and reactive ion etching, Kaji et al. have

integrated high aspect ratio (100–500 nm diameter and 500–5000 nm tall) nanopillars

inside a microchannel42 (Figure 7.4). Then, DNA fragments of 1–38 kbp were

separated into clear bands in a detection window of 1450mm from the entrance of

the nanopillar channel (25mm in width and 2.7mm in height) in only 170 s. The

technique has also been applied to the separation of long DNA molecules (l-phageDNA: 48 kbp, T4 DNA: 165.6 kbp) in less than 30 s under a DC electric field.

By using nanoimprint lithography, it is also possible to produce some devices.

Thermal nanoimprint lithographyhas beenfirst used for the fabricationof a pillar array

FIGURE7.4 (a and b) Photography of amicrofluidic chip for electrophoresis. (c) SEM image

of integrated quartz pillars. (d and e) Fluorescence images of migrating single l DNA and T4

DNAmolecules in the pillar region. (f) Electropherograms of lDNA and T4 DNA recorded at

two detection points.42


of 150 nm diameter and 320 nm period into microfluidic channels.43 Alternatively,

a nanoembossing technique has been developed to demonstrate the feasibility of one-

step replication of nanopillar arrays and microfluidic channel for DNA molecule

separation.44 More recently, a highly parallel and mix-and-match fabrication method

based on soft UV nanoimprint lithography and contact photolithography has been

proposed.26 Two pillar arrays with same period but different pore size are integrated

intomicrochannels. Single DNAmoleculemigration behavior has been demonstrated

and, consequently, l DNA and T4 DNAwere successfully separated in a couple of

minutes. The advantage of this method relies on the much improved process latitude

of the soft UV-NIL and easy implementation of whole fabrication steps. It is also

versatile and relevant to manufacture this type of nanodevice at low cost and high

throughput.

7.4.1.3 DNA and Protein ArraysCurrently, DNA and protein arrays are widely used for different purposes. For

example, the microarray chips are used to obtain new insights into how genes and

proteins work and how they are linked to the disease. They are also used for cancer

diagnostic and disease analyses. To improve the performance of themicroarray chips,

one can, for instance, increase the density of the array, increase the signal to noise ratio,

and develop technologies based on “lab-on-chip.”

Among other possibilities, surface patterning can be used to increase both array

density and signal to noise ratio. Previously, we performed a proof of concept

experiment by integrating high aspect ratio and high-density nanopillars in the

hybridization zones of a microarray chip, all surrounded by superhydrophobic

surfaces. The nanopillars were produced by soft nanoimprint lithography and reactive

ion-etching techniques. Then, by using the protocol described in Ref. 45, IgG

molecules could be immobilized on the surfaces of nanopillars and fluorescent-

labeled antigens could be easily attached through the bioaffinity interaction between

antibody and antigen. As a result, the nanopillar areas show a higher fluorescent

intensity than that in the surrounding areas (Figure 7.5a). By changing the size and

FIGURE 7.5 (a) SEM images of a SiO2 pillar array of 800 nm period and 1.6mm height

obtained by soft UV nanoimprint lithography. (b) Fluorescence image (lower) and intensity

profile (upper) of top view of a microchannel with integrated pillars, observed after a surface-

dependent chemical reaction.47


height of the nanopillars, a linear dependence of the fluorescence intensity versus the

effective surface could be obtained (Figure 7.5b). To increase the detection sensitivity,

the surrounding areas were also patterned, which become superhydrophobic to

prevent the undesirable deposition or the cross-contamination.46 For demonstration,

the fabricateddevice has beenused to show that thepresenceof the high-density pillars

and the superhydrophobic areas improved both the bioreaction efficiency and the

detection sensibility. The surface-dependent enzymatic reaction has also been studied

in microfluidic channels. Figure 7.5c shows that increasing the effective surface per

unit area (by etching) leads to an increase of fluorescence intensity,which corresponds

to the increase of the total amount of the enzyme attached to the microfluidic channel

wall.47

7.4.2 Pillars for Cellomics

Cells are the elementary building blocks of mammalians and many other living

systems. For in vitro studies, cells can be cultured, proliferated, and differentiated in

a culture dish or on other types of supports. In particular, patterned substrates can be

used for the control of cell organization and cell kinetics.

The size of a cell is typically several tens of micrometers so that their growth

behavior is largely influenced by the topographic or chemical patterns of the

comparable sizes. It is also known that many key functionalities of cells are also

regulated by nanoscale complexes of subcellular systems. For example, the building

blocks of both cytoskeleton (microtubules and actin filaments) and extracellular

matrix have dimensions in the range of 10 nm.Therefore, it is interesting to investigate

the influence of synthetic surface made of high-resolution patterns. Knowing that

patternedmicro- andnanostructures can be easily integrated intomicrofluidic devices,

a large variety of cell culture and manipulation techniques such as on-chip cell

sampling, trapping, sorting, characterization, and so on can be developed.

7.4.2.1 Cell TrappingPillar arrays can be used for size-dependent and/or affinity-dependent cell trapping.

Nagrath et al. presented isolation of circulating tumor cells (CTCs) in cancer patients

with a simple microfluidic platform embedded with pillar arrays (Figure 7.6).48 The

pillar arrays are first coatedwith antiepithelial cell adhesionmolecule antibody.Under

precisely controlled laminar flow conditions, viable CTCs can be isolated from

peripheral whole blood samples, mediated by the interaction of target CTCs with

antibody without requisite of prelabeling or processing of samples. Indeed, CTCs in

the peripheral blood of patients with metastatic lung, prostate, pancreatic, breast, and

colon cancer can be trapped in 115 of 116 (99%) samples with a concentration in the

rangeof 5–1281CTCspermilliliter and50%purity. In addition,CTCswere isolated in

7/7 patients with early stage prostate cancer.

Magnetic cell trapping can be done by using magnetic beads, also mediated by

antibody–antigen interaction. By using microfluidic devices with patterned magnetic

microcolumns, cancer cells could be most efficiently trapped.49 In this study,

a hexagonal array of nickel pillars was fabricated by standard photolithography


and electroplating. Then, they were integrated into microfluidic channels. Using

a solenoid coil with an iron core, a largemagnetic field gradient could be generated in

the pillar areas so that supermagnetic beads were trapped reversibly. After an in situ

surface modification with wheat germ agglutinin, A549 cancer cells were effectively

captured from cell mixtures with trapping efficiency in the range from 62% to 74%.

Cell trapping can also be done based on dielectrophoretic interaction. More

generally, the dielectrophoretic forces allow manipulating single cell by changing

the voltage, frequency, or phase of the electrical signal applied to the electrodes of the

devices. For example, a negative dielectrophoretic (nDEP) force can be applied to

generate traps at the stagnation points of cylindrical pillars arranged in a regular

array.50 By carefully controlling the dielectrophoretic and hydrodynamic forces, both

single-particle traps (capable of discriminating particles based on size) and multipar-

ticle traps (capable of controlling the number of particles trapped) could be achieved

with high precision. Although most of the previous studies were performed with

polystyrene beads, such a trapping mechanism is certainly applicable for cells.

7.4.2.2 Cell CultivationNot surprisingly, the most recent works of cell cultivation on patterned surface were

motivated by tissue engineering, regenerative medicine, and implantable medical

devices. It now becomes also important to integrate them into microfluidic chips. The

advantage of using microfluidic devices for cell culture is multifold. Apart from the

most evident facts of single-cell manipulation and high-throughput screening with

much reduced samples and biochemical reagents, microfluidics allows the creation of

a controllable and biomimetic microenvironment with high spatial and temporal

resolution. In addition to the control of soluble cell factor, microfluidic devices also

allow to integrate different types of materials and patterns as extracellular matrix. In

FIGURE 7.6 Pillar-based trapping of circulating tumor cells: (a) the workstation setup,

showing the manifold housing of CTC chip, a rocker for continuous mixing, and a pneumatic

pressure-regulated pump for controlling the sample flow; (b) SEM image of a captured

NCI-H1650 lung cancer cell spiked into blood. The inset shows a high magnification view

of the cell.48


particular, nanopillars cannowbeproduced at lowcost andhigh throughput, providing

anewdimension for enhancedcell adhesion, proliferation, anddifferentiation.Finally,

hybrid micro-, nano-, and chemical patterns are useful for the controlled formation of

cellular networks.

For example, Turner et al. have studied attachment of astroglial cells on smooth

siliconandsiliconpillarsandstripswithvariouswidthandseparations.51Fluorescence,

reflectance, and confocal light microscopes, as well as scanning electron microscopy,

were used to observe the cell morphology and the distribution of cytoskeletal proteins

suchasactin andvinculinondifferent surfaces (Figure7.7). It turnedout that bothactin

and vinculin distributions were highly polarized when cells were placed on the pillar

arrays. Scanning electron microscopy clearly demonstrated that cells made contact

with the top of the surface pattern and they did not reach the bottom even when the

patterned features were separated from each other with a distance as large as 5.0mm.

These experiments support the use of surface topography to direct the attachment,

growth, and morphology of cells.

Investigations on other types of materials also showed the very significant

differences of the cell behaviors on flat and patterned surfaces, indicating that the

conventional dish-based (two-dimensional) culture will not be sufficient for under-

standing or generating mimetic (three-dimensional) tissues. Nomura et al. have

reported cell culture on nanopillar arrays.52 By using nanoimprint lithography,

nanopillars could be produced as a new type of cell culture medium. Then, HeLa

cells were cultured and analyzed, showing that the cell division and proliferation on

nanopillars were significantly different from that found with culture Petri dishes. In

addition, the use of nanopillar substrates allows an easy subculture without conven-

tional trypsinization.

More recently, Kim et al. used nanopillar arrays made of a poly(ethylene glycol)

(PEG) hydrogel to guide a 3D construct of primary rat cardiomyocytes.53 They

observed that thePEGnanopillars not only guided the extensionof cellmembranes but

also led to the 3D growth of cardiomyocytes with a new topographical guiding

mechanism. In colonizing cardiomyocytes, the PEG nanopillars stimulated self-

assembled aggregates among the contacting cells, in comparison to those on the

bare PEG and the glass control. Themyocytes cultured on the PEG nanostructure also

FIGURE 7.7 LRM55 cells after 6 h grown on strips of pillars. (a) Reflectance image,

(b) fluorescence image, and (c) merged image of (a) and (b) (scale bar: 50 mm).51


exhibited substantial beating capability with higher amplitude when compared to

those cultured on controls (Figure 7.8).

Finally, protein patterns can be integrated into microfluidic chambers for cell

attachment at single-cell level and long-term cell culture.54 After patterning and

device integration, cells were seeded in and cultured for more than one week. In

comparison to more conventional culture, significant changes of the cell morphology

andprotein expressioncouldbeobserved: cells can formhighly elongated features that

cannot be observed with a protein pattern or in a microchamber alone.

7.4.2.3 Cell CharacterizationIn living systems, cells adhere and interact with the extracellular matrix via cell–cell,

cell–materials contacts. They can sense and respond to the signal arising in their local

environment through ion channels and receptors present in their membranes, which

are often associatedwith groups of proteins linked to the cytoskeleton. To obtainmore

clear insights into the signaling pathways activated by the membrane proteins, new

functionalities are integrated into microfluidic devices. These integrated functionali-

ties can be used not only to monitor the microenvironment or the inner parameters of

cells but also to generate local stimulations for more precise determination of cellular

reactivity.

For example, integrated microelectrodes are used for analysis of intercellular or

intracellular parameters such as pH, conductivity and the concentration of signaling

molecules, and the electrochemical signature of peroxynitrite oxidation.55 For

chemical sensing, cell adhesion behaviors could be regulated by mono- or mixed

chemical patterns defined by different patterning techniques.56–58 In addition, the

pattern size could be turned frommicrometers to nanometers. For cellular mechanical

properties, different techniques such as magnetic tweezers, atomic microscopy,

microplates, and so on are used.59–61 Pillar arrays are also used for cell mechanical

sensing. Elastomeric pillars made of PDMS are first used to measure the cell traction

forces byTan et al. (Figure 7.9).Based on the knownphysical parameters of the PDMS

FIGURE 7.8 (a) SEM of aggregated cardiomyocytes cultured on PEG nanopillars, (b and c)

electric properties of rat cardiomyocytes cultured on the pillar array and glass using the whole

cell patch clamp technique (scale bars: 20mV/100ms). (d and e) Representative trace of action

potentials recorded with PEG pillar substrate and glass (scale bars: 20mV/2 s).53


pillars, deflections can be measured and forces can be derived more accurately. Then,

they found that the intracellular force generated in a cell varied with cell spreading

such that well spread cells exerted more average force per post than the less spread

counterparts. They also confirmed earlier studies that the magnitude of the force

exerted by cells correlated with the size of adhesion formed by attaching cells.62 By

using the same technique, Roure et al. measured dynamic traction forces exerted by

epithelial cells migration.63 In these two examples, the pillars are used as probe to

measure the forces that were exerted by the cells. Compared to the microbeads or

cantilever-based sensing, the pillar-based sensing is simpler and more accurate,

allowing the determination over a large range the forces exerted on cells. Pillars

can also be used to generate forces on cells to mimic the forces encountered in vivo.

Sniadecki et al. employed magnetic nanowires embedded in the PDMS pillars to

activate and measure the traction forces in cells.64 External forces up to 45 nN were

achieved with a uniformmagnetic field. Also, it was verified that the magnetic pillars

can be driven at frequencies up to 5Hz without significant damping effects.

FIGURE 7.9 (a) SEM of a smooth muscle cell attached to an array of posts uniformly coated

with fibronectin. (b) Schematic ofmicrocontact printing of protein. (c) Differential interference

contrast (upper) and immunofluorescence (lower) graphs of the same region, only a 2� 2 array

of posts printed with fibronectin. (d) SEM of a smooth muscle cell attached to posts where only

the tops of the posts have been printed with fibronectin. (e–g) Confocal images of immunoflu-

orescence staining of a smooth muscle cell on posts. The force exerted by cells (white arrows)

was calculated through the position change of the posts. The force map was spatially correlated

to immunofluorescence localization of the focal adhesion protein vinculin. (h) Plot of the force

generated on each post as a function of total area of focal adhesion staining per post (scale bars

indicate 10mm).62


7.4.3 Pillars for Biosensing

In biomedical research, healthcare, and environmental monitoring, analysis tools are

generally expensive but not enough flexible. Microfluidic devices are developed for

fast and cost-effective analyses. Then, the integration of different sensing elements

becomes important and many efforts have been devoted to this particular area.

Optofluidics, for example, has been developed by integrating both active and passive

optical elements into microfluidic chips.65 It is known that fluorescence detection is

the most sensitive one but it requires specified molecular labeling. The interaction

between the target and recognition molecules could be characterized by the intensity

of the fluorescence.Although the detection limit can reach a single-molecule level, the

contaminating materials in the sample can frequently interfere with the detection

system. For this reason, highly sensitive, label-free detection is desired for both in vivo

and in vitro analyses. As nanopillar arrays can be designed to have very particular

electromagnetic or optical properties, they can be used to improve the sensibility

of label-free detection in a microfluidic device. Among many others, three types of

substrates are nowwidely studied for improved label-free detection, including surface

plasmon detection, surface enhanced Raman spectroscopy (SERS), and photonic

crystal-based analyses, all relying on the fabrication of high-density nanopatterns.

For surface plasmon-based sensors, Chen and Jiang introduced hybrid arrays of

metallic nanostructures obtained by depositing a silver film on the fused silica

nanopillars.66 Such a structure was used to monitor the evaporation process of the

absolute ethanol on the sample surface, showing two distinct peaks in the extinction

spectrum of the p-polarized incident light, one at 585.3 nm and the other at 493.6 nm.

With the addition of absolute ethanol on the sample surface, a redshift of 32.9 nmwas

observed for the higher peak while a blueshift of 42.3 nm was observed for the lower

peak. It was also found that narrowest extinction peak appeared at normal incidence,

while the polarization of the incident light did not affect the experimental result due to

the symmetrical distribution of the nanostructures. SERS-based biosensing also

depends on the shape of individual pillars. In general, sharp corners provide more

pronounced SERS signals.

Raman spectroscopy is also an ideal label-free optical detection technique for

chemical and biomolecules. Lee and Liu described a low-cost and ultrasensitive

substrate for SERS measurement in a microfluidic device (Figure 7.10).67 PDMS

nanopillars were fabricated by soft lithography. After selective deposition of Ag thin

film, the pillars were integrated into a glass-based microfluidic chip with a suitable

opticalwindow forSERSspectroscopic imaging.Rhodamine 6GandadenosineSERS

spectra were then obtained by using a 785 nmwavelength laser excitation. As a result,

the observed Raman scattering signal enhancement on the nanopillar-basedAg SERS

substrate is more than 107 times higher than the control sample.

Photonic crystals have recently been demonstrated as optical biosensors. Due to

the highly localized confinement of the coupled light, photonic crystal sensors can be

incorporated into microfluidic devices to facilitate localized measurements of the

change in refractive index.Wu et al. presented the concept of using three-dimensional

photonic crystals for refractive index sensing in a microfluidic channel.68 It was


demonstrated that a change in the refractive indexof the fluid in amicrochannel results

in a shift in the band gap or band gap defect position of the photonic crystal. According

to Fourier transform infrared spectroscopy of the photonic crystal sensor, a change of

6� 10�3 in the refractive index of the fluid can be detected.

7.4.4 Other Applications

Pillar arrays integratedmicrofluidic chips can havemany other applications. Even for

the most explored domain, only a few aspects have been investigated. Sometimes,

one can use the same material and the same design to study different properties of

a fabricated pillar array. For example, a nanopillar arraymade of zinc oxide (ZnO) can

beused for optical detection, piezoelectric sensing, field emission, and so on.Owing to

its high chemical stability, Bie et al. used ZnO nanopillars as gas sensor for hydrogen

and ethanol.70 Their samples exhibited large responses of 18.29 and 10.41–100 ppm

ethanol and hydrogen, respectively. Alternatively, Pradhan et al. obtained ZnO

nanopillars by electrodeposition on indium tin oxide (ITO) glass substrates without

any template catalyst or seed layer.69 They showed that these pillars can be used to

achieve an excellent field emission performance, with a low turn-on electric field of

3.2Vmm�1 for 1.0mAcm�2 and a threshold field of 6.6 V mm�1 for 1.0mA cm�2.

FIGURE 7.10 (a) Schematic diagram of the chip with integrated patterns and the Raman

imaging system. (b and c)Optical image of amicrochannelwith a laser focal spot illuminated on

the surface of Ag/PDMS substrate with and without patterns. (d) SERS spectra of 1mM R6G

molecules taken on the area with (upper) or without (lower) patterns (1 s integration time).

(e) SERS spectra of, from top to bottom, 1mM, 10 nM, 100 pM, 1 pM, and 10 fM adenosine

molecules taken from the Ag/PDMS nanowell SERS substrate, DI water, and the Raman

spectrum of 10mM adenosine molecules taken from the smooth Ag/PDMS substrate,

respectively.67


7.5 CONCLUSION AND FUTURE OUTLOOK

We have attempted to review the current research topics on pillars and pillar arrays

integrated into microfluidic devices. Examples were given to illustrate their high

potential in genomics, proteomics, and cellomics studies andwe expect a fast growing

interest in this emerging field. From the fabrication point of view, new materials and

newsurface protocols can be used to enhance the functionality of pillars and integrated

pillar arrays. From the application point of view, new topics are to be explored for

both fundamental research and commercial purposes. With rapid progress in cell

and molecular biology, nanofabrication and microdevice technologies will be spread

as a general undertaking in coming years.

ACKNOWLEDGMENTS

Thisworkwas partially supported the EuropeanCommission through project contract

NMP4-CT-2003-505311 (Nabis) and project contract CP-FP 214566-2 (Nanoscales).

The authors would also like to thank colleagues and students of CNRS-LPN

(Marcoussis, France) and ENS (Paris, France) for collaboration and assistance.

REFERENCES

1. Roco,M.C.;Williams, R.S.; Alivisators, P. Biological, medical and health applications. In

Nanotechnology Research Directions; Kluwer Academic Publishers: 2000; Chapter 8.

2. Niemeyer, C.M.; Mirkin, C.A. Nanobiotechnology: Concepts, Applications and

Perspectives, Wiley: 2004.

3. Roco, M.C. Nanotechnology: convergence with modern biology and medicine. Curr.

Opin. Biotechnol. 2003, 14, 337–346.

4. Whitesides, G.M. The “right size” in nanobiotechnology. Nat. Biotechnol. 2003, 18,

760–763.

5. Fortina, P.; Krick, L.J.; Surrey, S.; Grodzinski, P. Nanobiotechnology: the promise

and reality of new approaches to molecular recognition. Trends Biotechnol. 2005, 23,

168–173.

6. Ball, P. Natural strategies for themolecular engineer.Nanotechnology 2002, 13, R15–R28.

7. Huang, S. Gene expression profiling, genetic networks, and cellular states: an integrating

concept for tumorigenesis and drug discovery. J. Mol. Med. 1999, 77, 469–480.

8. Ito, Y. Surface micropatterning to regulate cell functions. Biomaterials 1999, 20,

2333–2342.

9. Girard, P.P.; Cavalcanti-Adam, E.A.; Kemkemer, R.; Spatz, J.P. Cellular chemomechanics

at interfaces: sensing, integration and response. Soft Matter 2007, 3, 307–326.

10. Sun, M.H.; Luo, C.X.; Xu, L.P.; Ji, H.; Qi, O.Y.; Yu, D.P.; Chen, Y. Artificial lotus leaf by

nanocasting. Langmuir 2005, 21, 8978–8981.

11. Kert�esz, K.; B�alint, Z.; V�ertesy, Z.; M�ark, G.I.; Lousse, V.; Vigneron, J.P.; Rassart, M.;

Biro, L.P.; Phys. Rev. E 2006, 74, 1539–3755.

REFERENCES 275

12. Duffy, D.C.; McDonald, J.C.; Schueller, O.J.A.; Whitesides, G.M. Rapid prototyping of

microfluidic systems in poly(dimethylsiloxane). Anal. Chem. 1998, 70, 4974–4984.

13. Squires, T.M.; Quakes, S.R. Microfluidics: fluid physics at the nanoliter scale. Rev. Mod.

Phys. 2005, 77, 977–1026.

14. Unger, M.A.; Chou, H.P.; Thorsen, T.; Scherer, A.; Quake, S.R. Monolithic microfabri-

cated valves and pumps by multilayer soft lithography. Science 2000, 288, 113–116.

15. Andersson, H.; van den Berg, A. Microfluidic device for cellomics: a review. Sens.

Actuators B 2003, 92, 315–325.

16. McClain, M.; Culbertson, C.T.; Jacobson, S.C.; Allbritton, N.L.; Sims, C.E.; Ramsey, J.M.

Microfluidic devices for the high-throughput chemical analysis of cells.Anal. Chem. 2003,

75, 5646–5655.

17. Gao, J.; Yin, X.F.; Fang, Z.L. Integration of single cell injection, cell lysis, separation

and detection of intracellular constituents on a microfluidic chip. Lab Chip 2004, 4,

47–52.

18. Galas, J.C.; Torres, M.; Belotti, M.; Kou, Q.; Chen, Y. Microfluidic tunable dye laser with

integrated mixer and ring resonator. Appl. Phys. Lett. 2005, 86, 264101–264104.

19. Galas, J.C.; Peroz, C.; Kou, Q.; Chen, Y. Microfluidic dye laser intracavity absorption.

Appl. Phys. Lett. 2006, 89, 224101–224103.

20. Xia, Y.N.; Whitesides, G.M. Soft lithography. Angew. Chem. Int. Ed. 1998, 37, 550–575.

21. Chou, S.Y.; Krauss, P.R.; Renstrom, P.J. Imprint of sub-25 nm vias and trenches in

polymers. Appl. Phys. Lett. 1995, 67, 3114–3116.

22. Haisma, J.; Verheijen, M.; Heuvel, K.V.D.; Berg, J.V.D. Mould-assisted nanolithogra-

phy: a process for reliable pattern replication. J. Vac. Sci. Technol. B 1996, 14,

4124–4128.

23. Chen, Y.; Carcenac, F.; Ecoffet, C.; Lougnot, D.J.; Launois, H. Mold-assisted near-field

optical lithography. Microelectron. Eng. 1999, 46, 69–72.

24. Chen, Y.; Roy, E.; Kanamori, Y.; Belotti, M.; Decanini, D. Soft nanoimprint lithography.

Proc. SPIE 2005, 5645, 283–288.

25. Roy, E.; Kanamori, Y.; Belotti, M.; Chen, Y. Enhanced UV imprint ability with a tri-layer

stamp configuration. Microelectron. Eng. 78/79, 2005, 689–694.

26. Shi, J.; Fang, A.P.; Malaquin, L.; Pepin, A.; Decanini, D.; Viovy, J.L.; Chen, Y. Highly

parallel mix-and-match fabrication of nanopillar arrays integrated inmicrofluidic channels

for long DNA molecule separation. Appl. Phys. Lett. 2007, 91, 153114–153117.

27. Kraus, T.; Malaquin, L.; Delamarche, E.; Schmid, H.; Spencer, N.D.; Wolf, H. Closing the

gap between self-assembly andmicrosystems using self-assembly, transfer, and integration

of particles. Adv. Mater. 2005, 17, 2438–2442.

28. Kosiorek, A.; Kandulski, W.; Glaczynska, H.; Giersig, M. Fabrication of nanoscale rings,

dots, and rods by combining shadow nanosphere lithography and annealed polystyrene

nanosphere masks. Adv. Mater. 2005, 1, 439–444.

29. Tang, Z.K.; Wong, G.K.L.; Yu, P.; Kawasaki, M.; Ohtomo, A.; Koinuma, H.; Segawa, Y.

Room-temperature ultraviolet laser emission from self-assembled ZnO microcrystallite

thin films. Appl. Phys. Lett. 1998, 72, 3270–3272.

30. Rensmo, H.; Keis, K.; Lindstrom, H.; Solbrand, A.; Hagfeldt, A.; Lindquist, S.E. High

light-to-energy conversion efficiencies for solar cells based on nanostructured ZnO

electrodes. J. Phys. Chem. B 1997, 101, 2598–2601.


31. Fujibayashi, T.; Matsui, T.; Kondo, M. Improvement in quantum efficiency of thin film Si

solar cells due to the suppression of optical reflectance at transparent conducting oxide/Si

interface by TiO2/ZnO antireflection coating. Appl. Phys. Lett. 2006, 88, 183508–183510.

32. Chen, G.F.; McCandless, G.T.; McCarley, R.L.; Soper, S.A. Integration of large-area

polymer nanopillar arrays into microfluidic devices using in situ polymerization cast

molding. Lab Chip 2007, 7, 1424–1427.

33. Christel, L.A.; Petersen, K.; McMillan, W.; Northrup, M.A. Rapid automated nucleic acid

probe assays using siliconmicrostructures for nucleic acid concentration. J. Biomech. Eng.

1999, 121, 22–27.

34. Witek, M.A.; Llopis, S.D.; Wheatley, A.; McCarley, R.L.; Soper, S.A. Purification and

preconcentration of genomic DNA from whole cell lysates using photoactivated polycar-

bonate (PPC) microfluidic chips. Nucleic. Acids Res. 2006, 34, e74.

35. Foote, R.S.; Khandurina, J.; Jacobson, S.C.; Ramsey, J.M. Microfabricated porous

membrane structure for sample concentration and electrophoretic analysis. Anal.

Chem. 2005, 77, 57–63.

36. Plecis, A.; Nanteuil, C.; Haghiri-Gosnet, A.M.; Chen, Y. Electropreconcentration with

charge-selective nanochannels. Anal. Chem. 2008, 80, 9542–9550.

37. Dauriac, V.; Descroix, S.; Chen, Y.; Peltre, G.; S�en�echal, H. Isoelectric focusing in an

ordered micropillar array. Electrophoresis 2008, 29, 2945–2952.

38. Harrison, D.J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C.S.; Manz, A. Micromachining

aminiaturized capillary electrophoresis-based chemical analysis system on a chip. Science

1993, 261, 895–897.

39. Tegenfeldt, J.O.; Prinz, C.; Cao, H.; Huang, R.L.; Austin, R.H.; Chou, S.Y.; Cox, E.C.;

Sturm, J.C. Micro- and nanofluidics for DNA analysis. Anal. Bioanal. Chem. 2004, 378,

1678–1692.

40. Fu, J.; Schoch, R.B.; Stevens, A.L.; Tannenbaum, S.R.; Han, J. A patterned anisotropic

nanofilter array for continuous-flow separation of DNA and proteins. Nat. Nanotechnol.

2006, 2, 121–128.

41. Han, J.; Craighead, H.G. Separation of long DNAmolecules in a microfabricated entropic

trap array. Science 2000, 288, 1026–1029.

42. Kaji, N.; Tezuka, Y.; Takamura, Y.; Ueda, M.; Nishimoto, T.; Nakanishi, H.; Horiike, Y.;

Baba, Y. Separation of long DNA molecules by quartz nanopillar chips under a direct

current electric field. Anal. Chem. 2004, 76, 15–22.

43. Studer, V.; Pepin, A.; Chen, Y. Nanoembossing of thermoplastic polymers for microfluidic

applications. Appl. Phys. Lett. 2002, 80, 3614–3616.

44. Pepin, A.; Youinou, P.; Studer, V.; Lebib, A.; Chen, Y. Nanoimprint lithography for the

fabrication of DNA electrophoresis chips. Microelectron. Eng. 61/62, 2002, 927–932.

45. Zhang, Z.L.; Crozatier, C.; Le Berre, M.; Chen, Y. In situ bio-functionalization and cell

adhesion in microfluidic devices. Microelectron. Eng. 78/79, 2005, 556–562.

46. Guo, S.S.; Sun,M.H.; Shi, J.; Liu, Y.J.; Huang,W.H.; Combellas, C.; Chen, Y. Patterning of

hydrophilic micro arrays with superhydrophobic surrounding zones.Microelectron. Eng.

2007, 84, 1673–1676.

47. Shi, J.; Peroz, C.; Peyrade, D.; Salari, J.; Belotti, M.; Huang, W.H.; Chen, Y. Tri-layer soft

UV imprint lithography and fabrication of high density pillars.Microelectron. Eng. 2006,

83, 1664–1668.

REFERENCES 277

48. Nagrath, S.; Sequist, L.V.;Maheswaran, S.; Bell, D.W.; Irimia, D.; Ulkus, L.; Smith,M.R.;

Kwak, E.L.; Digumarthy, S.;Muzikansky,A.; Ryan, P.; Balis, U.J.; Tompkins, R.G.;Haber,

D.A.; Toner, M. Isolation of rare circulating tumor cells in cancer patients by microchip

technology. Nature 2007, 450, 1235–1239.

49. Liu, Y.J.; Guo, S.S.; Zhang, Z.L.; Huang,W.H.; Baigl, D.; Xie,M.; Chen, Y.; Pang, D.W. A

micropillar-integrated smart microfluidic device for specific capture and sorting of cells.


50. Cui, H.H.; Lim, K.M. Pillar array microtraps with negative dielectrophoresis. Langmuir

2009, 25, 3336–3339.

51. Turner, A.M.P.; Dowell, N.; Turner, S.W.P.; Kam, L.; Isaacson, M.; Turner, J.N.;

Craighead, H.G.; Shain, W. Attachment of astroglial cells to microfabricated pillar arrays

of different geometries. J. Biomed. Mater. Res., Part A 2000, 51, 430–441.

52. Nomura, S.; Kojima, H.; Ohyabu, Y.; Kuwabara, K.;Miyauchi, A.; Uemura, T. Cell culture

on nanopillar sheet: study of HeLa cells on nanopillar sheet. Jpn. J. Appl. Phys. 2005, 44,

1184–1186.

53. Kim, D.H.; Kim, P.; Song, I.; Cha, J.M.; Lee, S.H.; Kim, B.; Suh, K.Y. Guided three-

dimensional growth of functional cardiomyocytes on polyethylene glycol nanostructures.

Langmuir 2006, 22, 5419–5426.

54. Wang, L.; Lei, L.; Ni, X.F.; Shi, J.; Chen, Y. Patterning biomolecules for cell attachment at

single cell levels in PDMS microfluidic chips. Microelectron. Eng., doi: 10.1016/j.mee.

2009.01.030.

55. Amatore, C.; Arbault, S.; Bruce, D.; De Oliveira, P.; Erard, M.; Vuillaume, M.

Characterization of the electrochemical oxidation of peroxynitrite: relevance to oxidative

stress bursts measured at the single cell level. Chem. Eur. J. 2001, 19, 4171–4179.

56. Singhvi, R.; Kumar, A.; Lopez, G.P.; Stephanopoulos, G.N.; Wang, D.I.C.; Whitesides, G.

M.; Ingber, D.E. Engineering cell shape and function. Science 1994, 264, 696–698.

57. Thery, M.; Racine, V.; P�epin, A.; Piel, M.; Chen, Y.; Sibarita, J.B.; Bornens, M. Adhesion

controls the division axis of animal cells. Nat. Cell Biol. 2005, 7, 947–953.

58. Guo, L.J.; Cheng, X.; Chou, C.F. Fabrication of size-controllable nanofluidic channels

by nanoimprinting and its application for DNA stretching. Nano Lett. 2004, 4,

69–73.

59. Amblard, F.;Maggs, A.C.; Yurke, B.; Pargellis, A.; Leibler, S. Subdiffusion and anomalous

local viscoelasticity in actin networks. Phys. Rev. Lett. 1996, 77, 4470–4473.

60. Ando, T.; Kodera, N.; Naito, Y.; Kinoshita, T.; Furuta, K.; Toyoshima, Y.Y. A high-speed

atomic force microscope for studying biological macromolecules in action.

ChemPhysChem 2003, 4, 1196–1202.

61. Thoumine, O.; Ott, A. Time scale dependent viscoelastic and contractile regimes in

fibroblasts probed by microplate manipulation. J. Cell Sci. 1997, 110, 2109–2116.

62. Tan, J.L.; Tien, J.; Pirone, D.M.; Gray, D.S.; Bhadriraju, K.; Chen, C.S. Cells lying on a

bed of microneedles: an approach to isolate mechanical force. PNAS 2003, 100,

1484–1489.

63. du Roure, O.; Saez, A.; Buguin, A.; Austin, R.H.; Chavrier, P.; Silberzan, P.; Ladoux, B.

Force mapping in epithelial cell migration. PNAS 2005, 102, 2390–2395.

64. Sniadecki, N.J.; Anguelouch, A.; Yang, M.T.; Lamb, C.M.; Liu, Z.; Kirschner, S.B.; Liu,

Y.; Reich, D.H.; Chen, C.S. Magnetic microposts as an approach to apply forces to living

cells. PNAS 2007, 104, 14553–14558.


65. Psaltis, D.; Quake, S.R.; Yang, C.H. Developing optofluidic technology through the fusion

of microfluidics and optics. Nature 2006, 442, 381–386.

66. Chen, X.; Jiang, K. A large-area hybrid metallic nanostructure array and its optical

properties. Nanotechnology 2008, 19, 215303–215308.

67. Liu, G.L.; Lee, L.P. Nanowell surface enhanced Raman scattering arrays fabricated by

soft-lithography for label-free biomolecular detections in integrated microfluidics. Appl.

Phys. Lett. 2005, 87, 074101–074103.

68. Wu, J.; Day, D.; Gu,M.Amicrofluidic refractive index sensor based on an integrated three-

dimensional photonic crystal. Appl. Phys. Lett. 2007, 92, 071108–071110.

69. Pradhan, D.; Kumar, M.; Ando, Y.; Leung, K.T. Fabrication of ZnO nanospikes and

nanopillars on ITO glass by templateless seed-layer-free electrodeposition and their field-

emission properties. ACS Appl. Mater. Interfaces 2009, 126, 604–608.

70. Bie, L.J.; Yan, X.N.; Yin, J.; Duan, Y.Q.; Yuan, Z.H. Nanopillar ZnO gas sensor for

hydrogen and ethanol. Sens. Actuator B 2007, 126, 604–608.

REFERENCES 279

8NANOCATALYSIS INMICROREACTOR FOR FUELS

SHIHUAI ZHAO1,2

AND DEBASISH KUILA1,3

1Institute for Micromanufacturing, Louisiana Tech University, Ruston, LA, USA2Tianjin University, Tianjin, China3Department of Chemistry, North Carolina A&T State University, Greensboro, NC, USA

8.1 INTRODUCTION

The first computer occupied several rooms when it was born. Thanks to the

development of microelectronic technology, small personal computers with dra-

matically improved processing powers and memory capacity have now become

common at home and in the workplace. Even the palm-sized computer is not a fairy

tale anymore. Chemists and chemical engineers are trying to bring the same kind of

revolution to the chemical plants. Traditionally, the chemical industry operates

reactors in a few large facilities to achieve economies of scale. Today, a small group

of research institutes and companies in the chemical process industries are working

in the opposite direction, that is, developing micrometer-scaled reactors, the so-

called microreactor, as well as integrating microstructured heat exchangers, pumps,

valves, and other devices to go with them. The idea of chemical processing on a chip

is brought step by step into reality by employing the microelectronic and micro-

machining technology. Microreactors exhibit many practical advantages when

compared with conventional reactors, not least is the demand for a high standard

of safety, such as the transportation and storage of toxic, explosive, or harmful

materials.1 In such cases, microreactors offer the capability to carry out production

on site at the point of demand. The removal of potentially significant large-scale

plant accidents associated with thermal runaway reactions could also be envisaged


281

due to inherent thermal dissipation possible in microreactor devices. Indeed, it has

been demonstrated that reactions can be performed beyond their current explosive

limits by adopting microreactor technology.

In addition to the environmental and safety benefits of dealing with smaller

quantities of material, microreactors offer many performance benefits compared to

traditional batch-scale reactors. Most of these advantages stem from the high surface

to volume ratio, which is a consequence of the decrease in fluid layer thickness in

microscale reactors. The decrease of linear dimensions in a microreactor changes the

properties of chemical processing, such as temperature, pressure, and so on, which

refers to better heat transfer and mass transport. The heat management in microscale,

enabling mass and heat transfer to be extremely rapid, leads inevitably to a higher

level of reaction control and reactant manipulation at any one point within a

microreactor. In particular, microreactors’ narrow channels (typical width of

5–500mm) and thin walls (typical thickness of 5–100 mm) make them desirable

candidates for studying potentially explosive reactions. The small channels and the

large surface to volume ratios serve to both inhibit gas-phase free radical reactions

and improve heat transfer for exothermic reactions. Microreactors exhibit different

fluid dynamics from conventional reactors due to operations at the microscale.

Microchannels have large length to depth ratios proportional to the number of

molecular collisions with channel walls. Thus, the reaction efficiency is improved

with the enhanced collision frequency. Microreactors are also able to promote the

performance of some reactions by shortening the residence time and thereby avoiding

side reactions.2

Microreactors promise a relatively simple and quick means for commercializing a

chemical process. Conventional scale-up needs going from laboratory scale to a single

large reactor unit through a series of costly laboratory experiments, pilot plant stages,

and simulations. In contrast, withmicroreactor system, the desired product output can

be achieved simply by combining a number of microreactors in parallel, allowing

quick fabrication with low cost. The functional unit of a microreactor, such as mixing

zone, is repeated, and fluid connection between these units can be achieved by using

distribution lines and flow equipartition zones. Adopting a scale-out philosophy

coupled with large-scale microreactor fabrication technology, it is possible to extend

the optimization of reaction conditions on a single microdevice.1 Hence, the reaction

efficiencyand throughput capacity allowing the productionofmaterial on a supply and

demand basis could be achieved without the need for redesigning the reaction

methodology. A parallel array of microreactors guarantees that desired features of

a single unit will remainwhile thewhole system size is increased.Meanwhile, a larger

number of units result in higher flexibility in adapting production range to varying

demand since a certain number of systems can be turned off or additional systemsmay

beadded to theproductionplant.Thus,microreactors offergreater reactioncontrol and

selectivity in the field of chemical and biochemical synthesis, which in turn can be

optimized through a scale-out methodology creating a safe and efficient approach to

discovery and production.

The development, optimization, and improvement of catalysts for use in commer-

cially viable processes provides major challenges for scientific research, both with

282 NANOCATALYSIS IN MICROREACTOR FOR FUELS

respect to the complexities of the systems being investigated and the practical

challenges of the researcheffort.3Commercial catalysts, particularly inheterogeneous

systems, are often complexmultielemental, multicomponent systems that are usually

prepared via multistep procedures with the conditions under which each intermediate

step is performed often known as impact performance. The advent of combinatorial

chemistry in drug discovery, driven by the need to synthesize and screen a large

number of candidates for a variety of applications, has provided a new opportunity for

catalyst development and screening. A parallel array of microreactors is also a novel

tool for screening inorganic materials, catalysts, and so on. These new parallel

approaches aim to address the fundamental need to more effectively and efficiently

evaluate complex systems to a degree that still provides useful results.

