Novel hybrid optical sensor materials for in-breath O2 analysis
-
Upload
independent -
Category
Documents
-
view
0 -
download
0
Transcript of Novel hybrid optical sensor materials for in-breath O2 analysis
Novel hybrid optical sensor materials for in-breath O2 analysis
Clare Higgins, Dorota Wencel, Conor S. Burke, Brian D. MacCraith and Colette McDonagh*
Received 19th October 2007, Accepted 5th November 2007
First published as an Advance Article on the web 28th November 2007
DOI: 10.1039/b716197b
This study focuses on the optimisation and characterisation of novel, ORganically MOdified
SILicate (ORMOSIL)-based, hybrid sensor films for use in the detection of O2 on a breath-by-
breath basis in human health monitoring applications. The sensing principle is based on the
luminescence quenching of the O2-sensitive ruthenium complex [Ru(II)-tris(4,7-diphenyl-1,10-
phenanthroline)], which has been entrapped in a porous sol–gel film. The detection method
employed is that of phase fluorometry using blue LED excitation and photodiode detection.
Candidate sensor films include those based on the organosilicon precursors,
methyltriethoxysilane, ethyltriethoxysilane, n-propyltriethoxysilane and phenyltriethoxysilane.
While it has been established previously by the authors that these films exhibit a stable, highly
sensitive response to O2, this study focuses on selecting the material most suited for use in a breath
monitor, based on the sensitivity, response time and humidity sensitivity of these films. Key
parameters to be optimised include the O2 sensitivity of the film and the film polarity, i.e. the
degree of hydrophobicity. These parameters are directly linked to the precursors used. In this
study a n-propyltriethoxysilane-derived O2 sensor platform was selected as the optimum material
for in-breath O2 analysis due to its short response time, negligible humidity interference and
suitable O2 sensitivity in the relevant range in addition to its compatibility with a single-point
calibration strategy.
Introduction
The diagnostic value of O2-in-breath analysis is well estab-
lished.1 Breath analysis is attractive as it permits non-invasive
monitoring of a patient’s state of health.2 Despite the benefits
of breath analysis, the use of breath gas analysers is typically
restricted to a laboratory environment due to the expensive,
cumbersome nature of most breath analysis systems. The
development of a portable, low-cost system is attractive as
such a system would be compatible with home healthcare and
advanced athletic performance monitoring applications. A key
requirement for such a breath monitoring device is a
disposable sensor element, capable of operating at a range of
breath rates in high humidity, which is incorporated into a
lightweight, compact sensing module.
Luminescence-based O2 sensors have been extensively
researched in recent years due to their advantages over the
Clark electrode.3 These sensors combine the intrinsic sensitiv-
ity of the luminescence process with the wide availability of
low-cost, discrete optoelectronic components, thereby enabling
a range of sensor configurations and facilitating design
features that are desirable in biomedical applications, such as
miniaturisation and disposability. Sol–gel materials are an
ideal sensor matrix, due to their ease of fabrication and the
versatility of the process. The sol–gel process is compatible
with the use of deposition techniques such as ink-jet- and pin-
printing,4–6 which enable deposition onto a range of planar
and non-planar substrates. This paper addresses some key
issues relevant to the successful implementation of a portable
in-breath O2 sensor platform. These issues include response
time, influence of humidity on sensor response and enhanced
sensitivity in the range of interest for O2 sensing. A range of
ORganically MOdified SILicate (ORMOSIL)-based O2
sensor materials are reported where the O2-sensitive complex
[Ru(II)-tris(4,7-diphenyl-1,10-phenanthroline)], [Ru(dpp)3]2+,
is entrapped in the ORMOSIL xerogel. The O2 sensitivity of
such systems is dependent on the choice of precursor, which
influences the relative hydrophobicity of the resulting material,
thereby impacting on O2 transport into the material. This
paper reports on the effect of various organosilicon precursors
on the response time, relative hydrophobicity and O2
sensitivity of the resulting sensor elements.
