Journal of Colloid and Interface Science 386 (2012) 51–59

Contents lists available at SciVerse ScienceDirect

Journal of Colloid and Interface Science

www.elsevier .com/locate / jc is

Hematite nanoparticle monolayers on mica preparation by controlled self-assembly

Magdalena Ocwieja ⇑, Zbigniew Adamczyk, Maria Morga, El _zbieta Bielanska, Adam WegrzynowiczJerzy Haber Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, Niezapominajek 8, 30-239 Cracow, Poland

a r t i c l e i n f o a b s t r a c t

Article history:Received 11 April 2012Accepted 17 June 2012Available online 2 August 2012

Keywords:Deposition of hematite particles on micaHematite nanoparticle filmsKinetics of hematite particle depositionMonolayers of hematite nanoparticles onmicaSelf-assembly of hematite particles

0021-9797/$ - see front matter � 2012 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.jcis.2012.06.056

⇑ Corresponding author. Fax: +48 124251923.E-mail addresses: ncocwiej@cyf-kr.edu.pl (M. Ocwi

(Z. Adamczyk), ncmorga@cyf-kr.edu.pl (M. Mor(E. Bielanska), vinnicki@gmail.com (A. Wegrzynowicz

A stable suspension of a-Fe2O3 (hematite) was synthesized according to the method of Matijevic andScheiner by an acidic hydrolysis of ferric chloride. The average size of the particles was determined bydynamic light scattering (DLS) and atomic force microscopy (AFM) and was 22 nm. The electrophoreticmobility and zeta potential of particles were determined as a function of ionic strength and pH. The zetapotential of the hematite particles was positive for pH < 8.9 (isoelectric point) and negative otherwise.Using the suspension, systematic studies of particle deposition kinetics on mica were carried out. Thecoverage of self-assembled particle monolayers was determined by AFM and SEM imaging. Particle depo-sition was diffusion controlled, with the initial rate proportional to the bulk concentration of particles. Onthe other hand, for long times, the saturation coverage was attained, increasing systematically with ionicstrength. The deposition kinetic runs were adequately reflected by the random sequential adsorption(RSA) model. Additionally, particle desorption kinetics, from previously formed monolayers, were studiedusing the AFM and SEM methods. It was confirmed that hematite particle desorption was practically neg-ligible within the time period of 60 h. Our experimental data proved, therefore, that it is feasible to pro-duce uniform and stable hematite particle monolayers of desired coverage in self-assembly processescontrolled by the bulk suspension concentration and the ionic strength.

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

Iron oxide nanoparticles are of great importance for advancedtechnologies and various industries. Hematite, (a-Fe2O3), isthermodynamically the most stable of iron oxides, showing weakferromagnetic properties at room temperatures [1]. It is a semicon-ducting material with a relatively narrow band gap of 2.0–2.2 eV[2]. Moreover, it can be easily doped by divalent or tetravalentadditives, for example, niobium, which converts hematite into anattractive material used in photovoltaic cells (water photo-oxida-tion and photo-electrochemical hydrogen production [3]).

Hematite is applied in industry in the form of suspensions asred pigment and anti-corrosion agent. It also has significance forcatalysis in the Haber process [1,4], the Fisher–Tropsch synthesis[5], and the desulfurization of natural gas [1].

Most practical applications of hematite involve thin filmsdeposited on various conducting surfaces, for example, thosewhich serve as gas, alcohol or humidity sensors [6,7], electrodesin lithium batteries [6], photo-anodes, etc.

Usually these hematite films are produced via physical orchemical methods such as sputtering [8], laser ablation [9],

ll rights reserved.

eja), ncadamcz@cyf-kr.edu.plga), ncbielan@cyf-kr.edu.pl).

electrodeposition [10], spray pyrolysis (SP) [11,12], ion beam in-duced chemical vapor deposition (IBICVD) [13], plasma enhancedvapor deposition (PECVD) [14], and the aerosol-assisted chemicalvapor deposition (AACVD) [15]. However, these ‘‘dry’’ methods re-quire sophisticated and expensive apparatus. The process of filmsformation is tedious, and the coverage degree and film structurecannot be regulated in a systematic manner.