On the basis of the advantages of microreactors discussed above, we could

obtain potential benefits, such as fast transfer from research results to production,

easier scale-up of production capacity, easier start of production at lower costs, and

more flexible response to market demand.4 The continuous process in microreac-

tors shortens the contact time of reactants and reaction times due to fast transfer in

thin fluid layers. Furthermore, conversion rates may increase due to short diffusion

distances, and selectivity may be improved due to more accurate control of contact

times in microreactors for gas and liquid-phase reactions covering heterogeneous

and homogeneous catalysis, catalytic oxidation, heterocyclic synthesis, and pho-

tochemical reactions. The chemical process miniaturization (CPM) technologies will

find applications in a variety of ways including environmental sensing and control,

improved operation of chemical processes, stronger economic performance through

reduced costs, and increased safety for processing hazardous materials, as well as

for research and teaching applications across a wide range of scientific disciplines.

One major application of microreactors is in battery replacement for portable

electric power devices. Various portable electronic devices have recently come into

use, and the performance of these devices has improved remarkably.5 However,

greater performance leads to greater consumption of electrical power, and it has

become difficult to secure a sufficiently long-lasting power source for these portable

electronic devices. This is often true even when a conventional lithium-ion battery is

used, despite the higher energy density available compared to other secondary

batteries. This energy demand is also expected to become more severe with the

advent of broadband network devices, and the higher power consumption contributes

to environmental pollution arising from the mass disposal of expended batteries. One

of the most promising fuel cell technologies appears to be the proton exchange

membrane fuel cell (PEMFC) fuelled by hydrogen. However, hydrogen fuel cells

present obvious problems with transportation, storage, and distribution. Because of

safety, onboard H2 generation by reforming a liquid fuel is being extensively studied

for commercialization. Among the liquid fuel to be reformed, methanol remains a

prominent candidate because of its lowcost, ease of handling, and high energydensity.

However, as a by-product of stream reforming of methanol (SRM), carbon monoxide

can poison the electrode catalyst (Pt) of fuel cells. Thus, a purification process,

preferential oxidation of carbon monoxide in H2-rich fuel cell feed, has to follow the

SRM process.

INTRODUCTION 283

The other important application of microreactors is on-site production or analysis

of chemicals. In petroleum engineering, the natural feedstocks are transported,

sometimes over long distances, to a central plant and converted therein to more

valuable products. The small size and remote location of a vast number of feedstocks

render exploitation not profitable since neither plant construction nor transport in

pipelines is economical.4 Processing in microreactors may be less expensive than

using conventional equipment. Gas-to-liquid (GTL) technology, for example, can

utilize microreactors on-site to produce higher alkanes from syngas (H2 and CO) for

easy transportation and storage. Installation and removal ofmicroreactor systemsmay

be sufficiently fast and flexible toward high productivity due to numbering up of the

single unit.

This chapter focuses on nanocatalyst development for industrially useful reactions

in silicon-based microreactors. Design and fabrication of microreactors were per-

formed using simulation and micromachining techniques used in microelectrome-

chanical systems (MEMS) and semiconductor industry. The use of cyclohexene for

hydrogenation and dehydrogenation was considered as a prototype reaction to

compare different catalyst coating methods in Si microreactors. The synthesized

catalysts for different reactions were characterized using modern tools of surface

science and chemical methods. Methanol steam reformer to produce H2 and CO

purifier is included for potential microreactor applications in next generation of

alternative energy for portable power devices.We have investigated Fischer–Tropsch

(FT) synthesis inmicroreactors, also knownasGTL technology, for syngas conversion

to higher alkanes. It could solve current difficulties of storage and transportation by

converting natural gas into liquid fuels. A parallel array system of microreactors was

designed and constructed for screening different nanocatalysts in GTL technology.

8.2 DESIGN OF MICROCHANNEL REACTORS: MICROMIXING

8.2.1 Principles and Types of Micromixing

Mixing has long been a critical issue in chemistry and chemical engineering. If the

mixing ispoor, the reactionprocessmaybe sloweddownby local shortage ofoneof the

reactants or thermal nonuniformities. Microreactors have the potential to be useful

tools for chemical syntheses and kinetics studies. The reduction of reactor dimensions

leads to a large surface to volume ratio of the reaction channel, which increases heat

and mass transfer efficiencies. This feature allows microreactors to suppress “hot

spots” and to be free from problems of mass transfer limitations. The reduction of

reactor dimensions also leads to a laminar flow with small Reynolds number at each

reactor channel. In other words, the mixing in microreactors is mainly driven by

molecular diffusion without any assistance of turbulence. The diffusion process is

almost accomplished between thin fluid layers that are formed by division of a

mainstream into many small substreams or reduction of the channel width along the

flow axis for one channel. Thus, large contact surfaces and small diffusion paths are

generated in microchannels.


The degree of mixing of reactants greatly influences the product composition for

multiple reactions and the kinetic measurements for a very fast reaction whose

reaction time is shorter than the mixing time. To achieve better mixing, there are

two basic principles to induce mixing at the microscale as defined by Hessel.6,7 In

active mixing, the external energy sources, such as ultrasound, acoustic, bubble-

induced vibrations, piezoelectric vibrating membranes, magnetohydrodynamic ac-

tion, and so on, are used to enhance themixingefficiency.However, this type ofmixing

requires additional design and construction of force-driving devices or structures,

which raise the complication and the cost of the whole mixing system. For passive

mixing, flow energy, such as pumping or hydrostatic potential, is applied to reform the

flowfor faster andbettermixing. Several types ofpassivemixinghavebeendeveloped,

such as chaotic mixing, microplume injection mixing, secondary flow mixing,

distributive mixing, and split-and-recombination mixing, and so on. However, these

techniques have their advantages and disadvantages.

In general, the influences of mixing on microchannel reactors’ performance are

investigated using computational fluid dynamics (CFD) simulations. Fluent� is an

example of such CFD codes and solves the conservation equations for mass,

momentum, and energy using the control volume method. The reduction of the

diffusion length is essential for fast mixing in microreactors, since mixing time is

proportional to the square of diffusion length. In many micromixers, which are

important parts ofmicroreactors, the reactant flow is split intomany laminar segments

to shorten the mixing time. However, few studies have quantitatively addressed the

relationship between the size of laminar segments in microreactors and the product

composition of multiple reactions, and how the size of laminar segments affects the

precision of the measurements of rate constants. These are very important factors for

establishing a design method of microreactors for industrial applications. Thus, it is

necessary to examine the beneficial effects of feeding reactants with a form of

lamination segments for multiple reactions and show the lamination width in micro-

reactors necessary to exhibit desired performance.

8.2.2 Omega Structure—A Novel Micromixing Technique

In this section, we discuss briefly about several microreactor designs, more specifi-

cally, a novel micromixing design termed omega channel mixing. Its mixing perfor-

mance is compared to that in straight and zigzag channels, the structures created for

secondary flow mixing developed by our group previously.

The dominant effect in micromixing is molecular diffusion, where the mixing

length is proportional to the mixing time and diffusion coefficient. Especially, in the

microscale reactor, the flow in microchannel is laminar; the mixing species is poorly

diffusive. Therefore, it is important to create chaotic and turbulent flows in micro-

channels. A novel omega-shaped microchannel reactor has been designed to create

profound chaotic flows in microchannels.8,9 A 3D fluidic dynamic simulation was

used to evaluate the mixing performance of the omega-shaped microchannel, which

is compared to common straight and zigzag microchannels. The 3D models of

microreactors with different microchannel structures were simulated using the

DESIGN OF MICROCHANNEL REACTORS: MICROMIXING 285

MemCFD module of CoventorWareTM software (Coventor Inc.). The flow of a fluid

through a fixed volume is governed by the Navier–Stokes equation. For fixed volume,

the flow is subjected to conservation of momentum.

The so-called omega channel microreactor consists of a network of omega-shaped

channels and they are integrated as shown in Figure 8.1. The channels adopt the

cellular structure of honeycombs. Figure 8.1 shows that the obstacles formed by the

shape of the omega channel induce a high velocity value. They force the flow

streamlines to comingle from center to the boundary channel wall and from boundary

channel wall to the center periodically. The flow velocity in the omega channel varies

significantly from point to point and the variance between the velocity values is large,

which is bigger than that in the straight and zigzag channels.

A gradual increase and decrease of path width in omega channel, the curve of

channel, and the changes of flow direction help uneven the local flow velocity and

generatevortices of the flow.Each omega channel has an impeding curve that impedes

the oncoming flow and split it into two streams, and the two streams will unite with

other flows from the adjacent omega channels. As a consequence, the chaotic motions

within the omega channel are created without active moving parts or external source.

The chaotic advection promotes the rapid changes (exponentially) of stretching and

foldingof themixing species interfaces byeither unsteadyfloworcomplex structureof

omega channel and rapidly increases the mixing efficiency.

The transport of vortices can be calculated from the conservation equations of

momentum. The incensement in vorticity is caused by promotion of stretching and

turning of the vortex lines and incensement of diffusion is caused by viscosity and

density inhomogeneity.Comparing thevorticity profile of the omega channelwith that

of zigzag channels in Figure 8.2 supportively shows that the omega channel generates

higher vorticity than the zigzag channels. The vortices generated in omega channel

tend to break up the streams into layers and each layer curls in a different manner.

These breaking and curling actions and the comingling of center streamline and the

edge streamline increase the chances ofdiffusionbetween themolecules of two liquids

in a mixing process. However, straight channels do not have such advantages.

In omega channel, the fluids flowing through channel are rotated as intended. There

are critical differences between the low and high Reynolds number flows. As fluids

pass through the omega channels with a periodically rotational shape, the fluids are

FIGURE 8.1 2D top view of velocity contours in omega channel microreactors.9


stretched and folded due to the inertial force. This convection widened the interfacial

area, inducing rapid mixing of fluids at a higher Reynolds number. As the Reynolds

number becomes bigger (but still in laminar domain), fluids are mixed more and the

interface of fluids becomes more distorted, recirculation occurs in the flow of omega

channel, and the streamline turns unsymmetrical. The backmixing as shown in

Figure 8.3 provides uniform distribution, and by increasing the residence time, a

nondiffusion character can be introduced by the obstacles in the microreactor. The

convection mixing action of the elements rapidly eliminates temperature gradients,

reducing thermal degradation.

To study the transitional properties of the mixing species, stochastic approach is

employed for presenting the cumulative probabilities of one particle exiting the

microreactor. The model presented is based on modeling the axial position of the

particle by means of aMarkov chain: the probability distribution of the axial position

of the particle at a given time step n depends only on its position at time step n�1 and a

set of transition probabilities. The transition probabilities are quantified in accordance

with the particle transport processes.9 As expected, a molecule in the omega channel

will take the longest time to exit the reactor among the three sets of reactors.Theomega

channel microreactor has a higher mean residence time than the other two micro-

reactors and the variance of the residence time is the biggest due to turbulence.

FIGURE 8.2 2D cross-sectional view of vorticities in (a) zigzag channel microreactor and

(b) omega channel microreactor.9

FIGURE 8.3 Unsymmetrical, recirculation flow in omega channels at Re> 200.8

DESIGN OF MICROCHANNEL REACTORS: MICROMIXING 287

To evaluate simulation results described above, the microreactors with omega-

shaped, straight, and zigzag channels were fabricated and tested using

Fischer–Tropsch synthesis (H2 þ CO) to higher alkanes, which is one of the major

gas-to-liquid technologies to convert natural gas or other gaseous hydrocarbons into

longer chain hydrocarbons for easy transportation and storage (discussed in

Section 8.6). The experimental results of conversion rates as shown in Figure 8.4

are in agreement with the previous simulation predictions. The conversion rates in

omega reactors are always higher than that of the straight and zigzag reactors under

various conditions. The conversion efficiency for the omega-shaped reactor is 17%

greater than that for the conventional straight channel microreactor and 12% greater

than that for the zigzag-shaped channel microreactor. The longer residence time

provides themolecules of different reaction speciesmoremixing time and breaking of

the C�O bond, thus promotes hydrocarbon chain (�CH2�)n growth, leads to more

collisions between reactant molecules and the catalyst, and encourages more CO

molecule consumption in the reaction. The advantage of omega reactor lies in the

mixing performances, and consequently this mixing privilege contributes to higher

conversion rate in omega reactor than that in the straight and zigzag channel reactors.

In short, as described above, a network of omega-shaped microchannels has been

designed and simulated by a Markov model for optimizing the channel geometry to

improve mixing efficiency in microreactors. The simulation results are in agreement

with the experimental data and show longer residence time and better mixing in

omega channels than in common straight channels and zigzag channels of micro-

reactors. In general, several mixing principles in microscale has been proposed. Most

of them have been verified by making devices and testing them with simple flow

visualization techniques and compared with predictions of CFD simulation. In future,

a detailed benchmarking may be needed for proper mixing comparison between the

different micromixers themselves, while dimensionless parameters are preferred.

Moreover, the microdevices need to be benchmarked to conventional equipment.

FIGURE 8.4 CO conversion in straight, zigzag, and omega microreactors with different

H2/CO gas ratios at H2/CO flow rate of 0.3/0.2, 0.2/0.1, and 0.3/0.1 sccm.8


Finally, mixing efficiency may not be the only factor when designing or fabricating

microreactors. The important factors, such as reactor functionality, reliability, and

cost, should also be considered to establish commercial chemical production

processes.

8.3 FABRICATION OF MICROCHANNEL REACTOR

8.3.1 Materials for Microreactor Construction

The advent of microreactor technology was initially supported by silicon micro-

fabrication techniques originally developed in the microelectronics and MEMS

industry. As discussed in the previous section, computer simulation studies are

performed prior to microfabrication. To date, a variety of materials and fabrication

methods are available for construction of microreactors and microprocess compo-

nents. In general, four types of materials are used as substrates for fabrication of

microreactors: metal, glass, polymer, and silicon.

Research at the Institut f€ur Mikrotechnik Mainz (IMM) was initially focused on

metal-based multichannel microreactors that could be produced using punching and

conventional machining techniques.4 Small channel dimensions and the relatively

high thermal conductivity of certain stainless steels result in extraordinary heat

transfer characteristics. Generally, the choice of a particular metal or alloy depends

on the application. Typical factors needed to be considered include resistance to

corrosion, thermal properties and thermal response, and ability to handle mechanical

stresses induced by the local process environment. Microfabrication techniques for

metal microreactors includemechanical micromachining, laser micromachining, wet

chemical etching, and selective laser melting.

The combined effects of both material cost and ease of fabrication create a

significant advantage for various polymers (plastics) overmetals.Microfluidic plastic

devices can be mass-produced using fabrication techniques such as hot embossing,

injection molding, and casting. The biggest disadvantage of plastic devices is the

maximum temperature that can be used without inducing material softening with

subsequent creep or flow, which is typically less than 200�C.10 However, for

applications where themaximum operating temperature is below the thermal stability

limits of the particular plastic materials used in the device, such as those used in

pharmaceuticals and bio-based material applications, plastics are an attractive alter-

native to metals.

A special-purpose photostructured glass called FOTURAN, which is based on

lithium aluminum silicate, is especially useful for creating microchannels and related

structureswith high aspect ratios, width ofwalls from50 to 500 mm, length ofwalls up

to 100mm, and diameter of holes down to 50mm. Sawing, grinding, polishing, and

edge-forming machining are the techniques used to produce the basic glass plates.11

Photolithographic methods (resist coating, exposure, and development) and the

subsequent chemical etching processes generate microstructures within the glass.

The etched components can be upgraded by different coating processes (sputtering,

FABRICATION OF MICROCHANNEL REACTOR 289

evaporation, screen printing, electroplating). Different bonding techniques (thermal

bonding, gluing, soldering) add to its technical potential. The disadvantages of glass

microreactors are that they are not suitable for high-temperature reactions (>400�C)and it is difficult to integrate them with heaters, sensors, and so on.

Compared to glass, silicon has played a prominent role in microreactor fabrication

due to the significant micromachining experience developed in both the integrated

circuit (IC) andMEMS industries. Silicon is an excellent material because of its large

operating temperature range and chemical inertness. It has been shown that the latter

feature can be improved further by modifying the surface. The high thermal conduc-

tivity of single-crystal silicon (236Wm�1 K�1) versus aluminum (157Wm�1 K�1)

also generates an advantage fromaconduction heat transfer perspective.10 In addition,

it is possible to integrate silicon-based microreactors with onboard sensors, such as

those used for measurement of flow, temperature, and pressure. This creates an

advantage of more precise control and performance information. This information

comes at a significant cost since the machining and fabrication require specialized

silicon processing equipment.

Mass production of silicon microreactors is possible, but high volumes are needed

to reduce the costs. Some research groups have focused on the use of silicon-based

microreactors because it allows sensors, actuators, and microstructures to be directly

integrated with microreactors.12 These features produce reactors with system char-

acteristics that cannot be duplicated either on the macroscale or on the microscale

using othermaterials, such as thermal time constants on the order ofmilliseconds, and

the ability to control temperature and size within a length scale of a fewmicrometers.

8.3.2 Microfabrication of Silicon Microchannel Reactors

The fabrication of silicon microreactors starts with cleaning of silicon surface by

trichloroethane, acetone, and isopropyl alcohol, followed by photolithography, which

is a well-developed microfabrication technique. The principle of photolithography is

similar to exposure and development of photos taken by traditional cameras. The

material to be exposed and developed to create intended patterns is called photoresist.

There are two types of photoresists: positive and negative. Photoresist is spun on to a

masking layer on a substrate, typically silicon oxide or nitride on a clean, polished

silicon wafer. A pattern is formed by placing a chromium-patterned glass mask

between a UV source and the silicon wafer. The resist is then developed to form a

patterned protected layer. The unprotected areas of the material are then etched using

either wet or dry etching when positive photoresist is used. There are two types of wet

etchants for silicon.The isotropic etchants etch at equal rates in all directions, resulting

in slightly rounded features. However, it is difficult to etch to high precision.

Anisotropic etchants etch at different rates in different crystallographic planes.13

Although precise etching is more easily obtained in anisotropic etching where etch

rates are low, it can result in rough surfaces.

As an alternative etching method, dry etching comprises plasma- or discharged-

based techniques. They etch accurately at small dimensions resulting in less under-

cutting and broadening of features to give good pattern transfer. The various methods


include chemical plasma etching, reactive ion etching, and ion beam etching.13,14

Recently, a deep reactive ion etching (DRIE) process was developed that utilizes a

chlorine-based system in conjunctionwith passivation to achieve structureswith large

aspect ratios. These processes are very sensitive to various parameters. For example,

temperature has a very strong effect and is one of the major causes of inconsistency.

Small amounts of contaminants can drastically affect the product by reacting with the

target or by changing the plasma chemistry. Combinations of dry and wet etching

techniques can allow fabrication of complicated structures.

One of theDRIE etching processes, inductively coupled plasma (ICP) etching,was

utilized in our study for micromachining the reactor channels on the silicon wafer.14

The process underlying this technology is the so-called Bosch process, which is well

developed andcurrentlywidely usedon ICPequipment forDRIE technology in silicon

micromachining. On the basis of Bosch process, Alcatel’s deep plasma etch technol-

ogy was designed to deliver superior process performance and to meet the needs of a

broad range of deep silicon etch applications. Its applications in MEMS and micro-

fluidic device fabrication include high aspect ratios, etch depths of as great as 500 mm,

etching through the wafer, etching into a buried cavity in the wafer, or etching onto

buried oxide. The etch technology uses a patented high-density ICP source and a

fluorine-based noncorrosive etch chemistry.

Thephotolithographyprocess canbe replacedbyLIGA, inwhichX-rays are used to

perform lithography on thick resist.4,10,13 After development, the gaps between the

remaining resist are filled by electroplating. The resist is then removed to leave ametal

mold insert for injectionor reactionmolding.Theuse of highly coherentX-rays results

in submicrometer resolution and large aspect ratios. Poly(methylmethacrylate)

(PMMA) was the traditional resist of choice due to its excellent contrast and stability.

However, due to its low sensitivity, a new negative resist based on a novolak resin was

developed. A large range of metals, alloys, and dispersion composites have been used

for electroplating, such as nickel, gold, copper, nickel, and cobalt. Suitable polymers

for molding include PMMA, polyoxymethylene (POM), polyvinyldenefluoride

(PVDF), polyaryl ether ether ketone (PEEK), and polycarbonate.

The microfabricated reactors along with different microchannel structures such as

straight, zigzag, and omega channels are shown in Figure 8.5. As mentioned above,

ICP etching allows high aspect ratios of the microchannels. It has been widely

embraced for MEMS processing due to its high etching selectivity, its ability to

precisely transfer photoresist patterns into silicon substrates, and its cleanliness and

compatibilitywithvacuumprocessing technologies.However, ICPetchinghas its own

specific problems that include “grass” formation, etching uniformity,mask selectivity,

and soon.Etchinggas cycle time, bias power, and chamber pressure are theparameters

adjusted to solve these unfavorable phenomena.14,15 The periodic change of different

gases for etching (SF6) and passivation (C4F8) can lead to very high aspect ratios and

very high etch rates. The surfacemorphologies of thewaferwith respect to the increase

of SF6 gas flow time change dramatically. The “grass” formation on the surface is the

result of particulate material sticking inadvertently on the silicon surface. This

material can locally mask the silicon during etching and can be formed due to

redeposition and growth of polymer material from the sidewall passivation step.


By increasing the cycle time of SF6, the “grass” formation was markedly decreased

(Figure 8.6).

As the SF6 gas flow time increased, the sidewall at the base of the reactor channel

assumed a slight retrograde profile. This is caused by thin sidewall polymer layers at

the bottom part of channel. When the cycle time of SF6 gas was increased, lateral ion

bombardment to the sidewall also increased.Thepassivatedpolymer layerwas thinner

at the base of channels than at the top of the channels. Thus, the etching profile of

trenches could be broadened at the bottom. It is useful to achieve high etch rates, but

often at the expense of sidewall broadening problems.

The etching characteristics at low chamber pressure (Table 8.1) show that by

increasing the SF6 cycle time to 10 s the overall etching rate increased, but the etching

uniformity dropped to 3.5%. The highest etching selectivity between the silicon and

the photoresist was determined to be 167:1 and the aspect ratiowas as high as 40:1. In

the long SF6 gas cycle time region, the overall etching rate and etching uniformity are

FIGURE 8.5 Fabricated microreactors (a) with SEM pictures of (b) straight channels,

(c) zigzag channels, and (d) omega channels.9

FIGURE 8.6 SEM photos after etching with different SF6 gas flow times.15


improved, and also the etching rate is nearly independent of the feature size. Increasing

bias power causes higher average etching rate but does not contribute to improving

etching uniformity to acceptable levels. Increasing bias power enhanced overall

etching rate, but the etching uniformitywas not improved. The surfacemorphology of

the silicon substrate was strongly affected by etching gas species and the gas flow

sequence time. However, the “grass” formation was not reduced at the low chamber

pressure. The formation of micro “grass” on the silicon surface could result from

several factors. Etching gas sequence time, bias power, and chamber pressure can all

greatly change the surface morphology of silicon surface. Grass formation was

dramatically reduced by increasing SF6 gas flow time. For the optimum conditions

of the ICP etching to fabricate siliconmicrochannel reactors, longer cycle time of SF6gas flow, high bias power, and low chamber pressure are all recommended.

Most of the fabrication methods can produce structures but cannot form sealed

structures that are imperative if the chemicals are to be contained in the channels. The

use of gaskets and a suitable housing that is kept tightly clamped is a common sealing

method, offering the advantage of easy assembly and disassembly. However, there are

no suitable procedures for microscale assembling and handling. The most popular

method for silicon microstructures is anodic bonding.14 Two wafers, one of which

must be at least semiconducting, are heated to 180–500�C and a direct current (DC)

voltage of 0.2–1 kV is applied across them. Silicon and Corning 7740 Pyrex glass are

most frequently used.Wafers must be cleaned and polished before the anodic bonding

procedure is started. The materials used must have similar thermal expansion

coefficients. Alternatively, silicon fusion bonding and lamination method also allow

the bonding of different materials.

In general, the composition of a specific material of construction will partially

dictate the preferred fabrication method. In many cases, a variety of materials may be

equally applicable due to their similar chemical and thermal compatibility character-

istics for the given process fluids as well as their ability to safely function over the

desired operating ranges for both temperature and pressure within the given process

safety settings.10 In these instances, the least expensive material with the most cost-

effective fabrication method is typically chosen. Ultimately, the cost and process

application requirementswill both dictate thefinal choice formaterials of construction

and fabrication methods. However, the enabling characteristics of miniaturization

TABLE 8.1 ICP Etching Characteristics with Variable Parameters15

SF6:C4F6(s)

Pressure

(Pa)

Bias Power

(W)

Etching rate

(mmmin�1)

Etching

Uniformity (%)

Grass

Formation

4:2 5 30 4.13 11.6 Lots

6:2 5 30 5.39 3.1 Little

8:2 5 30 5.90 2.4 None

10:2 5 30 6.40 3.5 None

4:2 5 40 3.91 8.7 Lots

4:2 5 50 4.86 8.3 Lots

4:2 3 30 3.16 6.8 Lots


must also be considered. If the selected material and fabrication technology cannot

produce a microreactor that has step change advantages over a larger conventional

reactor, the additional expense for miniaturization is not justified. Hence, when

choosing a fabrication technology, the following factors should be considered by

users: process cost, accuracy, process reliability, material choice, and process time. It

is worthwhile after finishing a conceptual design of a specificmicroreaction system to

compare the different technologies with respect to the mentioned parameters.

Commonly, decisions are made on an overall evaluation of technological potential.

8.4 NANOCATALYST DEPOSITION ON THE MICROCHANNELS

8.4.1 Pretreatment of the Substrate and Coating Methods

Structuredcatalysts and reactors aregainingmore andmore importance due to industrial

applications. Themicrochannels in amicroreactor have high surface to volume ratio and

its geometric surface may be directly used for performing catalytic reactions, which is

not the case in traditional reactors. However, to enhance activity and efficiency of a

microreactor, it is necessary to increase the specific surface area by treatment of the

microchannel walls, usually by applying porous coatings. The porous layer can be

catalytically active or serve as a support for a catalytic phase. Micropacked beds of

powder catalysts can sometimes be used, but, in general, a very thin layer of catalyst that

sticks to the reactor wall is preferred because ofmass and/or heat transfer improvement.

Different methods can be used to deposit a catalyst layer on a surface, depending on the

properties of the surface and the catalyst that has to be deposited.

The pretreatment of the substrate to coat is important because it increases adhesion

of the catalytic layer and thus the lifetime of the structured catalyst. Two common

pretreatment methods are anodic oxidation and thermal treatment.16 The anodic

oxidation method is generally applied to structures containing aluminum with the

objective to obtain a porous alumina layer at the surface. Themethod is used either as a

pretreatment before applying another coating or as a way to obtain a thin porous layer

that can bedirectly impregnated. Like anodic oxidation, thermal oxidation is not really

a deposition method but a surface modification technique. However, it can be used

either as a pretreatment step to increase the catalyst adhesion or as a catalyst support.

The chemical oxidation of the substrate is sometimes carried out by acid treatment to

form metal oxide layers, such as Al2O3 layer, or to form a pseudolayer accessible to

chemisorption of small charged particles. For silicon and titanium-based substrates,

etching and/or oxidation of the surface can be obtained using an alkali treatment

procedure.

In general, two types of coating methods, chemical and physical, are used for

nanocatalysts in the microchannels. The chemical methods include suspension,

sol–gel, hybrid method between suspension and sol–gel, electrophoretic deposition

(EPD), electrochemical deposition, electroless plating, and chemical vapor deposition

(CVD).14,16 The physical methods consist of a mechanical method such as plasma

vapor sputtering (PVD) and thermal methods such as evaporation and electron beam


evaporation.14,16 These fabrication techniques are used inMEMS and semiconductor

industry and often referred to as silicon coating procedures.

In plasma vapor sputtering, capacitive plasma is generated between the surface to

coat and a target made of the material to be deposited. Sputtering is performed under

vacuum, the structured surface is operated as the anode, and the coating material is

operated as the cathode that emits atoms to the surface.16,17 The catalyticmetal (Pt, Fe,

Co, etc.) is often sputtered without a prior oxide layer. In electron beam evaporation,

the electron beam with high kinetic energy is directed to the target material for

evaporation. Upon impact, the high kinetic energy is converted into thermal energy

allowing the evaporation of the material. In pulsed laser deposition process, also

known as pulsed laser ablation deposition, a laser is used to ablate particles from a

target in a deposition chamber under reduced pressure and at elevated temperature.

The number of laser pulses is related to the thickness of the film deposited on the

substrate. Forflame-assisted vapor deposition, the deposition process can take place in

an open atmosphere without requiring the use of complex deposition chamber and/or

vacuumsystem like that inCVDorPVDmethods.Theatomizedchemical precursor of

the catalyst is burned in a flame. Themethod can thus be considered as a “dry” way of

deposition for the substrate that is placed in the combustion zone at controlled

distances and temperatures.

CVD technique requires the use of chemical precursors of the desired deposited

material.17 The precursor can be the same as that used in sol–gel method (discussed

below), but no solvent is required.Only thevolatile precursor and the structured object

are present in the deposition chamber. To enhance the deposition rate, the use of low

pressures and high temperatures may be required. Plasma-assisted CVD (PACVD)

also allows to perform the deposition at lower temperature and higher deposition rate.

EPD is a colloidal process in which a DC electric field is applied across a stable

suspension of charged particles attracting them to an oppositely charged electrode.16

One electrode (cathode) consists of the substrate to coat, the anode being either an

aluminum foil or stainless steel. The thickness of the coating depends on the distance

between the two electrodes (ca. 10mm), the DC voltage (can vary from 10 to 300V),

the properties of the suspension (e.g., pH), and the duration. This technique is often

used to deposit a layer of aluminum oxide (by oxidation of an aluminum layer) as a

precoating, to favor the adhesion of a catalyst, deposited in the second time by dip

coating in a suspension.

Electrochemical deposition and electroless plating use ionic solutions.16 The first

method, also called “electroplating,” produces a coating of metal on a surface by the

action of electric current. The deposition of a metallic coating onto an object is

achieved by putting a negative charge on the object to be coated (cathode) and

immersing it into a solution that contains a salt of the metal to be deposited.When the

positively chargedmetallic ions of the solution reach the negatively charged object, it

provides electrons to reduce the positively charged ions to pure metal.

All methods based on the dispersion of a finished material (catalyst support or

catalyst itself) have been gathered under the term “suspension method.” In some

preparations, the difference with sol–gel method is small because the suspension

method often implies some gelification steps. It is the most widely used method,

NANOCATALYST DEPOSITION ON THE MICROCHANNELS 295

namely, for ceramic monoliths. Powder (catalyst support or catalyst itself), binder,

acid, and water (or another solvent) are the standard ingredients. The concentration of

all ingredients varies largely from one experiment to another and also depends on the

nature of the surface to coat and on the desired layer thickness. The size of the

suspended particles has a great influence on the adhesion to the substrate.

In sol–gel method, the starting point is a solution of a chemical precursor of the

material tobedeposited.14–16 It involves the transitionof a system from the liquid sol to

a solid gel and creates a three-dimensional networkof inorganicmatter knownas “gel”

from a colloidal or molecular solution of the precursor (sol) by low-temperature

polymerization. It mainly involves the production ofmetal oxides by the hydrolysis of

appropriate precursor compounds. The hydrolytic reaction is utilized to provide

oxides in the form of thin films on the substrates or powders. Hydrolysis of the metal

alkoxides to metal hydroxides followed by dehydration to metal oxides can be shown

in the following equations:

MðORÞnþ nH2O!MðOHÞnþ nROH; where M ¼ Si; Al; Ti; Zr

M�OHþHO�M0 !�M�O�H0 þH2O

One important factor in sol–gel technology is the aging time allowing the gelation

of the sol. It can vary from a few minutes to several weeks, depending on the

concentrations in the sol and the characteristic size of the object to coat. The conditions

during sol formation have to be chosen to obtain oligomers with desired degree of

branching. Sols with high viscosities, obtained after long aging time, allow depositing

thicker layer but easy to crack. A compromise has to be found for each preparation and

substrate to coat. Sol–gel can also, in certain cases, be used to deposit a primer on the

support to coat. In contrast, impregnation is often used (as a posttreatment) to deposit a

catalytic active phase on thewashcoat and does not differ from powder impregnation.

Hybrid method does not differ very much from suspension method and sol–gel

method. In the present case, a sol not only acts as the binder but also participates in the

chemical and textural properties of the final deposited layer.

In general, the suspension and the sol–gel methods are applied to the structured

object using dip coating. An alternative to dip coating is spray coating. In the case of

coatingmicroreactor channels, thedropsof the sol–gel canbedeposited (dropcoating)

withpossible simultaneousheatingof themicroreactor channels. Spin coating canalso

be used for a silicon microreactor. In this deposition method, the film thickness is

related to the sol viscosity and the spin speed.14,18

When the different methods for thin films are compared, it appears that sol–gel

technique produces layers less than 10 mm thick, whereas PVD methods yield layer

thinner than 1 mm. The suspensionmethods can generate layers from 1 to 100mm, but

it is in general used to obtain thicker layers than that obtained by sol–gel. The sol–gel

method allows porosity of the foam material, whereas the use of the suspension

technology can result in blocking pores.16 To avoid the penetration of the oxide

precursor in the pores, a hybrid method between suspension and sol–gel is preferred

over sol–gel alone. The hybrid procedure indeed combines the advantages of the sol


(precise control and tuning of the catalyst microstructure) and that of the suspension

(ease of deposition).

8.4.2 Comparison of PVD with Sol–Gel Method Using Pt Catalyst

We can compare the chemical method (e.g., sol–gel technology) with the physical

method (e.g., PVD) to obtain metal (or metal oxide) catalysts on the surface of

microchannels. Platinum, used as a catalyst in the hydrogenation anddehydrogenation

of cyclohexene to cyclohexane and benzene (see below), has been deposited and

studied in a siliconmicroreactor.14,18 Silica or alumina, prepared by a sol–gel process,

is chosenas support becauseof its high specific surface area.Wecompareddip coating,

drop coating, and spin coating by depositing SiO2 on the channels of themicroreactor.

The selective deposition of catalysts was applied to both PVD and sol–gel methods.14

The silica and alumina films were characterized by scanning electron microscopy

(SEM), X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM),

and Brunauer–Emmett–Teller (BET) surface area measurement techniques. The

microreactor was prepared by general lithography and ICP etching as discussed in

previous section. The conversion of cyclohexene was used for testing activity and

efficiency of the catalyst. Reaction results of supported Pt were compared to results

obtained in unsupported Pt microreactors.

The basic principle involved in the synthesis of the alumina coating solution is the

hydrolysis of aluminum alkoxide to form its hydroxide, peptization, and polymeriza-

tion to form the alumina. To allow the formation of a good three-dimensional network,

the reflux step was performed after acid addition to the resultant solution and before

coating on the channels of the microreactor. The silica sol–gel was prepared using

tetraethoxysilane (Si(OC2H5)4) as the precursor. We chose SiO2 acid instead of SiO2

basic because the surface area is expected to be approximately 10 times higher.19

Dip coating, spin coating, and drop coatingwere used to apply SiO2 orAl2O3 on the

channel region of the microreactor.14,18 Negative photoresist was chosen as the mask

for the selective deposition (Figure 8.7) because the unexposed channel part of

negative photoresist can be completely removed by developer, which minimizes

residual photoresist. Thephotoresistwasfirst spin coatedon the siliconwafer, and then

it was exposed in the UValigner. The development step opened the channel region of

microreactor. After coating SiO2 or Al2O3 solution, a 200�C, 20min baking step was

required to dry the sol–gel coating for good adhesion.

Ptwasdepositedon the catalyst support in the formofPtþ 2 by chemical deposition,

known as ion impregnation.16,20 Platinum(II)-2,4-pentanedionate (Pt(C5H702)2) was

dissolved completely in toluene and the solutionwas dropped into the channel area on

a hot plate at 70�C, with continued baking at the same temperature for 12 h. The

impregnated Ptþ 2 on the catalyst support was reduced to Pt0 in the presence of gas

mixture (40% H2 in N2, 0.2 Lmin�1) at 400�C for 4.5 h.

The chemical state and the atomic concentration of the elements present in the

coating of the catalyst support were studied using XPS. The ratio of Si:O in the silica

film prepared by the sol–gelmethod is 1:2, while the ratio of Al:O inAl2O3 is 1:1.242,

which is different from the stoichiometric ratio of 1:1.5. XPS analysis shows that the


peak for Ptþ 2 after reduction (Pt0) shifts from 73.125 to 70.9 eV (Figure 8.8a). The

distance between the separation of the two peaks (70.9 and 74.2 eV) after reduction is

3.3 eV,which is in agreementwith that of the standard peaks. TheXPS depth profile of

Pt shows the uniform concentration and equal distribution of Pt inside silica film.

SEM analysis was performed on the channels that were coated with silica or

alumina to image the catalyst support and also to verify thickness of the film.TheSEM

images in Figure 8.9 confirm that the adhesion of alumina to silicon sidewall is worse

than that of silica. Some parts of the alumina film peeled off sidewall to yield bad

coating. Figure 8.10a shows that the particles of silicon film in the microchannels are

around 100 nm in diameter,which is confirmed byAFM image shown in Figure 8.10b.

These nanoparticles create the pore structure of silica with the specific surface area

FIGURE 8.8 (a) Binding energy of platinum before and after reduction and (b) XPS depth

profile of Pt deposited by sputtering.14

Step 1: Before PR coating

Step 2: After PR coating

Step 3: After exposing photoresist

Step 4: After SiO2 coating

Step 5: Strip PR with acetone

and coat Platinum

Step 6: Bond reactor with Pyrex

FIGURE 8.7 The process for selective deposition of catalyst in the microchannels.18


(SSA) of �500m2 g�1. The SSA study at different aging times and calcination

temperatures show that longer aging time and higher calcination temperature

decrease the specific surface area. SiO2, which has different aging time, was coated

on 5 and 100 mm channel micreactors by dip-coating, drop-coating, and spin-coating

techniques. The reactors with dip coating have larger surface area than those obtained

with drop-coating and spin-coating methods. Under similar conditions, the reactor

with 5mm channels has higher surface area compared to that with 100 mm channels

because there are more channels in the former.

After the characterization was completed, the prototype reaction on conversion of

cyclohexene was studied to test the activity and efficiency of supported catalyst.

C6H12 $ C6H10 þH2 $ C6H6þ 3H2

From the reaction results, we conclude:

1. For 100mm channel reactor, the conversion increases with increasing residence

time. Effect of residence time on conversion is much smaller on 5 mm channel

than that on 100 mm, as shown in Figure 8.11a.

2. For 100mm channels, the conversion on silica-supported Pt is much higher than

that on sputtered Pt (Figure 8.11b) due to the high surface area of silica support

with porous structure shown in Figure 8.10a.

FIGURE 8.10 (a) SEM picture of silica particles and (b) AFM image of silica particles.14

FIGURE 8.9 (a) Silica on 100mm channels and (b) alumina on 100mm channels.14


3. The conversions with silica-supported Pt and 100mm channel are very close to

the conversion with sputtered Pt in 5mm channel when Figure 8.11a and b are

compared. Thus, wemay choose high reactant flow rate in 100 mmchannel with

sol–gel-supported Pt to receive more products at shorter time, which creates

high efficiency and high productivity.

4. Compared to silica-supported Pt, alumina-supported Pt gives higher conversion

of cyclohexene, but it deactivates faster. Moreover, the conversion on silica-

supported Pt increases when temperature rises, but keeps constant on alumina-

supported Pt and drops fast when the temperature is kept below 120�C.

We discussed previously two kinds of catalysts: supported (sol–gel method) and

unsupported (sputtering method). Sputter deposition is popular for coating thin,

uniform layers on the channels of microreactors. We can get different compositions

of catalyst layers by this method. But the disadvantage is that there is no surface area

enhancement. Supported catalysts can have high surface area, which provides more

reaction surface for high throughput. The sol–gel deposition process is cost-effective

since no expensive equipment is required. Due to its easy formulation and coating

processes, the high throughput of catalysts is the other benefit compared to sputter

deposition. In general, some methods concern only the oxide deposition (which can

further be impregnated by a catalyst precursor) and others concern the direct

deposition of a metal on substrate, without any oxide layer. The ease and the speed

of the PVD process are very advantageous in the case of parallel screening because it

allowsone toobtain an important catalyst library in a fewhours and thus provides rapid

information on the active metals to catalyze a reaction. However, due to their low

porosity, the activity of the obtained catalysts by PVD methods leads to low activity

catalysts compared to catalysts prepared by wet chemical procedures, such as sol–gel

method. The sol–gelmethod presents some advantages: it can be automated and it can

also be applied to closed microreactors.