Theory
Oxygen-sensitive materials comprised of an O2-sensitive
luminophore immobilised in an optically transparent
O2-permeable host matrix have been reported widely in the
literature.7–16 Sensing using these elements is attractive as the
sensor does not consume O2 and is not prone to electrical
interference. Such sensor elements are reported here. The
sensing mechanism centres on the dynamic quenching of
the excited-state luminescence lifetime (and, correspondingly,
the luminescence intensity) of the luminophore by O2
molecules. If the luminescence quenching is purely dynamic,
the excited-state lifetime and the intensity are related to the O2
concentration. This process is governed by the Stern–Volmer
relation,17 given in eqn (1),
Optical Sensors Laboratory, National Centre for Sensor Research,School of Physical Sciences, Dublin City University, Dublin 9, Ireland.E-mail: [email protected]; Fax: +353 1 700 8221;Tel: +353 1 700 5301
PAPER www.rsc.org/analyst | The Analyst
This journal is � The Royal Society of Chemistry 2008 Analyst, 2008, 133, 241–247 | 241
I0
I~
t0
t~1zKSV O2½ � (1)
where I is the luminescence intensity, t is the excited-state
lifetime of the luminophore, [O2] is the O2 concentration, KSV
is the Stern–Volmer constant and the subscript 0 denotes the
absence of O2.
KSV is a measure of the sensitivity of a sensor element and is
related to t0 by eqn (2) (where kq is the diffusion-dependent
bimolecular quenching constant) and to the diffusion coeffi-
cient, D, by eqn (3):
KSV = t0kq (2)
kq = 4pgRND (3)
In eqn (3), g is the spin statistical factor, R is the collision
radius, and N is Avogadro’s number. From eqn (2) and eqn
(3), it is clear that KSV is proportional to D, all other factors
being constant. Consequently, the sensitivity of the sensor
layer is increased by improving O2 transport through a
material.
From eqn (1), for an ideal, homogeneous environment, a
plot of the ratio of t0/t as a function of [O2] yields a straight
line with an intercept at 1 and a slope of KSV. In the case of
luminophores entrapped in a solid matrix, these plots deviate
from linearity.15,18–22 This behaviour is associated with the
distribution of the luminophore within the solid matrix. The
host micro-heterogeneity causes luminophore populations in
different sites to be quenched differently with a resultant
downward curve in the Stern–Volmer plot. Various models
have been developed to describe quenching data that deviate
from the Stern–Volmer ideal.23–25 For the sensor systems in
this work, the Stern–Volmer [eqn (1)] and two-site Demas
model [eqn (4)]7 were employed to analyse the O2 data. The
Demas model assumes that the luminophore population
distributed in the solid material resides at two different site
types, with each site exhibiting a particular quenching
constant.
I0
I~
t0
t~
f1
1zKSV1 O2½ �z
f2
1zKSV2 O2½ �
� �{1
(4)
In eqn (4), fi represents the fractional contribution of the
total emission from the luminophores located at site i (under
unquenched conditions) that exhibit a discrete Stern–Volmer
constant given by KSVi.
Excited-state lifetime measurements offer certain
advantages with respect to the performance of the sensors.
As lifetime is an intrinsic property of the luminophore,
instrumental fluctuations, leaching and photobleaching of
the luminophore do not affect the sensor performance.
Therefore, the long-term stability of such sensors is much
improved compared with that of luminescence intensity-based
sensors.