More convenient and flexible methods of hematite film produc-tion based on controlled self-assembly of nanoparticles from aque-ous phase hematite suspensions (sols). Obviously, the first step offilm preparation involves a synthesis of stable hematite sols ofwell-defined surface properties and appropriate particle size.

The most common method of preparing hematite nanoparticlesuspensions is the acid hydrolysis of iron salts [16–18], whichwas recently improved by introducing the gel–sol procedure de-scribed by Sugimoto et al. [19–22]. Other methods consist of thecalcinations of iron oxyhydroxide (b-FeOOH) [23,24] and the ionicliquid-assisted synthesis [25]. A major advantage of aqueoushematite suspensions is that the particles exhibit a positive charge(zeta potential) for pH < 9 [26], which facilitates their depositionon solid substrates promoted by electrostatic attraction.

Therefore, such aqueous phase hematite suspensions can beefficiently used for preparing hematite coatings and films on solidsupports. However, despite the major significance of wet methods,there are few systematic works devoted to this subject. As one ofthe few exceptions should be mentioned the work of Björkstèn

52 M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59

et al. [27] who described preparation of nanocrystalline hematitefilms directly from a colloidal a-Fe2O3 suspension for photoelectro-chemical studies. They used conducting glass, which was preparedby spray pyrolysis, and then, the hematite suspension was depos-ited on this substrate using a Pasteur pipette. Afterward, in order toobtain stable hematite film, the layer was dried for prolonged time.

Wang et al. [28] described the preparation of thin, uniformhematite monolayers from a suspension of iron oxide nanoparti-cles, using the Langmuir–Blodgett technique. Based on earlierworks [29,30], where this technique was applied to form a-Fe2O3

films, the authors modified the surface of hematite nanoparticlesusing three different procedures such as a solvent exchange pro-cess, a surfactant-assisted phase transfer process, and a thermalevaporation process.

Kuhnen et al. [31] studied deposition of large (120 nm) hema-tite nanoparticles on quartz sand media under flow rate. Hematitenanoparticles deposition dynamics was investigated by conductinga series of column transport experiments at different conditions.Authors show how the maximum surface concentration of parti-cles depends on the ionic strength, colloid concentrations, and flowvelocities. Experimental results were successfully interpreted interms of the random sequential adsorption (RSA) model.

In view of the deficit of experimental results, the goal of our pa-per is developing a reliable and efficient method of producinghematite monolayers by a colloid self-assembly carried out fromaqueous suspensions under diffusion transport conditions. Thecoverage and structure of hematite monolayers is systematicallycontrolled by the deposition time, the suspension concentration,pH, and ionic strength variations. An advantage of our method,as compared to the Wang’s work [28] and our previous studies de-voted to silver particle self-assembly [32], is that no modificationof particles or substrate surfaces are required prior to particledeposition.

Besides practical aspects, our measurements possess signifi-cance for basic colloid science, enabling one to test the validity oftheoretical approaches for predicting electrostatically driven depo-sition of nanoparticles.

2. Experimental

2.1. Materials

All materials were analytic reagents used without further puri-fication. The ferric chloride (FeCl3�6H2O), hydrochloric acid (HCl),sodium chloride (NaCl), and sodium hydroxide (NaOH) were pur-chased from POCH.

Natural ruby mica sheets supplied by Continental Trade Ltd.,Poland, were used as substrate surfaces for nanoparticle deposi-tion. Thin sheets were freshly cleaved and used in each experimentwithout any pretreatment.

Ultrapure water, used throughout this investigation, was ob-tained using the Milli-Q Elix & Simplicity 185 purification systemfrom Millipore SA Molsheim, France.

2.2. Synthesis of hematite (a-Fe2O3) nanoparticles

The hematite nanoparticles were synthesized according to themodified Matijevic method [16] based on forced hydrolysis of fer-ric chloride in an acid solution. A stock solution of ferric chloride(0.25 M) was first prepared by dissolving FeCl3�6H2O in 0.004 MHCl. Before starting the synthesis, the solution was purifiedthrough 0.22 lm Millipore filters to remove any particular contam-inations. Then, while stirring, 12 ml of this solution was rapidlyadded into a boiling solution of 0.004 M HCl to yield a final Fe3+

concentration of 5 mM. The reaction mixture was kept at boiling

temperature for 20 min and then cooled to room temperature.The colloid iron oxide suspension was purified from ionic excessusing a stirred membrane filtration cell with a regeneratedcellulose membrane (Millipore, NMWL: 100KDa). The washingprocedure was repeated until the electric conductivity of the sus-pension reached a minimum value of c.a. 15 lS/cm.