However, liquid-phase handling to coat a microreactor is not preferable due to

nonuniform removal of solvent. In other words, to overcome the problems of low

FIGURE 8.11 Reaction results on conversion of cyclohexene: (a) conversion of cyclohex-

ene on sputtered Pt in 5 and 100mm channels; (b) conversion of cyclohexene on sputtered Pt

and silica-supported Pt in 100 mm at the same design point of the reaction.18


activity of catalysts from CVD and PVD methods, the use of flame spray synthesis is

recommended since it yields porous catalysts without handling a liquid precursor.

The main data based on current research concern metal-on-oxide catalysts for

which many methods exist. Some consider physical treatment of the surface to coat

(anodization, plating, PVD, etc.), others involve a more or less complex chemical

preparation (suspension and sol–gel). The properties of the deposited layer vary to a

large extent, for example, the thickness, from nanometer (PVD) to near millimeter

scale (suspension). The textural properties of the oxide supports can in certain cases

reach that of traditional catalysts (suspension, sol–gel, powder plasma spraying

methods). “Physical” methods in general lead to more adhesive layers, but to less

activecatalysts. Themostwidely usedmethod involves direct synthesis on the surface.

The nanostructure and microstructure and the size of the supports affect conversion

efficiency, space availability, catalyst loading, and so on.We can simulate commercial

catalyst practice by sol–gel and impregnation methods to realize optimized coating

situations for future industrial implementation.

8.5 HYDROGEN PRODUCTION AND PURIFICATION

IN A MICROREACTOR

8.5.1 Hydrogen Production from SRM: Current Status

The demand for power sources with superior performance has increased as a result of

the rapid growth of the portable electronics market. Greater performance leads to

greater consumption of electrical power, and it has become difficult to secure a

sufficiently long-lastingpower source for these portable electronic devices, evenusing

lithium-ion batteries. The proton exchange membrane fuel cell (PEMFC), using

hydrogen as a energy source, has attracted much attention for portable electronic

devices since it has higher high volumetric energy density (for example, 2500WhL�1

for liquid hydrogen) than lithium-ion battery (400WhL�1).21 Thus, hydrogen pro-

duction is getting a lot of attention from today’s researchers, and also due to

consumption of gasoline and environmental concern.22–26

Steam reforming is an alternative process to produce hydrogen from organic

sources with the aid of a catalyst. Of many candidates being considered for hydrogen

fuel sources, methanol, ethanol/bioethanol, gasoline, and diesel are readily available

and currently being investigated. The use ofmethanol for steam reforming is attractive

due to its high energy density, low cost, easy transportation, and low reforming

temperature. The main reactions involved in steam reforming of methanol may be

presented by the following equations.

1. Steam reforming of methanol:

CH3OHþH2O Ð 3H2 þCO2; DHr ¼ 49:5 kJ mol�1

2. Methanol decomposition:

CH3OH Ð 2H2þCO; DHr ¼ 90:6 kJ mol�1

HYDROGEN PRODUCTION AND PURIFICATION IN A MICROREACTOR 301

3. Water–gas shift reaction:

COþH2O Ð H2 þCO2; DHr ¼ �41:2 kJ mol�1

Considerable work already exists in the literature on SRM for hydrogen

production using conventional macroscale reactors. It has disadvantages in terms

of size and weight that affect the compactness of fuel processor. The catalyst in

packed bed form also exhibits high-pressure drop and possible channeling of gases.

The low gas velocity is maintained in the reformer to achieve high conversion, but it

lowers the effective thermal conductivity of the catalyst bed. For an endothermic

reaction such as methanol steam reforming, temperature gradient in the packed bed

leads to lowered catalyst activity and falsified kinetics. In addition, hot and cold

spots are commonly encountered in the catalyst bed that results in overall poor

performance.

The use of microreactors for steam reforming of methanol can compete with the

conventional reactors due to the advantages of microreactor systems, including

lightweight, compactness, rapid heat and mass transport due to large surface to

volume ratio, and precise control of process conditionswith higher product yields.24–26

The high heat exchange efficiency in microreactors allows one to carry out reactions

under isothermal conditions. Also, microchannel reactors working under laminar flow

conditions show low pressure drop compared to randomly packed bed reactors. The

short radial diffusion time in microreactors leads to narrow residence time distribution

of reaction gases, which allows an optimum contact time between reactants and

catalysts avoiding the formation of unwanted by-products.

The catalysts for SRM can be divided into two types: non-noble metal (typically

nickel) andnoblemetals fromGroupVIII elements (typically platinum).Due to severe

mass and heat transfer limitations, conventional steam reformers are limited to an

efficiency factor for the catalyst, which is typically less than 5%. Therefore, kinetics

and thus the activity of the catalyst are rarely the limiting factors with conventional

steam reformers. The mass and heat transfer limitations have been shown to be

overcome by employing microchannel-based reactors, enabling intrinsic kinetics of

steam reforming to be exploited. In these systems, the noble Group VIII metals are

preferred since they exhibit much higher specific activities than nickel catalysts.

However, the high cost of Group VIII metals is driving some researchers to develop

alternative catalysts. The catalyst performance is greatly influenced by the type of

supports. A number of commercial and research derived noble metals (such as Pt and

Rh) loaded onto metal oxide supports (such as CeO2, ZnO, MgO, Al2O3, SiO2, ZrO2,

and TiO2) including more than one type of catalyst support have been tested for

hydrogenproduction frommethanol and ethanol.Although a number of catalysts have

shown promising results, problems with deactivation of the catalysts with an ensuing

decrease in hydrogen and carbon dioxide and an increase in carbon monoxide

production have been reported. Currently, there is little understanding of the behavior

of these catalysts in steam reforming reactions carried out in microreactors. Thus,

more basic research is necessary to find the optimized combination of catalyst and

support for hydrogen production.


8.5.2 Non-Noble Nanocatalysts for SRM Reactions to Produce Hydrogen

This section focuses on the development of Ni and Co non-noble nanocatalysts on

silica support in microchannel reactors for SRM reactions to produce hydrogen. The

SRMmicroreactor is a silicon-based microdevice with the dimension of 1.6 cm� 3.1

cm. It consists of vias, feed inlet, product outlet, and reaction zone with a volume of

9.6mm3 and 120 straight channels of 50mm width and 100 mm depth.27,28

The microreactors coated with silica-supported Co or Ni nanocatalysts were

prepared by a sol–gel procedure as described in Section 8.4. The precursors for Co

andNi catalysts are cobalt nitrate and nickel nitrate. The catalyst-coated microreactor

was dried by gentle heating and treated with 10%NH4OH solution for 30min to form

hydroxides ofmetal catalysts, followed bywashingwith DIwater to remove residuals

from ammonia treatment. Calcination completed the formation of oxides from

hydroxides. The oxides of cobalt and nickel were finally reduced with hydrogen to

active metals before packaging of the microreactor.

The calcination temperature for SiO2 supported catalyst was obtained by differen-

tial thermal analysis (DTA). For Ni/SiO2 (Figure 8.12), endothermic peaks are

observed at�100�C,which can be attributed to evaporation ofwater. Due to ammonia

treatment during synthesis of the sol–gel encapsulated catalysts, most of metal nitrate

salts are converted to metal hydroxides that are further converted to metal oxides by

heating in air. The broad endothermic peaks observed at 100–150�C can be attributed

to water loss and some of the metal nitrate salts getting decomposed. One exothermic

peakmaybeeither due to themetal hydroxidesgettingconverted to theoxidesordue to

structural changes of the surface species. Thus, we may conclude that all of the metal

hydroxides are converted to the metal oxides in temperature range of 350–400�C,which can be considered the minimum temperature for calcination.

To optimize the calcination temperature for Ni/SiO2, X-ray Diffraction (XRD)

analysis was performed by annealing the samples from 450 to 1000 �C.27 The resultsshow that the sharpness and intensity of the XRD peaks increased due to formation of

larger crystallites during the increase of the annealing temperature. This represents

evolution of the particle size and corresponds well to crystal growth. The higher

FIGURE 8.12 Differential thermal analysis of silica sol–gel-supported Ni catalyst.27


crystalline characteristics of the samples calcined at 1000 �C indicates that heating the

catalysts at very high temperatures may result in large crystallite sizes due to

aggregation. It has been reported that NiO crystallite in small size favors large Ni

surface area after reduction, which implies that higher calcination temperature is not

preferred for nanocatalysts. Similar observations of DTA and XRD were made for

Co/SiO2. Thus, we selected 450�C as the optimized calcination temperature for both

Ni and Co catalysts.

Since pure metallic Co and Ni are ferromagnetic, one can study the magnetic

properties of Co and Ni catalysts to understand the reduction efficiency during

hydrogenation and formation of chemical compounds of metal catalysts during

catalytic reactions. The saturation magnetization of the ferromagnetic component

in magnetic curves obtained fromVSMwas used along with the energy dispersive X-

ray (EDX) results to estimate the pure metallic Co and Ni in the catalysts.

Magnetization studies of the silica-supported Co catalyst before reduction with

hydrogen show paramagnetic behavior as cobalt is in its oxide forms (Figure 8.13),

which is also confirmed from the XRD and DTA results. Hydrogenation of the

catalyst reduces most, if not all, of the Co oxide to pure metal (the active phase for

SRM reaction), thus giving the catalyst the ferromagnetic behavior. The ferromag-

netic nature almost disappears in the postreaction catalyst sample as most of the

metallic Co yields nonferromagnetic species. Hence, the magnetization results

were used for ferromagnetic catalysts to estimate the pure metal percentage from

the saturation magnetization value of the ferromagnetic component obtained at

different stages.

The elemental analysis using EDX indicates both Ni and Co catalysts in the

microchannels show loadings of 5–6%. The EDX analysis at different locations of the

sample showsuniformdistributionof the catalyst in sol–gelmatrix.From transmission

electron microscopy (TEM) analysis shown in Figure 8.14, we can estimate the size

of the Co particles in silica sol–gel to be <10 nm. The BET analyses show that

–0.2

–0.15

–0.1

–0.05

0

0.05

0.1

0.15

0.2

1050–5–10

As-madeReducedAfter reaction

Applied Magnetic filed (kG)

Mag

neti

zati

on

(em

u g

–1)

FIGURE 8.13 Room-temperature magnetization curves of Co/SiO2 before reduction, after

reduction, and after SRM reaction using a vibrating sample magnetometer.27


specific surface area is 450m2 g�1 for Ni/SiO2 and 340m2 g�1 for Co/SiO2. The

Barrett–Joyner–Halenda (BJH) analyses show that the pore size of silica support is

32A�. With SiO2 porous structure, Ni and Co can be effectively distributed to obtain

larger surface area available for catalytic SRM reactions.

SRM reactions were conducted over Co or Ni nanocatalyst supported by silica

sol–gel in the temperature range of 180–240�C under atmospheric pressure. The flow

rates of 1:1 methanol/water between 5 and 20mLmin�1 were controlled using a

syringe pump (Figure 8.15). A cold trap was used to separate gaseous products from

aqueous methanol and water. Methanol conversion was calculated from volume

difference between the fed methanol–water mixture and the unreacted methanol–-

water mixture cooled with liquid N2 in the trap. The gaseous products containing H2,

CO2, and COwere diluted with helium and analyzed using a mass spectrometer (MS)

coupled with a residual gas analyzer (RGA) (QMS 200 gas analyzer from Stanford

Research Systems). Hydrogen selectivity was calculated on the basis of partial

pressures (proportional to moles) of different products.

FIGURE 8.14 TEM image of Co nanoparticles in silica synthesized by sol–gel method.27

FIGURE 8.15 Experimental setup for SRM reactions with Ni/SiO2 or Co/SiO2 catalyst.28


The main products of SRM reaction are hydrogen and carbon dioxide with a small

amount of carbon monoxide. The methanol conversion decreases as the flow rate is

increased (Figure 8.16) for both Co/SiO2 and Ni/SiO2 nanocatalysts with slightly

higher conversion for Ni/SiO2. While our studies with Ni/SiO2 show 53% methanol

conversion, Co/SiO2 shows only 37% conversion at 5mLmin�1 flow rate and 200�C.Thismay be attributed to larger specific surface area ofNi/SiO2 catalyst. However, the

catalysts’ behavior in microreactors for SRM is not fully understood. The decrease in

methanol conversion with increasing flow rate may be explained by lower residence

time of the reactants in the microreactor at higher flow rates. The hydrogen selectivity

(Figure 8.17) also decreases with increasing flow rate. The maximum hydrogen

FIGURE 8.16 Methanol conversion (CH3OH:H2O ratio of 1:1) as a function of flow rate at

200�C using silica sol–gel-supported Ni/SiO2 and Co/SiO2 nanocatalysts in 50mm channel

microreactor.27

FIGURE 8.17 Hydrogen selectivity (CH3OH:H2O ratio of 1:1) as a function of flow rate at

200�C using silica sol–gel-supported Ni/SiO2 and Co/SiO2 nanocatalysts in 50mm channel

microreactor.27


selectivity is �74% for Ni/SiO2 catalyst and �67% for Co/SiO2 at 5mLmin�1 flow

rate and 200�C.The temperature in the range of 180–240�Cdoes not have a significant

effect on methanol conversion for both catalysts, but the hydrogen selectivity is

maximum at 200�C for Ni/SiO2 catalyst and at 220�C for Co/SiO2 catalyst.

All the reactions described above were carried out within 8 h and no significant

deactivation of the catalysts was observed during this period. However, when these

reactionswere carried out over 10 h, deactivation of the catalystswas noticed,which is

consistent with the VSM analysis. The VSM results from the postreaction catalyst

sample (Figure 8.4) provide an estimate of�90%Co and�85%Ni being converted to

nonferromagnetic species after SRM reactions over 10 h. These species may be Ni or

Co compounds such as their oxides, carbonyls, and carbides.

8.5.3 Platinum Catalyst for Preferential Oxidation of CO in Hydrogen

As mentioned previously, there is a little amount of carbon monoxide (CO) produced

during steam reforming of methanol to produce H2. In general, CO poisons the

electrode catalyst of fuel cell. Thus, it is necessary to develop robust catalysts for

preferential oxidation of CO in an atmosphere of H2.29–31 The reactions for preferen-

tial oxidation of CO are shown below:

Desired reaction: COþO2�!Pt CO2

Undesired reaction: H2þO2�!Pt H2O

Platinumwas chosen as the catalyst, and a sol–gel supportwas utilized tomaximize

the surface area. High conversion of CO is required to reduce its concentration to a

level that is not detrimental to a proton exchange membrane (PEM)-based fuel cell.

High selectivity to CO2 is desired because hydrogen is used to generate electricity in

the fuel cell. The side reaction on oxidation of hydrogen to water reduces the energy

available for the fuel cell.

Themicroreactor used for CO oxidationwas 3.1 cm long and 1.6 cmwidewith 119

microchannels of 25mm in width and 100 mm in depth. The reactant gases were

premixed before being allowed to flow in the channels of the microreactor. An inlet

manifold was designed to promote a uniform distribution of flow among the reaction

microchannels. The inlet manifold consisted of two channels symmetrically off-axis

from the single outlet.32

The effect of the ratio of O2/CO onCOoxidationwas studied by keeping themixed

gas at 1 sccm and changing the airflow rate from 0.02 to 0.4 sccm at a constant

temperature of 200�C.As expected, the selectivity of oxidation of COdecreases as the

O2 to CO ratio is increased above 2. Surprisingly, the conversion of CO did not appear

to significantly increase as the ratio of O2 to CO was increased above 2. The effect of

temperature showed that conversion and selectivity were both maximized at the

minimum temperature tested, 120�C.The lower stoichiometric excess ofO2 displayed

a higher sensitivity of conversion with respect to increase in temperature. The

sensitivity of conversion and selectivity on reactor residence time was tested to


show that this particular catalyst and reactor combination favored the lowest space

velocity that could be achieved, 13 h�1. Additional experiments were then conducted

atWHSVof13 h�1 tofinalize theoptimal temperatureofoperation.The ratio ofO2/CO

was held constant at a stoichiometric ratio of 0.5. Figure 8.18 shows that 160�C is

optimal temperature of operation at these conditions. The catalyst deactivation study

showed that after 50 h of reaction both conversion of CO and selectivity to CO2

dropped significantly. We speculate that this could be due to two likely mechanisms:

(1) oxidation of the platinum and (2) sintering of the platinum nanoclusters into

larger particles. While limited recovery of the catalyst activity through repeated

reduction treatments suggests the former is more likely, the irreversible reduction in

exposed surface area through sintering of the platinumnanoclusters at the temperature

of operation is the likely cause for the remaining drop in activity. Carbon deposition is

not thought to be a major contributor due to the lack of a possible mechanism as

predicted by Chemkin simulations.33 Improvement of the catalyst activity and

efficiency will be a significant issue for future research and commercialization of

this technology in fuel cell applications.

The research described above for the development of catalysts for H2 production

and elimination of CO is very important for the development of microfuel cells.

Microfuel cells are high-energy-density sources for next-generation power portable

products such as personal digital assistant (PDA), laptop, cellular phone, and so on.

Fuel cells can provide more reliable, longer portable power than batteries. Microfuel

cell products compete with traditional power systems that utilize both direct and

indirect energy conversion methods. Direct methanol systems have the advantage of

room-temperature operation but offer only relatively low power density due to

methanol crossover through the proton exchange membrane and the low reaction

rate of methanol oxidation over the anode electrocatalyst. In contrast, reforming

systems generate electrical energy in the fuel cell from concentrated hydrogen

produced by steam reforming, for example, from methanol. Reforming systems

achieve high power density but are difficult to miniaturize due to the complexity

50.00%

55.00%

60.00%

65.00%

70.00%

75.00%

80.00%

85.00%

90.00%

95.00%

220200180160140120100

Temperature

Co

nvers

ion

or

sele

cti

vit

y

Conversion of CO

Selectivity to CO2

FIGURE 8.18 Temperature effect on the conversion of CO and selectivity to CO2 at a low

total flow rate of 0.2 sccm and O2/CO ratio of 0.5.32


of the required structure that includes not only a fuel reformer but also a vaporizer, a

CO removal unit, and various peripheral parts. However, themicrochannel fabrication

on a silicon substrate allows the various components of microfuel cell, such as steam

reformer, hydrogen purifier, driving units, sensors, and microvalves, to be integrated

by semiconductor technology andMEMS technology. Itmay also be possible, through

the use of these technologies, to achieve sufficient thermal insulation between the

reformer (with microreactor) and the periphery of the system, where the surface

temperature should be reasonable to touch. It is therefore expected that the small SRM

systems with high power density, sufficient to power portable electronic devices, will

be realized in the near future.

8.6 MICROREACTOR FOR GAS-TO-LIQUID TECHNOLOGY

Catalyst development for GTL technology using FT synthesis is a very active area of

research since it converts unconsumed natural gases into hydrocarbons that are

subsequently converted to fuels by hydrocracking for low-cost production and

easy transportation. Although FT synthesis has been developed by several worldwide

petroleum corporations such as Exxon Mobil, Shell, Sasol, Syntroleum, and Rentech

to help deliver the potential of the world’s untapped natural gas resources, a lot more

work is necessary to develop robust and stable catalysts in this regard.The catalysts for

FT synthesis generally contain metals such as cobalt, iron, nickel, and copper with

support materials of oxides and zeolites. More specifically, the influence of various

support materials such as titania, alumina, silica, and ceria on the activity of cobalt

catalysts forCOconversion have been studied.34–36Alumina is the primary choice as a

support for cobalt catalysts due to its thermal stability, although reduction of cobalt

oxide to cobalt is limited due to strong interaction between the support and cobalt

oxides. Addition of a second catalyst like iron may improve the selectivity to heavier

alkanes with less selectivity tomethane.37 Recent studies have shown that the catalyst

activity depends on the number of active sites located on the surface of the support

formed by the reduction process, determined by the particle size, loading, and the

degree to which the metal has been reduced.

8.6.1 Iron–Cobalt Mixed Catalysts for Fischer–Tropsch Synthesis

As discussed in earlier sections, microreactors are ideal systems for catalyst develop-

ment as they provide unique advantages such as low consumption of reactants, greater

speeds in catalyst characterization, and easy integration with other devices. More

significantly, high surface area ofmicrochannels to volume of the reactor inhibits gas-

phase free radical reactions and improves heat transfer for exothermic reactions, such

as FT synthesis to higher alkanes as shown below. Silicon microreactors with either 5

or 25 mmwide channels were coated with mixed metal Fe–Co in alumina sol–gel for

conversionof syngas (CO þ H2) to higher alkanes.Thevolumeof themicroreactors is

�9mm3 for both 5 mm (1200 in number) and 25mm (240 in number) wide channel

reactors.38–40

MICROREACTOR FOR GAS-TO-LIQUID TECHNOLOGY 309

COþH2�!200�300�C

Fe=Co on Al2O3 supportCH4þC2H6 þC3H8þ higher alkanes

Initial studies were performed in 5 mm channel reactors to study the conversion

rates at different ratios of H2:CO. The conversion of CO reachesmaximum (�32%) at

the H2:CO ratio of 3 with a total flow rate of 0.6 sccm and 230�C in 5mm channel

reactors. The lowCOconversion is likely due to insufficient coating of sol–gel in 5mmchannels of the reactor. This is indeed supported by higher conversion in 25mmchannel reactors (Figure 8.19) where sol–gel can penetrate easily (shown in

Figure 8.20b) allowing the reactants to have more interactions with the catalyst.

This may also explain the observed lower loading (�7%) than the intended loading

(12%) of supported catalysts from our EDX studies. We believe that the sol–gel

granules (less than 100 nm indicated by atomic force microscopy) and high viscous

nature of alumina sol–gel cause blocking of 5mm channels as shown in Figure 8.20a.

FIGURE 8.20 SEM image of alumina sol–gel encapsulated Fe/Co catalyst deposited on

(a) 5mmwide channelmicroreactor and (b) 25mmchannelmicroreactorwith 100mmdepth.38

FIGURE 8.19 Comparative studies on conversion of CO to alkanes in 5 and 25mm channel

reactors at different temperatures. Fixed H2/CO ratio of 3:1; 1 atm.38


However, the blockages of catalyst active sites in the microchannels may also occur

due to carbon formation from the water–gas shift reaction and can also decrease CO

conversion at higher temperature.

The study of temperature effect onCOconversion shows initial increasewith rising

temperature and decrease after 230�C due to exothermicity of the reaction as the

competition between activation and exothermicity progresses (Figure 8.19). The

effect of flow rate of syngas was also investigated. The reaction results show that the

conversion of CO decreases with increase in total flow rate as a result of decrease in

residence time of the reaction.

Although the conversion of CO is influenced by the width of the channels, the

selectivity toward different alkanes is not. Figure 8.21 shows the selectivity to

methane, ethane, and propane as a function of temperature in 5 and 25mm channel

reactors. Propane has been the major product of the reaction in both cases with

selectivity up to 78%. The residence time does not have any significant effect on

selectivity to alkanes in the microreactor.

The use of Fe–Co catalyst on alumina support for FT synthesis to higher alkanes is

basedon twoexperimental observations reported in the literature:while cobalt catalyst

inactivates water–gas shift reaction to some degree, iron has an advantage of

improving selectivity to higher alkanes with less selectivity to methane. Due to its

thermal stability, alumina with porous structure as a support for dispersing the metal

catalyst to a large total surface area with high catalytic conversion has been

investigated by Iglesia.34 However, reduction of cobalt oxide to cobalt is limited

due to strong interaction between alumina and cobalt oxide35 and formation of Co

aluminates.34 Furthermore, higher viscous nature of alumina solmakes the deposition

FIGURE 8.21 Selectivity of alkanes: methane, ethane, and propane in 5 and 25mm wide

channel reactors at 230�C; H2/CO ratio 3:1; 1 atm.38


process in Si microreactor difficult and results in nonuniform coating. Thus, alumina

has been replaced with silica as the catalyst support to improve catalyst coating

performance and to increase syngas conversion to alkanes due to good adhesion

between silica and Si microchannels (Figure 8.9). The particle size of silica in 25mmchannels is<100 nm. The field emission scanning transmission electron microscopy

(FE-STEM) analyses indicate that the particle sizes of the nanocatalysts in the silica

support is�6 nm (Figure 8.22). The CO conversions increase from 52%with alumina

(Figure 8.19) to 63% (Figure 8.23) with silica in 25mm channel microreactors.

30

35

40

45

50

55

60

65

70

75

80

250240230220210200190

Temperature (ºC)

% C

O c

on

versi

on

H2/CO 3:1, without RuH2/CO 2:1, with RuH2/CO 2:1, without RuH2/CO 3:1, with Ru

FIGURE 8.23 Comparative studies on CO conversion to alkanes at different temperatures

and two H2:CO ratios, using silica-supported Fe–Co catalysts with and without Ru at 1 atm and

a total flow rate of 0.4 sccm.41

FIGURE 8.22 FE-STEM image of silica-supported Co synthesized by sol–gel method in

25mm microchannel showing the particle size of �6 nm.41


8.6.2 Ruthenium Added as a Promoter to Iron–Cobalt Catalyst

for CO Conversion

Ruthenium as a promoter was added to the mixed Fe–Co catalyst to improve CO

conversion since it is one of the most active catalysts and can work at lower

temperatures.36,37,41 The highest conversion of CO on Ru–Fe–Co/SiO2 catalyst

reaches �78% at 220�C by adding only 0.4wt.% Ru at H2:CO ratio of 3:1. This is

much higher than that observed for Fe–Co/SiO2 catalyst (63%). The effect of H2:CO

ratio on CO conversion is less with Ru than that observed without Ru (Figure 8.24).

With Ru, CO conversion is almost same for both 3:1 and 2:1 ratios of H2:CO at 200�C,whereas the H2:CO ratio effect is significant without Ru (40% at 2:1 to 49% at 3:1).

This indicates that the role of Ru is more important in CO conversion than that of H2:

CO ratio. Ru can improveCOconversion at lower temperaturemore efficiently than at

higher temperature for both H2:CO ratios of 2:1 and 3:1 (Figure 8.23).

Addition of Ru not only increases CO conversion but also affects the selectivity to

alkanes.WhileRu–Fe–Co/SiO2 catalyst atH2:CO ratio of 3:1 yields�70%selectivity

to propane, �20% to ethane, and �10% to methane (Figure 8.24), these values with

Fe–Co/SiO2 catalyst are �80%, 17%, and 3%, respectively. The lower selectivity to

higher alkanes for Ru–Fe–Co/SiO2 may be due to possible side reactions to carbon

formation with all the FT catalyst metals (Fe, Co, Ru). For FT synthesis, at�200�Cor

below, the consecutive reaction of carbon hydrogenation to the CH2 monomer and its

consumption for chain growth inhibits CH4 formation and a carbon phase on the

catalyst is not allowed to grow.42 The selectivity changes mainly to methane at

elevated temperature, with possible deposition of carbon and catalyst deactivation

(especially with iron) reducing average chain length of the product molecules.41

The results from our studies are consistent with the steady-state isotopic transient

kinetic analysis (SSITKA), which indicates higher amount of CH4 in the presence of

0

10

20

30

40

50

60

70

80

90

3.532.521.5

H2:CO ratio

Sele

cti

vit

y (

%)

Methane with Ru Ethane with Ru

Propane with Ru Methane without Ru

Ethane without Ru Propane without Ru

FIGURE 8.24 Selectivity to various alkanes at different H2:CO ratios using silica-supported

Fe–Co catalysts with andwithoutRu in 25mmwide channel reactors at 230�C, 1 atm, and a total

flow rate of 0.4 sccm.41


Ru.41 Based on the increase in H2 chemisorption, Ru promotion increases the reaction

by increasing the Co metal dispersion, resulting in an increase in the number of

reaction sites. Similarly, Ru promotion of silica and titania-supported Co catalysts

exhibit a synergistic effect on CO hydrogenation, and higher hydrogenation ability of

Ru is due to a cleansing effect during COhydrogenation that prevents the formation of

carbon deposits on the catalyst surface. Meanwhile, it has been reported that Ru

promoter causes higher reducibility of Co ion and better dispersion of Cometal in the

catalyst support leading to higher CO conversion.34,35 In contrast, our magnetization

studies showaminimal effect of Ru onCo ion reduction and corroborate that Ru rather

has a synergistic effect on catalytic activities due to better dispersion of Co in the

catalyst support.41

8.7 PARALLEL MICROREACTOR SYSTEM FOR NANOCATALYST

SCREENING

Thedevelopment, optimization, and improvement of catalysts for use in commercially

viable processes face major challenges for scientific research with respect to the

complexities of the systems being investigated. Commercial catalysts, particularly

heterogeneous systems, are often complex multielemental, multicomponent

systems that are usually prepared via multistep procedures with the conditions under

which each intermediate step is performed often impact performance.3 Starting from

pharmaceutical research with respect to the identification and optimization of drugs,

combinatorial methods for synthesis and screening have become increasingly impor-

tant for other chemical and biological systems as well, for example, regarding

homogeneous or heterogeneous catalysts or other types of materials. The goal of

the development and application of parallel approaches to catalyst development is to

provide equivalent tools and techniques that significantly expand the ability to explore

the large complex parameter space addressed, while still generating results of an

adequate quality in terms of resolution and accuracy. The achievement of this goal

depends on a good understanding of how to integrate the parallel approach into the

target activities.

There are generally two approaches to combinatorial catalyst screening developed

among different research groups: (1) high sample throughput with less information

gained on each sample and (2) lower sample count with more complete information

obtained on each sample, depending on the preparation and screening techniques.43

The first approach normally pursues a large number of candidates (>100) to extract

10 or so “leads” based on relatively limited screening information. In contrast, the

second approach, which collects more detailed reaction information for a group of

leads, then facilitates further elimination of less effective candidates from this group.

In addition, the approach also yields information useful for process scaling and

deployment in actual industrial processes.

Specific advantages of using microreactors compared to conventional catalyst test

systems have been described previously. Due to unique mass and heat transfer

properties, as well as the uniform flow distribution, the microreactors are ideally


free of any dead zones. Micoreactors provide a well-defined setting of operating

conditions, and due to their small volume, fast changes in operating conditions can be

performed with minimal time required to reach equilibrium. More specifically,

microscale dimensions result in ultralow transport resistances such that the heat

and mass transfer are extremely fast. The capability of microreactors to test a number

of catalysts in separate reaction chambers, without any interference by flow mixing,

guarantees high accuracy and reliability of the results, while maintaining a high

analysis speed using compact devices. The microreactors consume reactant species

only very slowly, thus allowing extremely expensive or rare chemical systems to be

studiedmore economically. The small footprint of themicroreactor and its peripherals

need less infrastructure for operation, including floor space, energy supply, and

support personnel.43This compactness also allowsan experimental approachwhereby

anarrayofmicroreactors are connected to shareflowandanalysis equipment, allowing

a number of catalyst analysis experiments to be performed in parallel, increasing

experiment throughput with reducing time required for development. These technical

advantages are achievedwith added benefit that the small reactant volumes are highly

safe, being nearly exempt from explosion even when operating in what would

normally be considered explosive regimes. In addition, the environmental hazards

due to leakage are minimal.

Based on the combinatorial screening approach and our experience with micro-

fabricated reactor (microreactor) research described in previous sections, a silicon

microreactor-based parallel catalyst analysis system was developed for industrially

important reactions to dramatically decrease the catalyst development period and

reduce overall operating costs (Figure 8.25). Theflowof the reactantwas controlled by

Cole Parmer� mass flow controller. These flow meters also indicate the absolute

pressure in the line and are therefore useful for the alignment of the microreactors.44

Due to the low reactant flow rates required for themicroscale reactions, the pressure in

the system was built by carrier gas, helium, controlled by an Aalborg gas flow

FIGURE 8.25 A parallel array setup includes the following devices: (1) mass flow

controllers, (2) heating blocks with microreactors inside, (3) pressure gauges, (4) gas flow

controller of carrier gas, (5) multiline switching valve, (6) electrical connection, (7) mass

spectrometer, (8) gas chromatography, and (9) gases for calibration.44

PARALLEL MICROREACTOR SYSTEM FOR NANOCATALYST SCREENING 315

controller. This flow controller can provide a maximum of 10 sccm of flow rate. The

switching between the four lines was done by VALCO� multiposition valve that can

switch between amaximumof 16 reaction lines with the help of an electronic actuator

controlled both manually and automatically through the LabVIEW� program. The

pressure inside the setup is controlled by a Cole Parmer relief valve with the help of

LabVIEW and the inputs from the mass flow meter and pressure gauges are used to

open/close this relief valve such that the pressure inside the system ismaintained at the

set point provided by the user.

The analysis of the product stream can be performed with the help of Stanford

Research System� Residual Gas Analyzer/Mass Spectrometer or a Varian� Gas

Chromatograph (GC). The injection to the GC was done with help of an 8-port valve.

To achieve high-pressure synthesis, pressure has to be applied with the help of the

carrier gas controlled though an Aalborg valve (maximum 10 sccm) at downstream of

the multiposition valve. Since there is a pressure drop in nonselected streams by the

VALCOmultiposition valve, the downstream lines from the multiposition valvewere

vented out directly, and the carrier gas connection was made to the analyte line, thus

maintainingpressure for the active reaction line.The actual parallel array setup and the

complete process and instrument diagram (P&ID) are shown in Figures 8.25 and 8.26,

respectively.

The parallel array system of micoreactors with automatic control has been utilized

for FT synthesis (CO and H2) to alkanes. As discussed in the previous section, today’s

research groups in academia and industry have extensive interest to find a better

catalyst to increase productivity, to control hydrocarbon product distribution, and to

lengthen the catalyst life. The idea of using a parallel analysis system as a fast,

economic, and easy-to-scale-up solution for FT catalyst development fits these

requirements appropriately.

FIGURE 8.26 The P&ID of the parallel array setup of microreactors.44


To fully understand the activities of Fe/Co/Ru in FT synthesis described in the

previous section,we have tried to examine each catalyst in themicroreactor separately

using the parallel array system. Cobalt catalyst shows the highest CO conversion

(Table 8.2).44,45 From thevarious characterization studies,44 cobalt catalyst is found to

be well dispersed in the catalyst support matrix. This dispersion of catalyst allows a

larger surface area of the active metal to be in contact with the feed gases, and hence

higher conversion. The low conversion on the Ru catalyst (Table 8.2) can be attributed

to the low dispersion of the catalyst; also the agglomeration of the catalyst particles

reduced the amount of surface metal particles in contact with the feed gases. The

absence of water–gas shift reaction with cobalt catalyst compared to iron and

ruthenium is an indication of a different reaction mechanism. Iron and ruthenium

catalyst can be seen to promote the same reaction mechanism that can be the CO

insertion mechanism. The reaction temperature only increased the conversion of the

reaction and had no prominent effect on the selectivity to products formed. The

increase in residence time increased the production of long-chain compounds.

The smaller and better dispersed the metal particles are, the more surface area it

provides, and hence higherCOconversion is observed. From the above results, we can

conclude that in reactions involving nanocatalysis, dispersion, particle size, and

activity of supported metal catalysts play very significant roles.

In general, parallel catalyst screening approaches for heterogeneously catalyzed

gas-phase reactions have gained increasing popularity within the past years, as the

development of novel and better catalysts for chemical processes is still mainly an

empirical process utilizing existing technical know-how and experiences.

Simultaneous handling of many samples and large amounts of data impose a need

for the development of workflows and associated tool sets that minimize bottlenecks.

The key components of theworkflows are preparation system for synthesis, formation,

and treatment of arrays of samples, reactor system for evaluating the performance of

sample arrays for particularly target applications, characterization system for parallel

characterization of key properties of the sample arrays, and informatics system for

handling information and data flow between the various operations. Technical and

costs constraints have in general limited howeffectively themajor reactor sections can

be parallelized and for most groups have resulted in a systematic approach to the

implementation of parallel methodologies to catalyst screening.3

TABLE 8.2 COConversion on Three Different Catalysts in the Temperature Range

of 100–250�C44

Reaction Temperature (�C)

100 150 200 250

CO conversion on Co catalyst 31% 85% 92% 90%

CO conversion on Fe catalyst 74% 80% 82% 80%

CO conversion on Ru catalyst 31% 45% 52% 62%

PARALLEL MICROREACTOR SYSTEM FOR NANOCATALYST SCREENING 317

Typically, the highest level of parallelization functions as a primary screening tool,

providing evaluation of larger arrays of catalyst samples with a range of solutions

compromising on the degree to which they approximate conventional systems. The

secondary screening tools, which include the parallel array system described above,

aim to provide almost the same level of individual control of conditions and detailed

assessment of catalyst performance as typically performed on a bench-scale system.

The described parallel catalyst screening setup has the flexibility to optimize the

process conditions with accuracy equivalent to conventional units. This strategy of

integrating detailed catalyst screening and flexible process control within a large

parallel array system may help in the discovery of new commercial catalysts, the

results of which may be successfully validated at all levels from the miniaturized

parallel scale to final commercial operations.

8.8 SUMMARY

Microreactors are recognized in recent years as a novel tool for chemistry and

chemical process industry, such as fuel industry. The examples presented in this

chapter only represent a small fraction of the many studies for microreactors being

pursued by research groups worldwide. Invention of the next-generation process

technologies in fuel processing will strongly depend upon discovery of new catalyst

systems that result in attractive process economics with notably reduced environmen-

tal impact. The role of microreactor technology in the catalyst development and

commercialization is expected to gain more importance due to its various advantages

over conventional approaches. These advantages include improved safety character-

istics, enhanced rates of heat and mass transfer, reduced hardware footprint, lower

reagent costs, and ease of creating parallel systems for higher data throughput and

improved workflow efficiency. In contrast, micropacked-bed reactors are easy to

fabricate, butusually haveahigh-pressuredropduring thepassageofgases.Therefore,

thewalls ofmicroreactors aremore suitable for catalysis in this regard.Wehave used a

few reactions to illustrate the advantages of performing chemical reactions in

microreactors, which are particularly suited for highly exothermic and fast reactions.

However, some issues such as high cost of fabrication facilities, limitation to high-

pressure reactions, and catalyst coating still exist. Extensive efforts are being made to

address these problems.

Most of the new microreactor developments described here are the results from

studies of new chemistries with alternate synthetic routes, or as a result of scale-down

from a previous, more conventional multiphase reactor.10 As a result, a greater

opportunity exists to demonstrate the utility of microreactor as a robust enabling

technology for the discovery and development of newmultiphase catalyzed reactions.

Significant progress in scaling upmicrochannel process technology to commercial

scale has been made very recently. Focusing on solving challenges around device

fabrication, flow distribution and catalyst integration are the keys to success. Various

enabling technologies are allowing new microreactor designs to be fabricated.

Advances in MEMS and microelectronic industry from the perspectives of design


methodology, multifunctionality, new materials of construction, and fabrication

techniques have allowed the development of more sophisticated 3D geometries at

smaller length scales with increased spatial resolution. Adaptation of this knowledge

into microreactor technology will be an ongoing challenge, but it should create an

opportunity to combine various functionalities, such as onboard sensing and control

systems, with microreactors into an integrated package. Results from today’s flow

distribution models closely match experiments and are an integral tool for scaling

microchannel reactors to commercial capacities. Since the mechanism in micro-

fluidic flow is unique and distinct from that in conventional reactors, novel simulation

methods and tools are expected to be discovered and invented. Uniform coatings of

supported catalysts in the shape of thin films have been investigated extensively, but

the ideal solution is still unknown, which encourages chemists and chemical

engineers to pursue discovery of new technologies and effective modification of

current methods.

The larger task of integrating microreactors into functional microprocess systems

for small-scale fuel processing is in the early stages of development. Different

corporations and research centers, such as Sony, Casio, Exxon Mobil, and so on,

have made significant progress in microreactor technology. The microfuel processor

market can be divided into three device categories, portable electronics and portable

units for military and healthcare segments.23 There is a great potential for microfuel

processor to be integrated with microfuel cells to deliver more energy per volume or

weight than conventional batteries. The initial focus is to use it as a recharger. The use

of very small fuel cells in cell phones is a goal of many wireless carriers. When low

cost, high efficiency, and reasonably small size reach the required level in today’s

industry, and small enough to fit inside a cell phone, fuel cell powered electronic

devices would be the favorite of many electronics customers.

ACKNOWLEDGMENTS

We gratefully acknowledge the financial support of NSF-EPSCoR and Louisiana

BoR- RCS competitive Grant (to D.K.). We thank Drs. R. Besser, J. Palmer, and S.

Naidu for help and suggestions and the graduate students, V. S. Nagineni, A. Potluri,

W. Cao, Y. Liang, K. Shetty, and S.Mehta for their work described in this chapter.We

are indebted to these students and Mr. J. Fang who have helped us in simulation,

microfabrication, and experimental studies. We also thank Dr. K. Varaharamyan for

his support of this project.

REFERENCES

1. Haswell, S. J.; Skelton, V. Chemical and biochemical microreactors. Trends Anal. Chem.

2000, 19,6, 389–395.