In this work, phase fluorometry is employed to indirectly
monitor t. The excitation signal is sinusoidally modulated and
the luminophore’s emission is also modulated but is time-
delayed or phase-shifted relative to the excitation signal. The
relationship between t and the corresponding phase shift, w,
for a single exponential decay is given by eqn (5),
tan w = 2pft (5)
where f is the modulation frequency of the LED excitation
source. Therefore, phase fluorometry facilitates the indirect
monitoring of the luminescence excited-state lifetime, thereby
avoiding the high cost and time-consuming data processing
issues associated with direct lifetime measurements.26,27 This
capability can be provided using low-cost instrumentation and
straightforward data analysis, making it a more attractive
candidate for sensor development compared to direct lifetime
measurements.
The support medium in this work is produced via the sol–gel
process. Materials prepared using sol–gel technology can range
from simple inorganic glasses (e.g. when using tetraethoxysi-
lane, TEOS) to more complex inorganic–organic materials
called ORMOSILs.28–30 These materials facilitate the develop-
ment of a broad range of innovative materials and are of the
general form R4 2 xSi(OR9)x, where R represents an organic
group, such as –CH3, –C2H5, –C3H7, –C6H5, (–CH2)nNH2. R
is generally bound to the silicon via a Si–C bond that is not
hydrolysable and these silanes are often used as network
modifiers and network formers. They can combine the
advantages of the organic and inorganic constituents within
the same matrix and the physico-chemical properties can be
tailored through considered choice of R. Therefore, one can
fabricate materials with the desired hydrophobicity, flexibility
and stability. Sensor technology is an area in which the use of
these sensor matrices is becoming increasingly prevalent due to
the reliable, versatile, low-temperature nature of the sol–gel
technique.31,32 The organosilicon precursors selected for use
here have the effect of increasing the relative hydrophobicity of
the surface of the resulting material and also reducing the
connectivity of its microstructure. Both of these factors
contribute to improved O2 transport within the resulting
material, increasing D and thereby contributing to an
increased KSV. We will show that the increased hydrophobicity
also influences humidity sensitivity. This is of particular
relevance when selecting a material for use in a breath
monitoring application, due to the high level of humidity in
breath.
Materials and methods
Chemical reagents
RuCl3?3H2O, 4,7-diphenyl-1,10-phenanthroline ligand, abso-
lute ethanol (EtOH), 0.1 M HCl and the organosilicon
precursors methyltriethoxysilane (MTEOS), ethyltriethoxysi-
lane (ETEOS), n-propyltriethoxysilane (PTEOS) and phenyl-
triethoxysilane (PhTEOS), were purchased from Aldrich
Chemicals. All chemicals were used as received.
[Ru(dpp)3]2+was synthesised as described in the literature.33
Glass microscope slides were soaked for 24 h in HNO3, and
then rinsed with deionised water and EtOH before drying
under a N2 flow. All experiments were performed at room
temperature.
242 | Analyst, 2008, 133, 241–247 This journal is � The Royal Society of Chemistry 2008
Fabrication of sol–gel-based sensor platforms
All sol–gel-based films were prepared from sols containing the
relevant precursor, EtOH as co-solvent, HCl and the
[Ru(dpp)3]2+ complex at a concentration of 2.5 g L21 with
respect to the total volume of sol. For all sols, the final molar
ratio of silane : EtOH : water : HCl was 1 : 6.25 : 4 : 0.007. All
films used in this study were dip-coated in a controlled
environment using a computer-controlled dipping apparatus.
After deposition onto microscope slides, the films were dried at
110 uC for 18 h. Profilometry measurements showed that the
dip-coated sensor films were 400–500 nm in thickness.
Phase fluorometry instrumentation
The principle of phase fluorometry and the experimental
system used to examine the performance of the O2 sensors
have been described elsewhere.34,35 Briefly, the characterisa-
tion system consisted of a blue LED (Nichia, NSPE590,
Japan), which was modulated at a frequency of 20 kHz and
provides excitation of the luminescent sensors. A silicon
photodiode (Radionics, 194-290, Ireland) was used for the
detection of the O2-sensitive luminescence signal. In order to
examine their performance, O2-sensitive films were placed in a
flow cell into which controlled mixtures of O2 and N2 were
flowed using mass flow controllers (Celerity, Ireland). In order
to test the effect of humidity on O2 sensitivity, two gas wash
bottles containing deionised water were included in the gas line
before the gas mixture entered the flow cell, and a commercial
humidity probe (Testo 625, Germany) was employed to
monitor the humidity in the cell. To determine the effect of
temperature on the O2 sensitivity, a gas heater (Radionics, 200-
2496, Ireland) was included in the gas delivery setup.