2.3. Methods

Characterization of the hematite suspension and the monolay-ers on mica was performed using various techniques.

The weight concentration of the particle suspension was deter-mined using the high precision densitometer: DMA 5000M (AntonPaar).

The crystal structure and purity of a dried hematite sample wasdetermined by X-ray photoelectron spectroscopy (XRD), whichwas performed using a X’PERT PRO diffractometer (Cu Ka, 40 kV,30 mA), working in the Bragg–Brentano geometry, fixed slits,X’CELERATOR detector.

The size (hydrodynamic diameter) and the zeta potential of thenanoparticles were determined by dynamic light scattering (DLS)and microelectrophoresis, respectively, using the Zetasizer NanoZS Malvern instrument. The measurement range was 3 nm to10 lm for zeta potential and 0.6 nm to 6 lm for particle size. Addi-tionally, the zeta potential of particles was measured using theBrookhaven Zeta Pals apparatus.

The size and the morphology of nanoparticles were also inves-tigated using the JEOL JSM-7500F microscope working in transmis-sion mode. Samples for this examination were prepared bydispersing a drop of the hematite suspension on a copper grid cov-ered by a carbon film.

The surface concentration of hematite monolayers on the micasubstrate was quantitatively determined using the atomic forcemicroscopy (AFM) and the scanning electron microscope (SEM).AFM measurements were carried out using the NT-MDT SolverPro instrument with the SMENA SFC050L scanning head. Theimaging was done in semicontact mode using composite probespossessing a silicon body, polysillicon levers, and silicon high res-olution tips. Independently, SEM measurements of iron oxidemonolayers were carried out using the JEOL JSM-7500F FieldEmission microscope at 15 kV. To ensure a sufficient conductivityduring the measurement, the hematite nanoparticles were treatedwith a thin layer of chromium before the examination. The num-ber of particles per unit area of the substrate was determinedfrom these AFM pictures and SEM micrographs using the imageanalysis software.

Zeta potential of mica was determined via streaming potentialmeasurements using a home-made cell previously described[33,34]. The main part of the cell was a parallel plate channel, hav-ing dimensions of 2bc � 2cc � L = 0.027 � 0.29 � 6.2 cm, formed bymica sheets separated by a perfluoroethylene spacer. The stream-ing potential Es was measured using a pair of Ag/AgCl electrodesas a function of the hydrostatic pressure difference DP, whichwas driving the electrolyte flow through the channel. The cell elec-tric conductivity Ke was determined using Pt electrodes. Knowingthe slope of the Es vs. DP dependence, the apparent zeta potentialof substrate surface (fi) can be calculated from the Smoluchowskirelationship

fi ¼gL

4ebcccRe

DEs

DP

� �¼ gKe

eDEs

DP

� �ð1Þ

where g is the dynamic viscosity of the solution, e is the dielec-tric permittivity, and Re is the electric resistance of the cell gov-erned mainly by the specific conductivity of the electrolyte inthe cell.

Table 1The Bragg reflections peaks and the assigned latticeplanes (hkl).

Peak position 2H (�) hkl

23.9 01233.2 10435.5 11039.8 00649.6 02453.9 11662.8 21464.1 30071.9 101075.6 220

Cou

nt10

20

30

40

M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59 53

3. Results and discussion

3.1. Hematite nanoparticles characterization

The chemical composition and crystallinity degree of hematitenanoparticles were examined using powdered XRD measurements.The hematite suspension was dried at 100 �C for a prolonged timeperiod, and the obtained powder was examined using a X’PERTPRO diffractometer, working in the Bragg–Brentano geometry,fixed slits, X’CELERATOR detector. The measurement parameterswere 40 kV, 30 mA.

The XRD spectrum of the powder is shown in Fig. 1. All of theobserved peaks can be attributed to the rhombohedral structureof a-Fe2O3 (space group R3c). The narrow sharp peaks suggest thatthe obtained product is highly crystalline. No characteristic peaksindicating impurities like c-Fe2O3 and other inorganic ions wereobserved.