2. Gavriilidis, A.; Angeli, P.; Cao, E.; Yeong, K. K. Y.; Wan, S. S. Technology and

applications of microengineered reactors. Trans. IChemE 2002, 80, 3–30, Part A.

REFERENCES 319

3. Karlsson, A.; Akporiaye, D. E.; Plassen, M.; Gillespie, R.; Holmgren, J. S. Parallel

heterogeneous reactor systems for catalyst screening. Microreactor Technology and

Process Intensification; ACS Symposium Series; Oxford University Press, 2005,

Chapter 4.

4. Ehrfeld, W.; Hessel, V.; L€owe, H.; State of the art of microreaction technology.

Microreactors: New Technology for Modern Chemistry; Wiley-VCH: Weinheim, 2000;

p 1–12.

5. Kawamura, Y.; Ogura, N.; Yamamoto, T.; Igarashi, A. A miniaturized methanol reformer

with Si-based microreactor for a small PEMFC. Chem. Eng. Sci. 2006, 61, 1092–1101.

6. Hessel, V.; L€owe, H.; Sch€onfeld, F. Micromixers—a review on passive and active mixing

principles. Chem. Eng. Sci. 2005, 60, 2479–2501.

7. Hessel, V.; L€owe, H. Mixing principles for microstructured mixers: active and passive

mixing. Microreactor Technology and Process Intensification; ACS Symposium Series;

Oxford University Press, 2005, Chapter 21.

8. Liang, Y.; Nassar, R.; Fang, J.; Kuila, D.; Varahramyan, K. Investigation of a novel

microreactor for enhancing mixing and conversion. Chem. Eng. Commun. 2008, 195(7),

745–757.

9. Liang, Y. Development of a novel microreactor for improved chemical reaction conver-

sion, PhD dissertation, 2005.

10. Mills, P. L.;Quiram,D. J.; Ryley, J. F.Microreactor technology andprocessminiaturization

for catalytic reactions—a perspective on recent developments and emerging technologies.

Chem. Eng. Sci. 2007, 62, 6992–7010.

11. http://www.mikroglas.com/index.php?PAGE_ID¼538&LANG_ID¼9.

12. Jensen, K. F. Microreaction engineering—is smaller better? Chem. Eng. Sci. 2001, 56,

293–303.

13. Fang, J.; Wang, W.; Zhao, S.; Fabrication of 3D microfluidic structures. Encyclopedia of

Microfluidics and Nanofluidics; Springer, 2008.

14. Zhao, S., Nano-scale platinum and iron-cobalt catalysts deposited in microchannel

microreactors for use in hydrogenation and dehydrogenation of cyclohexene, selective

oxidation of carbon monoxide and Fischer–Tropsch process to higher alkanes, PhD

dissertation, 2003.

15. Shin, W. C.; McDonald, J. A.; Zhao, S.; Besser, R. Etching characteristics of a micro-

machined chemical reactor using inductively coupled plasma. Proceedings of the 6th

International Conference onMicroreaction Technology (IMRET VI), p. 357, AIChE, New

Orleans, LA, 2002.

16. Meille, V. Review on methods to deposit catalysts on structured surfaces. Appl. Catal. A

2006, 315, 1–17.

17. Madou, M. Pattern transfer with additive techniques. Fundamentals of Microfabrication;

CRC Press, 1997, Chapter 3, p 89–144.

18. Zhao, S.; Besser, R. Selective deposition of supported platinum catalyst for hydrogenation

in a micromachined reactor. Proceedings of the 6th International Conference on

Microreaction Technology (IMRET VI), p. 289, AIChE, New Orleans, LA, 2002.

19. Okuzaki, S.; Okude, K.; Ohishi, T. Photoluminescence behavior of SiO2 prepared by

sol–gel processing. J. Non-Cryst. Solids 2008, 265, 61–67.

20. Campanati, M.; Fornasari, G.; Vaccari, A. Fundamentals in the Preparation of

Heterogeneous Catalysts. Catal. Today 2003, 77, 299–314.


21. Balat, M. Potential importance of hydrogen as a future solution to environmental and

transportation problems. Int. J. Hydrogen Energy 2008, 33, 4013–4029.

22. Holladay, J. D.; Hu, J.; King, D. L.; Wang, Y. An overview of hydrogen production

technologies. Catal. Today 2009, 139, 244–260.

23. Kundu, A.; Jang, J. H.; Gil, J. H.; Jung, C. R.; Lee, H. R.; Kim, S.-H.; Ku, B.; Oh, Y. S.

Micro-fuel cells—current development and applications. J. Power Sources 2007, 170,

67–78.

24. TaegyuKim, T.; Kwon, S. MEMS fuel cell system integrated with a methanol reformer for

a portable power source. Sens. Actuators A 2008, 204–211, available online.

25. Kundu,A.; Park, J.M.; Ahn, J. E.; Park, S. S.; Shul, Y. G.; Han, H. S.Micro-channel reactor

for steam reforming of methanol. Fuel 2007, 86, 1331–1336.

26. Men, Y.; Gnaser, H.; Zapf, R.; Hessel, V.; Ziegler, C.; Kolb, G. Steam reforming of

methanol over Cu/CeO2/g-Al2O3 catalysts in a microchannel reactor. Appl. Catal. A 2004,

277, 83–90.

27. Shetty, K.; Zhao, S.; Seetala, N. V.; Kuila, D. Synthesis and characteriztion of non-noble

nano-catalysts for hydrogen production in microreactors. J. Power Sources 2007, 163(2),

630.

28. Shetty, K.; Zhao, CaoW. S.; Seetala, N. V.; Kuila, D. Silica sol-gel supported nickel nano-

catalyst for hydrogen production using microreactors.Mater. Res. Soc. Symp. Proc. 2006,

885, 259.

29. Nikolaidis, G.; Baier, T.; Zapf, R.; Kolb, G.; Hessel, V.; Maier, W. F. Kinetic study of CO

preferential oxidation over Pt–Rh/g-Al2O3 catalyst in a micro-structured recycle reactor.

Catal. Today, 2008, available online.

30. Srinivasa, S.; Dhingrab, A.; Imb, H.; Gularia, E. A scalable silicon microreactor for

preferential co oxidation: performance comparison with a tubular packed-bed micro-

reactor. Appl. Catal. A 2004, 274, 285–293.

31. Snytnikov, P. V.; Popova,M.M.;Menc, Y.; Rebrov, E. V.; Kolb, G.; Hessel, V.; Schouten, J.

C.; Sobyanin, V. A. Preferential CO oxidation over a copper–cerium oxide catalyst in a

microchannel reactor. Appl. Catal. A 2008, 350, 53–62.

32. Zhao, S.; Hu, J.; Kuila, D.; Besser, R.; Nassar, R.; Palmer, J. Nano-scale platinum catalyst

in microreactors for preferential oxidation of CO amelioration in hydrogen fuel cell feeds.

Microreactor Technology and Process Intensification; ACS Symposium Series; Oxford

University Press, 2005, Chapter 8.

33. Ouyang, X.; Bednarova, L.; Ho, P.; Besser, R. S. Preferential oxidation of carbonmonoxide

in a thin-film catalytic microreactor: advantages and limitations. AIChE J. 2005, 51,6

1758–1772.

34. Iglesia, E. Design, synthesis, and use of cobalt-based Fischer–Tropsch synthesis catalysts.

Appl. Catal. A 1997, 161, 59–78.

35. Iglesia, E.; Soled, S. L.; Fiato, R. A.; Via, G. H. Bimetallic synergy in cobalt-ruthenium

Fischer–Tropsch synthesis catalysts. J. Catal. 1993, 143, 345–368.

36. Diehl, F.; Khodakov, A. Y. Promotion of cobalt Fischer–Tropsch catalysts with noble

metals: a review. Oil Gas Sci. Technol. Rev. IFP 2009, 64,1, 11–24.

37. Li, S.; Krishnamoorthy, S.; Li, A.;Meitzer, G. D.; Iglesia, E. Promoted iron-based catalysts

for the Fischer–Tropsch synthesis: design, synthesis, site densities, and catalytic proper-

ties. J. Catal. 2002, 206, 202–217.

REFERENCES 321

38. Nagineni, V. S.; Zhao, S.; Siriwardane, U.; Seetala, N. V.; Fang, J.; Palmer, J.; Kuila, D.

Microreactors for syngas conversion to higher alkanes: characterization of sol–gel

encapsulated nano-scale Fe–Co catalysts fabricated in the microchannels. Indus. Eng.

Chem. Res. 2005, 44(15), 5602.

39. Nagineni, V. S.; Zhao, S.; Indukuri, H.; Liang, Y.; Potluri, A.; Siriwardane, U.; Seetala, N.;

Fang, J.; Kuila, D. Characterization of alumina and silica sol–gel encapsulated Fe/Co/Ru

nanocatalysts in microchannel reactors for F–T synthesis of higher alkanes. Mater. Res.

Soc. Symp. Proc. 2004, 820, P51.

40. Zhao, S.; Nagineni, V. S.; Hu, J.; Liang, Y.; Fang, J.; Nassar, R.; Siriwardane, U.; Naidu, S.

V.; Varahramyan, K.; Palmer, J.; Kuila, D. Microreactor research and development at LA

Tech University: fabrication of silicon microchannel reactors for catalyst studies on

conversion of cyclohexene as a prototype and syngas to alkanes. Microreactor

Technology and Process Intensification; ACS Symposium Series; Oxford University

Press, 2005, Chapter 5.

41. Zhao, S.; Nagineni, V. S.; Seetala, N. V.; Kuila, D. Microreactors for syn-gas conversion to

higher alkanes: the promoter effect of ruthenium on silica supported iron–cobalt catalysts.

Ind. Eng. Chem. Res. 2008, 47(5), 1684.

42. Zhang,Y.;Hou, L.; Tierney, J.W.;Wender, I. Addition of acetylene to the Fischer�Tropsch

reaction. Energy Fuels 2007, 21(2), 640–645.

43. Ouyang, X.; Besser, R. S. Development of a microreactor-based parallel catalyst analysis

system for synthesis gas conversion. Cataly. Today 2003, 84, 33–41.

44. Mehta, S.; Thesis, M. S. Comparative studies of silica sol–gel supported iron, cobalt and

ruthenium catalysts for Fischer–Tropsch reaction in microreactors, Louisiana Tech

University, 2007.

45. Mehta, S.; Zhao, S.; Fang, J.; Kuila, D. Comparative studies of silica sol–gel supported Fe,

Co, and Ru catalysts for F–T reaction in Si-microreactors. Preprints Am. Chem. Soc., Div.

Fuel Chem. 2008, 53(2), 584–585.


9MICROFLUIDIC SYNTHESIS OFIRON OXIDE AND OXYHYDROXIDENANOPARTICLES

ALI ABOU-HASSAN, OLIVIER SANDRE, AND VAL�ERIE CABUIL

Laboratoire de Physicochimie des Electrolytes Collo€ıdes et Sciences Analytiques (PECSA),UMR 7195, Equipe Collo€ıdes Inorganiques, Universit�e Paris 6, Paris Cedex 5, France

9.1 INTRODUCTION

Iron oxides arewidespread in nature1 and are present almost everywhere in the global

system, even in Mars� soil.2 They are used for various applications in industry as

colored pigments, magnetic materials, ferrofluids, catalysts, and so on. There are 16

species of iron oxides, hydroxides, or oxyhydroxides, which will be collectively

referred to as iron oxyhydroxides in this chapter (Table 9.1). For more details on

different iron oxyhydroxides, the reader can refer to the book by Cornell and

Schwertmann.3

All the ironoxyhydroxides are ofgreat interest andhavenumerous applications, but

this chapter will focus only on materials widely studied in the past few years, that is,

magnetic iron oxides such asmagnetite Fe3O4 andmaghemite g-Fe2O3 nanoparticles.

These materials have wide-ranging technological applications when they are divided

into nanoparticles, ranging from navigation with magnetite (or Lodestone) to modern

high-density magnetic recording media and read head devices.

Magnetite Fe3O4 is a black ferrimagneticmineral containing both FeII and FeIII and

has an inverse spinel structure. Maghemite g-Fe2O3 is a red-brown ferrimagnetic

material, isostructural with magnetite but with cation-deficient sites.When the size of


323

magnetite or maghemite particles is reduced below about approximately 15 nm, the

particles are magnetic monodomains. It means that they have a permanent magnetic

moment whose intensity is proportional to their volume, but the direction depends

on the spontaneous fluctuations inside the grain. This magnetic moment is due to the

crystalline order, a characteristic of the spinel-like structure. For ultrasmall particles

(diameter smaller than a few nanometers), the surface disorder leads to a very

significant decrease in the moment.

Thus, nanometric ferro- or ferrimagnetic particles behave very differently from the

corresponding bulk materials and their magnetic behavior is called superparamag-

netism.4 The main characteristic of superparamagnetism is the spontaneous fluctua-

tion of the direction of themagnetic moment in the small magnetic grain, which is due

to the fact that, for very small ferromagnetic particles, themagnetic anisotropy energy

(KV) responsible for keeping the magnetization oriented in the easy axis of magneti-

zation is comparable to the thermal energy (kT). This results in zero magnetization in

zero field if the fluctuations are averaged over a timescale larger than their typical time

t.5 The fluctuation time t generally varies over a very broad timescale, depending on

the size of the particles (Figure 9.1)

Superparamagnetism also refers to the extremely large magnetic moments that

these nanoparticles bear (typically a few tens of thousand Bohr magnetons mB)

compared to themoment of isolated ions (5.4mB for Fe2þ and 5.9mB for Fe

3þ ).When

placed in external magnetic fields, the magnetization of a superparamagnetic suspen-

sion of nanoparticles is about 104 times larger than the magnetization of a paramag-

netic solution with an equivalent iron salt concentration.

Ferromagnetic bulk materials, once magnetized, show remanence (i.e., remain

partially magnetized even in the absence of an applied field) and therefore are used as

recording materials. In contrast, superparamagnetic materials differ from ferromag-

netic bulk substances because they do not retain any magnetization once the external

field is removed.6

Among others, superparamagnetic nanoparticles are largely used in magnetic

storage media,7 for biosensing applications,8 for medical applications, such as

TABLE 9.1 The Main Iron Oxyhydroxides

Oxide hydroxides and hydroxides Oxides

Goethite a-FeOOH Hematite a-Fe2O3

Lepidocrocite g-FeOOH Magnetite Fe3O4

Akagan�eite b-FeOOH Maghemite g-Fe2O3

Schwertmannite Fe16O16(OH)y(SO4)z � nH2O b-Fe2O3

d-FeOOH e-Fe2O3

Feroxyhyte d0-FeOOH W€ustite FeO

High-pressure FeOOH

Ferrihydrite Fe5HO8 � 4H2O

Bernalite Fe(OH)3Fe(OH)2

Readapted from Ref 3.

324 MICROFLUIDICSYNTHESISOF IRONOXIDEANDOXYHYDROXIDENANOPARTICLES

targeted drug delivery,9 as contrast agents in magnetic resonance imaging,10 and as

ferrofluids.11–14Formost of these applications, it is necessary to control theproduction

of magnetic nanoparticles, their monodispersity, and their states of aggregation, as

these physical parameters control their physical and physicochemical properties.

In the past few years, several research groups have proposed to use microfluidic

systems as a promising strategy for obtaining high-quality nanoparticles with high

monodispersity in a single-shot process without any subsequent size selection. This

chapter reviews the recent scientific literature concerning the use of microfluidics for

the synthesis of the iron oxide nanomaterials over the past 5 years. After a review of

the main synthesis methods used to prepare these materials in bulk chemistry, a few

works related to the synthesis of ferric oxide nanoparticles in microfluidics will be

introduced.

9.2 MAIN BULK PROCEDURES FOR THE SYNTHESIS

OF IRON OXIDE NANOPARTICLES

The chemistry of iron oxyhydroxides is very diversified and rich. Almost all the

species can be formed from solutions by a polycondensationmechanism, which is the

FIGURE9.1 (a) For very small ferromagnetic (FM) particles, themagnetic anisotropy energy

(responsible for keeping the magnetization oriented in certain direction) is comparable to

the thermal energy (kT). When this happens, the particles become superparamagnetic,

as thermal fluctuations randomly flip the magnetization direction between parallel and

antiparallel orientations. (b) Typical magnetization curve for superparamagnetic nanoparticles

(Langevin�s curve). Under a zeromagnetic field, the magnetic moments are randomly oriented,

but they progressively align parallel to the field direction when a magnetic field is applied.

When all the magnetic moments are aligned with the magnetic field, the curve attains a

saturationvalueMs, which is the product of the volume fractionF by the specificmagnetization

ms of the material (e.g., 3� 105Am�1 for colloidal maghemite, which corresponds to 33Bohr

magnetons per nm3).

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES 325

main topic of this section. For more details on the mechanisms and kinetics of the

precipitation from ionic solution, the reader can refer to Refs 15 and 16.

9.2.1 Metallic Cations in Solution and Polycondensation

Metal cationsMzþ inwater are solvatedbydipolarwatermolecules giving rise to aquo

cations [M(OH2)6]zþ .17 In the particular case of iron salts (chloride, nitrates, etc.),

dissolution inwater produces hexacoordinated aquo complexes [Fe(OH2)6]zþ , where

z¼ 2 or 3. The polarization of coordinatedwater molecules in the coordination sphere

strongly depends on the oxidation state and size of the cation. Charge transfer occurs

via the Fe–OH2 s-bond and electron density is transferred from the bonding 3a1molecular orbital of coordinated water molecules toward empty orbitals of the metal

cations.18

This charge transfer results in weakening of the O–H bond within the water

molecule, and the aquo complexes manifest Brønsted acid–base properties leading to

deprotonation of coordinated water molecules:

½FeðOH2Þ6�zþ þ hH2O!½FeðOHÞhðOH2Þ6�h�ðz�hÞþ þ hH3Oþ

The higher the oxidation state of the cation, the lower its size and the higher the

acidity of the complex. Thismakes the ferric aquo complexesmore acidic than ferrous

complexes and hydroxylation of the cations occurs at very distinct ranges of pH,

as indicated in Figure 9.2.

The hydroxylation ratio h of a complex increases when the pH increases and

aquohydroxo or oxohydroxo complexes are formed. In general, hydroxylated cation

monomers are instable in solution. They spontaneously condense because of the

nucleophilic character of the OH� ligands and the electrophilic character of cations.

Depending on the nature of the coordination sphere, two basic mechanisms of

the cations are proposed for the condensation of hydroxylated complexes.19

Aquohydroxo complexes condense through a nucleophilic substitution that proceeds

FIGURE9.2 Speciation of [Fe(OH)h(OH2)6�h](z�h)þ complexes of (a) Fe(II) and (b) Fe(III).

Reprinted with permission from Ref. 15. Copyright 2004 the Royal Society of Chemistry.


when the coordination number of the hydroxo ligand is increased andwatermolecules

are eliminated. This mechanism is called olation.

For oxyhydroxo complexes, there is no water molecule in the coordination sphere

of the complexes and therefore no leaving group. The condensation mechanism

proceeds in that case by two steps:

. First, association of the oxohydroxo complexes.

. Then, elimination of water molecule and formation of oxo bridges.

As long as charged complexes exist in solution, the condensation is limited and

polycationic species of 10–20 cationic atoms are formed. For iron complexes, which

are very reactive, ferric species condense very rapidly as soon as pH� 1 and it is

difficult to isolate polycationic species. On the contrary, ferrous complexes condense

only above pH 6, so some ferrous polycationic species have been isolated. When

only zero-charge complexes exist, the condensation is unlimited and a solid phase

precipitates:

n�MðOHÞzðOH2ÞN�z

�0 !½MðOHÞz�nþ nH2O

The precipitation is accompanied by the elimination of all the coordinated water

molecules and results in hydroxide formation. When the precipitated hydroxide is

unstable, it dehydrates spontaneously to form oxides and oxyhydroxides. For

example, alkalinization at room temperature of an aqueous solution of ferric ions

quasi-instantaneously leads to a poorly defined highly hydrated phase called ferri-

hydrite. Depending on the pH of the solution, the ferrihydrite suspension evolves

either to the oxyhydroxide phase a-FeOOH (goethite) or to the oxide phase a-Fe2O3

(hematite).


9.2.2 Kinetic Steps for the Precipitation Process

To understand why precipitation leads to the formation of nanometric particles and

how microfluidics can be used as a tool to elucidate nanoprecipitation mechanisms,

we examine in this section the kinetics of the polycondensation process.

We refer to the solid precursor as the zero-charged complex [Fe(OH)z(OH2)N�z]0

obtained by hydroxylation of the iron salt solution by an alkaline solution, as we

introduced previously. In the following, we abbreviate the zero-charged complexes

by P.

Three steps are usually considered to describe the formation of solid particles:

nucleation, growth (primary growth), and aging (secondary growth). Nucleation can

be homogeneous nucleation, heterogeneous nucleation, or secondary nucleation.20

We consider here the simplest case of a homogeneous nucleation, also called classical

nucleation theory (CNT), which occurs in the absence of a solid interface and consists

in combining solute molecules to produce nuclei.

This step leads to the formation of small clusters with a number of iron atoms

sufficiently large to overcome the nucleation barrier. The global rate of such a process

can be written as v¼ k[P]a, where the values of a can range between 4 and 10, as

proposed by Nielsen.16 These high a-values mean that the nucleation process is not

an elementary reaction but the result of many chemical elementary steps.

As the concentration of the precursor P generated by hydroxylation increases,

and possibly reaches a critical concentration, the condensation rate increases,

leading to the formation of many nuclei. This induces a decrease in the precursor

concentration and in the condensation rate, which can be annulled if the concen-

tration of P is very low. For any chemical process, the driving force behind the

homogeneous nucleation is the total free energy of the supersaturated solution DG.The overall free energy of the nucleation phenomena can be written as

DG¼DG1 þ DG2, where DG1 is the volume contribution resulting from the

difference between the chemical potential of ions in the nuclei (mPn) and in solution

(mP) and DG2 is the contribution of the interfacial energy (g) when a solid–liquid

interface of surface area (A) is created.20

If we suppose that the global reaction leading to a nuclei formation from p

precursors is

pP , Pp

then DG1 and DG2 can be written as

DG1 ¼ pðmPn�mPÞ ¼ �pRT ln S and DG2 ¼ gA

S is called “supersaturation,” which represents the ratio of the precursor concen-

tration in the solution to the solubility Cb of the macroscopic (bulk) solid, that is,

S¼ [P]/Cb.

Spontaneous nucleation can occur if S> 1 (DG1< 0), while no nuclei can form

when S< 1 (DG1> 0). However, even if S> 1, the nuclei can disappear if their size is


not sufficient to overcome the energy barrier due to the competition with interfacial

energy DG2.

DG2 ¼ gp2=3ð36p�v2Þ1=3

The total free enthalpy is thus DG ¼ �pRTLnSþ gp2=3ð36p�v2Þ1=3The variation in the total free enthalpy with respect to p reaches a maximum when

q(DG)/qp¼ 0. This allows to define a critical number of precursor molecules p�, acritical spherical radius r� beyond which the growth of nuclei is spontaneous, and theenergetic barrier that the system should overcome to reduce its surface energy and to

minimize the total free energy DG, DG�:

p* ¼ 32p�v2g3

3ðRTLnSÞ3 ; r* ¼ 2�vgRTLnS

; and DG* ¼ 16

3

p�v2g3

ðRTLnSÞ2 ¼p*

2RTLnS

From the above equations, it follows that the higher the saturation ratio S, the

smaller the critical nuclei size r� and the higher theDG�. Indeed, for a givenvalue of S,all particles with r> r� will grow and all particles with r< r� will dissolve. Figure 9.3illustrates this thermodynamic approach for the nucleation process for several cases

of supersaturation.

When theconcentrationof theprecursor reduces below theminimumconcentration

for nucleation, the latter stops, whereas the growth continues until the saturation

equilibrium concentration of the precipitated species is reached (i.e., the solubilityCb

of the bulk solid). In the classical ion-mediated crystal growth, growth occurs by

addition of soluble species to the solid phase. The uniformity of the size distribution

can be achieved through a short nucleation period that generates all the particles

obtained at the end of the nucleation followed by a self-sharpening growth process. At

this stage, the system is under kinetic control; the smaller particles growmore rapidly

than the larger ones because the free energy driving force is larger for smaller particles

than for larger ones if the particles are slightly larger than the critical size r. Figure 9.4

shows the variation in the precursor concentration with time during the precipitation

in the ideal case when growth successively follows the nucleation step. This is the

(a)

(b)

(c)

ΔG

p

ΔG *

p*

FIGURE 9.3 Illustration of the overall free energy DG as a function of the number of the

precursors p in the nuclei. (a) S< 1; (b and c) S> 1 and Sc> Sb.


famous model proposed first by LaMer and Dinegar to explain the mechanism of

formation of sulfur sols.21,22 But in most systems, depending on the concentration

of the precursor and the relative rates of the precursor formation and nucleation,

nucleation and growth can occur successively or at the same time.

Ideally, a requirement to achieve the monodispersity of the nanoparticles is that

nucleation and growth are separated, in time or in space (in separate vessels). In

practice, nearly monodisperse size distribution can be obtained by quickly stopping

the nucleation growth (thermal quenching) or by supplying a reactant source to keep

saturated conditions during the whole reaction.24

Growth processes are traditionally referred to as ripening or coarsening. Two

primary growthmechanisms, illustrated in Figure 9.5, are commonly active to varying

degrees during the ripening process.25 In the first growth mechanism, known as

Ostwald ripening, larger particles grow at the expense of smaller ones, which are less

stable because the solubility of a particle depends on its dimension, according to the

Gibbs–Thomson equation26

Cr ¼ Cb expð2sVm=rRTÞwhere Cr and Cb are the solubility values of the nanocrystals and the corresponding

bulk solid, respectively, s is the interfacial tension, Vm is the molar volume of the

materials, r is the radius of particles, R is the gas constant, and T is the temperature.

The coefficient 2sVm=RT called “capillary length” is usually on the order 1 nm.27 The

FIGURE 9.4 Cartoon illustration of nucleation and growth during the preparation of

monodisperse nanoparticles. Reprinted with permission from Ref. 23. Copyright 2004 the

Royal Society of Chemistry.


solubility based on the Gibbs–Thomson equation describes the solubility of colloidal

particles whose radius is larger than about 20 nm. For nanoparticles with r¼ 1–5 nm,

the value of the capillary length approaches the particle radius and the particle

solubility ln(Cr) becomes strongly nonlinear against r�1,28 presumably because the

interfacial tension s for a particle with a small number of atoms can no longer

be approximated by the value of the macroscopic solid phase. In addition, it is not

necessary that the nuclei are made of the thermodynamically stable crystalline phase;

they could be made either of an amorphous phase or of a metastable allotropic phase.

The kinetics of Ostwald ripening crystal growth can usually be described by the

following power law:29

DðtÞ ¼ D0þ k � t1=n

whereD0 is the initial particle size (diameter),D(t) is the size at time t, and k is a rate

constant for the limiting step. The exponent n is determined by the nature of the rate-

limiting step. It is equal to 1 when the rate of the growth is controlled by diffusion in

solution, equal to 2 when it is controlled by diffusion at the particle surface, and equal

to 3 when it corresponds to the interface dissolution/precipitation step.

In the secondgrowthmechanism,knownasSmoluchowski ripening, particles grow

by coalescence through convection or active mixing.

The nanoparticles themselves act as the building blocks for crystal growth. An

oriented pair of nanoparticles aggregates and fuses (like in thermal sintering),

eliminating the crystalline defects at their interface and releasing interfacial energy.31

Nucleation clusters

Crystal growth

Primary nanoparticles

> 3 nm

Amplifications

Single crystal

Mesoscale assembly

Iso-oriented crystal

FIGURE 9.5 Alternative mechanisms of growth for single crystals. The classical crystal

growthmodel is described by the path on the left. The path on the right involves the arrangement

of primary nanoparticles into an iso-oriented crystal via oriented attachment, which can form

a single crystal upon fusion of the nanoparticles and elimination of the grain boundaries.

Partially adapted from Ref. 30.


This less known mechanism for nanoparticles growth through oriented attachment

wasfirst discovered in the samples ofhydrothermally treatednanocrystallineTiO2.32–34

Natural nanoparticles of iron oxy/hydroxyl oxides were also found to grow by this

mechanism under certain geochemical conditions.35

Despite its great popularity (565 citations according to ISIWebofScience inMarch

2009), the classicalmodel proposed byLaMer seems to be rigorously appropriate only

to the system it was developed for (sulfur sols) and other closely analogous systems

of inorganic molecular crystals.36 Shortly after LaMer�s work, Turkevich called it a

“theory of great tradition” but ultimately rejected it to interpret the kinetics of the

synthesis of gold sols by reduction of chloroauric acid with citrate ions.37 Another

issue in generalizing the LaMer mechanism to other systems is that at the end of the

nucleation period, nuclei aggregation or Ostwald ripening may have already started

to generate larger nuclei that are more stable than smaller ones, resulting in a range of

nuclei sizes.36,38 As another limitation of LaMer�s theory and its variants, diffusion isnot always the rate-determining step in particle growth.39

Microfluidics appears as a very relevant tool to study the kinetics of the synthesis of

particles. Indeed, in a reactor as the one illustrated in Figure 9.6, due to the steady-state

laminar flow, an almost linear relation exists between the position in the reactor and the

time.40 A direct observation of the reaction mixture at several points of the reactor, if

any suitable detection method is available, provides information about the kinetics of

the nucleation and growth processes.41 Mixing and observing at the same time will

reduce the dead time that is, even for highly efficientmixers, very long compared to the

rate of precipitation reactions. Mixing in continuous-flow microreactors operating

under laminar flow occurs by diffusion of the species at the point of confluence.42

Compared to bulk chemistry, where mean concentrations are used to describe

the chemical system, chemistry in laminar flows needs to account for the local

FIGURE 9.6 Cartoon showing a typical Y-shaped continuous-flow microfluidic reactor

operating under laminar flow, where the reagents mix by diffusion. At the interface, nano-

particles nucleate and grow.


concentration of species and their resulting gradients. In the case of the nucleation

process, a local precursor concentration can be defined at every point of the reactor,

giving rise to a local supersaturation and then to different nucleation phenomena. It

means that different nuclei with different sizes will be formed depending on the local

conditions. This normally should favor the increase in polydispersity of nanoparticles,

unless a fast mixing of the reagents can occur, creating a burst nucleation followed by

a fast growth.

Anyway, microfluidics appears as a versatile tool to screen the effect of several

parameters on the size and shape of particles. Indeed, mixing and residence times can

be easily manipulated.

9.2.3 Case of Magnetite and Maghemite Nanoparticles

Numerous bulk chemical methods can be used to synthesize magnetic nanoparticles:

coprecipitation of iron salts,43–46 sol–gel synthesis,47 hydrothermal reactions,48

hydrolysis and thermolysis of precursors,49 synthesis in microemulsions,50 flow

injection synthesis,51 and electrospray synthesis.52 Contrary to the case of silica or

titania, for which a large variety of organometallic precursors exist, providing a good

control of the precipitation kinetics, organometallic iron precursors are less abundant

and highly reactive; thus, synthesis involving these precursors cannot be used for the

synthesis of iron oxide particles.

Until now, only one process that has been extended to microfluidics for the

synthesis of magnetic nanoparticles is the most used technique in bulk chemistry,

that is, coprecipitation of ferrous and ferric salts in alkaline medium. As polyol

processes and thermal decomposition processes can be easily (with some conditions)

extended to microfluidics, they will also be discussed here.

9.2.3.1 CoprecipitationCoprecipitation is a facile andaconvenientway to synthesize ironoxides (eitherFe3O4

or g-Fe2O3) inwater froma stoichiometric aqueousFe2þ /Fe3þ salt solutionby adding

a base under inert atmosphere at room temperature or elevated temperature. The

chemical reaction of Fe3O4 formation may be written as

Fe2þ þ 2Fe3þ þ 8OH� ! Fe3O4þ 4H2O

This procedure is in fact a polycondensation process, with nucleation and growth

steps. Quantitative data on nucleation and growth of hydrous metal oxides or

hydroxides are rather limited. The reason is that during the precipitation of solids,

competing reactions such as hydrolysis, condensation, and anion coordination take

place concurrently. The elucidation of the processes is even more difficult when

several solute complexes become involved in solid phase formation.53 In the case of

iron oxides such as magnetite and maghemite, due to the high reactivity of Fe(II) and

Fe(III), fast hydrolysis and condensation occur, leading to concurrent nucleation

and growth (primary and secondary) and thus wide size distribution.


According to thermodynamics of the precipitation reaction, complete precipitation

of Fe3O4 should be expected at a pH between 8 and 14, with a stoichiometric ratio of

2:1 (Fe3þ /Fe2þ ) in a nonoxidizing media.54 Experimental results show that the size,

shape, and composition of the magnetic nanoparticles strongly depend on the type of

anions associated with the ferric and ferrous cations (e.g., chlorides, sulfates, and

nitrates), themolar ratio (Fe3þ /Fe2þ ), the reaction temperature, the pH value, and the

ionic strength of the synthesis medium.

Magnetite Fe3O4 nanoparticles are sensitive to oxidation. Magnetite evolves into

maghemite g-Fe2O3 in the presence of oxygen. The latter is chemically stable in

alkaline and acidic media. During oxidation of magnetite to maghemite, various

electronor ion transfers are involveddependingon thepHof the suspension.Oxidation

under alkaline conditions involves the oxidation of the surface of particles, while

under acidic and anaerobic conditions, surface Fe2þ ions are desorbed as hexa-aqua

complexes in solution. Rapid and complete oxidation can be achieved in acidic

medium, as described by Massart and Cabuil.13

The main advantage of the coprecipitation process is that a large amount of

nanoparticles can be synthesized, without any surfactant. However, achieving by

this process a narrow particle size distribution without performing any size sorting is

still a challenge.

The first controlled preparation of superparamagnetic iron oxide particles by

alkalinization of an aqueous mixture of FeCl3 and FeCl2 salts was performed by

Massart in the 1980s.12 The synthesized nanoparticles were roughly spherical and

XRDmeasurements showed a diameter of 8 nm. Different parameters of this process

were largely studied to demonstrate the influence of the pH value, the base (ammonia,

CH3NH2, andNaOH), the added cations (N(CH3)4þ , CH3NH3

þ ,Naþ , Liþ ,Kþ , andNH4

þ ), and the Fe2þ /Fe3þ ratio, denoted x, on the coprecipitation yield, the

diameter, and the polydispersity of the nanoparticles. By modulating the different

parameters, magnetic nanoparticles with a mean diameter ranging between 16 and

4 nmwere prepared with a good reproducibility.13 The same results were obtained by

Vayssi�eres et al.55 and Jolivet et al.54,56–58 The latter explained the shape tailoring

by the variation of the electrostatic surface density of the nanoparticles determined

by the chemical composition of the crystal surface, the pH, and the ionic strength.

Babes et al.59 investigated the effect of iron concentration and the molar ratio x.

When x increased, the mean size of particles increased but the synthesis yield

decreased.

The particles synthesized by Massart�s process have been coated with a wide

range of molecular species such as amino acids, a-hydroxyacids (citric, tartric, andgluconic acids),60 hydroxamate (arginine hydroxamate),61dimercaptosuccinic acid

(DMSA),44,62 or phosphoryl choline.63 Bee et al.45 investigated the effect of the

concentration of citrate ions on the size of maghemite particles synthesized by

Massart�s process. Increasing the amount of citrate ions allows a decrease in the

diameter of citrate-coated nanoparticles from 8 to 3 nm. The authors explained these

results by the chelation effect of the citrate on the ferric and ferrous cations, preventing

nucleation, and by the adsorption of citrate on the nuclei, inhibiting the growth of the

latter. Also, the authors took advantage from the adsorbed citrate species to stabilize


the nanoparticles in aqueous dispersion at neutral pH.64 The effect of citrate during

the synthesis of ironoxidewasalso studiedbyLiu andHuang.65The crystallinityof the

synthesized nanoparticles decreased when the concentration of citrate was increased

during the synthesis, and the presence of citrate induced changes in the surface

geometry of the nanoparticles. Similarly, Barker et al.25 showed that by capping

the magnetite nanoparticles during the synthesis with heptanoic acid in trioctylamine

solvent, they were able to slow the ripening process, thus reducing the defects in the

nanoparticles.

Thus, the size and shape of the nanoparticles can be tailored by adjusting the pH, the

ionic strength, the temperature, the nature of the iron salts, and the Fe2þ /Fe3þ molar

ratio or by adding chelating organic anions (caboxylate, citric, gluconic, or oleic acid).

Other factors such as the mixing rate or the mixing manner can also affect the size

and polydispersity of particles. For example, a decrease both in the size and in the

polydispersity is observed when the base is added to the aqueous solution of metallic

salts compared to the opposite process where the solution of iron salts is added to the

alkaline solution.13 Surprisingly, injection flux rates do not seem to have a prepon-

derant influence on the nanoparticle synthesis.59

For magnetite nanoparticles, there were much more studies on the growth mecha-

nism and its consequences on the magnetic properties than on the nucleation step.

In the case of iron oxyhydroxide particles, dissolution–crystallization plays an

important role in growth mechanisms. It depends on several parameters such as

particle size, pH, ionic strength, presence of additives, and so on. Different values are

provided for the solubility products (Ks) of the several iron oxyhydroxides, according

to the authors. This may be due to differences in the characteristics of particles (size,

shape, surface state, etc.). In general, the Ks values of different iron oxyhydroxides

range from 10�44 to 10�34.3 Concerning the Fe3O4 solubility, there are in fact large

discrepancies in its solubility especially in alkaline medium.66 This is probably due to

the dissolution mechanism that involves the reduction of FeIII to FeII.67 As a result,

the solubility is a function of the reduction potential of the system, which is a real

problemunder alkaline conditions as dissolvedO2 is an extremely oxidizing agent and

kinetic effects may be important.

Ferrihydrite precipitation and aging in solution illustrates nicely how pH controls

the solubility and thus themechanisms of evolution of a population of nanoparticles in

suspension (Figure 9.7). The evolution of the small amorphous nuclei of ferrihydrite

obtained by alkaline precipitation of iron(III) salts (nitrate, chloride, etc.) strongly

depends on pH (thus on solubility): in the range 5� pH� 8, the insoluble ferrihydrite

germs transform by in situ dehydration and local rearrangement into very small

acicular particles of hematite a-Fe2O3, whereas for a higher solubility in acidic

(pH< 4) or alkaline (pH> 8) media, the transformation proceeds more easily via a

dissolution–crystallization process, leading to large goethite needles.

It seems that both Ostwald ripening and coalescence are involved in the growth of

the magnetite nanoparticles. Vayssi�eres et al.55 showed that the size of magnetite

precipitated in aqueous solution can be adjusted and stabilized against ripening by

controlling the pHand the ionic strength, the latter being imposed by anoncomplexing

salt in the precipitation medium.


9.2.3.2 Thermal DecompositionDecomposition of organometallic compounds in high-boiling organic solvents con-

taining stabilizing surfactants is a procedure that has been widely used to produce

magnetic nanoparticles because these nanoparticles are obtained with a high level of

monodispersity and size control. The iron organic precursors are [Fe(acac)3] (acac¼acetylacetonate), Fe(Cup)3 (Cup¼N-nitrosophenylhydroxylamine, C6H5N(NO)O

�),or Fe(CO)5. Hexadecylamine, oleic acid, and fatty acids are often used as surfactants.

The size and morphology of the nanoparticles can be controlled by adjusting the

reaction times, as well as the aging period, the temperature, the concentration and

ratios of the reactants, the nature of the solvent and the precursors, and the addition

of seeds. The decomposition of iron pentacarbonyl (Fe(CO)5) in a mixture of octyl

ether and oleic acid and at 100�C, followed by oxidation by trimethylamine oxide

((CH3)3NO) at elevated temperature, resulted in the formation of monodisperse

maghemite nanocrystals with a size of approximately 13 nm.68 The decomposition

of [Fe(acac)3] in the presence of 1,2-hexadecanediol, oleylamine, or oleic acid in

phenol ether leads directly to oxides.69 The use of iron(III) chloride salts as a iron

FIGURE 9.7 Influence of pH on the solubility of iron and ferric (hydro) oxide crystal

structure. Reprinted with permission from Ref. 15. Copyright 2004 the Royal Society of

Chemistry.


source has been proposed for the preparation of magnetic nanoparticles.70,71 For

details, the reader can refer to the reviews by Tartaj and Sato et al.72,73 The

nanoparticles obtained by this procedure are dispersible in different organic solvents

(hexane and toluene) but not in water, and sophisticated postpreparative methods are

needed to make these nanocrystals water soluble.