Other characterisation techniques
In order to measure the intrinsic response time of the sensor
films and not that of the gas handling system (i.e. the time
taken to fill the gas cell and lines), the gas exchange times must
be minimised. To this end, measurement instrumentation was
developed that incorporated a fast solenoid valve capable of
5 ms switching times. Details of this characterisation system has
been reported previously.36
The oxygen diffusion coefficient within the films was
determined in order to establish the origin of the observed
oxygen sensitivity. This was achieved by measuring the
response time for films of known thickness. Sensing layers
with a thickness of 1 mm were used as this was the minimum
thickness compatible with the detection efficiency of the
response time measurement apparatus. Diffusion coefficients
were then obtained during a step change in oxygen concentra-
tion, using the technique reported previously in ref. 36.
The excited-state lifetime of [Ru(dpp)3]2+ immobilised in
each ORMOSIL film was determined using a pulsed laser
system.37 The samples were excited with 15 ns pulses from a
Nd:Yag laser (l = 355 nm). Samples were degassed with N2
prior to measurements. The emitted decay was detected using a
photomultiplier tube and captured using a digital oscilloscope.
The excited-state luminescence lifetime was calculated by
analysing the decay trace using MicroCal Origin.
Water contact angle measurements were used to quantify
the relative hydrophobicity of the surface of ORMOSIL films.
These measurements were carried out using a static contact
angle analyser (FTA 200, USA).
The sensor film thickness was measured using a white light
interferometer (WYKO, NT1100, Veeco, USA).
A commercially available metabolic calibration unit
(VacuMed, USA) was employed to characterise the response
of the sensor film to changes in O2 concentration on a breath-
by-breath basis. This device is essentially a lung simulator,
which mimics O2 consumption through the dilution of an
inspired volume of ambient air with a mixture of carbon
dioxide in nitrogen (referred to as the calibration gas). This
apparatus is described in more detail in ref. 38.
For the work described, a module containing a sensor chip
coated with a PTEOS-based sensor film38 was attached to the
outlet of the lung simulator and its response to changes in O2
concentration was recorded.
With the exception of the sensor platforms used for response
time measurements, all other sensor film characterisation was
carried out using dip-coated films with a thickness in the range
of 400–500 nm.
Results and discussion
All MTEOS-, ETEOS-, PTEOS- and PhTEOS-based O2
sensor platforms exhibit excellent repeatability, reversibility
and short response times. Fig. 1 shows typical phase response
for all sensor elements tested in this study.
The key requirements in the successful implementation of a
sensor platform for in-breath O2 analysis are discussed in the
following sections. Table 1 is a compilation of all character-
isation data of the O2 sensor elements produced in this work.
Response time
Rapid sensor response is an important consideration in
designing a device for breath-by-breath monitoring. To
maximise the diagnostic possibilities and ensure the breath
monitor is suited to many medical applications, a range of
breath rates must be accommodated. A shorter response time
allows for the detection of fast breath rates, thereby increasing
the breathing range over which the sensor can be used. The
response time of a sensor membrane has been defined here as
t90, the time taken to achieve 90% of the optical signal change,
as the environment switches from vacuum to atmospheric
pressure. The sensor response time is dependent on the
diffusion path length of the O2 molecules to the luminophore,
which is directly related to film thickness. Sensing layers with a
thickness of 1 mm were used for the measurement of sensor
response time.