A list of Bragg reflections and assigned lattice planes (hkl) wasshown in Table 1.

The hematite weight concentration in the stock suspensionafter the cleaning procedure was determined by the densitometerproviding relatively precise of measurements 5 � 10�6 g cm�3. Thedensity of the stock hematite suspension qs and the supernatantsolution qe acquired by membrane filtration were measured usingthis device. Then, the weight fraction w of colloid hematite in thesol was determined from the formula

0

w ¼

qpðqs � qeÞqsðqp � qeÞ

ð2Þ

dH [nm]0 10 20 30 40 50

Fig. 2. The histogram of hematite particle size distribution determined from theAFM image, scan size 2 � 2 lm, see the inset. The average diameter of particles is22 ± 5 nm.

where qp = 5.26 g cm�3 is the specific density of hematite particles[1].

It was determined in these experiments that the weight fractionof hematite in the stock suspension, after the cleaning procedure,was 2.8 � 10�3, which corresponds to 2800 ppm (parts per millionby weight). The well-characterized stock suspension was dilutedprior to each experiment to a desired concentration, usually vary-ing between 10 and 20 ppm in the deposition experiments and 50–100 ppm in the DLS and microelectrophoretic measurements.

The size distribution of hematite nanoparticles was determinedfrom AFM images (a typical example of hematite monolayer isshown in Fig 2). The particle diameter were determined from theirheight with respect to the surface, the values were calculated withthe use of Nova 1152 software that is coupled directly with AFMmicroscope. A few hundred particles were counted in this way to

2θ

20 30 40 50 60 70 80

Cou

nt

0

50

100

150

200

250

300

Fig. 1. The XRD pattern of the a-Fe2O3 (hematite) nanoparticles sample.

obtain a histogram also shown in Fig. 2. The mean diameter of par-ticles was 22 nm with the standard deviation of 5 nm.

Hematite nanoparticle size was also determined via diffusioncoefficient (D) measurements performed using the DLS technique.Knowing the diffusion coefficient, one can determine the hydrody-namic diameter of particles using the Stokes–Einstein relationship:

dH ¼kT

3pgDð3Þ

where dH is the hydrodynamic diameter, k is the Boltzman constant,and T is the absolute temperature.

The hydrodynamic diameter can be interpreted as the size of anequivalent sphere having the same hydrodynamic resistance coef-ficient as the particle. The advantage of using this quantity in com-parison with the diffusion coefficient is that it is independent oftemperature and liquid viscosity, so it is an appropriate parameterfor analyzing suspension stability under various conditions.

The dependence of the hydrodynamic diameter of hematite par-ticles on ionic strength varied between 3 � 10�5–10�1 M andpH = 5.5 is shown in Fig. 3a. As can be seen, for ionic strength upto 3 � 10�2 M, the hydrodynamic diameter was practically con-stant, attaining an average value of 22 nm with a standard devia-tion of 5 nm, that is, identical to that previously determined byAFM. A small increase in the hydrodynamic diameter was only ob-served for ionic strength approaching 0.1 M.

In Fig. 3b, the dependence of dH on pH regulated by the additionof an appropriate amount of HCl or NaOH for a fixed ionic strengthof 10�2 M is shown. A considerable increase in the hydrodynamicdiameter of hematite particles was observed for pH above 8,

t [min]0 300 600 900 1200 1500 1800 18000 19000 20000 21000

d H [n

m]

0

10

20

30

40

Fig. 4. The dependence of the hydrodynamic diameter of hematite nanoparticles onthe time, determined by DLS for pH = 3.5 I = 3 � 10�2 M and bulk concentration40 ppm.

I [M]

10-5 10-4 10-3 10-2 10-1

d H [

nm]

0

10

20

30

40

pH2 4 6 8 10 12

0

10

20

30

40

50

60

70

80

90

100

200

250

d H [

nm]

(a)

(b)

Fig. 3. (a) The dependence of the hydrodynamic diameter of hematite nanoparticleson ionic strength, determined by DLS for pH = 5.5, the dashed line represents aconstant value of the hydrodynamic diameter of 22 nm. (b) The dependence of thehydrodynamic diameter of hematite nanoparticles on pH determined by DLS forI = 10�2 M, (cb = 40 ppm); the solid line represents non-linear fits of experimentaldata.