Until now, microfluidic reactors have not been used for this kind of synthesis, but

the possibility to manipulate small volumes could be of great interest as soon as the

structure of microreactors and the volumetric rate flow of different reagents permit

the control of the reaction times and the aging periods of the chemical reactions. Also,

due to the small dimensions of the channels, a precise control of the temperatures and

the temperaturegradients could bemore precise than in bulk.However, even if thermal

decomposition enables the synthesis ofmonodisperse nanoparticles, this processmust

be largely improved to be suitable formicrofluidic preparation, especially because the

different organic solvents usually used in this chemistry and the high temperatures

needed for the decomposition of the precursors are both incompatible with typical

PDMS channels and will require the use of quartz microreactors. And, even if quartz

microreactors can be used, the extensionof this chemistry inmicrochannels has to face

the problem of the bubbling of the boiling solvents and the elimination of effluents

(sometimes toxic) resulting from the decomposition of organometallic precursors.

Finally, even if nontoxic precursors can be used, the nanoparticles generated through

this process are dispersible in organic solvents, although the main applications of

magnetic nanoparticles nowadays require water-soluble particles, for example, in

biotechnology.

9.2.3.3 The Polyol ProcessThepolyol process refers to the use of polyols (e.g., ethyleneglycol, diethyleneglycol,

etc.) as solvents for the synthesis ofmetal ormetal oxide nanoparticles. Owing to their

high dielectric constants, polyols act as solvents able to dissolve inorganic com-

pounds. They offer a wide range of operating temperature for producing inorganic

compounds due to their relatively high boiling points.74 They also play the role of

reducing agents, producing the metal particles from the precursor, and of stabilizers,

allowing control of the growth of particles and preventing interparticle aggregation.75

In this method, the metal precursor is suspended in a liquid polyol and the solution is

heated to a temperature close to its boiling point. This chemical approach has been

described for the preparation of well-defined shapes and controlled sizes of oxides

nano- and microparticles.76–83

Cai and Wan84 successfully synthesized magnetite nanoparticles in several

polyols (ethylene glycol, diethylene glycol, triethylene glycol, and tetraethylene

glycol) from Fe(acac)3 at high temperatures (Figure 9.8). The mixture was slowly

heated to 180�C and kept at that temperature for 30min, and then quickly heated to

reflux (280�C) and kept at reflux for another 30min. Joseyphus et al.85 reported the

synthesis of Fe nanoparticles in polyols, their magnetic properties, and the influence

of the nature of polyols on the formation of Fe nanoparticles, but no precisions

either on the temperature gradients or on the final temperatures were given by the

authors.


Caruntu et al.86 described a new method for the synthesis of nanocrystalline

iron(II,III) oxide, based on the elevated temperature hydrolysis of chelate ion alkoxide

complexes in solutions of corresponding alcohol, diethylene glycol (DEG), and

N-methyl diethanolamine (NMDEA).

Polyol processes seem easy and efficient for the synthesis of iron and iron oxide

nanoparticles, but all the publications describe the manipulation and control of

temperature and temperature gradients. This control can be facilitated by using

microfluidic reactors due to the high surface to volume ratio. But at the same time,

if a polyol process has to be transposed in microreactors, one has to take into account

several important points. The precursor and the polyol have to be chemically

compatible with the reactor materials, which must also accept the high boiling

temperatures needed for this procedure as in the thermal decomposition of metal

complexes.

9.2.3.4 Synthesis in Constrained EnvironmentsDue to the importance of producingmagnetic monodisperse nanoparticles, numerous

methods havebeendeveloped toobtainnanoparticles ofmoreuniformdimensions and

well-defined size in constrained environments. These constrained environments

include reversed micellar structures of surfactants in nonpolar solvents,87–89 vesi-

cles,90 dendrimers,91 and cyclodextrins,92 and so on.

Here, we present a few examples of the synthesis in reverse micelles as they are

based on the same idea as the digital microfluidics, opposite to the synthesis in direct

micelles, using surfactants forwhich the counterion is themetallic cation.93The idea is

that the imprisonment of the reactions in smallmicro/nanoreactors can impose kinetic

FIGURE 9.8 Formation of metal-chelated complexes and their decomposition yielding

colloidal transition metal ferrites. Reprinted with permission from Ref. 86. Copyright 2004 the

American Chemical Society.


and thermodynamic constraints on particle formation and provide a confinement that

limits particle nucleation and growth.90

The first synthesis of magnetic nanoparticles in micelles was reported by Inouye

et al.,94 who prepared g-Fe2O3 and Fe3O4 by oxidation of Fe2þ salts. Recently, Lee

et al.95 described the use of the reverse micelles technology for the large-scale

synthesis of uniform and highly magnetic nanocrystals. The particle size is tuned

by varying the relative proportion of the iron salts, the surfactant, and the solvent.

Vidal-Vidal et al. presented a one-pot microemulsion method to produce monodis-

perse and coated small nanoparticles. The nanoparticles were formed by the copre-

cipitation reaction of ferrous and ferric salts with two organic bases cyclohexylamine

and oleylamine in a water-in-oil microemulsion. As the last example, the synthesis of

magnetic nanoparticles inside phospholipidic vesicles was reported in the literature

(magnetovesicles). Magnetoliposomes of 25 nm were prepared directly using the

phospholipid vesicle encapsulating FeII ions. The slow diffusion of the hydroxide ions

inside the vesicles causes the formation of magnetic nanoparticles.90

9.3 MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDENANOPARTICLES

g-Fe2O3 superparamagnetic iron oxide nanoparticles have been prepared for the first

time by our group in a continuous coaxial-flow microreactor that achieves small

diffusion distances and fast mixing times.1 In the same year, Frenz et al.96 reported the

use of droplet-basedmicroreactor for the synthesis of g-Fe2O3 nanoparticles. Later on,

the initial coaxial-flow setupwas improved by separating a nucleation reactor from an

aging reactor to synthesize another iron oxide, the antiferromagnetic goethite nano-

lathsa-FeOOH.97 This section reviews the different methods used for the preparation

of these nanoparticles in microfluidic reactors.

9.3.1 Synthesis of g-Fe2O3 Nanoparticles in Microfluidic Reactors

9.3.1.1 Synthesis in Continuous-Flow MicroreactorAs previously mentioned, there are several processes for the synthesis of magnetic

nanoparticles. Among them, only coprecipitation has been transposed in microreac-

tors, certainly because reactions occur in aqueous solution at room temperature.

It allows using PDMSmicroreactors without any sophisticated chemical engineering.

The chemical reaction summarizing the synthesis of magnetite nanoparticles is

Fe2þðaqÞ þ 2Fe3þðaqÞ þ 8OH�ðaqÞ ! Fe3O4ðsÞ þ 4H2OðlÞ

Thefirst trials of the synthesiswere run in a typical two-dimensionalY- orT-shaped

microreactors, made by lithography in polydimethyl siloxane (PDMS, Sylgard 184)

(Figure 9.9). In suchmicroreactors, evenwhen the concentration and the contact times

of the different reagents were varied, a magnetic precipitate appears at the interface

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES 339

and clogs the channels, leading to noncontinuous synthesis process. Clogging is

probably due to the adsorptionof themagnetic nanoparticles,which are in contactwith

the PDMS walls on the top and the bottom of the 2D channel.

To avoid the technical problems of adsorption and clogging, a 3D coaxial-flow

microreactor to mix the two coaxial flows of miscible fluids, one containing the iron

“precursor salts” and the other one a strong base, has been designed (Figure 9.10). It

offers the opportunity to enable a precision positioning of the precursors flow at the

center of the channel in both longitudinal and lateral dimensions, and on the other

hand, it avoids adsorption of any precipitate species onto the PDMSwalls as the latter

are totally wetted by the alkaline outer flow.1

The length of the capillary from the confluence region to the outlet was 3 cm. A

polytetrafluoroethylene (PTFE) tube (500mm ID and 10 cm long) leading to a sample

FIGURE 9.10 Coaxial flow device operating under laminar regime. The inset shows the

outlet of the inner capillary with the solution of iron(II) and iron(III) flowing into the stream of

TMAOH alkaline solution. Reprinted with permission from Ref. 1. Copyright 2008 the Royal

Society of Chemistry.

FIGURE 9.9 Synthesis of magnetic nanoparticles inside a 2D Y-shaped microreactor

showing the clogging at the interface. In a typical test, a solution of total iron salts with

different concentrations and 0.5 as molar ratio Fe(II)/Fe(III) was injected in one microreactor

arm and a solution of the alkaline solution tetramethylammonium hydroxide ((CH3)4NOH,

TMAOH) in the other arm.


vial was connected to the reactor outlet. Depending on the two flow ratesQin andQout,

the residence times ranged between 10 and 48 s. The outer capillary with 1.7mm

diameter (d) was shaped by the molding of a cylindrical tubing (Upchurch Scientific)

in a Petri dish with PDMS (Sylgard 184) and subsequent removal when the resin is

cured. The central capillary with 150mm ID and 360mm OD was obtained by fixing

a glass capillary (Plymicro�, usually used for capillary electrophoresis) inside the tip

of a micropipette (Gilson), the conical shape of which enables a precise centering.

The iron (II/III) solution of total concentration 10�2 mol L�1 (Fe(II)/Fe(III)¼ 0.5)

was injected into the inner flow with a volumetric rate flow Qin (1<Qin<100mLmin�1). The alkaline solution of TMAOH of concentration 0.172mol L�1

was injected in the outer flow with a volumetric rate flow Qout (100<Qout<400mLmin�1). TMAOH was chosen prior to any other base as the TMAþ cations

afford enhanced stability of colloidal oxide dispersion.43

9.3.1.2 Flow and Transport Modeling in the Continuous Flow MicroreactorMicroreactors� modeling is frequent in chemical engineering to deduce the right

hydrodynamic and chemical parameters needed for the chemical synthesis. The

objective of the present section is to describe some aspects of the flow behavior in

the microreactor that are useful for the synthesis of iron oxide nanoparticles. Indeed,

there were a few studies on the flow and mass transport in coflow mixers and they

reported fundamentally the analytical aspects of mixing under laminar flow, rather

than the influenceofmixingon thephysicochemical parameters of a reaction.Andreev

et al.98 developed a mathematical model to describe the hydrodynamic and mass

transfer during an acid–base reaction and showed that themaximumof themeanvalue

product is obtainedwhen the inner flow is considerably higher than the outer flow rate.

The same group99 studied again the mixing in a coflow mixer for injection flow

analysis and deduced that the mixing time is independent of the total volumetric rate

flowQtot after the point of confluence but depends on the volumetric rate flow ratio of

the outer flow to the inner flow a, a¼Qout/Qin. Mixing times decreased when a was

increased.

Confocal laser scanning microscopy compared to the numerical model for pH

mapping in the microreactor has been described in detail in our submitted work.100

Here, we review in brief some of the results important for the synthesis of the

nanoparticles.

Transport Modeling: pH Gradients Studies in bulk (see Section 9.2) have shown

that for the synthesis of high-quality magnetic nanoparticles and the good reproduc-

ibility of the results, parameters such as the ratio Fe(II)/Fe(III), the pH, the mixing

manner, the temperature, and so on need to be controlled.

pH is the most important parameter, as it controls the hydrolysis, the polyconden-

sation, and then the precipitation of the iron oxides. As Fe(II) and Fe(III) hydrolysis

occurs in two different pH ranges, and as Fe(II) absorption is responsible for the

magnetic properties of the magnetite,18 a very fast elevation of the pH (and thus a fast

mixing) in the alkaline zone would minimize the formation of nonmagnetic iron(III)

oxyhydroxides and increase the yield in magnetic nanoparticles.101


To understand how different volumetric rate flows affect pH gradients in a

microreactor, a theoretical approach based on solving the underlying mass transport

for different chemical species in the reactor, coupled with the comparison to

experimental images from confocal laser scanningmicroscopy (CLSM) experiments,

was performed.

First, as precipitation reactions are complex, pH variations occurring during the

mixing in the microreactor described above due to a model reaction that is the

neutralization of a strong acid, HCl, by the strong base, TMAOH, were studied. HCl

and TMAOH concentrations were the same as those used in the typical chemical

synthesis of magnetic nanoparticles, that is, 0.794 and 0.172mol L�1, respectively.

The experimental methodology consists in using a micromolar concentration of

pH-dependent dye (fluorescein) in the inner stream containing the HCl solution to

map the pH changes in the central jet stream where the acid–base reaction proceeds.

The experimental results were compared with those of the modeling, the pH

distribution in the reactor and the local fluorescein concentration being predicted

by modeling the underlying mass transport of the various species in the system.

The evolution of the concentration Ci ¼ Ciðr; tÞ of solute particles i follows thecontinuity equation

qCi

qtþ divðCiuþ jiÞ ¼ si

where u is the hydrodynamic velocity in the channel, calculated from the

Navier–Stokes equation together with the compressibility condition. For a dilute

system, it can be identified as the local averagevelocity of the solvent.si is the creation

term and represents the local creation rate of particles due to the chemical reactions.

ji is the diffusive flux of species i. The general expression of ji can be obtained fromnonequilibrium thermodynamics. For a dilute solution, cross-correlations are negli-

gible and it reduces to the Nernst–Planck expression:

ji ¼ �Di grad Ci þ Di

kBTCiZieE

The electric field E is chosen to satisfy the Henderson field condition.

Consequently, at any time the charge distribution has the time to relax and the local

electroneutrality conditionX

iZiCi ¼ 0 is valid.

The various solute species i are Hþ(aq), OH

�(aq), Cl

�, TMAþ , and fluorescein.

Diffusion coefficients were estimated from the values at infinite dilution: DHþ ¼9:2� 10�9 m2 s�1, DCl� ¼ 2:03� 10�9 m2 s�1, DOH� ¼ 5:28� 10�9 m2 s�1, and

DTMAþ ¼ 2� 10�9 m2 s�1.

To simulate the different fluorescein species during the acid–base reaction between

HCl and TMAOH, the following assumptions were made: (i) Fluorescein is diluted

in the hydrochloric acid, so no need to account for its charge. Convection diffusion

equation with no chemical reaction can be solved to calculate the local concentration

map of the total fluorescein denoted Flu. (ii) Fluorescein is diluted compared to Hþ


concentration in the acidified fluorescein solution and cannot interfere during the

acid–base reaction. (iii) Different ionic species of the fluorescein have the same

diffusion coefficient D 0.2� 10�9m2 s�1.

After calculating the local pH in the microreactor, the convection diffusion

equationwas solved for the fluorescein dye. Since the acid–base reactions arevirtually

instantaneous, it has been assumed that chemical equilibria are locally achieved for

different fluorescein species.

Confocal laser scanning microscopy experiments were used to confirm these

modeling results. Fluorescein is a well-known pH-sensitive dye that progressively

deprotonates when pH increases, as shown in Figure 9.11, producing the dianion F2�,which is the only fluorescent species (monoanion F� is quite less fluorescent). As the

pH rises in aqueous solution, the fluorescence signal increases starting at pH

approximately 5 and saturates at a maximum value above pH approximately 7. To

verify the pH dependence of the signal, a stock solution of disodium fluorescein was

prepared and diluted at 6mMin different pH-buffered solutions. These solutions were

perfused in both the inner and the outer capillaries, and the correspondingfluorescence

imageswere capturedwith the CLSMat themedian plane of the channel (i.e., far from

the walls that produce an artifact due to possible adsorption of fluorescein onto

PDMS). The images appeared uniform across a 300mm� 300mm view field. By

Cation Neutral Monoanion Dianion

0

20

40

60

80

100

12108642

% L

igh

t In

ten

sit

y

pH

HO O O H

COOH

HO O O

COOH

HO O O

COO

O O O

COO

OH

H

OH OH

H H

F+1F0 F-1 F-2

(a)

pKa0 = 2.08 pKa1 = 4.31 pKa2 = 6.27

(b)

FIGURE 9.11 (a) Chemical structures of fluorescein based on different pH values.

Fluorescein is a cation at pH< 2.08, neutral at 2.08< pH< 4.31, an anion at 4.31< pH<6.27, and a dianion at pH> 6.27. (b) Percentage light intensity relative to the value at pH 11 as

a function of pH for a solution of 6mM disodium fluorescein. Copyright 2009 the American

Chemical Society.


averaging the intensity over this constant area, a calibration plot has been drawn up in

Figure 9.11b. The results are in good agreement with cited work.102

Figure 9.12 shows the steady-state 2D fluorescence profile obtained by confocal

slicing in the middle of the outlet of the inner capillary where acidified fluorescein

solution (6mM)flows in the center surrounded by the outer alkaline TMAOH solution

upon application of a volumetric ratio a¼ 400.

The fluorescence burst in the central jet was interpreted as to have stemmed from

the deprotonation of fluorescein by the hydroxide ions diffusing toward the center

that remain in excess after reaction with Hþ . The hollow cylindrical shape on the

fluorescence image (cross sections (1) and (2) of the central stream) illustrates

the transition from a still acidic core (below pH 4) of the stream to a neutralized

“skin” near the pKa of 6.2 of fluorescein. It is exactly in this boundary region near the

pH equivalence that the coprecipitation of the iron salts in a synthesis experiment is

expected.

To illustrate the good agreement between the CLSM experiment and the model,

Figure 9.13 compares the 2Dfluorescence intensitymap of fluorescein represented by

a scale from cold (low intensity, acid) to warm colors (high levels, basic). The

symmetrybetween the experimental part of the imageon the left and thecalculatedone

on the right is a clear evidence that the assumptions made to calculate the diffusion

of the fluorescein species and the acid–base reaction are acceptable.

Finally, the flow rates ratioawas tuned between 40 and 400 and the corresponding

fluorescence profiles along the r¼ 0 axis were compared in Figure 9.14a. The good

matching between the predicted and the experimental intensity curves for several

values of a clearly validates the proposed simulation method.

Fluorescein is just a pH reporter dye, and the shape of the pH curves is in good

agreementwith the titrationof a strong acid (HCl) by a strongbase (TMAOH).Whenaincreases, the squeezing effect of the inner stream by the outer stream increases; the

inner stream containing the fluorescein dye and HCl is focalized, thus OH� diffusion

pathways to the Hþ ions are reduced. This decrease in the diffusion distance between

the reagent species implies a faster mixing and a steeper jump of pH near the

equivalence point (pH 7). In view of these results and in order to synthesize magnetic

nanoparticles, a¼ 400 was found to be the best suitable case because it offers the

FIGURE9.12 (a) CLSM image of the acid–base reaction in the presence of 6mMfluorescein

solution in the inner flowwith an applied volumetric rate flowa¼ 400. The imagewas recorded

in X–Y plane. (b) The three images are X–Z images, taken at the locations indicated by (1), (2),

and (3), constructed from a “Z-stack.” The Z-stack consisted of 20 slices, which were spaced at

20mm intervals. Copyright 2009 the American Chemical Society.


advantage of a fast mixing and a sharp pH jump. The choices a¼ 80 and 40 would

lead to a decrease in the yield in magnetic nanoparticles due to the precipitation of

antiferromagnetic iron hydroxides. These results were used for the preparation of the

stable colloidal and magnetic nanoparticles reported in our work1 using the coaxial-

flow microreactor and that will be resumed in the next section.

Application for the Synthesis of g-Fe2O3 Nanoparticles For nanoparticle synthesis

experiments, the inner solutionwas amixtureof iron saltswith a total ferric and ferrous

salt concentration of c¼ 10�2mol L�1 and amolar ratio Fe(II)/Fe(III) of 0.5, prepared

by mixing FeCl3 and “fresh” FeCl2�4H2O salts in diluted and degassed hydrochloric

acid (pH 0.10). The outer flow was an alkaline solution of tetramethylammonium

(TMAOH,0.172mol L�1),whichwas injectedwith an outer volumetric rate flowQout.

The reaction was “quenched” by fast solvent extraction (using didodecyl dimethyl

ammonium bromide in cyclohexane) to prevent any aging of the nanoparticles in the

aqueous solution.

The suspensions obtained in cyclohexanewere always stable and the nanoparticles

produced in the channel were fairly spherical with an average size around 7 nm. The

evidence of their crystallinity was provided by the electronmicrodiffraction pattern in

the inset of Figure 9.15, which shows the presence of the maghemite phase g-Fe2O3.

Although suspensions obtained in cyclohexanewere stable in a zeromagnetic field,

they sediment in the presence of amagnetic field gradient (e.g., on a strong permanent

FIGURE9.13 Comparison of the experimental and predicted fluorescence intensities inX–Y

plane with (a) the left half-plane representing the experimental fluorescence intensity and (b)

the right half-plane with the predicted fluorescence intensity for a¼ 400. Copyright 2009 the

American Chemical Society.


magnet), which suggests a magnetic character. This observation was confirmed by

magnetization measurements (using a vibrating sample magnetometer) on a stable

unquenched (aqueous) suspension: the magnetization curve (Figure 9.16) followed

the Langevin law typical of superparamagnetism, calculated for an assembly of

nanoparticles with a rather narrow distribution of diameters fitted by a log-normal law

of parameters d0¼ 6 nm and s¼ 0.2.

Bymeasuring both the volume fraction of nanoparticles f¼ 5.7� 10�5 (from iron

titration by atomic spectroscopy) and the saturation magnetizationMsat¼ 7.9Am�1

for the suspension, the specific magnetization of the materials was deduced ms¼Msat/f¼ 1.4� 105Am�1, which is much below the bulk value ofmaghemite g-Fe2O3

(3.5� 105Am�1), but not so far from thems value of about 2.6� 105Am�1 usually

obtained for nanoparticles of approximately the same sizes preparedwith the standard

large-scale synthesis. Therefore, it can be deduced that nanoparticles prepared within

few seconds in a millifluidic channel exhibit only a small decrease in ordering of their

magnetic moments compared to particles obtained within about 30min in bulk.

FIGURE 9.14 (a) Evolution of the experimental fluorescence profiles (&) when a¼ 40, 80,

and 400 along the symmetry axis (r¼ 0). The lines represent the predicted fluorescence

intensities calculated for each case from the simulated F2� concentrations. (b) Simulated pH

profiles along the symmetry axis (r¼ 0) for a¼ 40, 80, and 400. Copyright 2009 the American

Chemical Society.


FIGURE 9.15 TEM image of nanoparticles prepared in the channel (for flow rates

Qin¼ 100mLmin�1 and Qout¼ 400mLmin�1). The inset shows the electron microdiffraction

pattern with the Miller indices of g-Fe2O3. Reprinted with permission from Ref. 1. Copyright

2008 the Royal Society of Chemistry.

FIGURE9.16 Magnetization curve of a stable suspension inwater of nanoparticles produced

in the millifluidic device. The inset curves represent the fitting log-normal laws for the number

distribution (solid line) and the volume distribution (dotted line) of diameters. Reprinted with

permission from Ref. 1. Copyright 2008 the Royal Society of Chemistry.


9.3.2 Synthesis in Microdroplet Reactor

The use of a droplet-basedmicroreactor for the synthesis ofmagnetic nanoparticles by

coprecipitation of iron(II) and iron(III) by an alkaline solution of ammonium

hydroxide was reported by Frenz et al.96 The microfluidic device consisted of two

hydrodynamically coupled nozzles (Figure 9.17). During droplet formation in one

of the nozzles, the aqueous stream blocks the oil coming from the central channel,

leading to an increased oil flow through the secondnozzle.Once thedroplet is released,

the oil flow switches back to the first channel, allowing droplet pairing at various

flow rates.

Iron chloride solution was flushed into one arm of the nozzle and ammonium

hydroxide into the second arm, which led to droplet pairs containing the two reagents

(Figure 9.18). To start a reaction, the droplet pairs can be coalesced by applying an

electric field between the two on-chip electrodes.

Transmission electron microscopy (TEM) and electron microdiffraction pattern

showed that synthesized nanoparticles are monocrystalline and that the phase is

g-Fe2O3. The average particle size deduced from TEM images is smaller for the fast

compoundmixing (4� 1 nm) than for bulkmixing (9� 3 nm). The superparamagnetic

character of the nanoparticles is confirmed by the absence of hysteresis in the

magnetization curve.

The authors present their methods as a reliable way to produce magnetic nano-

particles. However, this method uses oils and surfactants to achieve the formation of

the droplets and their fusion. These “additives” can affect the nucleation and growth

mechanisms of the particles. Compared to microdroplet reactors, continuous-flow

reactors are easier to handle and are more representative of the conditions of the bulk

synthesis, with improved homogeneity, thus offering a better reproducibility of the

synthesized particles. The authors defend their use of microdroplets by the

FIGURE9.17 (a) Pairingmodule. Two aqueous phases are injected by the outer channels and

are synchronously emulsified by the central oil channel. The flow rates areQo¼ 800mLh�1 for

the oil, and Qx¼ 400mLh�1, Qy¼ 100mLh�1 for the aqueous phases. (b) Fusion module.

Paired droplets can be coalesced by applying an electrical voltage U between the two

electrodes. Qo¼ 650mLh�1, Qx¼ 100mLh�1, Qy¼ 60mLh�1. Reprinted with permission

from Ref. 96. Copyright 2008 Wiley.


enhancement inmixingby convectionand thedecrease in the reagent dispersion due to

the droplets that act as spatially isolatedmicroreactors. This concept is in fact the same

as the one evoked for the synthesis of nanoparticles inside vesicles or microemul-

sions.93–95 The faster mixing time reported in this system is 2ms, which is far larger

than the nucleation time. Moreover, this time is difficult to define as it is totally

arbitrary and depends on the concentration of reagents.

9.3.3 Synthesis of a-FeOOH Nanoparticles in Microfluidic Reactors

Another interesting iron oxyhydroxide phase is goethite (a-FeOOH), which is widelyfound in iron-rich soils.35 This claymineral constitutes the natural ochre pigment, and

because of its elongated shape, synthetic goethite is often used as a precursor of a-Fe“hard magnet” particles for magnetic recording.103 Because of this elongated shape,

suspensions of antiferromagnetic goethite/plate-like nanostructures (nanolaths) ex-

hibit an original magneto-optical effect and spontaneously self-assemble into a

nematic liquid–crystal phase above a threshold concentration.104 The importance

of particle shape for the improvement of magnetic properties, or the control of the

particle assembly, requires control of the synthetic conditions of these particles.105

The bulk methods reported for the synthesis of acicular (needle-like) goethite

particles are based on the agingof ferrihydrite nanoparticles obtainedbyalkalinization

of iron(III) salt solutions.106 They are indeed easily transferable to microfluidic

devices as illustrated by Abou-Hassan et al.97

FIGURE 9.18 Characterization of the iron oxide particles produced. (a) TEM image of the

nanoparticles. Inset: HRTEM image of a particle showing (220) spinel planes. (b) Electron

diffraction pattern indicating different planes of the spinel structure. (c) Magnetization M/Ms

(Ms is the saturation magnetization) as a function of the magnetic field H. Reprinted with

permission from Ref. 96. Copyright 2008 Wiley.


As discussed earlier, the alkalinization at room temperature of a solution of ferric

salts byanalkaline solution leads to theprecipitationof anamorphousoxidehydroxide

precipitate of ferrihydrite. At its minimum of the solubility, ferrihydrite can evolve to

hematite through an internal dehydration process, while in high-solubility domains

(very acidic or very alkaline solutions), the transformation via a dissolution–preci-

pitation mechanism is possible and leads to the formation of goethite. Goethite is the

most thermodynamically stable phase, but due to the low solubility of iron oxides, the

transformation into goethite is very slow, offering the possibility of a good separation

between nucleation and growth. Another possible mechanism for the formation of

goethite nanoparticles is the oriented attachment of iso-oriented ferrihydrite nano-

particles and then crystallization into goethite nanoparticles.107

Acomplexmicrofluidic device (Figure 9.19)was proposedbyAbou-Hassan et al.97

in order to physically separate the process of nucleation of the ferrihydrite nanopar-

ticles from their growth, leading to goethite particles. The nucleation of the primary

ferrihydrite nanoparticles is induced by diffusive mixing at room temperature in a

microreactor that is based on coaxial-flow geometry (R1). This mixing reactor is the

same as described for the synthesis of magnetic nanoparticles and is based on a three-

dimensional coaxial-flow device of two streaming reagents. At the outlet of this

micromixer, the suspended ferrihydrite nanoparticles are directly injected into the

microtubular aging coil R2,which consists of a transparent PTFE tube of 1.7mm inner

diameter and 150 cm total length continuously heated in a water bath at 60�C.Temperature profiles were calculated to determine the tubing length (and thus the

time) required for thefluid to reach a steady state.At the outlet ofR1 (before aging) and

FIGURE 9.19 The experimental setup used for the preparation of the ferrihydrite and

goethite nanoparticles. TMAOH¼ tetramethylammonium hydroxide. Reprinted with permis-

sion from Ref. 97. Copyright 2009 Wiley.


R2 (after aging), the resulting suspension is collected and analyzed by TEM and by

high-resolution TEM (HRTEM).

TEM pictures of the particles obtained after R1 show well-defined spherical

ferrihydrite nanoparticles (nanodots) about 4� 1 nm in size (Figure 9.20a).

HRTEM measurements (Figure 9.20b) show that the nanoparticles are monocrystal-

line, exhibiting atomic planes with an interplanar distance of about 2.5A�, which is

consistent with ferrihydrite nanoparticles.

Under the givenflow rate, the ferrihydrite solution reaches 60�C in about 1 s, that is,

within the first centimeter after it has entered the heated zone of the tubing. The

effective residence time is about 15min, as estimated from the length of the tubing

along which the fluid has reached the stationary temperature of 60�C.After aging for 15min under continuous flow in the aging coil R2, goethite plate-

like nanostructures were observed with an average length L¼ 30� 17 nm and width

w¼ 7� 4 nm (Figure 9.21a). This short aging time appeared to be sufficient for the

growth of crystalline and anisotropic goethite nanoparticles that differ only in smaller

sizes compared to those obtained after complete aging (1 day at 60�C).108

Moreover, the presence of few remaining ferrihydrite nuclei undergoing aggrega-

tion in the batch after 15min and even after 24 h at 60�C (data not shown) supports the

idea that goethite nanoparticles were formed by the aggregation mechanism rather

than by dissolution/reprecipitation.

Thus, the use ofmicrofluidic device allows to significantly accelerate the synthesis

of goethite nanoparticles from ferrihydrite nuclei. The novelty of this approach lies in

the separation of the nucleation of the primary particles (ferrihydrite) and the growth

of the goethite nanoparticles in two independent microreactors operating in different

conditions. In the nucleation microreactor, the streaming reagents are mixed by

molecular diffusion at room temperature in a flow-focusing geometry. The homoge-

neity of the mixture is ensured by the fast mixing time and the technical difficulty of

microchannel clogging owing to the fact that precipitation onto the walls is avoided

FIGURE 9.20 (a) TEM picture of the sample taken after precipitation in the microreactor R1

(before aging), showing ferrihydrite nuclei of 4� 1 nm diameter. The selected area diffraction

pattern (inset) is typical of two-line ferrihydrite. (b)HRTEM image of an individual ferrihydrite

nanoparticle with a 2.5A�lattice fringe. Reprinted with permission from Ref. 97. Copyright

2009 Wiley.


by 3D geometry. The use of a microfluidic device for aging, by minimizing local

temperature gradients, ensures a regular laminar flow, and finally leads to crystalline

plate-like nanostructures. These particles have approximately the same values of

aspect ratio and polydispersity index as those obtained in bulk synthesis (bulk

synthesis usually yields goethite nanoparticles with a typical length of about

250 nm and a width of 40 nm, with a polydispersity index of about 50% for both

dimensions109) but are smaller in size. The time required for aging decreases to 15min

(for a velocity of 0.1 cm s�1) compared to bulk synthesis (several hours or days). This

may originate from the small diameter of the aging reactor, causing a shear stress

that prealigns the primary ferrihydrite nanoparticles and speeds up their oriented

aggregation process.

9.4 PERSPECTIVES

Among all the ferric oxide nanoparticles, superparamagnetic ones (superparamag-

netic iron oxide nanoparticles, SPIONs) are of special interest because of their

applications in the field of imagery and therapy. At present, the screening of the

relation properties/structure is neither very easy nor economic in bulk, and continuous

microfluidic systems, enabling to add reagents along the entire length of the channel,

can be a very useful tool for screening the different parameters (size, surface

functionalization, and aggregation) allowing to optimize given properties. This

idea is illustrated in Figure 9.22 and can be summarized by three operations: adding,

mixing, and reacting.

Indeed, if several microreactors Ri are associated to form a series of microunit

operations, the synthesis ofSPIONnanoparticles and their surfacemodificationwould

FIGURE 9.21 (a) TEM image of the nanolaths after aging for 15min in the microtubular

loop R2, produced at pH 13 and under laminar flow. (b) HRTEM image of a nanorod particle.

Lattice fringe spacing is consistent with goethite. The dashed lines serve to highlight the

morphology and texture of the particle. Reprinted with permission from Ref. 97. Copyright

2009 Wiley.


be possible in an unique online process. The surface coating of the SPIONs by silica is

a good example. Silica has been extensively exploited as a coating material for

magnetic nanoparticles110–112 in order to get a protective, biocompatible, inert, and

hydrophilic surface with excellent anchoring points for derivatizing molecules.113

Several methods have been reported in the literature for the formation of super-

paramagnetic iron oxide silica nanocomposites including reactions performed under

St€ober conditions,114,115 microemulsions,116 emulsions,117 and aerosol pyrolysis.118

It seems that core–shell SPION silica nanoparticles and SPION luminescent silica

nanoparticles canbeobtained inmicrofluidic reactorswithout theuse of anysurfactant

(Abou-Hassan, unpublished results).

In the field of on-chip magnetic separation, there were many works devoted to the

separation of micrometric magnetic particles, but to date no separation of nanometric

particles on the microfluidic scale has been reported.119 A continuous-flow method

capable of both separating magnetic from nonmagnetic particles and separating

different magnetic particles from each other can be very helpful in synthetic

chemistry.

But the most important challenge in the field of nanomaterials� synthesis in

microfluidic lies in developing online characterization methods. For quantum

dots or metallic nanoparticles, optical characterizations allowing establishment of

a simple relation with particle size are available. That is not the case for ferric oxide

nanoparticles. As the latter have magnetic properties, online magnetic measurements

can perhaps be designed.On the samekind of idea, but of course generally for anykind

of materials, online characterizations using small-angle X-ray or neutron scattering

have to be developed. These are thus important in designing online nanoparticle

characterization techniques.

FIGURE 9.22 Cartoon illustrating the idea of online synthesis of functional nanomaterials.

PERSPECTIVES 353

REFERENCES

1. Abou-Hassan,A.; Sandre,O.;Cabuil,V.; Tabeling, P. Synthesis of iron oxide nanoparticles

in a microfluidic device: preliminary results in a coaxial flow millichannel. Chem

Commun. 2008, 1783.

2. http://marsrover.nasa.gov/.

3. Cornell, R.M.; Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions,

Occurrences and Uses, Wiley-VCH: 2003.

4. Dormann, J.L.; Firoani, D. Magnetic Properties of Fine Particles, Amsterdam: North-

Holland, 1992.

5. Eisenmenger, J.; Schuller, I.K.Magnetic nanostructures: overcoming thermal fluctuations.

Nat Mater. 2003, 2, 437.

6. Ferrucci, J.T.; Stark, D.D. Iron oxide-enhancedMR imaging of the liver and spleen: review

of the first 5 years. Am J. Roentgenol. 1990, 155, 943.

7. Sun, S.; Murray, C.B.; Weller, D.; Folks, L.; Moser, A. Monodisperse FePt nanoparticles

and ferromagnetic FePt nanocrystal superlattices. Science 2000, 287, 1989.

8. Miller, M.M.; Prinz, G.A.; Cheng, S.F.; Bounnak, S. Detection of amicron-sized magnetic

sphere using a ring-shaped anisotropic magnetoresistance-based sensor: a model for a

magnetoresistance-based biosensor. Appl Phys. Lett. 2002, 81, 2211.

9. Chourpa, I.; Douziech-Eyrolles, L.; Ngaboni-Okassa, L.; Fouquenet, J.F.; Cohen-

Jonathan, S.; Souce, M.; Marchais, H.; Dubois, P. Molecular composition of iron oxide

nanoparticles, precursors formagnetic drug targeting, as characterized by confocal Raman

microspectroscopy. Analyst 2005, 130, 1395.

10. Martina,M.S.;Fortin,J.P.;Menager,C.;Clement,O.;Barratt,G.;Grabielle-Madelmont,C.;

Gazeau, F.; Cabuil, V.; Lesieur, S. Generation of superparamagnetic liposomes revealed as

highlyefficientMRIcontrastagents for invivo imaging.JAm.Chem.Soc.2005,127, 10676.

11. Massart, R.;French Patent 79-188-42, 1979.

12. Massart, R. Preparation of aqueous magnetic liquids in alkaline and acidic media. IEEE

Trans. Magn. 1981, 17, 1247.

13. Massart, R.; Cabuil, V. Synth�ese en milieu alcalin de magn�etite collo€ıdale: controle durendement et de la taille des particules. J. Chim. Phys. 1987, 7, 84.

14. Massart, R.; Dubois, E.; Cabuil, V.; Hasmonay, E. Preparation and properties of

monodisperse magnetic fluids. J Magn. Magn. Mater. 1995, 149, 1.

15. Jolivet, J.-P.; Chaneac, C.; Tronc, E. Iron oxide chemistry. From molecular clusters to

extended solid networks. Chem. Commun. 2004, 481.

16. Nielsen, A.E.; Kinetics of Precipitation, Oxford: 1964.

17. Baes, C.F.; Mesmer, R.E. The Hydrolysis of Cations, Wiley: New York, 1976.

18. Jolivet, J.P.Metal Oxide Chemistry and Synthesis. From Solution to Solid State, Wiley:

Chichester, 2000.

19. Jolivet, J.P.; De la solution �a l�oxyde, EDP Sciences: 1998.

20. Kashciev,D.NucleationBasic Theorywith Applications; Butterworth-Heinemann; 2000.

21. LaMer, V.K.; Dinegar, R.H. Theory, production and mechanism of formation of mono-

dispersed hydrosols. J Am. Chem. Soc 1950, 72, 4847.

22. KerKer, M.; LaMer, V.K. Particle size distribution in sulfur hydrosols by polarimetric

analysis of scattered light. J. Am. Chem. Soc 1950, 72, 3516.


23. deMello, A.J.; deMello, J.C. Microscale reactors: nanoscale products. Lab Chip 2004,

4, 11N.

24. Zhang, Q.; Liu, S.-J.; Yu, S.-H. Recent advances in oriented attachment growth and

synthesis of functional materials: concept, evidence, mechanism, and future. J Mater.

Chem. 2009, 19, 191.

25. Barker, A.J.; Cage, B.; Russek, S.; Stoldt, C. Ripening during magnetite nanoparticle

synthesis: resulting interfacial defects and magnetic properties. J Appl. Phys. 2005, 98,

063528.

26. Sugimoto, T. Preparation of monodispersed colloidal particles.Adv Colloid Interface Sci.

1987, 28, 65.

27. Kabalnov, A.S.; Shchukin, E.D. Ostwald ripening theory: applications to fluorocarbon

emulsion stability. Adv Colloid Interface Sci. 1992, 38, 69.

28. Talapin, D.V.; Rogach, A.L.; Haase, M.; Weller, H. Evolution of an ensemble of

nanoparticles in a colloidal solution: theoretical study. J Phys. Chem. B. 2001, 105, 12278.

29. Borg, R.; Dienes, G.; The Physical Chemistry of Solids, Academic Press: Boston, MA,

1992.

30. C€olfen, H.; Mann, S. Higher-order organization by mesoscale self-assembly and trans-

formation of hybrid nanostructures. Angew. Chem., Int. Ed. 2003, 42, 2350.

31. Gilbert, B.; Zhang, H.; Huang, F.; Finnegan, M.; Waychunas, G.; Banfield, J. Special

phase transformation and crystal growth pathways observed in nanoparticles. Geochem

Trans. 2003, 4, 20.

32. Penn, R.L.; Banfield, J.F.Morphology development and crystal growth in nanocrystalline

aggregates under hydrothermal conditions: insights from titania. Geochem Cosmochim.

Acta 1999, 63, 1549.

33. Penn, R.L.; Banfield, J.F. Oriented attachment and growth, twinning, polytypism, and

formationofmetastablephase: insightsfromnanocrystallineTiO2.AmMiner.1998,83,1077.

34. Oskam, G.; Hu, Z.; Penn, R.L.; Pesika, N.; Searson, P.C. Coarsening of metal oxide

nanoparticles. Phys Rev. E 2002, 66, 11403.

35. Banfield, J.F.; Welch, S.A.; Zhang, H.; Ebert, T.T.; Penn, R.L. Aggregation-based crystal

growth and microstructure development in natural iron oxyhydroxide biomineralization

products. Science 2000, 289, 751.

36. Watzky, M.A.; Finke, R.G. Transition metal nanocluster formation kinetic and mecha-

nistic studies. A new mechanism when hydrogen is the reductant: slow, continuous

nucleation and fast autocatalytic surface growth. J. Am. Chem. Soc. 1997, 119, 10382.