From Table 1 the shortest t90 was found for ETEOS- and
PTEOS-based sensors. The response times of these sensors
were 232 and 223 ms, respectively. The required response time
for this breath monitoring application is 50 ms and can be
obtained by reducing the sensor layer thickness to approxi-
mately 500 nm, meaning that the sensor films produced for all
other characterisation stages reported here were adequately
thin for this application.
This journal is � The Royal Society of Chemistry 2008 Analyst, 2008, 133, 241–247 | 243
The longest t90 was found for the PhTEOS-based sensor
platforms (t90 = 2687 ms) due to the relatively low value of D
for this film. This response time was an order of magnitude
greater than that of any of the other materials tested.
Accordingly, PhTEOS-based sensors were deemed to be
unsuited for this application and were not investigated further.
Influence of relative humidity
Fig. 2 shows the humidity dependence for MTEOS-, ETEOS-
and PTEOS-based O2-sensitive films. Clearly, the more
hydrophobic ETEOS- and PTEOS-based films exhibit negli-
gible sensitivity to humidity. This is consistent with the water
contact angle (CA) data in Table 1. The most hydrophobic
films in this work are ETEOS- and PTEOS-based sensor
elements with a water CA of 97 ¡ 2u and 100 ¡ 1u,respectively. As such, these coatings should be less prone to
humidity interference than, for example, MTEOS-based coat-
ings. These hydrophobic sensor films have a reduced sensitivity
to ambient moisture compared to hydrophilic films.
Since considerable humidity interference was observed for
the MTEOS-based sensor films, this material was consequently
eliminated as a possible sensor matrix for this application.
Sensitivity
It has been demonstrated previously that films with increased
hydrophobicity exhibit higher O2 sensitivity, both in gas and
dissolved phase.39 Table 1 shows that there is a good
correlation between the Stern–Volmer constant (KSV, O2
sensitivity) and both the measured D and water CA. The
parameters KSV, D and CA increase in magnitude as the alkyl
chain length increases. The sensitivity of the MTEOS-based
sensor element is less than that of either the ETEOS- or
Fig. 1 Phase response of MTEOS-, ETEOS-, PTEOS- and PhTEOS-based sensor platforms.
Table 1 Effect of organosilicon precursor on the characterisation data of the O2 sensor platforms
Organosilcon precursor t0 (accuracy ¡ 10%)/ms KSV (0–30% O2)/[O2 %]21 Dw (= w25 2 w15) D(61026)/cm2 s21 t90a/ms Contact angle/u
MTEOS 4.92 0.062 ¡ 0.003 2.7 9.86 432 ¡ 17 85 ¡ 1ETEOS 4.91 0.072 ¡ 0.002 2.4 62.1 232 ¡ 9 97 ¡ 2PTEOS 5.11 0.108 ¡ 0.003 2.6 67.3 223 ¡ 16 100 ¡ 1PhTEOS 5.81 0.023 ¡ 0.001 2.9 0.03 2687 ¡ 161 90 ¡ 1a Normalised response time, representative of a sensor layer thickness of 1 mm.
244 | Analyst, 2008, 133, 241–247 This journal is � The Royal Society of Chemistry 2008
PTEOS-based sensor elements, with the sensitivity of the
PTEOS-based sensor layer being twice that of the MTEOS-
based sample. As mentioned earlier, O2 sensitivity, represented
by KSV, is dependent on both t0 and D. From Table 1, t0 was
found to remain essentially constant for MTEOS-, ETEOS-
and PTEOS-based sensor materials. While all samples were
stored in a N2 environment, minor deviations in the lifetime
data are thought to be due to slight fluctuations in O2
concentration within the measuring environment due to leaks
in the sample cell. Therefore, we can attribute the enhanced
sensitivity to the increase of O2 transport within the ETEOS-
and PTEOS-based O2 films, as evidenced by the diffusion
coefficients recorded for these films. PhTEOS-based sensor
elements exhibit longer excited-state lifetime (ca. 18%) than
other materials tested but exhibit ca. 56 lower sensitivity than
the most sensitive PTEOS-based films, which is due to the
steric effect of the bulky phenyl group.