54 M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59

suggesting that the suspension became unstable. However, forpH > 10, the hydrodynamic diameter again attained the value of22 nm, pertinent to a stable hematite suspension. This significantgrowth of hydrodynamic diameter observed for 8 < pH < 10 sug-gests that the hematite particles aggregated within this pH range.

It is interesting to mention that analogous results were ob-tained by Schudel et al. [35] for two hematite samples having anaverage particle size of 69–84 nm (determined by time-resolvedDLS). However, the range of pH above 9 was not studied it hiswork.

Additionally, the stability of the hematite suspension upon pro-longed storage times was checked. The results shown in Fig. 4(pH = 3.5, I = 3 � 10�2 M and the bulk suspension concentrationof 40 ppm) indicate that the suspension was stable up to the timeof 300 h, even at such a high ionic strength. Obviously, the sametrend was observed for lower ionic strength if the pH of the sus-pension was kept below 8.

To fully characterize the hematite suspension, the electropho-retic mobility of particles le was also determined as describedabove. It is a quantity of major significance for characterizing par-ticle stability and interactions with surfaces leading to deposition.Knowing the electrophoretic mobility one can calculate the zetapotential of particles using the Henry–Smoluchowski formula

fp ¼3g

2ef ðjdpÞle ð4Þ

where fp is the zeta potential of particles, f(jdp) is the correctionfunction of the dimensionless parameter, jdp, j�1 ¼ ekT

2e2 I

� �1=2is the

thickness of the electric double layer, e is the elementary charge,I ¼ 1

2

Piciz2

i is the ionic strength, ci is the ion concentration, zi isthe ion valency, and dp is the characteristic dimension of the parti-cle, for example, the hydrodynamic diameter dH.

The dependencies of electrophoretic mobility and zeta poten-tial of hematite nanoparticles on the pH, determined for a fixedionic strength of 10�3 M and 10�2 M, are shown in Fig. 5. As canbe seen, the mobility is highly positive for the low pH range,attaining the value of 4.01 lm cm (V s)�1 for pH = 3.5(I = 10�3 M) and 3.43 lm cm (V s)�1 for pH = 3.5 (I = 10�2M). Thiscorresponds to the zeta potential of 47 and 38 mV, respectively.For higher pH values, the electrophoretic mobility and the zetapotential of hematite decrease monotonically and vanish atpH = 8.8–9 (for I = 10�3 M and 10�2 M, respectively). For higherpH, both values become negative. Thus, the zeta potential at-tained the value of �47 mV for I = 10�3 M and �42 mV forI = 10�2 M (pH = 10.7). Our data indicate that hematite particlesexhibit a clearly defined isoelectric point (i.e.p.) at pH 8.9. This isthe value at which the electrophoretic mobility of the particlesvanishes. It should be mentioned that our result agrees with thatreported by Zhang and Buffle [26] (9.2) and Schudel et al. [35](8.8–9.2). However, other values were also reported. For exam-ple, He and coworkers [36] reported pH = 7.8–8.8 as the isoelec-tric point of hematite nanoparticles of various sizes. Even a lowervalue of pH = 7–7.5 was reported by Delgado and Gonzalez-Cabá-llero [37] and Bentaleb et al. [38]. Variations in the isoelectricpoint value for various hematite sols were also observed by Mat-ijevic and Scheiner [16].

It is interesting to mention that the electrophoretic mobilitydata (Fig. 5) correlate well with the hematite suspension stabilityrange above shown in Fig. 3, which suggest that the stability ismainly governed by electrostatic interactions.

Additionally, in a separate series of experiments, the depen-dence of the particle zeta potential on the ionic strength variedwithin 3 � 10�5 and 3 � 10�2 M for a fixed pH = 5.5 was deter-mined (Fig. 6). As can be seen, the zeta potential of hematite par-ticles decreased slightly with ionic strength, from 52 mV forI = 3 � 10–5 M to 38 mV for I = 3 � 10�2 M.