37. Turkevich, J.; Stevenson, P.C.; Hillier, J. A study of the nucleation and growth processes

in the synthesis of colloidal gold. Discuss Faraday Soc. 1951, 11, 55.

38. Matijevic,E.Preparationandpropertiesofuniformsizecolloids.ChemMater.1993,5,412.

39. Overbeek, J.Th.G. Monodisperse colloidal systems, fascinating and useful. Adv Colloid

Interface Sci. 1982, 15, 251.

40. Salmon, J.-B.; Dubrocq, C.; Tabeling, P.; Charier, S.; Alcor, D.; Jullien, L.; Ferrage, F. An

approach to extract rate constants from reaction–diffusion dynamics in a microchannel.

Anal Chem. 2005, 77, 3417.

41. Sounart, T.L.; Safier, P.A.; Voigt, J.A.; Hoyt, J.; Tallant, D.R.; Matzke, C.M.; Michalske,

T.A. Spatially-resolved analysis of nanoparticle nucleation and growth in a microfluidic

reactor. Lab Chip 2007, 7, 908.

REFERENCES 355

42. Knight, J.B.; Vishwanath, A.; Brody, J.P.; Austin, R.H. Hydrodynamic focusing on a

silicon chip: mixing nanoliters in microseconds. Phys Rev. Lett. 1998, 80, 3863.

43. Cabuil, V.; Massart, R.; Bacri, J.C.; Perzynski, R.; Salin, D. Ionic ferrofluids: towards

fractional distillation. J Chem. Res. 1987, S, 130.

44. Fauconnier, N.; Pons, J.N.; Roger, J.; Bee, A. Thiolation of maghemite nanoparticles by

dimercaptosuccinic acid. J Colloid Interface Sci. 1997, 194, 427.

45. Bee, A.; Massart, R.; Neveu, S. Synthesis of very fine maghemite particles. J Magn.

Magn. Mater. 1995, 149, 6.

46. Neveu-Prin, S.; Cabuil, V.; Massart, R.; Escaffre, P.; Dussaud, J. Encapsulation of

magnetic fluids. J Magn. Magn. Mater. 1993, 122, 42.

47. Biddlecombe, G.B.; Gun�ko, Y.K.; Kelly, J.M.; Pillai, S.C.; Coey, J.M.D.; Venkatesan,

M.; Douvalis, A.P. Preparation of magnetic nanoparticles and their assemblies using a

new Fe(II) alkoxide precursor. J Mater. Chem. 2001, 11, 2937.

48. Wan, J.; Cai, W.; Feng, J.; Meng, X.; Liu, E. In situ decoration of carbon nanotubes with

nearly monodisperse magnetite nanoparticles in liquid polyols. J Mater. Chem. 2007,

17, 1188.

49. Kimata, M.; Nakagawa, D.; Hasegawa, M. Preparation of monodisperse magnetic

particles by hydrolysis of iron alkoxide. Powder Technol. 2003, 132, 112.

50. Chin, A.B.; Yaacob, I.I. Synthesis and characterization of magnetic iron oxide nano-

particles via w/o microemulsion and Massart�s procedure. J Mater. Process. Technol.

2007, 191, 235.

51. Salazar-Alvarez, G.; Muhammed, M.; Zagorodni, A.A. Novel flow injection synthesis

of iron oxide nanoparticles with narrow size distribution. Chem Eng. Sci. 2006, 61,

4625.

52. Basak, S.; Chen, D.R.; Biswas, P. Electrospray of ionic precursor solutions to synthesize

iron oxide nanoparticles: modified scaling law. Chem Eng. Sci. 2007, 62, 1263.

53. Bell, A.; Matijevic, E. Growth mechanism of hydrous chromium(III) oxide spherical

particles of narrow size distribution. J Phys. Chem. 1974, 78, 2621.

54. Jolivet, J.P.; Froidefond, C.; Pottier, A.; Chan�eac, C.; Cassaignon, S.; Tronc, E.; Euzen, P.Size tailoring of oxide nanoparticles by precipitation in aqueous medium. A semi-

quantitative modelling. J. Mater. Chem. 2004, 14, 3281.

55. Vayssi�eres, L.; Chan�eac, C.; Tronc, E.; Jolivet, J.P. Size tailoring of magnetite particles

formed by aqueous precipitation: an example of thermodynamic stability of nanometric

oxide particles. J Colloid Interface Sci. 1998, 205, 205.

56. Jolivet, J.P.; Belleville, P.; Tronc, E.; Livage, J. Influence of Fe(II) on the formation of the

spinel iron oxide in alkaline medium. Clays Clay Miner. 1992, 40, 531.

57. Jolivet, J.P.; Vayssi�eres, L.; Chan�eac, C.; Tronc, E. Synthesis of iron oxide-based

magnetic nanomaterials and composites. Mater Res. Symp. Proc. 1997, 432, 145.

58. Jolivet, J.P.; Chan�eac, C.; Tronc, E. Synthesis of iron oxide-basedmagnetic nanomaterials

and composites. Comp. Rend. Series 2002, 5, 659.

59. Babes, L.; Denizot, B.; Tanguy, G.; Le Jeune, J.J.; Jallet, P. Synthesis of iron oxide

nanoparticles used as MRI contrast agents: a parametric study. J. Colloid Interface. Sci.

1999, 212, 474.

60. Fauconnier, N.; Bee, A.; Roger, J.; Pons, J.N. Adsorption of gluconic and citric acids on

maghemite particles in aqueous medium. Prog Colloid Polym. Sci. 1996, 100, 212.


61. Fauconnier, N.; Bee, A.; Roger, J.; Pons, J.N. Synthesis of aqueous magnetic liquids by

surface complexation of maghemite nanoparticles. J Mol. Liq. 1999, 83, 233.

62. Roger, J.; Pons, J.N.; Massart, R.; Halbreich, A.; Bacri, J.C. Some biomedical applica-

tions of ferrofluids. Eur Phys. J. Appl. Phys. 1999, 5, 321.

63. Denizot, B.; Tanguy, G.; Hindre, F.; Rump, E.; LeJeune, J.J.; Jallet, P. Phosphorylcholine

coating of iron oxide nanoparticles. J Colloid Interface Sci. 1999, 209, 66.

64. Fauconnier, N.; B�ee, A.; Roger, J.; Pons, J.N. Synthesis of aqueous magnetic liquids by

surface complexation of maghemite nanoparticles. J Mol. Liq. 1999, 83, 233.

65. Liu, C.; Huang, P.M. Atomic force microscopy and surface characteristics of iron oxides

formed in citrate solutions. Soil Sci Soc. Am. J. 1999, 63, 65.

66. Tremaine, P.R.; LeBlanc, J.C. The solubility of magnetite and the hydrolysis and

oxidation of Fe2þ in water to 300�C. J Sol. Chem. 1980, 9, 415.

67. Sweeton, F.H.; Baes, C.F. The solubility of magnetite and hydrolysis of ferrous ion in

aqueous solutions at elevated temperatures. J Chem. Thermodyn. 1970, 2, 479.

68. Hyeon, T.; Lee, S.S.; Park, J.; Chung, Y.; Na, H.B. Synthesis of highly crystalline and

monodisperse maghemite nanocrystallites without a size-selection process. J Am. Chem.

Soc. 2001, 123, 12798.

69. Sun, S.; Zeng, H. Size-controlled synthesis of magnetite nanoparticles. J Am. Chem. Soc.

2002, 124, 8204.

70. Jana, N.R.; Chen, Y.; Peng, X. Size- and shape-controlledmagnetic (Cr,Mn, Fe, Co, Ni)

oxide nanocrystals via a simple and general approach. Chem. Mater. 2004, 16,

3931.

71. Park, J.; An,K.;Hwang,Y.; Park, J.-G.;Noh,H.-J.; Kim, J.-Y.; Park, J.-H.;Hwang,N.-M.;

Hyeon, T. Ultra-large-scale syntheses of monodisperse nanocrystals. Nat Mater. 2004,

3, 891.

72. Tartaj, P.; Serna, C.J. Synthesis of monodisperse superparamagnetic Fe/silica nano-

spherical composites. J Am. Chem. Soc. 2003, 125, 15754.

73. Sato, S.;Murakata, T.; Yanagi, H.;Miyasaka, F.; Iwaya, S. Hydrothermal synthesis of fine

perovskite PbTiO3 powders with a simple mode of size distribution. J Mater. Sci. 1994,

29, 5657.

74. J�ez�equel, D.; Guenot, J.; Jouini, N.; Fie�evet, F. Submicrometer zinc oxide particles:

elaboration in polyol medium and morphological characteristics. J. Mater. Res. 1995,

10, 77.

75. Feldmann, C.; Jungk, H.-O. Polyol-mediated preparation of nanoscale oxide particles.

Angew. Chem., Int. Ed. 2001, 40, 359.

76. Tzitzios, V.K.; Petridis, D.; Zafiropoulou, I.; Hadjipanayis, G.; Niarchos, D. Synthesis

and characterization of L10 FePt nanoparticles from Pt–Fe3O4 core–shell nanoparticles.

J Magn. Magn. Mater. 2005, 294, e95.

77. Chow, G.M.; Kurihara, L.K.; Kemner, K.M.; Schoen, P.E.; Elam,W.T.; Ervin, A.; Keller,

S.; Zhang, Y.D.; Budnick, J.; Ambrose, T. Structural, morphological, and magnetic study

of nanocrystalline cobalt–copper powders synthesized by the polyol process. J Mater.

Res. 1995, 10, 1546.

78. Viau, G.; Ravel, F.; Acher, O.; Fivet-Vincent, F.; Fivet, F. Preparation and microwave

characterization of spherical and monodisperse Co20Ni80 particles. J. Appl. Phys. 1994,

76, 6570.

REFERENCES 357

79. Viau, G.; Ravel, F.; Acher, O.; Fivet-Vincent, F.; Fivet, F. Preparation and microwave

characterization of spherical and monodisperse Co–Ni particles. J Magn. Magn. Mater.

1995, 140, 377.

80. Viau, G.; Fivet-Vincent, F.; Fivet, F. Nucleation and growth of bimetallic CoNi and FeNi

monodisperse particles prepared in polyols. Solid State Ionics 1996, 84, 259.

81. Mercier, D.; Lvy, J.C.S.; Viau, G.; Fivet-Vincent, F.; Fivet, F.; Toneguzzo, P.; Acher, O.

Magnetic resonance in spherical Co–Ni and Fe–Co–Ni particles. Phys. Rev. B: Condens.

Matter 2000, 62, 532.

82. Feldmann, C. Preparation of nanoscale pigment particles. Adv Mater. 2001, 13, 1301.

83. Giri, A.K.; Chowdary, K.M.; Majetich, S.A. ACmagnetic properties of compacted FeCo

nanocomposites. Mater. Phys. Mech. 2000, 1, 1.

84. Cai, W.; Wan, J. Facile synthesis of superparamagnetic magnetite nanoparticles in liquid

polyols. J Colloid Interface Sci. 2007, 305, 366.

85. Joseyphus, R.J.; Kodama, D.; Matsumoto, T.; Sato, Y.; Jeyadevan, B.; Tohji, K. Role of

polyol in the synthesis of Fe particles. J. Magn. Magn. Mater. 2007, 310, 2393.

86. Caruntu, D.; Caruntu, G.; Chen, Y.; O�Connor, C.J.; Goloverda, G.; Kolesnichenko, V.L.Synthesis of variable-sized nanocrystals of Fe3O4 with high surface reactivity. Chem

Mater. 2004, 16, 5527.

87. Dresco, P.A.; Zaitsev, V.S.; Gambino, R.J.; Chu, B. Preparation and properties of

magnetite and polymer magnetite nanoparticles. Langmuir 1999, 15, 1945.

88. Santra, S.; Tapec, R.; Theodoropoulou, N.; Dobson, J.; Hebard, A.; Tan,W. Synthesis and

characterization of silica-coated iron oxide nanoparticles in microemulsion: the effect of

nonionic surfactants. Langmuir 2001, 17, 2900.

89. Meldrum, F.C.; Heywood, B.R.; Mann, S. Magnetoferritin: in vitro synthesis of a novel

magnetic protein. Science 1992, 257, 522.

90. Sangregorio, C.; Wieman, J.K.; Connor, C.O�.; Rosenzweig, Z. A new method for the

synthesis of magnetoliposomes. J Appl. Phys. 1999, 85, 5699.

91. Strable, E.; Bulte, J.W.M.; Moskowitz, B.; Vivekanandan, K.; Allen, M.; Douglas, T.

Synthesis and characterization of soluble iron oxide–dendrimer composites. Chem

Mater. 2001, 13, 2201.

92. Bonacchi, D.; Caneschi, A.; Dorignac, D.; Falqui, A.; Gatteschi, D.; Rovai, D.;

Sangregorio, C.; Sessoli, R. Nanosized iron oxide particles entrapped in pseudo-single

crystals of g-cyclodextrin. Chem Mater. 2004, 16, 2016.

93. Pileni, M.P.; Duxin, N. Micelle technology for magnetic nanosized alloys and compo-

sites. Chem Innov. 2000, 30, 25.

94. Inouye, K.; Endo, R.; Otsuka, Y.;Miyashiro, K.; Kaneko, K.; Ishikawa, T.Oxygenation of

ferrousionsinreversedmicelleandreversedmicroemulsion.JPhys.Chem.1982,86,1465.

95. Lee, Y.; Lee, J.; Bae, C.J.; Park, J.G.; Noh, H.J.; Park, J.H.; Hyeron, J.H. Large-scale

synthesis of uniform and crystalline magnetite nanoparticles using reverse micelles as

nanoreactors under reflux conditions. Adv Funct. Mater. 2005, 3, 503.

96. Frenz, L.; El-Harrak, A.; Pauly,M.; B�egin-Colin, S.; Griffiths, A.D.; Baret, J.-C. Droplet-based microreactors for the synthesis of magnetic iron oxide nanoparticles. Angew.

Chem., Int. Ed. 2008, 47, 6817.

97. Abou-Hassan, A.; Sandre, O.; Neveu, S.; Cabuil, V. Synthesis of goethite by separation of

the nucleation and growth processes of ferrihydrite nanoparticles using microfluidics.

Angew. Chem., Int. Ed. 2009, 48, 2342.


98. Andreev, V.P.; Koleshko, S.B.; Holman, D.A.; Scampavia, L.D.; Christian, G.D.

Hydrodynamics and mass transfer of the coaxial jet mixer in flow injection analysis.

Anal Chem. 1999, 71, 2199.

99. Scampavia, L.D.; Blankenstein, G.; Ruzicka, J.; Christian, G.D. A coaxial jet mixer for

rapid kinetic analysis in flow injection and flow injection cytometry. Anal Chem. 1995,

67, 2743.

100. Abou-Hassan, A.; Dufreche, J.F.; Sandre, O.; M�eriguet, G.; Bernard, O.; Cabuil, V.Confocal laser scanning microscopy for pH mapping in a coaxial flow microreactor:

application in the synthesis of superparamagnetic nanoparticles. J. Phys. Chem. C 2009,

113, 18097.

101. Kim, D.K.; Mikhaylova, M.; Zhang, Y.; Muhammed, M. Protective coating of super-

paramagnetic iron oxide nanoparticles. Chem Mater. 2003, 15, 1617.

102. Margulies, D.;Melman, G.; Shanzer, A. Fluorescein as amodelmolecular calculator with

reset capability. Nat Mater. 2005, 4, 768.

103. Nunez, N.O.; Morales, M.P.; Tartaj, P.; Serna, C.J. Preparation of high acicular and

uniform goethite particles by amodified-carbonate route. JMater. Chem. 2000, 10, 2561.

104. Vroege,G.J.; Thies-Weesie, D.M.E.; Petukhov,A.V.; Lemaire, B.J.; Davidson, P. Smectic

liquid-crystalline order in suspensions of highly polydisperse goethite nanorods. Adv.

Mater. 2006, 18, 2565.

105. Lemaire, B.J.; Davidson, P.; Ferr�e, J.; Jamet, J.P.; Petermann, D.; Panine, P.; Dozov, I.;

Jolivet, J.P. Physical properties of aqueous suspensions of goethite (a-FeOOH) nanorods.Part I. In the isotropic phase. Eur. Phys. J. E 2004, 13, 291.

106. Schwertmann, U.; Murad, E. Effect of pH on the formation of goethite and hematite from

ferrihydrite. Clays Clay Miner. 1983, 31, 277.

107. Penn, R.L.; Banfield, J.F. Imperfect oriented attachment: dislocation generation in defect-

free nanocrystals. Science 1998, 279, 1519.

108. Krehula, S.; Popovic, S.;Music, S. Synthesis of aciculara-FeOOHparticles at a very high

pH. Mater Lett. 2002, 54, 108.

109. Thies-Weesie, D.M.E.; de Hoog, J.P.; Hernandez Mendiola, M.H.; Petukhov, A.V.;

Vroege, G.J. Synthesis of goethite as a model colloid for mineral liquid crystals.

Chem Mater. 2007, 19, 5538.

110. Stjerndahl, M.; Andersson, M.; Hall, H.E.; Pajerowski, D.M.; Meisel, M.W.; Duran, R.S.

Superparamagnetic Fe3O4/SiO2 nanocomposites: enabling the tuning of both the iron

oxide load and the size of the nanoparticles. Langmuir 2008, 24, 3532.

111. Alcala, M.D.; Real, C. Synthesis based on the wet impregnation method and characteri-

zation of iron and iron oxide–silica nanocomposites. Solid State Ionics 2006, 177, 955.

112. Gushikem, Y.; Rosatto, S.S. Metal oxide thin films grafted on silica gel surfaces: recent

advances on the analytical applicationof thesematerials. JBraz.Chem. Soc.2001,12, 695.

113. Sharma, P.; Brown, S.; Walter, G.; Santra, S.; Moudgil, B. Nanoparticles for bioimaging.

Adv Colloid Interface Sci. 2006, 123, 471.

114. Barnakov, Y.A.; Yu, M.H.; Rosenzweig, Z. Manipulation of the magnetic properties of

magnetite–silica nanocomposite materials by controlled St€ober synthesis. Langmuir

2005, 21, 7524.

115. Deng, Y.H.; Wang, C.C.; Shen, X.Z.; Yang, W.L.; An, L.; Gao, H.; Fu, S.K. Preparation,

characterization, and application of multistimuli-responsive microspheres with fluores-

cence-labeled magnetic cores and thermoresponsive shells.Chem Eur. J. 2005, 11, 6006.

REFERENCES 359

116. Lu, A.H.; Salabas, E.L.; Schueth, F. Magnetic nanoparticles: synthesis, protection,

functionalization, and application. Angew. Chem., Int. Ed. 2007, 46, 1222.

117. Xu, Z.Z.; Wang, C.C.; Yang, W.L.; Fu, S.K. Synthesis of superparamagnetic Fe3O4/SiO2

composite particles via sol–gel process based on inverse miniemulsion. J. Mater. Sci.

2005, 40, 4667.

118. Tartaj, P.; Gonzalez-Carreno, T.; Bomati-Miguel, O.; Serna, C.J.; Bonville, P. Magnetic

behavior of superparamagnetic Fe nanocrystals confined inside submicron-sized spheri-

cal silica particles. Phys Rev. B: Condens. Matter 2004, 69, 094401.

119. Pamme, N. Magnetism and microfluidics. Lab Chip 2006, 6, 24.


10METAL NANOPARTICLE SYNTHESISIN MICROREACTORS

PETER MIKE G€uNTHER, ANDREA KNAUER, AND JOHANN MICHAEL K€oHLERDepartment of Physical Chemistry and Microreaction Technology, Institute of Micro- and

Nanotechnologies, Ilmenau University of Technology, Ilmenau, Germany

10.1 INTRODUCTION

Metal nanoparticles are the oldest nanomaterials used by mankind for achieving

specific functional material properties. Colloidal gold was generated in ancient glass

matrix for coloring vine glasses or windows in middle age churches. Therefore, the

tendency of gold to form homogeneous distributed nanoparticles under moderate

reducing conditions and the efficient specific plasmon absorption were used. The

optical properties of a nanocomposite material were the first reason for the synthesis

and application of metal nanoparticles.

Liquid colloidal solutions of noble metals such as gold were also known since the

middle age. In the nineteenth century, Michael Faraday discovered basic laws of

electrochemistry, investigated colloidal solutions of gold, and explained their nature

based on the fundamental of understanding redox processes and chargedparticles. The

particles inside the colloidal solutions were stabilized by charges that led to a high

thermodynamic stability. For this reason, colloidal solutions were stable over decades

and centuries not only in a rigid inorganic glassmatrix but also inside a highlymovable

liquid.

Metal nanoparticles are characterized by high surface to volume ratios, special

chemical activities of surface sites, and as possessing different shapes and sizes like

nanoparticles of othermaterial classes. The particular interest inmetal nanoparticles is

due to the delocalization of electrons inmetals. In contrast to dielectric solids,metallic


361

solids can be described as a system consisting of cations fixed in a crystal lattice and a

cloud of highly movable electrons compensating the positive charges of the multi-

cationic lattice andmediating the bonds inside thewhole system. The typical metallic

properties such as metallic shine and reflectivity, high thermal and electrical conduc-

tivity, and high refractive index are caused by the long-range delocalization of outer

electrons. The particular properties of electrons, their high movability, and energy

states make metallic nanoparticle completely different from nanoparticles of other

materials.

The small dimensions of nanoparticles restrict the delocalization range and the

number of interacting electrons. This leads to a significant change of electronic and

optical properties if the size of a metallic solid is shrinked down to the submicron and

further down to the nanometer scale. Separated metal nanoparticles are charge

confinements. Electron density exchange between metal nanoparticles depends on

barriers at the particle surface. Thin adsorbate films, monomolecular capping layers,

or a narrow vacuum gap can act as barriers that can be passed by electrons using the

tunneling effects. Metal nanoparticles inside a chain, a matrix, or a network of such

tunnel contacts act as chargeable islands in a charge transfer system controlled by

Coulomb blockade effects. Thicker insulating surface films or the embedding ofmetal

nanoparticles can suppress the charge transport between particles completely.

The cloud of movable electrons inside the metal nanoparticles can be polarized by

electricfields.This polarizationoccurs by static fields aswell as byalternatingfields of

lower and higher frequencies. A temporal or oscillating polarization with very high

frequency is also caused by interaction of metal nanoparticles with the photons of

electromagnetic fields. The restricted number of atoms andmovable electrons inside a

metal nanoparticle lead to the formation of distinct energy levels in contrast to the

energetic continuumrepresentedby theband structureof abulkyconductive solid.The

number of distinct energy levels decreases with decreasing number of atoms and

electrons inside the particle. The specific interaction of nanoparticlewith photons and

the distinct energy levels for electron states are responsible for the specific optical

properties of metal nanoparticles and their dependence on elementary composition,

size, and shape.

The strong dependence of electronic and optical properties on the particle para-

meters is an important reason for all recent demands formethods of producing particle

populations of very high homogeneity. This demand is valid for nanoparticles of pure

metals, but still more important for composite nanoparticles, where the spatial

distribution of components represents an additional parameter with high impact on

the electronic behavior of the particles and their interaction with the electromagnetic

field.

The demand for high-quality metal nanoparticles with narrow size and shape

distribution aswell as homogeneous composition can be satisfied by highly controlled

preparation techniques. Continuous-flow processes are particularly suited for realiz-

ing constant mixing, reaction, and quenching conditions. The introduction of micro-

reaction technology gives the possibility for realizing mixing, thermal activation, or

cooling down within shortest time intervals. Consequently, microcontinuous-flow

processes are of particular interest for the synthesis of metal nanoparticles in research

362 METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

and development and for using this technique for production of high-quality metallic

nanomaterials, too.

10.2 MECHANISM OF METAL NANOPARTICLE FORMATION

10.2.1 Nucleation

For the synthesis of metal nanoparticles in the liquid phase, solutions of metal salts or

complex compounds (as sources for the nanoparticle formation) are used. In these

reactants, themetal is always present in an oxidized state (positive oxidation number).

Therefore, the formation of metal nanoparticles from the liquid phase must always

include a reduction step. One or more electrons have to be transferred from the

reducing agent of the suited redox potential to themetal ion or the complex compound.

So, in particular the first step of particle formation, namely, the nucleation, is a

rather complex process. This step realizes the transition from the molecular disperse

state (chemical solution) to the initial phase of the solid state (colloidal solution). It

includes the following partial processes:

. Electron transfer from the reducing agent to the metal ion

. Exchange or removal of ligands from the primary coordination sphere of the

metal ion

. Aggregation of a small group of metal atoms forming a nucleus

All threeprocessesmustoccur simultaneously.So, the reaction conditionsmust suit

to all these different reactions. On the one hand, the ligand exchange removal is

normally strongly dependent on pH value. Low pH values lead to the destruction of

many complexes since the protons can successfully compete with the metal ions. For

example, ammonium or nitrogen-containing ligands that are applied frequently are

protonatedat lowpHvalues resulting ina strongdecrease in thecoordinationability.At

higher pH values, the complexes are more stable. On the other hand, the reduction

power of reducing agents increases frequently with increasing pH. Thus, the nucle-

ationprobability decreases inmanycaseswith loweringpHdue to the increase in redox

potential of the reducing agent. Furthermore, the pH value, together with existing

ligands, surfactants, and the solution ionic strength, is responsible for the charge of

forming metal nuclei. Uncharged nuclei will come together fast and can bind,

aggregate, and sediment. As a consequence, large particles and irregular precipitates

are formed. Charging is responsible for negative electrostatic interaction between the

forming particles and contributes to the suppression of nanoparticle aggregation.

The formation of small, well-defined, stable nuclei frequently depends on specific

ligands supporting the complex mechanism of nucleation. It has been shown that

appropriate derivatives of phosphoric acid can be applied to selectively form cluster

species with exactly 11 or 55 gold atoms.1 Also, in case of particles of other sizes,

ligands adsorbed at the particle surface support the electrostatic stabilization of the

particles in the colloidal-dissolved state.

MECHANISM OF METAL NANOPARTICLE FORMATION 363

In contrast to nucleation, the growth of metal nanoparticles can be described by an

electrochemical process, a mechanism well known from macroscopic heterogeneous

chemistry: The electron transfer for conversion of metal ions or complex compounds

into the neutral metal atoms of the solid state is divided into two partial processes that

are electrically coupled by the electrochemical potential of the forming nanoparticle.

The whole system has no electron exchange by a power supply. The formed

electrochemical potential results from the superposition of an anodic and a cathodic

process, which electrically compensate each other. The nanoparticle acts as a mixed

electrode. The cathodicpartial process is representedby the reductionof themetal ions

to metal. Its potential depends on the position of the metal in the electrochemical

series. It is influenced by the ligands of an original complex compound and by other

species acting as ligands in the cathodic partial process. The anodic partial process is

represented by the oxidation of the reducing agent. The electrode potential of this

process is determined by the electrochemical standard potential of this oxidation

process, the concentration of the reducing agent in the solution, and the pH.

Despite the general importance of redox properties of involved species in both

processes, the reaction rates for nucleation and nanoparticle growth may differ

considerably. Nucleation is normally strongly nonlinear dependent on the reactant

concentration. It goes on only above a critical concentration. The consumption of

metal ions and reducing molecules leads to a decrease in the concentrations of these

essential reaction partners, so that they can drop below the nucleation threshold. In

contrast, the rate of the electrochemical processes in crystal growth of the forming

nanoparticles follows the electrochemical laws and monotonously falls with decreas-

ing reactant concentrations.

The formation of nanoparticles of equal size can be expected if the nucleation

occurs rapidly. After a short time interval with nucleation, the nucleation threshold

should be achieved and then the further process is marked by the nanoparticle growth

exclusively.Ahomogeneousgrowth of all nanoparticles can be expected if the time for

crystal growth is much longer than the time for nucleation and if the local reactant

concentrations are equal for all parts of the reaction solution during nanoparticle

growth.

Therefore, techniques are required for ensuring a very fast mixing of the reactants.

This requirement is met by microreaction technology. Under homogeneous condi-

tions, static micromixers can be used to realize fast mixing. So, multilamination

mixers possess very short diffusion lengths, which results in short diffusion times. A

certain disadvantage of processes under homogeneous-phase conditions is the fluid

dispersion, leading to abroaddistributionof residence times.A seconddisadvantage is

the wall contact, leading to wall adhesion of particles, in particular under the high

surface to volume ratios of microchannels. Both problems are overcome by introduc-

tion of the segmented flow technique for nanoparticle synthesis (see below).

10.2.2 Particle Growth

The microreaction technology is suited for a combination of fast nucleation and

growth initiation and moderate or slow nanoparticle growth. Different shapes of


nanoparticles can be addressed by choosing between transport or surface-controlled

metal deposition during the growth of nanoparticles: A low electrochemical potential

at the nanoparticle surface caused by high concentrations of a strong reducing agent

possessing a low electrochemical overpotential at the particle surface leads to a fast

reduction of metal ions on the surface. This results in concentration gradients in the

environment of the nanoparticles and a transport control in the anodic partial process.

Under these conditions, local differences in the surface energy are negligible formetal

deposition. As a consequence, the particles grow as spheres.

A slower growth of nanoparticles can be realized by weaker reducing agents and

lower concentration as well as by the application of ligands forming more stable

complexeswith themetal cations or are preferentially bonded to certain bindingplaces

at the particle surface. The electrochemical open circuit potential of the growing

nanoparticles is higher and nearly no concentration gradients appear under these

conditions. So, small differences in local chemical properties and surface energyof the

single binding places at the particle surface play an important role for the cathodic

partial process, with implications for themetal deposition. As a result, crystallograph-

ic growth can be observed and the possibility of formation of triangles, hexagons, and

cubes increases. The shape of nanoparticles and the ratio of different nanocrystal types

can be influenced by the choice of reducing agent, ligands, solvent, and temperature.

10.2.3 Surface Capping

The surface state of nanoparticles is important for the energetic state of the whole

nanoparticle as well as for its chemical behavior. So, the ligand shell of a nanoparticle

or adsorbed polymer molecules influence the plasmon absorption spectrum of the

particle and modify the charge transfer behavior, the electrochemical properties, and

the specific binding behavior between the particle and other particles, molecules, or

surfaces. In addition, surface capping is always related to a change in the electrical

particle charge. So, the aggregation behavior, the stability of colloidal solutions, the

solubility properties, and precipitation are altered by surface modification.

Surface capping is very important for the chemical and physical properties of

nanoparticles and can be used for the extension of tuning ranges for nanoparticle

parameters. The deposition of monomolecular layers, molecular bilayers, and multi-

layers allow controlling precisely the electronic and energetic coupling between a

metallic nanoparticle and its environment. So, well-defined barriers for electron

tunneling can be realized, which are of interest, for example, for single-electron

devices. Different electronic properties of capping layers can be generated by

application of monomolecular films containing organic conductors and semiconduc-

tormolecules. The orientated incorporation of asymmetrically functionalized organic

molecules into capping layers allows the construction of diode-like or transistor-like

systems with metal nanoparticles. The integration of dye molecules can be used for

addressing of electronic energybetween themetal core and the organic chromophores.

This is used for resonance Raman spectroscopy. Plasmon energy of themetal core can

be transferred into electronically excited states of organic chromophores and vice

versa. In the future, such systems could also be of interest for nanoparticle-based

MECHANISM OF METAL NANOPARTICLE FORMATION 365

optoelectronic devices for nanometer-sized light sources, for solar cells, or for

artificial photosynthesis. But all these applications require a high reproducibility

in the formation of capping films and, therefore, well-controlled formation

procedures.

Micromixing is a convenientmethod for fast additionof surface capping reagents to

colloidal solutions. It is of particular interest for two-step andmultistep reactions with

short residence times for fast switching of reaction conditions. So, nanoparticles,

ligands, or parameters such as salt concentrations, pH, and ion strength can be varied

quickly by fastmixing.Microreaction technologyoffers best possibilities for fastmass

transfer to get highest homogeneity in the formation of capping layers.

In addition, the thermal activation and the temperature dependence of adsorption

equilibriums can be easily and quickly addressed by fast heat transfer and temperature

changes under microfluidic conditions. Fast local convection, small channel dia-

meters, and thin walls support the fast heat transfer in microreactors or in micro-

segmented flow.

The synthesis of capping layer is often connected with a complete change in the

solubility of nanoparticles. Frequently, metal nanoparticle capping is coupled with a

transfer of nanoparticles between different solvents. Surfactants are used as phase

transfer catalysts and form shells of surfactants or support the formation of shells of

additional molecules.2 There are several approaches of microreaction technology for

such phase transfer procedures. They can be realized by microcontactors, by seg-

mented flow, and in particular by segmented flow three-phase systems consisting of an

aqueous and a water-immiscible organic phase, both embedded in an inert carrier

phase of a perfluorinated liquid. A third strategy for phase transfer processes in

nanoparticle synthesis is based on the application of staticmicromixers for generation

of emulsions and following reaction and phase separation.

10.3 NP PRODUCT AND PROCESS CHARACTERIZATION

10.3.1 Spectrometry

The plasmon absorption ofmetal nanoparticles, first theoretically have been described

byMie,3 provides an opportunity for optical spectrometry of nanoparticle dispersions.

Plasmons are density oscillations of charge carriers in metals or semimetals. In

quantum mechanics, they act as quasiparticles (quantums) with energy EPl (given by

equation (10.1)), where vPl is the frequency of the plasmon and h the reduced Planck

constant. The resulting energyEPl is valid for the so-called volume plasmons of solids.

EPl ¼ vPl�h ð10:1Þ

For small particles, the plasmon is described by the Mie theory and called Mie

plasmon.4Mie plasmons are surface plasmonsof a sphere andpossess discrete spectra.

Plasmon absorptions can be described as resonance of oscillating electrons of the

conduction bandwith the exciting electromagnetic field (e.g., radiation in the range of


visible light). Therefore, it is possible to calculate the absorbanceof colloidal solutions

ofmetal nanoparticles if the size of nanoparticle ismuch lower than thewavelength of

exciting light radiation (e.g., fulfilled by Au-NP with diameter of 25 nm and less) and

known dielectric function of the nanoparticle material. For these small particles, the

dipole approximation is valid5 and leads to equation (10.2).

el ¼ 18pNPVe3=2D e2

l e1 þ 2eDð Þ2 þ e22ð10:2Þ

where Np is the number of nanoparticles, V is the volume of nanoparticles, eD is the

dielectric constant of dispersionmedium, l is thewavelength of absorbed light, and e1and e2 are real part and imaginary part of dielectric function of the nanoparticle

material, respectively.

The extinction coefficient el of nanoparticle dispersions grows with the volume of

particles and can reach values much higher than organic dyes, but low nanoparticle

numbers in dispersions reduce el to the known values of metal organic compounds of

about 3000–4000 Lmol�1 cm�1.

A very important part of plasmon absorptions is the fact of dependency from the

particle diameter (Figure 10.1). This means that the particle size influences the

absorption behavior, leading to shift of radiation absorption.6

Gold nanoparticles with diameter in the order of magnitude of 20 nm show a

characteristic plasmon absorption with a maximum at about 526 nm. Larger gold

particles and particle aggregates are characterized by an increase in the absorption at

wavelengths above 600 nm. A shift in the optical properties can also be used for the

detection ormonitoring of the surface state of particles. Changing particle charges and

the formation of adsorbate films influencing the electronic state of the particle result in

changed plasmon absorption. Thereby, frequently a bathochromic shift can be

observed.

Silver nanoparticles as well as the formation of silver shells on gold cores can be

easily detected bymeans of the sharp plasmon absorption at about 410 nm.Additional

FIGURE 10.1 Au and Ag nanoparticle dispersions with different nanoparticle sizes showing

different colors.

NP PRODUCT AND PROCESS CHARACTERIZATION 367

longer wavelength shoulders or maxima indicate larger or nonspherical silver

nanoparticles. Alloy-like binary particles have absorption maxima between the

gold and the silver peak. Core/shell particles or other composite particles show, in

general, broader peaks, double peaks, or shoulders.

The size dependency of plasmon absorption opens an easy way to determine

quickly the particle dimension or to observe the kinetics of particle growth by using

commonUV–Vis spectrometer.But the plasmon spectra are strongly dependent on the

nanoparticle composition,7 the morphology of the particles,8 the chemical composi-

tion, the particle surface7,9 (e.g., type of surface ligands), the dispersion agent itself,

and the grade of aggregation.6

A further interesting spectrometric method of nanoparticle characterization is the

MALDI-TOF (time-of-flight) mass spectrometry,10 especially for ligand characteri-

zation on the particle surface. However, this method is limited to very small Au

nanoparticles (<2 nm) as well as Au clusters due to mass restriction. Nanoparticles

with higher diameter are too heavy to fly and reach the detectors of TOF mass

spectrometer.

10.3.2 Optical Microscopy

Abbe�s theory of imaging (equation (10.3)) describes the “problem” of optical

microscopes with nanoparticles. In the range of visible light (400–780 nm), optical

microscopes are not able to resolve dimensions of nanoparticles with diameters

smaller than or equal to 100 nm.

s ¼ ln sin a

¼ lNA

ð10:3Þ

where s is the resolving capacity of amicroscope, l thewavelength of used light, n therefractive index of medium, and NA the numerical aperture.

In the best case (l¼ 400 nm, high NA of about 2), it is possible to detect

nanoparticle with minimum size greater than or equal to 200 nm as single object.

A special method to detect nanoparticlewith sizes smaller or equal to 100 nm is the

dark-field microscopy. This technique uses the ability of very small objects to scatter

light perpendicular to the beaming direction. To get dark-field images, a special dark-

field condenser lens system is used that leads the incoming light in very small angle to

the microscope objective. Due to this small angle, only scattered light from the

nanoparticles passes the objective lens and is detectable. Nanoparticle scatters light

very well due to the Mie scattering. The obtained images (Figure 10.2) show a dark

background and very bright scattering centers. But these bright spots do not give

correct information about the nanoparticle size and morphology.

It is, however, important that the microscopic sample possess a high purity since

impurities act as strong scattering centers. The detection and analysis of the scattered

light from a single metallic nanoparticle (single plasmon resonance methods) with a

spectrometer provides information about the particle size, morphology, and the

chemical environment of the nanoparticle.11–14


10.3.3 Ultramicroscopy

Typical methods to determine the nanoparticle size and morphology are electron

microscopic methods such as SEM (scanning electron microscope) and TEM (trans-

mission electron microscope), as well as scanning probe microscopic methods like

AFM (atomic force microscopy).

Electron microscopic methods use a beam of electrons with certain energy to scan

the object (e.g., Au nanoparticles in Figure 10.3) through interactions of the electron

FIGURE 10.2 Dark-field image of Au/Ag core/shell nanoparticles embedded in a thin

polymer film on glass substrate (glass slide).

FIGURE 10.3 SEM image from Au nanoparticles on silicon substrate.


beamwith thematerial of the sample. These electron beam interactions depend on the

material density, the chemical composition of the sample, electric properties of the

sample material, and the energy of the electron beam.

Electron microscopic methods can be used only for electroconductive samples. In

the case of metal nanoparticle, this requirement is fulfilled.

The difference between SEM and TEM, besides resolution varieties, is the manner

of sample illumination. TEMs transmit the electronbeam through the sample,whereas

SEMs illuminate the sample surface (compare thedifference inFigures 10.3 and10.4).

The sample preparation is therefore different and in case of TEMs quite laborious. For

this reason, the SEM is the preferred method for routine analyses. TEMs and SEMs

deliver two-dimensional pictures of samples (Figures 10.3 and 10.4).

AFMs are further important microscopes for the determination of nanoparticle

properties such as size (Figure 10.5) and morphology. They work with a mechanical

probe (cantilever) in several working modes. The most important modes are the

tappingmodewith anoscillatingcantilever tip innanometer distance to sample surface

and the contact mode where the cantilever pin gets in direct contact with the sample

surface.

In contrast to electron microscopes, the scanning speed is much lower, but the

AFMgives real three-dimensional information of samples with a resolution down to

few nanometers. Furthermore, AFMs can determine interactions between nano-

particles and their chemical environment.15 With chemically modified AFM tips, it

should be possible to get information about the ligand sphere on the nanoparticle

surface.

FIGURE 10.4 TEM image of Au/Ag core/shell nanoparticles. The Au cores are distinguish-

able in bigger particle (darker than Ag shell) due to stronger interactions with the electron

beam.


10.3.4 Differential Centrifugal Sedimentation

The differential centrifugal sedimentation (DCS) is used to determine the size

distribution of particle-supported systems such as dye pigment suspensions. The

size of typical detectable particle ranges from 20 nm up to 30 mm. Formaterials with a

high density like Au nanoparticles, the lower detection limit is about 3–5 nm

(Figure 10.6).

Detection times depend on sample�s size distribution and density. Typical detectiontimes are between some minutes (e.g., for metal nanoparticle) and hours (e.g., for

polymer nanoparticles).