While KSV data are presented here over the full O2 range to
facilitate comparison with other O2 work in the literature, for
in-breath O2 analysis, the concentration range of interest is 16–
21%. Therefore, it was necessary to compare the sensitivity of
the films studied over this range. The parameter Dw, was used
as a relative sensitivity parameter to denote the change in
phase response as O2 concentration changed from 15 to 25%
(Dw = w25 2 w15). From the Dw data in Table 1, the most
sensitive film, in this range, was the PhTEOS-based sensor
element. However, this material had been eliminated as a
possible candidate for the application due to its unacceptably
long t90. Bearing this in mind, the most suitable matrix was
clearly PTEOS-based with a Dw value of 2.6 compared with 2.4
for an ETEOS-based sensor. Having identified the optimum
sensor matrix, all further analysis was limited to this matrix
only.
Fig. 3 presents the best fit to the Stern–Volmer and Demas
models for PTEOS-based sensor elements. The fit to the
Demas model is superior (r2 = 0.999). The Stern–Volmer plot
for PTEOS-derived sensor platforms are also very well
described by eqn (1) (r2PTEOS = 0.997), which would allow
for a simple one-point calibration of these sensors if desired.
Influence of temperature on O2 sensor response
Temperature effects on the response of the PTEOS-sensor
films have been tested. The phase response was recorded as
Fig. 2 Influence of humidity on O2 response for MTEOS-, ETEOS- and PTEOS-based sensor platforms. The data were recorded at 0% and 95%
RH.
This journal is � The Royal Society of Chemistry 2008 Analyst, 2008, 133, 241–247 | 245
described earlier, using gas at temperatures of 20, 30 and 40 uC.
This temperature range was selected as it reflects the working
range required of a breath monitor. A change in temperature
results in a change in t0 of the [Ru(dpp)3]2+ complex. The
diffusion coefficient of the sol–gel matrix is also temperature
dependent, with higher temperatures resulting in faster
diffusion, leading to an increase in collisional quenching. The
results in Fig. 4 show that sensor performance does not deviate
greatly and the sensors still operate sensitively within the tested
temperature range. However, as expected, temperature cali-
bration will be required for these sensor platforms.
Optimum material for O2-in-breath sensor platform
PTEOS-based sensor platforms were selected as the optimum
material for use in an O2-in-breath monitor, as they exhibited
sufficient O2 response in the relevant O2 concentration range,
from 15 to 25%, coupled with negligible humidity interference
and short response time. The error bars associated with the
films shown in Fig. 3 are very small (0.03u at 30% O2) and
demonstrate that film-to-film reproducibility is excellent. A
good fit to the linear Stern–Volmer model facilitates one-point
calibration procedures. The O2 response of such a PTEOS-
based sensor in the flow-cell-based instrumentation in this
laboratory yielded an LOD of 0.09% (of full scale) and a
resolution of 0.09% (in the range 10–20% O2). Fig. 5 presents
the phase data obtained using a PTEOS-based sensor element
to monitor the changing O2 levels delivered by the lung
simulator described previously. The simulator was configured
to deliver a 2% reduction in the expired O2 concentration
relative to ambient O2 levels at a breathing rate of 10 breaths
per minute. The ability of the sensor to respond to changes in
O2 concentrations on a breath-by-breath basis is clearly
demonstrated and a resolution of approximately 0.03% O2
was achieved in this concentration range.
While PTEOS has been selected here as the optimum
organosilicon precursor for the production of materials for an
O2-in-breath monitor, it is worth noting that the criteria for
that application are similar to those essential for O2-monitors
in bioprocessing applications. As such, PTEOS-based materi-
als are suited to a broad range of sensor solutions in many
industrial applications.