From these measurements, one can conclude that the hematiteparticles exhibit a high positive zeta potential for a broad range ofpH and ionic strength, which is expected to promote their efficientdeposition on negatively charged substrates.

pH3 4 5 6 7 8 9 10 11

ζ [m

V]

-140

-120

-100

-80

-60

-40

-20

0

20

Fig. 7. The dependence of the zeta potential of mica determined by streamingpotential measurements on pH. The points denote experimental results obtainedfor: (1) (s) I = 3 � 10�2 M, (2) (h) I = 10�2 M, (3) (}) I = 10�3 M.

-5

-4

-3

-2

-1

0

1

2

3

4

5

pH2 3 4 5 6 7 8 9 10 11 12

ζ [m

V]

-60

-40

-20

0

20

40

60

μ e [(μ

m c

m)(V

s)-1

]

-5

-4

-3

-2

-1

0

1

2

3

4

5

μ e [(μ

m c

m)(V

s)-1

]

pH2 3 4 5 6 7 8 9 10 11 12

ζ [m

V]

-60

-40

-20

0

20

40

60

(b)

(a)

Fig. 5. The dependence of the electrophoretic mobility and zeta potential ofhematite particles on pH. (a) I = 10�3 M, (b) I = 10�2 M.

ζ [m

V]

34

36

38

40

42

44

46

48

50

52

54

I [M]10-5 10-4 10-3 10-2 10-1

2.6

2.8

3.0

3.2

3.4

3.6

3.8

4.0

μ e [(μ

m c

m)(V

s)-1

]

Fig. 6. The dependence of the electrophoretic mobility and zeta potential ofhematite particles on ionic strength for pH = 5.5.

M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59 55

3.2. Mica substrate characterization

In our hematite particle deposition experiments, mica sheetswere used as a well-defined substrate. There are many advantagesconnected with using mica because it is a chemically stable andmolecularly smooth material, characterized by a uniform andhomogeneous surface charge distribution [39].

Zeta potential of freshly cleaved mica sheets was determined bythe streaming potential method according to the proceduredescribed in previous works [33,40]. The dependence of the zeta

potential of mica on pH (adjusted with the addition of HCl orNaOH), for three different ionic strengths (10�3 M, 10�2 M and3 � 10�2 M), is shown in Fig. 7. As can be noticed, the zeta potentialwas negative for the entire range of pH studied (3–11) for both io-nic strength values. For pH = 3.5, it was �20 mV and �40 mV, forI = 3 � 10�2 and I = 10�2 M, respectively. The zeta potential de-creased monotonically with pH, attaining �45 mV and �70 mV,for I = 3 � 10�2 and 10�2 M, respectively (pH = 7.4).

3.3. Kinetics of hematite particle deposition

After a thorough characterization of the system, a series of mea-surements were performed with the aim of quantitatively deter-mining the kinetics of hematite particle deposition.

In the first stage of these measurements, freshly cleaved micasheets were immersed in a thermostated diffusion cell containinga hematite suspension of a desired concentration. The diffusion cellwas a cylindrical glass vessel with a capacity of 6 ml placed inthermostated chamber. The vessel was equipped with a glass ele-ment that allows for vertical immersion of mica sheets in thehematite suspension. Particle deposition was allowed to proceedover a prescribed time under diffusion-controlled transport. Next,the sheets were rinsed with distilled water and air dried. The sur-face concentration of deposited particles was determined by a di-rect enumeration carried out using AFM and SEM imaging. Thenumber of particles deposited on 8–10 equal-sized surface areas(typically having the dimensions of 2 lm � 2 lm or2.3 lm � 1.7 lm for AFM and SEM, respectively) was determined.The overall number of particles counted was larger than 1000,which ensured a relative precision of these measurements betterthan 5%. Knowing the average number of particles and the surfacearea DS, the particle coverage was calculated as the number of par-ticles per square micrometer, denoted hereafter by NS.

Typical AFM and SEM images of hematite particle monolayersdeposited from a suspension of the concentration equal to20 ppm (I = 10�2 M and pH = 3.5) are shown in Figs. 8 and 9,respectively. As can be noticed, NS increased monotonically withdeposition time.