FIGURE 10.5 Three-dimensional AFM image of Au nanoparticles.

FIGURE 10.6 DCS spectrum of Au nanoparticles with sizes below 5 nm (device: DC 20000,

CPS Instruments, USA).


Sedimentation methods use the fact of particle separation by g-forces (through the

earth gravity or centrifuges) in media with known density and viscosity. For particle

size determination, parameters such as the migration time or velocity are important to

know. The active principle of sedimentation methods is described by the Stokes

equation (10.4). The sedimentation velocity ns is dependent on particle diameter Dp,

particle density rp, sedimentation medium density rs, and viscosity hs, as well as the

acting g-forces a.

ns ¼aD2

p rp�rs

� �18hs

ð10:4Þ

Low differences between sedimentationmedium density and particle density (e.g.,

polymer nanoparticle) lead to a decrease of the sedimentation speed, whereas very

high-density differences in case of metal nanoparticle dispersions increase the

sedimentation rate.

The Stokes equation is valid in case of smooth, spherical, and inelastic particles if

the volume of sedimentationmedium ismuch greater than the particle volume and the

molecules of sedimentation medium are much smaller than the particle, as well as

under nonturbulent or convective conditions.

To separate particles smaller than 100 nm, the g-forces have to be higher than

9.81m s�2 of the earth gravity force. For this reason, the nanoparticle sedimentation

can be done in special centrifuges with several ten thousands rpm. In this case, the

Stokes equationneeds amodification because theg-forces change thevalue depending

on the distance from the rotation center. If the separation medium temperature and

angular speed vc of the centrifuge are constant over measurement time and the start

point R0t, as well as end point RE of sedimentation is known, the particle diameter Dp

depends on the sedimentation time ts (equation (10.5)).

Dp ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi18hsln RE=R0ð Þ

rp�rs

� �v2c

vuut 1ffiffiffiffits

p ð10:5Þ

Adensity gradient is used inside the centrifuge to support the sedimentationprocess

and suppress sedimentation instabilities like the so-called streaming.

10.4 METAL NP SYNTHESIS IN HOMOGENEOUS FLUIDS

In the past 5–6 years, an increasing interest in nanoparticle syntheses inmicroreactors

or microfluidic devices is observed.16–18 Several types of nanoparticles were synthe-

sized in microstructure containing arrangements ranging from semiconductor parti-

cles such as CdS,19,20 inorganic pigments21,22, and metallic nanoparticles such as Ag

and Au nanoparticles with several morphologies23,24,26–28 to Au/Ag core/shell

nanoparticles25 as well as Pt nanoparticles.29


But what is the most important reason to use the microreaction technology? There

are some known advantages of nanoparticle syntheses in microreactor in contrast to

batch syntheses. Microreactors or microfluidic channels allow more effective heat as

well as mass transfer due to larger surface to volume ratios. Smaller dimensions of

microchannels result in shorter mixing times compared to common batch reactors or

streaming tubes, and in shorter mean diffusion length ofmolecules. Inmicrochannels,

the conditions for nanoparticle formation (nucleation and growth) are fulfilled as

discussed in Sections 10.2.1 and 10.2.2 required. Due to better mixing, the nucleation

step is preferred instead of the growth step. This results in a higher amount of

nanoparticles in contrast to batch syntheses. Furthermore, the half-width of size

distributions of microreactor produced nanoparticle was lowered to 50% of batch

produced nanoparticles (Figure 10.7). This effect depends on the flow rates in the

microchannels. In general, higher flow rates result in smaller size distributions and

normally smaller nanoparticles (Figure 10.8).

The size distribution and the nanoparticle size are affected not only by the flow rate

but also by chemical conditions (e.g., pH value of reaction solution) or additives such

as polymers (e.g., polyvinyl pyrrolidone (PVP)), or surfactants (e.g., sodium dodecyl

sulfate (SDS)),which influence the diffusion behavior of the reactants and particularly

the surface tension.

But an important aspect using microreactors to form nanoparticles is the so-called

reactor fouling.Thismeans theprecipitationor depositionof thenanoparticle-forming

materials on the reactor channel walls. One reason is the formation of nucleation

FIGURE 10.7 Size distributions of Au nanoparticle synthesized by reduction of tetrachloro-

auric acid with ascorbic acid in a batch experiment and in a microreactor; cHAuCl4 ¼0:001 mol L�1, cascorbic acid ¼ 0:020 mol L�1, pH 9.5, and 0.25% PVP.

METAL NP SYNTHESIS IN HOMOGENEOUS FLUIDS 373

centers not only inside the reaction mixture but also on the channel surface. To avoid

the reactor fouling, a few approaches are known.

The simplest way is to minimize the residence time of reactants through high flow

rates. In this case, the microreactor or micromixer just provides a goodmixing but the

nuclei formation takes place outside. The procedure works if the residence time is

much lower than the nucleation kinetics.

A second approach is to use hydrophobic channels for aqueous solutions. This can

be realized by using hydrophobic materials (e.g., silicon chips made of PDMS) or by

means of hydrophobization of glass channels. In this case, the formation of nucleation

centers on the channel walls is reduced by the high interface energy (surface tension).

The increased value of surface tension avoids the contact of the reactants with the

channel surface. However, one disadvantage is low chemical resistance of hydro-

phobized glass channels.

A third approach is to adjust the chemical behavior of the reactant solution or

formed nanoparticles and the channel walls in such amanner that no nucleation on the

walls can occur. One example for this is the preparation of Au nanoparticles by

reduction ofHAuCl4with citric or ascorbic acid in glassmicroreactors. Using high pH

values during the formation of the negative stabilized Au nanoparticles (due to the

negatively charged surface ligands ascorbate or citrate), the reactor fouling is reduced.

The reason is thedeprotonationof the superficialOHgroupsof theglass surface,which

causes a negative surface charge. In this case, the negatively charged nanoparticles are

repulsed by the negative charge of the glass channel surface. Another method to avoid

reactor fouling is using the segmented flow principle (see Section 10.5).

As per description in Section 10.3.1, the characteristic plasmon absorption of

nanoparticles offers a convenient way for noninvasive monitoring of the metal

nanoparticle formation in microcontinuous-flow synthesis. Experiments with a

time-resolved multichannel monitoring show that the formation and growth of Au

FIGURE 10.8 Particle diameter and half-width of size distribution of Au nanoparticles

prepared in continuous-flow microreactor syntheses: cHAuCl4 ¼ 0:001 mol L�1, cascorbic acid ¼0:020 mol L�1, pH 9.5, and 0.025% PVP.


and Au/Ag nanoparticles cannot be simply described by a continuous increase of a

certain plasmon absorption.

The nucleation and the growth of nanoparticles is always connected with a change

in the electronic and optical properties of the particles. This effect is due to the size

dependence of the electronic states of particles. The situation becomes complicated by

the formation of binary or ternary metal nanoparticles because the changing compo-

sition and the growing size of the particles affect the change of the plasmon absorption

during the particle formation. The response of optical properties of the reaction

solution on the changing particle properties is well suited for the monitoring and

detection of intermediate states.Changing optical properties including shifting aswell

as increasing plasmon absorption, for example, were measured by a six-channel

arrangement of microflow-through spectrometers (see Figure 10.9) in a continuous-

flow synthesis of Au/Ag core/shell particles under segmented flow regime.

10.5 METAL NP SYNTHESIS UNDER SEGMENTED FLOW

CONDITIONS

10.5.1 Specificity of Segmented Flow Conditions

Liquid/liquid two-phase systems as emulsions are well known for long time and

applied in a lot of special materials and technical processes. Normally, they are

marked by a more or less statistical distribution of the single phases. In contrast, a

well-organized series of segments connected with a homogeneous size and regular

distance between the single phases was introduced for FIA analysis about 30 years

ago.30,31 Droplets of analyte were embedded in a phase of a chemically inert and

immiscible carrier liquid. So, well-defined portions of analyte solution can be

transported by an ideal plug flow and the sequence of droplets can be used for

precise analytical measurements. This principle became particularly important for

FIGURE 10.9 Scheme of setup for onlinemonitoring by flow-throughmicrospectrophotome-

try in Au/Ag core/shell nanoparticle syntheses under continuous segmented flow conditions.

METAL NP SYNTHESIS UNDER SEGMENTED FLOW CONDITIONS 375

serial sample handling in high-throughput biomolecular analytical procedures,32–34

for miniaturized cellular screenings,35–38 and is of particular interest for a minia-

turized chemistry.39–42

The microsegmented flow is marked by several specific features making this

approach particularly suited for nanoparticle synthesis. An ideal decoupling in the

reaction solution can be achieved if the inner surface of the wall has well-adapted

properties to optimal interaction of the carrier liquid and the wall and the interaction

of the embedded phase with the wall is suppressed. In case of mineral oil or

aliphatic separation phase, an alkyl silanization is a possibility to reduce the wall

wetting by an aqueous phase considerably. The silanization supports a good wall

wetting by the lipophilic phase. The wetting behavior determines the geometry of

the liquid/liquid interface. High wettability for the organic phase and low wettabil-

ity for the aqueous phase result in a high contact angle for the aqueous phase.

Segments can be completely released from the channel surface under fluid actuation

and segment motion.

A second important advantage of segmented flow is the residence time behavior.

Homogeneous fluids moving through a channel form a laminar flow in case of lower

Reynolds numbers, which is typical in the case of microfluidics. The consequence of

laminar flow is that fluid volumes near the wall are moved much slower than fluid

volume elements at the center of the channel. This effect leads to a strong fluidic

dispersion. A sharp rectangular concentration signal applied to a microfluidic

channel—for example, a reactant for nanoparticle formation—is reduced and spread

over a larger region during the fluidic transport. As a consequence, large distributions

of residence time are observed. The residence time distribution increases with

increasing channel length. Segmentedflowshows the opposite behavior. The transport

of droplets or slugs is an ideal plug-like motion. The residence time behavior is very

sharp and independent of the length of the transport path. The order of segments

remains constant. All processed segments have the same residence time.

10.5.2 Mixing in Segmented Flow

A third important advantage is the convective behavior related to the plug-like

transport of fluid segments. The constant motion of the segment and suppression

of normal laminar flow structuremust be compensated by a radial motion component.

Themotion of segments induces a strong segment-internal conversion. Inside the fluid

segments, pairs of vortices are formed. This increased convection must be compen-

sated by an increase in fluidic pressure drop.43–45 However, the energy for segment-

internal flow-induced convection represents a unique opportunity for fast mixing

inside fluid segments. Segment-internal convection leads to very efficient mixing, in

particular at higher flow rates.

Besides mixing, the heat transfer from the walls into the liquid is also strongly

supported by the segment-internal convection. The strong radial component of liquid

motion inside segments contributes veryefficiently to the transport of heat between the

inner part of the liquid and thewalls. The heat transfer by segment-internal convection

also increases with increasing flow rate.


The application of microsegmented flow for nanoparticle synthesis is a compro-

mise between a maximum mixing rate or heat transfer efficiency and a high

reproducibility in segment formation. In general, higher homogeneity in segment

formation can be achieved at lower flow rates because the hydrodynamic conditions

are more stable than in case of higher flow rates. In contrast, the transfer rates during

mixing as well as the convection-induced heat transfer are intensified by increasing

flow velocities. A homogeneous segment formation, dosing, or fusion behavior is an

essential precondition of homogeneous product quality and narrow distribution of

particle properties. Fluctuations in segment size and segment distances cause a shift in

the concentration ratios of educts after dosing or droplet fusion, which results in

changing conditions in particle formation. These problems are well illustrated by the

consequences of formation of alternating droplet sizes or alternating droplet distances

in some experiments at high flow rates. Such a behavior is responsible for formation of

segments with periodically changing reactant ratios. The colloidal product solutions

of such synthesis experiments are frequently marked by a bimodal or a multimodal

particle size distribution (Figure 10.10) depending on droplet size distribution.

Changing droplet sizes are a result of inhomogeneous droplet formation

(Figure 10.12), which leads to longer segmentswith higher concentrations and shorter

segments with lower concentrations of the added compound.

10.5.3 Process Homogeneity and Product Quality

The control of segment parameters is important for the optimization of flow rate and

reaction conditions inside the microfluidic segments. The quality of segment se-

quences can be characterized by fast microphotometric measurements applied for a

direct monitoring of segment formation. The key parameters such as “segment size”

and “segment distance” are determined by an automatic procedure. The quality of the

whole process can be characterized by segment size/segment distance plots. They

show the modality and the distribution width of the generated segments for the

FIGURE 10.10 Bimodal size distribution of Au/Ag core/shell nanoparticles due to alternat-

ing droplet sizes.


complete segment sequence generated and used in a synthesis experiment. Examples

are given in Figure 10.11a and b. Alternating segment distances result in alternating

particle sizes in the single segments (Figure 10.12), which is well reflected in the

centrifugal sedimentation spectra (Figure 10.10).

High homogeneity of particles can be achieved if the conditions of formation of

microfluid segments, dosing, and mixing are well controlled. Monomodal particle

distributions were obtained if the segments are formed regularly and with constant

frequency. The product quality can be enhanced by increasing flow rate. High total

flow rates of carrier liquid are always connected with high segment-internal convec-

tion. An increase of the local convection reduces the time needed for diffusivemixing

considerably. So, optimal conditions for fast and homogeneous nucleation, fast

termination of nucleation, and homogeneous particle growth can be realized.

900700500300100

400

900

1400

1900

2400

2900

(a)

Seg

men

t d

ista

nce (

ms)

Segment length (ms)

99.0 %

55045035025015050

200

500

800

1100

1400

1700(b)

Seg

men

t d

ista

nce (

ms)

Segment length (ms)

80.3%

7%

7%

FIGURE 10.11 Scatterplot of segment size and segment distance with (a) monomodal

segment distribution and (b) trimodal segment distribution. Percentage shows the amount of

segments in the respective scatterplot.


The specific advantage ofmicrofluidics and, in particular, of segmentedflow for the

formation of nanoparticles is the control of fast mixing for process initiation and for

process quenching. So, fast reaction steps undermicrocontinuous-flow conditions can

be combined with very precise and narrow residence time for growth processes. One

strategy for composite metal nanoparticle synthesis can be the combination between

fast continuous-flow nucleation or growth initiation and a following slow growth

process in batch. Another strategy is the combination of fast steps with longer

residence loops for the realization of a combination of fast initiation and a slow

surface-specific depositionprocess under completemicrocontinuous-flowconditions.

10.5.4 Core/Shell Metal Nanoparticles Generated by Use

of Microsegmented Flow

Themicrosegmented flow is of particular interest for the formation of core/shell metal

particles with high homogeneity. Fast nucleation of the core material and the fast

initiation of deposition of the shell material can be achieved by fast mixing in

microfluidic segments. It is very important to silanize the internal device surfaces

if glass/glass or glass/silicon double injector chips are used. Then, very homogeneous

segment formationandhomogeneousdosing intoperformed segments canbe realized.

Core/shell metal nanoparticles can also be synthesized in tube arrangements.

T-junctions of PTFE are well suited for segment generation and for injection and

dosing. The reproducibility of segment formation is not as high as in the case of

silanized chip devices, but it is, however, sufficient for forming high-quality core/shell

nanoparticles.

The flow conditions and the rate of mixing not only are important for the initial

nucleation and growth of nanoparticles but also influence the quality of core/shell

particles. Flow rate-dependent particle diameters were found if gold seeds are

introduced into microfluid segments and are mixed with reducing agent and silver

FIGURE 10.12 Schematic of segment formation with alternating composition caused by

alternating segment distances at constant injector flow rate.


FIGURE 10.13 Au/Ag core/shell nanoparticles prepared by reduction of tetrachloroauric

acid and silver nitrate in aqueous solution at ambient temperature by application of micro-

continuous-flow synthesis for starting silver shell formation: (a) initial Au/Ag core/shell

particles with size of about 60 nm, (b) enforcement of silver shell at lower silver salt

concentration resulting in particle diameters of about 120 nm, (c) enforcement of silver shell

at higher silver salt concentration; Au seed particles with a diameter of 50 nmwere enforced by

1.4mmol L�1 AgNO3 solution by reduction with ascorbic acid (4.0mmol L�1) in the presence

of thiourea (1.0mmol L�1) resulting in larger particles of about 250 nm diameter.


salts. This effect is obviously caused by a selection between gold seeds for silver

deposition. The percentage of gold particles covered by silver is enhanced with

increasing flow rate. At lower flow rates, larger particles were obtained while smaller

particles were found at high flow rates. It is assumed that a fast mixing of the colloidal

solution of gold seeds with the silver salt is responsible for a homogeneous start of

silver deposition on the gold seeds. Different thicknesses of silver shells can be

realized by enforcement of primary formed gold particles through further silver

deposition (Figure 10.13). Slower mixing in case of lower flow rates results in

concentration gradients inside the microfluid segments in timescales comparable

to the silver shell growth,which could explain the observed thicker shells at lower flow

rates.

The development of microfluidic synthesis for binary and ternary nanoparticles in

microfluid segments always demands for an adaptation of flow conditions for high

segment quality and fast mixing at the same time. Unfortunately, the regularity of

segment formationoftendecreaseswith increasingflowrate.This fact is obviouslydue

to a bifurcation in the flow and droplet formation behavior, convective fluctuations,

and also transition to turbulent conditions at very high flow rates. It leads to changes in

segment size and in particular in segment distances causing differences in concentra-

tions and also in concentration ratios if dosing by injection of reactants into preformed

segments is applied. Larger particle size distribution and bimodal or multimodal

nanoparticle peaks are obtained. Thus, a compromise for flow rates has to be found

allowing sufficient high segment-internal convection on the one side and stable and

regular segment formation on the other side.

10.6 CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION

TECHNOLOGY

10.6.1 Addressing of Morphological Classes and Yield Improvement for

Nonspherical Particles

Future demands on the technical development of microreactors, microfluidics, and

new miniaturized procedures for synthesis of metal nanoparticle result from the

general requirements and vision of nanoparticle fabrication and from the specific

properties andpossible applications ofmetal nanoparticles.Microreaction technology

should contribute by their specific features as fast heat and mass transfer and fast

switching of reaction conditions to the further development of nanomaterials. So,

microreaction technology will probably become one of the most important classes of

technological procedures for production of high-quality nanoparticles.

Microreaction technology has tomeet the requirements for differentmetals, alloys,

and compounds for a large spectrum of new composite functional materials. Besides

the elementary composition, the addressing of specific nanoparticle shapes and the

reproducibility of morphology and size of the generated particles are a big challenge.

Despite the recent progress in synthesis of metal and semiconductor nanoparticles

by microcontinuous-flow synthesis, the development of synthesis protocols and

CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION TECHNOLOGY 381

microreaction tools for a larger spectrum of metals and even less noble elements must

be intensified. Protocols and reactor arrangements must be qualified in order to

become suited for the continuous-flow generation of a large spectrum of metals and

alloys and a large spectrum of size classes and morphological types.

A particular need is the generation of nonspherical nanoparticles with high yield.

The advantages of microreaction technology should be utilized consequently for

realizing a selection of particle types by suitablemixing, reaction, and flow conditions

in order to become able to synthesize a certain shape type exclusively. A high

selectivity in nanoparticle shape is necessary to avoid any later selection or separation

steps. In principle, microreaction technology should allow to find reaction conditions

andoptimal process chains resulting ina free addressability of shapeand size classes of

nanoparticles.

10.6.2 Improvement of Homogeneity, Shape-Identical, and Size-Identical

Particle Populations

The technical application of nanomaterials is mainly dependent on the properties of

single components. So, nanoparticle production should be able to ensure constant

quality in the particle parameters. The demands result from the homogeneity

requirements in production, processing, and application. The properties and the

parameter distribution of nanoparticles determine the chemical behavior, the

density, and the mechanical, optical, and other physical parameters of composite

materials. For automated processing and larger technical applications, constant

properties of nanoparticles and constant distribution are absolutely needed.

Longtime stability, aging effects, and response of materials against any kind of

stress depend on the properties of the nanoparticles, their spatial distribution, and

the distribution of their properties.

It is well known that the application of suitable ligands can be used for the synthesis

of metal clusters consisting of an exactly defined number of metal atoms. These metal

clusters are a special type of small nanoparticles with absolutely monodispersity in

composition, size, and shape. The definition of atom number and position of atoms

inside these particles make them an analogue to molecular objects. In contrast, the

overwhelming number of metal nanoparticles is formed by a more or less statistical

deposition ofmetal atoms on the surface of a growing core resulting always in a certain

size and shape distribution.

A strict control of local transport and reaction conditions could help to achieve a

nucleation and particle growth that goes on always following the same atomic

interactions and the same scheme of deposition of atoms at the particle surface.

Such a process can be thought in analogy to the epitactic growth of single-crystalline

films known from solid-state semiconductor techniques using gas phase deposition. It

is imaginable that the strict control of fluid motion, diffusion, and rates of surface

reactions at different crystallographic planes of a growingparticle can result in a strong

reduction of statistical effects and an important improvement of reproducibility in

growth of the individual nanoparticles.


The ultimate demand would be the synthesis of nanoparticles of a certain type of

shape with an absolute identical number and arrangement of the involved atoms. The

idea is to synthesize nanoparticles having the quality similar to that of molecules of a

pure substance. Nanoparticles of this quality are small solids with the character of a

three-dimensional macromolecule from the point of view of the definition of spatial

organization. It is assumed that the goal of obtaining nanoparticle populations of

complete identical shape and size can only be realized by the application of micro-

reaction technology.

10.6.3 Continuous-Flow Synthesis of Composite Nanoparticles

by Two and Multistep Reactions

The architecture of composite nanoparticles is influenced by the character of the

components, the applied precursors and reagents, and the temporal order of particle

forming reactions. So, the reaction protocol can decide if an alloy or doped nanoparti-

cle or a nanoparticle consisting of two regionswith different elementary compositions

is formed.

Complex composite nanoparticleswith different regions are obtained, for example,

by deposition of one material after the completion of nucleation and deposition of the

first material. Such synthesis succeeds if the nucleation of the second material occurs

only at a selected surface areas of the primary particle formed by the first material.

Normally, these surface-sensitive processes are slow reactions. It is not the

transport but the surface reaction that must be the rate-determining step. A high

reproducibility can be expected if the secondmaterial is deposited only on planeswith

a specific crystallographic orientation.

Slow processes are normally carried out in batch synthesis. A combination of fast

steps of reaction initiation and formation of stable nuclei must be combined with a

surface control of primaryparticle growth.Therefore, it couldbeof importance to have

well-defined conditions of local convection and particle motion inside the reaction

mixture in order to have optimal condition of particle growth and to avoid aggregation,

wall adsorption, and sedimentation. That is why microreaction technology could also

be interesting for the slow steps of thewhole process chain. These reaction steps must

be realized by longer residence loops. This demand corresponds well with the

principle of slowly moved fluid segments for the synthesis of different types of

nanoparticles in the so-calledSFTR (segmentedflow tube reactor).46This technique is

the key for slow ongoing microcontinuous-flow reactions that can be combined with

fastmixing stepsor fast temperature changes toeitherget nucleationswithhigh rates in

a small time frame or change quickly the conditions of surface binding of effector

molecules modulating the deposition of metals on different crystallographic planes.

The definition of nucleation and growth phases of a second, third, or further

material on the surface of preformed initial nanoparticles in a microcontinuous-flow

process demand for good insights into the mechanisms. Changes in pH, temperature,

or ligand concentration as well as the application of reducing agents with different

redox potentials are tools for controlling the anodic and cathodic partial processes of

open circuit metal deposition. Under certain conditions, the dissolution of metal


particles or a competition between metal deposition and dissolution can occur. The

control of outer currentless electrode potential of nanoparticles is essential for giving

the formation process of composite metal nanoparticles the right direction and

performance. Obviously, the application of ligands binding selectively to certain

crystallographic planes can control the deposition or redissolution of metal in the

critical potential range. Besides the study of the elementary reaction steps, process

monitoring can help to optimize the reaction conditions, the temporal order, and the

speed of the single reaction steps.

The types of obtained composite particles can vary considerably depending on the

deposition conditions. Same crystallographic planes on the surface of the primary

formed nanoparticle should undergo identical nucleation and growth for the second

material if statistical effects are not dominating for the nucleation.Asymmetric shapes

can be obtained if a nucleation of one crystallographic plane will alter the nucleation

conditions drastically enough for inhibiting a nucleation of another surface area of the

same crystallographic type. Such conditions can result from stronger changes in the

electrochemical potential of the nanoparticle due to the first nucleation step.

10.6.4 Regional Functionalization and Three-Dimensional Self-Assembling

of Nanoparticle

A second possibility for generating composite nanoparticles is the connection of two

or more preformed nanoparticles. In principle, nanoparticles can be moved in liquids

by the thermally actuated Brownian motion. They stick together if the binding energy

is higher than the thermal activation. The formation of stable chemical bonds leads to

irreversible aggregation. In principle, all nanoparticles tend to aggregate.Aggregation

takes place if solvation and electrostatic repulsion are too weak to inhibit a direct

contact of nanoparticles. Metal nanoparticles are particularly sensitive against

aggregation due the possibility of forming metal bonds. Metal bonds are much

stronger than van der Waals interactions, for example, and can take place in all cases

of direct contact between “naked” metal surfaces.

Thus, the problem for building of nanoparticle assemblies is not to aggregate them,

but to avoid undesired aggregation and enable the particles to a highly selective

bindingbehavior.The constructionofwell-definedparticle aggregatesdemands stable

colloidal solutions of reactants on the one side and specific chemical interaction

between particles on the other side.

The simplest strategy of binary assembling was found by the principle of

complementary surface functionalization. This principle was realized, for example,

by covering two populations of gold nanoparticles with complementary chains of

thiolated oligonucleotides.47–51 The cooperative effect of hybridization with a

sufficient high number of hydrogen bonds leads to an overcoming of repulsion

due to negative excess charge between the colloidal particles. The surface functio-

nalization of nanoparticles can be performed in batch as well as in microreactors.

Besides the formation of particle dimers, more types of aggregates can be found due

to the coverage of the complete particle surface by the binding molecules

(Figure 10.14).


A well-defined geometric condition for particle binding requires a regional

functionalization. This demand means that the most important challenge for nano-

particle self-assembling is the realization of regional complementary functionalized

nanoparticles. Nanoparticle pairs are formed by two particle types with one comple-

mentary region.Twocomplementary regions on the sameparticle or double regionally

functionalized particles of two complementary types are forming chains or rings.

Three or more binding domains will allow the formation of three-dimensional

nanoparticle assemblies (Figure 10.14).

A regional functionalization can be based in principle on selective reaction of

different crystallographic planes of surfaces of a nanocrystal, but spherical nanopar-

ticles or particles with low differences should be functionalized differently in two or

three surface regions.

How microreaction technology can support the formation and the assembling of

regionally functionalized nanoparticles? Regional functionalization could be sup-

ported by enrichment of nanoparticles at liquid/liquid interfaces. A two-region

functionalization could be realized if the affinity to both phases is regionally enhanced

by two different chemical surface reactions in the two different liquid phases. The

surface modification will enhance the stability of particles at the interface

(Figure 10.15). The process could be realized under continuous-flow conditions if

emulsions or microfluid segments are applied as liquid/liquid two-phase system.

A three-region functionalization ismore difficult to achieve. It could be realized by

application of two different succeeding liquid pairs. Two differently functionalized

surface regions are generated in the first step. One of the two regions must then

additionally be functionalized by a third reaction modifying only a part of this region

by application of a second liquid pair. Therefore, the wetting conditions in all three

liquids must be well adapted. The three-region functionalization allows in principle

the construction of anykind of two- or three-dimensional objects by self-assembling if

a suitable complementarity of surface functionalization is chosen (Figure 10.16). So,

FIGURE 10.14 Formation of nanoparticle aggregates depending on binding molecules.


more or less compact objects consisting of nanoparticles can be realized and more or

less dense three-dimensional nanoparticle networks can be formed.

10.6.5 Nanoparticle-Based and Catalytic Nanomachines

The vision of nanoparticle-based nanomachines could be realized if the self-assem-

bling of nanoparticles leads to combinations of rigid and flexible structure elements in

one nanosystem. In analogy to biomacromolecules, such nanoparticlemachines could

be actuated by molecular energizers and operated in dependence on substrates,

initiators, promoters, and inhibitors.

From a chemical point of view, such systems would work like a catalyst. From a

mechanical point of view, they would work as a mechanical machine. Metals and

metal compounds are well known for a lot of different catalytic effects. They could

FIGURE 10.15 Stabilization of particles at the liquid/liquid interface by phase-specific

surface modification.

FIGURE10.16 Examples of different assembly types in case of particles with three different

functions.


support, for example, redox processes or the transfer of atoms or atom groups in

molecules or from one molecule to another. In principle, different metals or alloys

can be combined in composed nanosystems by nanoparticle self-assembling. The

construction of nanoparticle-based catalytic assemblies and the connection of

chemical activity on specific reaction sites with movability of the nanoenvironment

of this reaction site would open a lot of possibilities for well-designed artificial

catalysts.

The so-formed catalytic nanosystems are thought to operate similarly to biological

enzymes. Many highly efficient biocatalysts contain metal complexes in their active

centers. It could be imagined to combine nanoparticle self-assembling alsowith more

or less movable molecular structures. So, a large spectrum of new nanomachines for a

huge number of different functions is imaginable. Additional electrochemical effects

can be achieved by using the metal nanoparticles inside the particle assemblies for

redox processes with local charge transfer via the nanoparticle surface. The nanopar-

ticle acts as an open circuit electrodewithmixed potential in this case. Photochemical

processes and localized photothermal processes can be supported, in addition, by

using the plasmon absorption band of suitable metal nanoparticles that are incorpo-

rated into the nanoparticle assembly.

All these processes require well-controlled chemical and electrochemical condi-

tions in the environment of the nanoparticle-based machines. Microfluidics and

microreaction technology are challenged to supply the needed local fluidic conditions

for operating the nanomachines. Microreaction technology is also required for the

realizationof compartments and for switchingoperation conditions to realize different

catalytic processes in parallel and the desired temporal order.

10.6.6 Electronic and Optoelectronic Devices Formed by Self-Assembled

Metal Nanoparticles

Metal nanoparticles are ofparticular interest, in general, due to their specificelectronic

and optical properties. It is assumed that combinations of metal nanoparticles and

molecules will play an important role after the era of integrated solid-state electronics

using semiconductors. Future electronic devices will probably have to meet require-

ments of highly locally movable electrons, fast switching, well-defined tunneling

barriers, and strong localization. The basics for one kind of future electronics were

already created in the frame of single-electron tunneling devices. Networks of metal

nanoparticles, dielectric shells, and molecular components will probably best suit for

the realization of this type of future information processing system.

Metal nanoparticles are also of interest for new classes of optoelectronic devices.

The higher number of possible electron states distinguishes the metal nanoparticles

from molecules and the discretization of electron states instead of a continuous band

structure that distinguishes them from bulk metals or semiconductors. So, resonant

interactions ofmetal nanoparticles and nanoparticle/molecule architectureswith light

lead to phenomena ofmolecular excitation and relaxation aswell as solid-state effects,

as known from photoelectrochemistry. So, nanoparticle systems including metal

components seem to be suitable for conversion of light into electrical energy and


vice versa. This is of interest for light emitters and optical sensors as well as for

nanolocal energy conversion and for development of new information processing

systems integrating the handling and conversion of single photons, excitons, plas-

mons, bipolarons, and electric charges.

10.6.7 Microfluidic Strategies for Particle-Based Information Processing

and Multichannel Interfaces

The connection of three-dimensional nanoparticle assemblingwith fluidics could also

open the way for highly integrated three-dimensional systems for information

processing and storage.On theonehand, nanoparticle assemblingcouldhelp construct

nanochannels and compartments for nanofluidics. On the other hand, nanofluidics can

support the transport and addressing of nanoparticles. Probably, nanoparticle func-

tionalization, nanoparticle self-assembling, and nanofluid-based particle handling

will become key elements for future information processing systems.

A simple three-dimensional operation of information in microfluidic networks

could be realizedby encoding informationbynanoparticle combinations inmicrofluid

segments. The whole system could consist of a hierarchy of information storage and

handling including the following levels:

. Fluid strings: segments of liquid embedded in an immiscible carrier liquid inside

micro- or nanochannels

. Larger nanoparticles inside the fluid segments

. Small nanoparticles assembled in the larger particles

. Molecules bond on the surface of smaller nanoparticles

Presently, the typical segment size in microfluidics is in the range between about

1mL and some dozen of nanoliters. But there are many reports on smaller liquid

portions handled by the principle of microsegmented flow. So, it is possible to scale

down this principle to the femtoliter range. A femtoliter segment is large enough for

containing 1000 nanoparticles with a diameter of 50 nm. Each of these nanoparticles

can be composed of hundreds of smaller nanoparticles with diameters of 5 nm, which

can carry hundreds ofmolecules on their surface. So, each femtoliter segment could, in

principle, contain 107–108 elementary information units corresponding to about

100MB. The space need for a femtoliter fluid segment in the plane is 10mm2.

That means that about 1000 TB could be stored in a planar microchannel system

of 1 cm2. A three-dimensional microchannel system with 1000 fluid microchannel

levels would possess an estimated storage capacity of about 1 billion TB.

New strategies in liquid handling and nanoparticle management are necessary for

writing and reading in this nanoparticle-based information handling system and for

logic operations. Probably, the liquid handling operations will be similar to recent

strategies of generation, switching, splitting, and fusion in segmented flow micro-

fluidics. It can be expected that the development of microreaction technology for

synthesis and modification of nanoparticles will prepare the first steps into a

hypothetic future “fluidoparticle informatics” (Figure 10.17).


10.7 CONCLUSIONS

There are a lot of very promising opportunities and challenges for the application of

microreaction technology for the synthesis and manipulation of nanoparticles. This

technique is under development for the generation of nanoparticles with very high

homogeneity in particle properties. The possibility of sharp residence time distribu-

tion, fast mixing, and fast heating offers the best conditions for a homogeneous

nucleation and short nucleation phases during the nanoparticle formation. Thus,

narrow distributions of particle sizes can be achieved. The high quality of micro- and

nanoparticles generated by the segmented flow tubular reactor (SFTR) is mainly

attributed to these effects. Further effort in understanding mechanisms of nucleation

and control of nucleation processes should help improve this critical process step for

different types of nanoparticles. In all cases, the nucleation should start under same

conditions in each volume element of a reaction mixture, go on under the same

conditions, and be finished after the same time interval. This time interval should be as

short as possible to avoid different starting conditions and different times for the

growth of nanoparticles following the nucleation. So, for the different nanoparticles,

special protocols are realization for realizationof simultaneous initiationofnucleation

and for very fast nucleation processes.

The possibility of generating homogeneous particle qualities is not restricted on

chemically uniform nanoparticles only. The advantages of microfluidics can also be

used for producing binary and other more complex nanoparticles with constant

structural and composition features. So, themixing and reaction activation conditions

can be optimized under microcontinuous flow, so that only one of several possible

nanoparticle types is formed.Forexample, the formationofbinarymetal nanoparticles

might be restricted by reaction condition in such a way that alloy nanoparticles, core/

shell particles, twin particles, or higher aggregates are formed.

FIGURE 10.17 Schematic of a possible “fluidoparticle informatics.”

CONCLUSIONS 389

The application of different strategies for microreactor-based nanoparticle synthe-

sis opens a wide spectrum of reaction conditions for addressing different particle

properties with narrow parameter distribution. It is expected that these possibilities

will be used in future for further improvement in synthesis methods for metal

nanoparticleswith specific catalytic, optical, and electronic properties; for specifically

functionalized metal nanoparticles with selective aggregation and binding behavior;

and for very different applications integrating the specific advantages of molecule-

like, solid-like, and mesoscale behavior of nanoparticles. These specific features are

very important for new generations of electronic and optoelectronic devices as well as

for new approaches in catalyst design and nanochemistry. It is assumed that micro-

fluidicswill help in future to qualify the automatedmicrosynthesis andmodification of

metal nanoparticles into new principles of nanoengineering, combinatorial develop-

ment of particle-based nanoarchitectures, developing of nanodevices, and nanoparti-

cle-based information handling.

ACKNOWLEDGMENT

Wewould like to thank J.Wagner,M.Brust, andT.R. Tshikhudo for their cooperation,

S. Schneider and F.M€oller for their support of the experimental work, and F. Jahn and

H. Romanus for ultramicroscopic images.

REFERENCES

1. Foos, E. E.; Twigg, M. E.; Snow, A. W.; Ancona, M. G. Competition between thiol and

phosphine ligands during the synthesis of Au nanoclusters. J. Cluster Sci. 2008, 19,

573–589.

2. Levy, R.; Thanh,N. T.K.; Doty, R. C.; Hussain, I.; Nichols, R. J.; Schiffrin, D. J.; Brust,M.;

Fernig, D. G. Rational and combinatorial design of peptide capping ligands for gold

nanoparticles. J. Am. Chem. Soc. 2004, 126, 10076.

3. Mie, G. Ann. Phys. 1908, 25(3), 377–455.

4. Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles,

Wiley–Interscience, 1983.

5. Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters, Springer: Berlin, 1995.

6. Link, S.; El-Sayed, M. A. Size and temperature dependence of the plasmon absorption of

colloidal gold nanoparticles. J. Phys. Chem. B 1999, 103, 4212–4217.

7. Sun,Y.G.; Xia,Y.N. Increased sensitivity of surface plasmon resonance of gold nanoshells

compared to that of gold solid colloids in response to environmental changes. Anal. Chem.

2002, 74(20), 5297–5305.

8. Link, S.; El-Sayed, M. A. Spectral properties and relaxation dynamics of surface plasmon

electronic oscillations in gold and silver nanodots and nanorods. J Phys. Chem. B 1999,

103, 8410–8426.

9. Linnert, T.; Mulvaney, P.; Henglein, A. Surface chemistry of colloidal silver: surface

plasmon damping by chemisorbed I-, SH-, andC6H5S-. J. Phys. Chem. 1993, 97, 679–682.


10. Dass, A.; Stevenson, A.; Dubay, G. R.; Tracy, J. B.; Murray, R. W. Nanoparticle

MALDI-TOF mass spectrometry without fragmentation: Au25(SCH2CH2Ph)18 and

mixed monolayer Au25(SCH2CH2Ph)18�x(L)x. J. Am. Chem. Soc. 2008, 130,

5940–5946.

11. S€onnichsen, C.; Franzl, T.; Wilk, T.; von Plessen, G.; Feldmann, J.; Wilson, O.; Mulvaney,

P. Drastic reduction of plasmon damping in gold nanorods. Phys. Rev. Lett. 2002, 88(7),

0774021–0774024.

12. Mock, J. J.; Oldenburg, S. J.; Smith, D. R.; Schultz, D. A.; Schultz, S. Composite plasmon

resonant nanowires. Nano Lett. 2002, 2(5), 465–469.

13. Kalkbrenner, T.; Hakanson, U.; Sandoghdar, V. Tomographic plasmon spectroscopy of a

single gold nanoparticle. Nano Lett. 2004, 4(12), 2309–2314.

14. Song, Y.; Zhang, T.; Yang,W.; Albin, S. Correlating plasmon resonance spectra with three-

dimensional morphology of single silver nanoparticles. J Phys. Chem. C 2008, 112,

10767–10772.

15. McNamee, C. E.; Pyo, N.; Higashitani, K. Atomic force microscopy study of the specific

adhesion between a colloid particle and a living melanoma cell: effect of the charge and

the hydrophobicity of the particle surface. Biophys. J. 2006, 91(5), 1960–1969.

16. deMello, A. J. Control and detection of chemical reactions inmicrofluidic systems.Nature

2006, 442(7101), 394–402.

17. Watts, P.; Wiles, C. Recent advances in synthetic micro reaction technology. Chem.

Commun. 2007, 5, 443–467.

18. Song, Y.; Hormes, J.; Kumar, C. S. S. R. Microfluidic synthesis of nanomaterials. Small

2008, 4(6), 698–711.

19. Edel, J. B.; Fortt, R.; deMello, J. C.; deMello, A. J. Microfluidic routes to the controlled

production of nanoparticles. Chem. Commun. 2002, 10, 1136–1137.

20. Uehara, M.; Nakamura, H.; Maeda, H. Preparation of ZnS/CdSe/ZnS quantum dot

quantum well by using a microfluidic reactor. J. Nanosci. Nanotechnol. 2009, 9(1),

577–583.

21. Kockmann, N.; Kastner, J.;Woias, P. Reactive particle precipitation in liquidmicrochannel

flow. Chem. Eng. J. 2008, 135(S1), S110–S116.

22. Su, Y. -F.; Kim, H.; Kovenklioglu, S.; Lee, W. Y. Continuous nanoparticle production by

microfluidic-based emulsion, mixing and crystallization. J. Solid State Chem. 2007, 180

(9), 2625–2629.