Conclusions
Oxygen-sensitive xerogel systems based on hydrophobic
ORMOSILs have been fabricated and tested. These sensor
platforms exhibit increased O2 sensitivity with increasing
length of alkyl group and there is a strong correlation between
O2 sensitivity and film hydrophobicity. The sensor response
has been measured using phase fluorometry. The humidity
sensitivity of the O2 sensor films was investigated and it was
shown that the influence of humidity can be virtually
eliminated by increasing film hydrophobicity. Temperature
interference has been examined and we conclude that, while
temperature correction must certainly be implemented for
Fig. 3 Best fit to the Stern–Volmer and Demas model for PTEOS-
based sensor platforms. The calibration is based on the performance of
four sensor films. The slope represents KSV, O2 sensitivity.
Fig. 4 Influence of the temperature of the gas on O2 response for
PTEOS-based sensor platforms. The data were recorded at gas
temperatures of 20, 30 and 40 uC.
Fig. 5 Response of PTEOS-based sensor element. DO2 = 2%,
10 breaths per minute.
246 | Analyst, 2008, 133, 241–247 This journal is � The Royal Society of Chemistry 2008
ORMOSIL films, it is clear that these sensor films operate to a
sufficiently high sensitivity within the working temperature
range required of a breath monitor. Overall, we demonstrated
that PTEOS-based O2 sensor platforms exhibit characteristics
such as: excellent reproducibility, short response time,
humidity insensitivity, enhanced O2 sensitivity and compat-
ibility with single-point calibration procedures, making them
an ideal choice for breath monitoring applications. Future
work will address the evaluation of sensor performance in
clinical trials.
Acknowledgements
The authors wish to acknowledge Dr Mary Pryce, School of
Chemical Sciences, Dublin City University for assistance with
lifetime measurements.
References
1 H. Albrecht, in Sensors: A Comprehensive Survey, ed. W. Gopel,J. Hesse and J. N. Zemel, VCH, Weinheim and New York, 1992,vol. 3, pp. 1047–1093.
2 H. Lord, Y. Yu, A. Segal and J. Pawliszyn, Anal. Chem., 2002, 74,5650.
3 O. S. Wolfbeis, Fiber Optic Chemical Sensors and Biosensors, CRCPress, Boca Raton, FL, 1991, vol. 1.
4 C. J. Brinker and G. W. Scherer, Sol–Gel Science: The Physics andChemistry of Sol–Gel Processing, Academic Press, San Diego andLondon, 1990.
5 B. E. Yoldas, J. Sol–Gel Sci. Technol., 1998, 13, 147.6 E. J. Cho and F. V. Bright, Anal. Chem., 2002, 74, 1462.7 E. R. Carraway, J. N. Demas, B. A. DeGraff and J. R. Bacon,
Anal. Chem., 1991, 63, 337.8 J. N. Demas, B. A. DeGraff and W. Xu, Anal. Chem., 1995, 67,
1377.9 B. D. MacCraith, C. M. McDonagh, G. O’Keeffe, E. T. Keyes,
J. G. Vos, B. O’Kelly and J. F. McGilp, Analyst, 1993, 118, 385.10 A. K. McEvoy, C. M. McDonagh and B. D. MacCraith, Analyst,