Knowing NS as a function of deposition time, particle deposi-tion kinetic for various experimental conditions can be deter-mined. Typical examples of such kinetic runs obtained from thebulk suspension concentration of 10 and 20 ppm (I = 10�2 M NaCl,pH = 3.5) are shown in Fig. 10. Note that the square root of depo-sition time t1/2 was used as an independent variable rather than

Fig. 9. SEM images of hematite nanoparticles monolayers on mica obtained fromcolloidal hematite suspension: 20 ppm, I = 10�2 M NaCl, pH = 3.5, (a) particlesurface concentration Ns = 163 lm�2, (b) particle surface concentrationNs = 765 lm�2.

Fig. 8. AFM images (scan size 2 lm � 2 lm) of hematite particles monolayers onmica. Deposition conditions: 20 ppm, I = 10�2 M NaCl, pH = 3.5, (a) particle surfaceconcentration Ns = 175 lm�2, (b) particle surface concentration Ns = 258 lm�2.

t1/2 [min1/2]

0 5 10 15 20 25 30

NS

[ μm

-2]

0

100

200

300

400

500

Fig. 10. The kinetics of hematite particle deposition on mica determined by AFM(full points) and SEM (hollow points) on the square root of adsorption time t1/2.Deposition conditions: pH = 3.5, I = 3 � 10�2 M, T = 293 K, (1) (d,s) cb = 20 ppm, (2)(j,h) cb = 10 ppm. The lines denote linear fits of experimental results.

56 M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59

the deposition time. As shown in our previous work [32], this isan appropriate transformation to analyze diffusion-controlleddeposition processes.

As can be seen in Fig. 10, for adsorption time t1/2 < 30 min1/2

(t < 900 min), particle deposition was described by a linear func-tion of t1/2, with the slope two times larger for cb = 20 ppm andfor 10 ppm. This behavior was in accordance with an analytical for-mula, pertinent to bulk controlled diffusion transport [41,42]

NS ¼ 2Dp

� �1=2

t1=2nb ð5Þ

where nb = cb/mp is the number concentration of particles in thebulk, mp ¼ p

6 d3mqp is the effective mass of one particle.

An agreement of our results with Eq. (5) proves that the particledeposition process was solely controlled by diffusion and the sur-face blocking effects were negligible. As a result, Eq. (5) can be usedfor selecting the appropriate bulk concentration and depositiontime to obtain hematite monolayers of desired coverage. Moreover,the results shown in Fig. 10 can be used to determine the

‘‘diffusion’’ diameter of hematite particles, which is given by thepreviously derived expression [32]

M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59 57

dp ¼12

qpp2sD

!2=7kT3g

� �1=7

ð6Þ

where sD is the slope of the empirical dependence of NS/cb on t1/2.From the results shown in Fig. 10, one can calculate that the aver-

age slope sD = 1.34 ppm�1 lm�2 min�1/2 = 1.42 � 1013 cm g�1 s�1/2.Substituting this value in Eq. (6) and T = 293 K, g = 0.01 g (cm s)�1,qp = 5.26 g cm�3, one obtains from dp = 21.6 nm. This is in agree-ment with the hydrodynamic diameter determined by DLS andthe average particle size determined by AFM.

However, as demonstrated in previous work concerning silverparticle deposition [43] for longer times, the surface blocking ef-fects start to play a significant role reducing particle depositionrates and the maximum concentration of deposited particles. In or-der to study this effect, long-lasting deposition experiments forhematite nanoparticles were also performed.

In Fig. 11 results obtained for cb = 20 ppm, pH = 3.5 and variousionic strength are shown. The surface concentration of particleswas determined by both AFM (full points) and SEM (hollowpoints). Characteristic features of these kinetic runs are a linear in-crease in NS with t1/2 for shorter times and then, after reachingsome critical time, an abrupt stabilization of the surface concentra-tion at a constant value. The latter is referred to as the maximumsurface concentration of particles, denoted by NSmx . It is interestingto observe that the maximum coverage of hematite particles in-creased significantly with the ionic strength assuming 469 lm�2

for I = 10�3 M, 853 lm�2 for I = 10�2 M and 1033 lm�2 forI = 3 � 10�2 M. It is convenient, as is commonly done, to expressthese maximum concentration values in terms of the dimension-less coverage defined as

Hmx ¼ pd2pNSmx=4 ð7Þ

Taking dp = 22 nm, as determined above, one obtains Hmx = 0.18for I = 10�3 M, Hmx = 0.33 for I = 10�2 M and Hmx = 0.41 forI = 3 � 10�2 M. Hence, deposition of hematite nanoparticles at anionic strength of 3 � 10�2 M creates the most favorable conditionsfor producing uniform monolayers of the highest coverage. How-ever, at higher ionic strengths, the hematite suspension becomeless stable (see Fig. 3a); therefore, it is impractical to attain surfacecoverage higher than Hmx = 0.41.