23. Wagner, J.; K€ohler, J. M. Continuous synthesis of gold nanoparticles in a microreactor.

Nano Lett. 2005, 5(4), 685–691.

24. Boleininger, J.; Kurz, A.; Reuss, V.; Soennichsen, C. Microfluidic continuous flow

synthesis of rod-shaped gold and silver nanocrystals. Phys. Chem. Chem. Phys. 2006,

8(33), 3824–3827.

25. K€ohler, J. M.; Held,M.; H€ubner, U.;Wagner, J. Formation of Au/Ag nanoparticles in a two

step micro flow-through process. Chem. Eng. Technol. 2007, 30(3), 347–354.

26. Wagner, J.; Tshikhudo, T. R.; K€ohler, J. M. Microfluidic generation of metal nanoparticles

by borohydride reduction. Chem. Eng. J. 2008, 135(1), 104–109.

27. Shalom, D.; Wootton, R. C. R.; Winkle, R. F.; Cottam, B. F.; Vilar, R.; deMello, A. J.;

Wilde, C. P. Synthesis of thiol functionalized gold nanoparticles using a continuous flow

microfluidic reactor. Mater. Lett. 2007, 61, 1146–1150.

REFERENCES 391

28. Tsunoyama, H.; Ichikuni, N.; Tsukuda, T. Microfluidic synthesis and catalytic

application of PVP-stabilized, similar to 1 nm gold clusters. Langmuir 2008, 24(20),

11327–11330.

29. Niesz, K.; Hornyak, I.; Borcsek, B.; Darvas, F. Nanoparticle synthesis completed with in

situ catalyst preparation performed on a high-pressure high-temperature continuous flow

reactor. Microfluid. Nanofluid. 2008, 5(3), 411–416.

30. Ruzicka, J.; Hansen, E. H. Flow injection analyses. 1. New concept of fast continuous-flow

analysis. Anal. Chim. Acta 1975, 78(1), 145–157.

31. Ruzicka, J.; Hansen, E. H. Flow injection analyses. 10. Theory, techniques and trends.

Anal. Chim. Acta 1978, 99(1), 37–76.

32. Petrlova, J.; Mikelova, R.; Stejskal, K.; Kleckerova, A.; Zitka, O.; Petrek, J.; Havel, L.;

Zehnalek, J.; Adam, V.; Trnkova, L.; Kizek, R. Simultaneous determination of eight

biologically active thiol compounds using gradient elution-liquid chromatography with

Coul-Array detection. J. Sep. Sci. 2006, 29(8), 1166–1173.

33. Alonso, A.; Almendral, M. J.; Curto, Y.; Criado, J. J.; Rodriguez, E.; Manzano, J. L.

Determination of the DNA-binding characteristics of ethidium bromide, proflavine, and

cisplatin by flow injection analysis: usefulness in studies on antitumor drugs. Anal.

Biochem. 2006, 355(2), 157–164.

34. Chen, X. W.; Chen, M. L.; Chen, S. A.; Wang, J. H. Flow-based analysis: a versatile,

powerful platform for DNA assays. Trends Anal. Chem. 2008, 27(9), 762–770.

35. Ivnitski, D.; Abdel-Hamid, I.; Atanasov, P.; Wilkins, E. Biosensors for detection of

pathogenic bacteria. Biosens. Bioelectron. 1999, 14(7), 599–624.

36. Borland, L. M.; Kottegoda, S.; Phillips, K. S.; Allbritton, N. L. Chemical analysis of single

cells. Annu. Rev. Anal. Chem. 2008, 1, 191–227.

37. Funfak, A.; Br€osing, A.; Brand, M.; K€ohler, J. M. Micro fluid segment technique

for screening and development studies on Danio rerio embryos. Lab Chip 2007, 7,

1132–1138.

38. Martin, K.; Lemke, K.; Henkel, T.; Grodrian, A.; K€ohler, J. M.; Metze, J.; Roth, M. Neue

Kultivierungstechnologie zur Suche nach seltenen Mikroorganismen. Biospektrum 2006,

7, 743–745.

39. K€ohler, J. M.; Henkel, Th.; Grodrian, A.; Kirner, Th.; Roth, M.; Martin, K.; Metze, J.

Digital reaction technology by micro segmented flow—components, concepts and appli-

cations. Chem. Eng. J. 2004, 101(1–3), 201–216.

40. Fukuyama, T.; Rahman, T.; Sato,M.; Ryu, I. Adventures in inner space:microflow systems

for practical organic synthesis. Synlett 2008, 2, 151–163.

41. Salazar-Alvarez, G.; Muhammed, M.; Zagorodni, A. A. Novel flow injection synthesis of

iron oxide nanoparticles with narrow size distribution. Chem. Eng. Sci. 2006, 61(14),

4625–4633.

42. Teh, S. Y.; Lin, R.; Hung, L. H.; Lee, A. P. Droplet microfluidics. Lab Chip 2008, 8(2),

198–220.

43. Vanapalli, S. A.; Banpurkar, A. G.; van den Ende, D.; Duits, M. H. G.; Mugele, F.

Hydrodynamic resistance of single confined moving drops in rectangular microchannels.

Lab Chip 2009, 9(7), 982–990.

44. Adzima, B. J.; Velankar, S. S. Pressure drops for droplet flows in microfluidic channels. J.

Micromech. Microeng. 2006, 16(8), 1504–1510.


45. Groß, G. A.; Thyagarajan, V.; Kielpinski, M.; Henkel, T.; K€ohler, J. M. Viscosity-

dependent enhancement of fluid resistance in water/glycerol micro fluid segments.

Microfluid. Nanofluid. 2008, 5, 281–287.

46. Donnet, M.; Bowen, P.; Jongen, N.; Lemaıtre, J.; Hofmann, H. Use of seeds to control

precipitation of calcium carbonate and determination of seed nature. Langmuir 2005, 21,

100–108.

47. Ding,Y.;Gu,G.;Xia,X.H.;Huo,Q.Cysteine-grafted chitosan-mediated gold nanoparticle

assembly: from nanochains to microcubes. J. Mater. Chem. 2009, 19(6), 795–799.

48. Dillenback, L. M.; Goodrich, G. P.; Keating, C. D. Temperature-programmed assembly of

DNA: Au nanoparticle bioconjugates. Nano Lett. 2006, 6(1), 16–23.

49. Hussain, I.; Brust, M.; Barauskas, J.; Cooper, A. I. Controlled step growth of molecularly

linked gold nanoparticles: from metallic monomers to dimers to polymeric nanoparticle

chains. Langmuir 2009, 25(4), 1934–1939.

50. Park, S. Y.; Lytton-Jean, A. K. R.; Lee, B.;Weigand, S.; Schatz, G. C.; Mirkin, C. A. DNA-

programmable nanoparticle crystallization. Nature 2008, 451(7178), 553–556.

51. Lee, J. S.; Lytton-Jean, A. K. R.; Hurst, S. J.; Mirkin, C. A. Silver nanoparticle-

oligonucleotide conjugates based on DNA with triple cyclic disulfide moieties. Nano

Lett. 2007, 7(7), 2112–2115.

REFERENCES 393

INDEX

A549 cancer cells, 269

Abbe’s theory of imaging, 368

acid-base reaction, 344

CLSM image, 344

Acrylamide copolymers, 149

N-Acryloylsuccinimide, 149

Activated silica column, 147

Active nanodrug carriers (NDCs), 188

Affinity biosensors, 99, 107

Alginate microparticles/nanoparticles, 246

diagram, 246

Alzheimer’s disease, 1

Amine-terminated nanoparticles, 225

7-Aminoclonazepam (7-ACZP), 110

Aminolysis, 150

Aminopropyl-modified nanoparticles, 226

3-Aminopropyltriethoxysilane (APTES),

24, 139

Ammonium/nitrogen-containing

ligands, 363

Anodic partial process, 364

Anodization, 54–56

Anodized aluminum oxide (AAO)

membranes, 52

Antibodies, 99–100

Antibody-antigen (Ab-Ag) interactions, 99

Anticancer drugs, 197

camptothecin, 197

daunorubicin (DaunoXome), 197

doxorubicin (Doxil), 197

Aplysia neurons, 24

Apoptosis, 16, 18

Aquagels, See Hydrogels

Astrocytes, 3, 18, 20, 26, 28, 36

Atomic force microscopy (AFM), 2, 3,

369, 370

AuNPs, see Gold nanoparticles (AuNPs)

a-N-Benzoyl-l-arginine ethyl ester(BAEE), 167

Bio-barcode assay (BCA), 105, 243

implementation, 243

Biodegradable nanoparticles,

characteristics, 227


395

Biological/biomedical microelectromecha-

nical system (Bio-MEMS), 74, 187,

191, 193, 209

applications, 200–208

components/flow chart, 192

experimental/computational simulation

aspects, 188

family of, 200

goal, 201

use, 187

Bioluminescent reaction, 172

Biopolymer-SiO2 nanocomposite aerogel, 64

Biosensing, 126

pillars for, 273

Biosensors, 49–51

for liver diagnosis, 81

for membrane-based labeled detection, 67

for membrane-based label-free detection,

67–68

impedance/capacitance, 70–71

mechanical detection, 73–74

optical detection, 71–73

potentiometric, 69–70

voltammetry, 68–69

functionality of membrane in, 64–66

Biotin-avidin-biotin coupling, 132

Biotinylated polylysine, 130

Block copolymer, 260

Blood-brain barrier, 2

Bottom-up approach, 257, 260

Bovine knee chondrocytes (BKCs), 227

fixed charge density, 227

zeta potential, 227

Bovine serum albumin (BSA), 6, 101

Bradford protein assay, 222

Brownian motion, 384

Caenorhabditis elegans, 4

Capacitance, 70, 72, 74

Capillary electrophoresis (CE), 131, 213

ESI/TOF-MS analysis of protein

mixtures, 159

microreactor for peptide mapping, 167

nanomaterial applications, 214

pseudostationary phase, 214

separation buffer additives, 214

using nanoparticles, 213

Carbon nanotubes (CNTs), 91, 195, 226, 238

bovine serum albumin-conjugated, 238

multiwalled nanotubes (MWNTs), 195,

226, 227

single-wall nanotubes (SWNTs), 195, 226

use, 226

Carbonyldiimidazole (CID) activation, 151

Catalytic nanosystems, 387

Cathodic partial process, 364, 365

role, 365

Cell-based biosensor, 14

Cellomics, pillars for, 268

cell characterization, 271–272

cell cultivation, 269–271

cell trapping, 268

Cell viability, 10

Ceramic nanoparticles, 195

Cetyltrimethylammonium bromide

(CTAB), 222

Chemically sensitized field effect transistor

(CHEMFET), 99

Chemical vapor deposition (CVD), 261

Chemiluminescence (CL), 72

detection, 96

Chip-based immunoassays, 139

Chitosan (CTS), 130, 155

Cholera toxin subunit B (CTB), 108

Circulating tumor cells (CTCs), 268

isolation, 268

pillar-based trapping, 269

Clonazepam (CZP), 110

CMOS-based semiconductor industry, 256

Colloidal self-assembly strategy, 241

Commercial multielectrode arrays, 34

Competitive binding assay formats, 103

Composite nanoparticles, 383, 384

architecture, 383

continuous-flow synthesis, 383

formation process, 384

Computational fluid dynamics (CFD)

simulations, 285

Computer-aided design (CAD), 7

Conductingpolymer nanowires (CPNWs), 96

Confocal laser scanning microscopy

(CLSM), 341–343

Continuous-flow processes, 362

Conventional lithography methods, 263

Conventional silica-based monolithic

columns, 147–148

Coulometric efficiency, 247

definition, 247

396 INDEX

CoventorWareTM software, 286

Critical nanoparticle concentration (CNC),

definition, 225

Cross-linking agent, concentration, 245

Crystal lattice, metallic properties, 362

Curie temperature, 141

Current density, 261

b-Cyclodextrin (CD)-modified

nanoparticles, 227

Cyclohexene, 299–300

Cytochrome P450 (CYP)-based immobilized

enzyme reactors, 170

based IMERs in drug metabolism, 171

Dendrimers, 196

properties, 196

water-soluble dendrimer, 196

Designs and devices. See also PDMS

microfluidic design

for neuroscience applications

for generating gradients of green and red

dyes, 13

gradient-generating designs, 12–13

integrated electrophysiology, 14–15

Diagnostic biosensors, See Biosensor

Dialdehyde activation, 150

Didodecyl dimethylammonium bromide

(DDAB), 222

capped gold nanoparticles, 222

Diethylene glycol (DEG), 338

Differential centrifugal sedimentation

(DCS), 371

Diffusion coefficient, 285, 342, 343

Dipole-dipole repulsion, 145

Direct simulation Monte Carlo (DSMC), 190

Disuccinimidyl suberate (DSS), 154

DNA, 216, 217, 229, 240

bioassay, nucleic acids structures, 111–113

dsDNA fragments, 242

electrophoresis, 217

hybridization, 111

mobility, 216

molecules, 217

role, 266

preconcentration, 239

on-chip preconcentration method, 239

separation/electrochemical detection,

240

probes, 114–115

protein arrays, 267

sequencing, 4

DNA biosensors, 113–114

electrochemical, 116–118

hybridization, 115

detection, 115–116

reaction, 115

optical, 118–120

probe, 114–115

Drug, 198

conventional nanoparticle carriers, 198

Drug delivery systems (DDSs), 188, 208

Drug-loaded system, 198

Electrical biosensors, See Biosensors

Electrochemical DNA biosensor, 116–118

Electrochemical immunosensors, 104–109

Electrochemical impedance spectroscopy,

231

Electrochemical transduction, 99

Electron beam interactions, 370

Electron beam lithography, 257, 266

Electron microscopic methods, 369, 370

scanning electron microscope (SEM), 369

transmission electron microscope (TEM),

369

Electroosmotic flow (EOF)-driven system,

214, 217, 218, 225

Electrophoresis, microfluidic chip, 266

Electrophysiology

and mass spectrometry, 24

and microfluidics, 26–28

change in, 33

effect of toxins on, 32

integrated, 14–15

Electroporation, 16

Electrospray ionization mass spectroscopy,

162

Electrostatic forces, 101

Entropy balance equation, 206

Enzyme immobilization techniques,

126–127

biospecific (affinity) adsorption, 130–131

covalent immobilization, 127–129

layer-by-layer assembly, 130

physical adsorption, 129–130

Enzyme immunoassay (EIA), 102

Enzyme-linked immunosorbent assay

(ELISA), 102

INDEX 397

Enzymes, 125

immobilized magnetic nanoparticles, 145

modified fused silica microreactor, 164

Escherichia coli, 64, 107

biosensor, 75–76

Esoteric driving forces, surface tension, 190

Extinction coefficient, 367

Extracellular matrix (ECM), 256

Fabricated high-density dot/pillar arrays, 259

SEM images, 259

Fabricated urea biosensor, 82

Faraday’s constant, 247

Faraday’s equation, 247

Fast diffusion, 3

Femtoliter segment, 388

Fiber-optic localized plasma resonance

(FO-LPR) 109

microfluidic chip, 109

Field-amplified sample injection (FASI), 239

Field-amplified sample stacking (FASS), 239

Flow cytometry, 4

Fluidoparticle informatics, 389

schematic presentation, 389

Fluorescein, chemical structures, 343

Fluorescence detection, 273

Fluorescence resonance energy transfer

(FRET), 71

Fluorescent immunoassay (FIA) analysis,

102, 375

Fluorinated ethylene propylene (FEP) mem-

brane, 18

Focused ion beam (FIB), 61

Gel-based electrophoresis, 265

Gene therapy, 29–30

Genomics/proteomics, pillars for, 264–268

biomolecule preconcentration, 264–265

DNA and protein arrays, 267–268

DNA stretching and separation, 265–267

g-forces, 372

Glicidoxypropyltrimethoxysilane, 139

Globulins, 100

Glucose biosensor, 79, 80

Glucose detection, See Porous membrane-

based biosensor

Glucose oxidase (GOx), 80

biosensor, 79

g-Glutamyl transpeptidase, 147

Glutaraldehyde (GA), 24, 144

Glycidoxypropyltrimethoxysilane

(GLYMO), 144

Glycidyl methacrylate copolymers, 149

GOD-CPG particles, 141

Goethite (a-FeOOH), nanoparticles, 349Gold nanoparticles (AuNPs), 95, 141,

214, 221–225, 229–232, 244, 367

core/shell, 370, 375

bimodal size distribution, 377

flow-through microspectrophotometry,

375

preparation, 380

TEM image, 370

dark-field image of, 369

DCS spectrum, 371

dispersions, 367

electrophoretic mobility, 236

nanomosaic network, 231

percentage, 381

probes, 243, 244

SEM image, 369

size distributions, 373, 374

thiol surface displacement, 243

three-dimensional AFM image, 371

use, 214, 229

Gold pillar arrays, 262

Green fluorescent protein (GFP), 227

3D structure, 228

mutant, 227

Growth factor, 28–29

Harrison’s approach, 242

Heat flux, 205

influence on minimal uniformity length,

205

HeLa cells, 16, 270

High refractive index, 362

High-throughput method, 257

High-throughput screening, 269

Homogeneity, improvement, 382

Horseradish peroxidase (HRP), 132

Human genomics project, 264

Hybridization, 115–116

Hybrid nanofluidic-microfluidic devices, 243

characteristic features, 243

Hydrogels, 196

Hydrogen production. See also Platinum

catalyst

398 INDEX

from SRM, 301–302

non-noble nanocatalysts for SRM

reactions to, 303–307

Hydrophobic-hydrophobic bonding

interaction, 235

Hydrophobic PMMA microchannels, 135

3-Hydroxybutyrate dehydrogenase, 139

Hydroxypropyl cellulose (HPC) matrix, 239

Hydroxysulfosuccinimide (sulfo-NHS), 133

Iminodiacetic acid (IDA), 129

based impedance biosensor, 108

Immobilization techniques, 64

Immobilized capillary acetylcholinesterase

(AChE) reactor, 169

Immobilized enzymes, 168

biosensors, 168–171

Immobilized glucose oxidase (GOD)

particles, 140

Immobilizedmicrofluidic enzymatic reactors

(IMERs), 126

application, peptide mapping, 156–158

capillary electrophoresis system,

162–168

liquid chromatography system,

160–162

MS, 158–160

online coupling, 164

Immobilized protease P, 147

Immobilizing sites, of proteins, 127

Immunoassays, formats, 100, 102–103

Immunoglobulin G (IgG), 100

Immunosensors, 103–104

Impedance/capacitance, 70–71

Indium tin oxide (ITO) electrodes, 21,

274

glass substrates, 274

Integrated sensors, 15

Interdigitated ultramicroelectrode array

(IDUA), 109

Interference localized surface plasmon

resonance (iLSPR) biosensor, 109

Iron oxide (Fe2O3) nanoparticles, 323

g-Fe2O3 nanoparticles, 339, 345

for magnetite and maghemite

nanoparticles, 333

coprecipitation, 333–335

influence of pH, 336

polyol process, 337–338

synthesis in constrained environments,

338–339

thermal decomposition, 336–337

synthesis, 325

metallic cations and polycondensation,

326–327

precipitation process, kinetic steps,

328–333

Iron oxyhydroxide nanoparticles, 323–324

microdroplet reactor, synthesis in,

348–349

microfluidic synthesis, 339

synthesis of a-FeOOH nanoparticles,

349–352

synthesis of g-Fe2O3 nanoparticles,

339–348

Isoelectrical focusing (IEF), 265

Knudsen number, 189

Label-free biosensor, 49–50

Lab-in-a-cell technology, 4

Lab-on a-chip devices, 214

mass production, 214

Labyrinth like structure, 145

Laminar flow, 3, 12

Langmuir adsorption isotherm model, 235

Langmuir kinetic models, 243

Laser-Doppler electrophoresis, 238

Laser-induced fluorescence (LIF), 110

detection, 131

Lattice Boltzmann method (LBM), 190

Lennard–Jones (L-J) potential, 191

Liftoff technique, 260

Linear polyacrylamide (LPA), 224

Lipid-based liquid crystalline nanoparticles,

227–228

Liposomes, 97, 196, 197

property, 197

Lithography, 59–60

Localized surface plasmon resonance

(LSPR), 95

Locked nucleic acids (LNA�), 99, 114

Lodestone, 323

Lower critical solution temperature (LCST),

233

Low-pressure chemical vapor deposition

(LPCVD), 62

LRM55 cells, 270

INDEX 399

Macroscopic heterogeneous chemistry, 364

Magnetic anisotropy energy, 325

Magnetic cell trapping, 268

Magnetic microparticles (MMPs), 105

Magnetic nanoparticles, 232–235

Magnetite, 323

Fe3O4 nanoparticles, 334

Matrix-assisted laser desorption/ionization

mass spectrometry (MALDI-MS),

156

probe, 153–154

TOF (time-of-flight) mass spectrometry,

235, 368

for the digestion and peptide mapping,

155

Mass spectrometry (MS), 158

Membrane-based biosensor, 48

Membrane-based technology, 52. See also

Nanoporous membrane

Mercaptopropyltrimethoxysilane

(MPTMS), 117

Metal-chelated complexes, 338

Metal nanoparticles, 362, 364, 366, 382, 384

colloidal solutions, 366

core/shell, generated by microsegmented

flow use, 379–381

formation mechanism, 363

nucleation, 363–364

particle growth, 364–365

surface capping, 365–366

growth, 364

interaction, 362

plasmon absorption, 366

synthesis, 372, 375

in homogeneous fluids, 372–375

in microreaction technology,

challenges, 381

under segmented flow conditions,

375–381

overwhelming number, 382

Methyl methacrylate (MMA), 155

Methylmethacrylate-sulfopropylmethacrylate

(MMA-SPM) nanoparticles, 227

Micellar electrokinetic chromatography

(MEKC), 215, 217–219

advantage, 219

separation principle, 218

Micelle, 197

vs. drug carrier systems, 197

Michaelis–Menten constant (Km), 137, 171

Microchannels

and microfluidics, 189

blockages of catalyst active sites in, 311

coated, 217

degree of microbeads immobilized on, 140

enzymatic reactions in, 126

flow rates in, 373

fluid-particle flow in, 200, 206

high surface area of, 309

nanocatalyst deposition on, 294

nanodrug transport phenomena in, 189

network comprising of, 84

network of omega-shaped, 288

optical image of, 140

packing with, 137

selective deposition of catalyst in, 298

silica and alumina on, 299

turbulent flows in, 285

unusual behavior of fluid traversing, 3

Microchip-based bio-barcode assay, 243

Microchip electrophoresis (MCE), 213

microfluidic devices, 214–219

advantages and applications, 214

separation techniques, 215

size, 213

using nanoparticles, 213

Microchip enzymatic microreactor, 232

Microchip for cancer biomarkers, 106

Microchip gel electrophoresis (MGE), 215

separation of DNA, 216

Microchip-immobilized magnetic enzyme

reactor system, 235

compatibility, 235

Microchip zone electrophoresis (MZE),

215–216

Microdroplet reactor, 348

Microelectromechanical systems

(MEMS), 261

rf-interrogated biosensor, 74

Microemulsions, 197

Microfluidic channels, 255

application examples, 264–268, 274

pillars for biosensing, 273–274

pillars for cellomics, 268–272

pillars for genomics and proteomics,

264–268

introduction, 255–257

other fabrication aspects, 261–264

400 INDEX

integration into microdevices, 263

material choice, 261–262

surface fuctionalization, 262

patterning techniques, 257–261

auto assemblage, 260

growth, 261

lithography and related techniques,

257–259

pillars and pillar arrays, 255

Microfluidic chips, 15, 29, 255

chip based biosensor, 108

combining a biobarcode with, 119

fabrication, 7

FO-LPR, 109

for analyzing benzodiazepines, 110

for electrophoresis, 266

for microelectronic circuitry, 2

glass-based, 273

ideal systems for, 256

interfaces with ESI, 159

manipulation of Ca-alginate microspheres

using, 245

miniaturization of the sol-gel structures

on, 132

multiple active sites on, 51

neurophysiology experiments using,

See Neurophysiology

pillar arrays integrated, 274

PMMA for enantioseparation, 238

to improve neuron survival, 11

Microfluidic devices, 187, 191, 214–219,

266, 268

advantages and applications, 214

Bio-MEMS system, 187

categories, 191–194

development, 191

lab-on-a-chip (LOC) systems, 187

mechanics, 191

microfluidics and microsystems, 189–190

microsystem modeling assumptions,

190–191

R&D areas, 189

separation techniques, 215–219

DNA separation by microchip gel

electrophoresis, 216–217

micellar electrokinetic chromatography,

217–219

microchipzoneelectrophoresis, 215–216

use, 191

Microfluidic platforms, 4

architectural designs, 8–9

Microfluidics system, 83, 201, 241, 388, 389

advantages, 389

analysis systems, 264

based cell studies, 4

based colloidal self-assembly technique,

239

based culture platform, 31

based microchips, 231

biosensor systems, 82–84

chambers for gene therapy, 6

for MALDI protein analysis, 160

in bioassays, 93–94

operation of information, 388

self-assembled colloidal arrays,

fabrication/characterization, 241

sorting device, 233

synthesis, development, 381

technology, defined, 2

to synthesize cDNA, 17

Micro-high-performance liquid

chromatography (m-HPLC), 156

Micromixer

baffle-slit, 203

influence, 202

pumping power, 202

Micromixing, 366

omega structure, technique, 285–289

principles and types, 284–285

Microreaction technology, 364, 366, 381,

385, 387

advantages, 382

approaches, 366

development, 388

Microreactors, 361, 381

based nanoparticle synthesis, 390

construction, materials for, 289–290

for gas-to-liquid technology, 309–314

iron-cobalt mixed catalysts, 309–312

ruthenium as promoter, 312–314

for kinetic studies, 171–172

for steam reforming of methanol, 302

hydrogen production and purification, 301

metal nanoparticle synthesis, 361

technical development, 381

Microscaffold system, 19

Micrototal analysis systems (m-TAS), 214Mie plasmon, 366

INDEX 401

Mie theory, 366

Migration time window, 218

Mild anodization (MA), 54

Mix-and-match lithography methods, 259,

263, 264

advantages, 263

Mobility

apparent mobility, 221

electroosmotic mobility, 221

Molecular dynamics simulation (MDS), 190

Monolithic phases, 145–147

Multichannel interfaces, 388

microfluidic strategies for, 388

Multielectrode arrays (MEAs), 5

Multiple sclerosis, 1

Nanocatalyst

calcination temperature, 304

deposition on microchannels, 294

coating methods, 294–297

CVD technique, 295

electrochemical deposition, 295

pretreatment of substrate, 294–297

PVD vs. sol-gel method, 297–301

sol-gel method, 296

in GTL technology, 284

non-noble, for SRM reactions to produce

hydrogen, 303

particle sizes of, 312

screening, parallel microreactor system

for, 314–318

SRM reactions conducted over, 305

Nanocrystalline iron (II, III) oxide, 338

Nanocrystals, 62, 75, 76, 96, 195, 330, 336,

337

Nanodevice system, 231

Nanodrug delivery systems, 187, 194–200

bio-MEMS applications, 200–208

device optimization, 203

nanofluid flow simulations, 201–203

conclusions/future perspective, 208

introduction, 187–190

microfluidic devices, 188–194

categories, 191–194

microfluidics and microsystems,

189–190

microsystem modeling assumptions,

190–191

strategies, 194

Nanodrugs, 194–200

carriers, passive drug targeting, 199

colloidal soft matter, 196–198

development, 211

nanodrug carriers, desirable


solid nanoparticles, 194–196

targeting, 198–200

Nanoelectrodes, 2

Nanoimprint lithography, 258, 263, 266

Nanolaths, TEM image, 352

Nanomaterials, application, 382

Nanometric ferro/ferrimagnetic

particles, 324

Nanoneedles, 2

Nanoparticle-mediated capillary


carbon nanotubes, 226

gold nanoparticles, 221–225

lipid-based liquid crystalline nanoparticles

227–228

silica nanoparticles, 225

TiO2 226

Nanoparticle-mediated electrophoresis

(NME), 220, 221

Nanoparticle-mediated microchip


Au microelectrodes, in-channel

modification, 247

AuNP-mediated on-chip preconcentration,

239

biopolymer micro/nanoparticles,

microfluidic fabrication, 244

carbon nanotubes, 238–239

colloidal Au self-assembly, 239–243

gold nanoparticles, 229–232

magnetic nanoparticles, 232–235

microchip-based bio-barcode

assay, 243

polymer nanoparticles, 236–238

surface displacement reactions, 243

TiO2 nanoparticles, 236

zeolite nanoparticles, 235–236

Nanoparticles (NPs), 213, 219, 362, 364, 383

aggregation, 248

aggregates formation, 385

application in CE and MCE, 219

nanoparticle-mediated capillary

electrophoresis, 220–228

402 INDEX

nanoparticle-mediated microchip


based catalytic assemblies, construction,

387

based/catalytic nanomachines, vision, 386

based information handling system, 388

carrier, 198

size, 198

surface charge, 198

colloidal suspensions, characterization, 247

definition, 213

filled capillary electrophoresis, 222

formation, 364

growth, 365

plasmon absorption, 374

polybutylcyanoacrylate (PBCA), 227

product/process characterization, 366–372

differential centrifugal sedimentation,

371–372

optical microscopy, 368–369

spectrometry, 366–368

ultramicroscopy, 369–371

properties, 220

regional functionalization, 384

shape, 365

silica-based, 96

small dimensions, 362

surface capping, 365

surface state, 365

synthesis, 373, 383

advantages, 373

TEM image, 347

three-dimensional self-assembling, 384

used in bioassays, 94–97

Nanopillars, 260

arrays, 267

Nanoporous membrane, 47

aluminum oxide membrane (AOM), 84

basedmicrofluidic biosensors, application,

50–51

efficient size sorting, 52

molecular sorting, design considerations

for, 52–53

fabrication and integration into

microfluidic device, 54

anodization, 54–56

focus ion beam etching, 61–62

ion track etching, 56–58

lithography, 59–60

phase separation, 58–59

rapid thermal annealing (RTA)

technique, 62

sol-gel technology, 63–64

need for, 51–52

Si3N4/SiO2 membranes, 82

types of, 53

Nanotechnology hardware, 3

Nanowires, 2, 91, 272

Navier–Stokes equation, 286, 342

Negative dielectrophoretic (nDEP) force, 269

Nerve growth factor (NGF), 29

Nervous system, 1

Neural lineage cells, 3

Neural prosthesis development, 18

Neuroblastoma-glioma hybrid cells, 16

Neurodegenerative processes, 1

Neurons, 1

Neuropeptides, 20

Neurophysiology

experiments using microfluidic chips, 19

axonal isolation, 30, 32–33

cell separation tools, 19–20

electrophysiology, 26–28

gene therapy, 29–30

growth factor effects, 28–29

neuropeptide release, 20–23

physical and chemical guidance cues,

23–26

Newton’s second law of motion, 191

N-hydroxysuccinimide (NHS), 134

Nickel-nitrilotriacetic acid (Ni-NTA), 139

N-methyl diethanolamine (NMDEA), 338

NMR spectroscopy, 16

Noncross-linking interaction mechanism,

231

Nonspherical particles, 381

generation, 382

morphological classes, 381

yield improvement, 381

Nucleation mechanism, 363–364, 383

definition, 383

centers, formation, 374

Nucleic acids structures, 111–113

Oligonucleotides (ODNs), 114

Omega channel microreactors, 286–288

On-chip microreactor, 235

Open-channel reactor designs, 132

INDEX 403

Open-ended microchannel, 216

Optical detection techniques, 71–73

Optical DNA biosensor, 118–120

Optical immunosensor, 109–111

Optical lithography, 258

Organic polymer monoliths, 148–152

Organogels, 197

Organomercaptans, 243

surface displacement kinetics, 243

Ovalbumin (OVA), 101

Oxidase-based amperometric biosensors, 78

Packing microchannels, with micro/

nanoparticles, 137–138

magnetic supports, 141–145

nonmagnetic supports, 138–139

organic supports, 139–141

Parkinson’s disease, 1

Particles, 386

assembly types, 387

based information processing, 388

microfluidic strategies for, 388

capture and release scheme, 234

stabilization, 386

Particle-supported systems, 371

dye pigment suspensions, 371

Passive multifunctional nanoparticle systems

(MFNPSs), 188

Patch clamp array, 15

Patterning techniques, 257–261

auto assemblage, 260

growth, 261

lithography and related techniques,

257–259

P�eclet numbers, 201, 202

Penicillin G acylase (PGA), 137

Pepsin, 156

Peptide mapping, 126

microfluidic enzymatic reactors for, 132

Phase transfer processes, 366

Photonic crystal sensor, 274

Fourier transform infrared spectroscopy,

274

Photopolymerization, 148

Physical vapor deposition (PVD), 261

Platinum catalyst

for preferential oxidation of CO

in hydrogen, 307–309

in PDMS, 11

Polyacrylamide, 149

Polycarbonate (PC), 52, 58, 94

Poly(diallyldimethylammoniumchloride)

(PDDA), 134

filled CE, 223

made device, 134

Poly(diallyldimethylammonium chloride)

(PDDC), 221

adsorption, 221

Polydimethylsiloxane (PDMS), 6, 82, 258,

271

architectural designs, 8–9

based microdevices, 246

characterstics, 6–7

chip fabrication protocol, 7–8

culture plate, 18

elastomer, 94

growth chambers, 34

microchannels, 21

microchip channel, 229

microfluidic design, 6–7

practical considerations and limitations,

10–12

tools, 9–10

use, 258

Poly(ether ether ketone) (PEEK), 147

Polyethylene, 94

terephthalate, 58

Poly(ethylene glycol) (PEG) hydrogel,

7, 270


nanopillars, aggregated cardiomyocytes,

271

Poly(ethylene oxide) (PEO), 222, 229

Poly(ethylene terephthalate) (PET)

microchannels, 141


photoablated microchannel, 232

SEM images, 232

Poly(ethylenimine) (PEI), 29

based DNA, 30

Poly(glycidyl methacrylate-co-ethylene

dimethacrylate), 146, 150

Polyimide (Kapton) membranes, 58

Poly-L-lysine (PLL), 24

Polymerase chain reaction (PCR), 264

Polymer-based microparticles, advantages,

244

Polymer encapsulation, 142–143

404 INDEX

Polymeric membranes, incorporation into

microfluid, 152–153

Polymerization-induced phase separation

(PIPS), 59

Polymer (s), 195, 262

advantage, 262


fabrication, 262

nanoparticles, 236

pillars, 262

Polymethylmethacrylate (PMMA), 94, 135,

235

Poly-N,N dimethylacrylamide (PDMA), 224

Polyols, 337

Polypropylene, 94

Polystyrene, 94

Polystyrene-block-poly(methyl methacrylate)

(PS-b-PMMA) copolymer, 53

Polysulfone (PSf) porous films, 60

Poly(vinyl alcohol) (PVA), 139

Poly(vinylidenefluoride) (PVDF)membrane,

59, 153

Polyvinyl pyrrolidone (PVP), 373

Porous membrane-based biosensor

as cholesterol biosensor, 80

E. coli biosensor, 75–76

for detection of S. enteritidis, 76–77

for treatment of diabetes mellitus, 78–80

for virus detection, 77–78

new diverse sensors, 81–82

Post-translational modified proteins, 162

Potassiumchloride (KCl), 21

Potentiometric sensors, 69–70

Pressure drop vs. pumping power, 207

Programmed cell death, See Apoptosis

Prostate-specificmembrane antigen (PSMA),

236

Protein digestion, 139, 153

enzymes for, 157

Protein encapsulation techniques, 132

Proteolysis analysis, 233

Proton exchange membrane fuel cell

(PEMFC), 301

Pyrex� 7740, 116

Quantum dots (QDs), 2, 96

CdSe-based, 96

Radioimmunoassay (RIA), 102

Raman imaging system, 274

chip with integrated patterns, schematic

diagram, 274

Raman spectroscopy, 273

Rapid thermal annealing (RTA), 62

Rayleigh criterion, 258

Reactive ion etching technique, 260

Real-time biosensor, 49

Real-time measurements, 48–49

Retention factor, definition, 218

Reynolds number, 3, 191, 202, 209, 210, 284

ratio vs. minimal uniformity length, 203

vs. system entropy generation, 206, 207

Ruthenium dye particles, 16

S/V ratio, 133

S-adenosylhomocysteine (SAH), 135–136

Salmonella enteritidis, 76–77

Scanning electron microscope (SEM), 270,

369, 370

images of microchannel modified with

silicalite-1, 142

micrographs of ID monoliths, 163

picture of typical porous structure, 147

Scanning probe microscopic methods, 369

atomic force microscopy (AFM), 369

Schiff’s base, 127

Sedimentation methods, 372

principle, 372

Segmented flow conditions, 375

advantage, 376

mixing in, 376

process homogeneity/product quality, 377

scatterplot, 378

segment formation, 379

specificity, 375

Segmented flow tube reactor (SFTR), 383,

389

Self-assembled metal nanoparticles, 387

electronic/optoelectronic devices, 387

Self-assembled monolayers (SAMs), 12

analysis, 135

Silica, See Silicon dioxide (SiO2)

Signal-to-noise (S/N) ratio, 51

Silicalite-1, 141

Silicon

based materials, 261

microchannel reactors, 290–294

microstructures, 265

INDEX 405

Silicon (Continued )

nanopillars, 261

porous membrane-based optical label-free

biosensor, 73

Silicon dioxide (SiO2), 109, 242, 261, 262,

299

AFM image, 299

encapsulation, 143–145

nanoparticles, 225

photonic thin-film multilayers, 109

pillar array, SEM images, 267

SEM picture, 299

Silver nanoparticles, 367

core/shell, 370

bimodal size distribution, 377

flow-through microspectrophotometry,

375

preparation, 380

TEM image, 370

dark-field image of, 369

dispersions, 367

film deposition, schematic diagram, 237

Simulation, in vivo tissueswithmicrofluidics,

17–19

Single plasmon resonance methods, 368

Single-stranded DNA (ssDNA), 232

sequencing ladders, 265

Smart drug delivery system (SDDS), 192

Smooth muscle cell, SEM, 271

Sodium dodecyl sulfate (SDS), 218, 373

denatured protein markers, 242

Soft lithography process, 233

Soft nanoimprint lithography, 267

UV lithography, 267

Sol-gel encapsulation, 132

Solid lipid nanoparticles (SLNs), 196, 227

Solid-phase immunoassay, 99

antibody in sensor applications, 100–101

antigen/antibody interaction, 101–102

Solid-state semiconductor techniques, 382

Solid tumors, chemotherapeutic

agents, 199

SPION nanoparticles, 352

S-ribosylhomocysteine (SRH), 136

Staphylococcus aureus, 131

Stem cells, 20

Stokes equation, 372

Streaming process, 372

Streptococcal protein G, 131

Superparamagnetic nanoparticles,

196, 324

Superparamagnetism, 324

Surface-dependent enzymatic

reaction, 268

Surface enhanced Raman spectroscopy

(SERS), 273

based biosensing, 273

Surface immobilized biobarcode assay

protocol, 106

Surface plasmon-based sensors, 273

Surface plasmon resonance (SPR), 99

Surface-sensitive processes, 383

Surfactants, 366

Synaptogenesis on chip, 22

Synergism, 92

Tetraethyl orthosilicate (TEOS), 63

Tetramethyl orthosilicate (TMOS), 63

Thermal energy (kT), 324

Thermally induced phase separation (TIPS),

59

Tin-doped indium oxide (ITO), 117

TiO2-based nanoparticles, 226, 236

Top-down techniques, 257

L-1-Tosylamido-2-phenylethyl chloromethyl

ketone (TPCK)-trypsin, 128, 158

Transmission electron microscope (TEM),

348, 369, 370

Traumatic injuries, 1

Tris(2,20-bipyridyl)dichlororuthenium(II)

(RuBpy), 96

Trypsin

encapsulation, 131

in tetramethoxysilane-based hydrogel,

167

immobilized MALDI probe, 153

linked magnetic nanospheres, 155

membrane reactor, 153

modified magnetic nanoparticles, 144

Tumor cell receptors, 199

Tumor markers, detection, 107

Two-dimensional polyacrylamide

gel electrophoresis (2D PAGE),

160

Ultraviolet (UV)-modified PMMA

surface, 133

UV lithography, 258

406 INDEX

UV nanoimprint lithography (UV-NIL),

258–259

UV spectroscopy, 265

UV-Vis spectrometer, 368

van der Waals bonds, 65

van der Waals forces, 101, 242

Vinylazlactone copolymers, 151–152

Voltammetric sensors, 68–69

Water-in-oil (w/o) emulsions, 245

X-ray photoelectron spectroscopy, 238

Zeolite nanoparticles, 235

Zinc oxide (ZnO), 274

INDEX 407

MICROFLUIDIC DEVICES IN NANOTECHNOLOGY

Documents

Transcript of MICROFLUIDIC DEVICES IN NANOTECHNOLOGY