1996, 121, 785.11 O. S. Wolfbeis, I. Klimant, T. Werner, C. Huber, U. Kosch,
C. Krause, G. Neurauter and A. Durkop, Sens. Actuators, B, 1998,51, 1.
12 C. Malins, M. Niggemann and B. D. MacCraith, Meas. Sci.Technol., 2000, 11, 1105.
13 H. S. Vorberger, H. Kreimaier, K. Biebernik and W. Kern, Sens.Actuators, B, 2001, 74, 179.
14 A. N. Watkins, B. R. Wenner, J. D. Jordan, W. Y. Xu, J. N. Demasand F. V. Bright, Appl. Spectrosc., 1998, 52, 750.
15 Y. Tang, E. C. Tehan, Z. Tao and F. V. Bright, Anal. Chem., 2003,75, 2407.
16 Z. Tao, E. C. Tehan, Y. Tang and F. V. Bright, Anal. Chem., 2006,78, 1939.
17 J. R. Lakowicz, Principles of Fluorescence Spectroscopy, KluwerAcademic/Plenum Publishers, New York, 1999.
18 J. R. Bacon and J. N. Demas, Anal. Chem., 1987, 59, 2780.19 I. Klimant and O. S. Wolfbeis, Anal. Chem., 1995, 67, 3160.20 D. Wencel, C. Higgins, A. Klukowska, B. D. MacCraith and
C. McDonagh, Mater. Sci., 2007, 25, 767.21 M. T. Murtagh and M. R. Shahriari, Chem. Mater., 1998, 10, 3862.22 P. Hartmann and M. J. P. Leiner, Anal. Chem., 1995, 67, 88.23 W. Xu, R. C. McDonough, B. Langsdorf, J. N. Demas and
B. A. DeGraff, Anal. Chem., 1994, 66, 4133.24 A. Mills, Sens. Actuators, B, 1998, 51, 69.25 J. N. Demas and B. A. DeGraff, Sens. Actuators, B, 1993, 11, 35.26 G. O’Keeffe, B. D. MacCraith, A. K. McEvoy, C. M. McDonagh
and J. F. McGilp, Sens. Actuators, B, 1995, 29, 226.27 Q. Chang, J. R. Lakowicz and G. Rao, Analyst, 1997, 122, 173.28 O. Lev, M. Tsionsky, L. Rabinovich, V. Glezer, S. Sampath,
I. Pankratov and J. Gun, Anal. Chem., 1995, 67, 22A.29 M. M. Collinson, Microchim. Acta, 1998, 129, 145.30 P. Lavin, C. M. McDonagh and B. D. MacCraith, J. Sol–Gel Sci.
Technol., 1998, 13, 641.31 D. Anvir, Acc. Chem. Res., 1995, 28, 328.32 M. M. Collinson and A. R. Howells, Anal. Chem., 2000, 72, 702A.33 R. J. Watts and G. A. Crosby, J. Am. Chem. Soc., 1971, 93, 3184.34 C. McDonagh, C. Kolle, A. K. McEvoy, D. L. Dowling,
A. A. Cafolla, S. J. Cullen and B. D. MacCraith, Sens.Actuators, B, 2001, 74, 124.
35 D. Wencel, C. Higgins, A. Guckian, C. McDonagh andB. D. MacCraith, in Opto-Ireland 2005: Optical Sensing andSpectroscopy, ed. H. J. Byrne, E. Lewis, B. D. MacCraith,E. McGlynn, J. A. McLaughlin, G. D. O’Sullivan, A. G. Ryderand J. E. Walsh, Proceedings of SPIE, vol. 5826, SPIE Press, 2005,p. 696 (http://adsabs.harvard.edu/abs/2005SPIE.5826..696W).
36 C. McDonagh, P. Bowe, K. Mongey and B. D. MacCraith, J. Non-Cryst. Solids, 2002, 306, 138.
37 B. S. Creaven, M. W. George, A. G. Ginzburg, C. Hughes,J. M. Kelly, C. Long, I. M. McGrath and M. T. Pryce,Organometallics, 1993, 12, 3127.
38 C. S. Burke, J. P. Moore, D. Wencel, A. K. McEvoy andB. D. MacCraith, J. Biomed. Opt., accepted September 2007.
39 C. McDonagh, B. D. MacCraith and A. K. McEvoy, Anal. Chem.,1998, 70, 45.
This journal is � The Royal Society of Chemistry 2008 Analyst, 2008, 133, 241–247 | 247