Using the dimensionless coverage also facilitates a comparisonof experimental results with theoretical models developed to

Fig. 11. The kinetics of hematite particle deposition at mica determined for variousionic strengths using AFM (full points) and SEM (hollow points). Particle depositionconditions: pH = 3.5, T = 293 K, cb = 20 ppm. The points denote experimental resultsobtained for: (1) (d,s), I = 3 � 10�2 M, (2) (j,h) I = 10�2 M, (3) (�,}) I = 10�3 M, (4)(.) I = 10�4 M. The solid lines denote the theoretical results calculated from the RSAmodel.

quantitatively describe nanoparticle deposition processes [42,44].In this case, the most convenient model is the random sequentialadsorption (RSA) model, successfully applied before for describingirreversible adsorption (deposition) of colloid microparticles (poly-styrene latexes) [45,46], nanoparticles [32,47–51], and proteins[40,52]. The method is also extensively discussed in the book[42] and recent reviews [39,53–55].

According to the generalized RSA model, the constitutive equa-tion describing the adsorption flux of particles to interfaces for anarbitrary adsorption mechanism is given by

ja ¼1Sg

dhdt¼ kanðdaÞBðhÞ �

kd

Sgh ð8Þ

where ja is the adsorption flux, Sg is the characteristic cross-sectionarea of the particle, t is the adsorption time, ka and kd are theadsorption and desorption kinetic constants, n(da) is the concentra-tion of particles at the adsorption boundary layer of the thicknessda, and B(h) is the surface blocking function.

It should be mentioned that Eq. (8) is coupled with the convec-tive diffusion equation describing particle transport from the bulkto the diffusion boundary layer [42,43].

For spherical particles, the RSA blocking function is described bythe expression [42]

BðhÞ ¼ 1þ a1�hþ a2

�h2 þ a3�h3� �

1� �h� �3 ð9Þ

where �h ¼ h=hmx is the normalized coverage of particles, hmx is themaximum coverage of particles usually dependant on the ionicstrength [32,48,55], and a1–a3 are the dimensionless coefficientsequal to 0.812, 0.426 and 0.0717, respectively.

The maximum coverage increases with ionic strength becauselateral electrostatic interactions among particles are screened[42,43]. Exploiting the RSA approach one can formulate a usefulanalytical expression connecting the maximum coverage withmeasured quantities

hmx ¼ h11

1þ 2h�=dp� �2 ð10Þ

where H1 is the maximum coverage for hard (non-interacting) par-ticles equal to 0.547 for spheres [42,52] and h� is the effective inter-action range characterizing the repulsive double-layer interactionsparticles, which can be calculated from the formula

2h�=dp ¼1

jdpln

/0

2/ch� ln 1þ 1

jdpln

/0

2/ch

� �� ð11Þ

where /0 ¼ 16pedpkTe

� �2tanh2 fpe

4kT

�is the characteristic interaction

energy of particles and /ch is the scaling interaction energy, closeto the kT unit.

As discussed in Ref. [42], Eq. (11) is a good approximation of ex-act numerical results for relatively thin double-layers, wherejdp > 2.

In our case of diffusion-controlled transport, Eqs. (8)–(10) to-gether with the bulk transport equation can be integrated numer-ically using the efficient finite-difference method [32,42,43]. Thetheoretical results obtained in this way are shown in Fig. 11 as so-lid lines. The agreement with experimental data is satisfactory gi-ven the irregular shape of hematite particles and the polydispersityof the suspension. Additionally, in Fig. 12, the maximum coveragescalculated from Eq. (10) for various ionic strengths are comparedwith experimental data. Again, one observes an agreement of the-oretical and experimental results which suggests that Eq. (10) canbe used for predicting the coverage of the hematite monolayer foran arbitrary ionic strength range, exceeding 10�4 M. Therefore, byadjusting the ionic strength, one can in a fully controlled wayproduce hematite nanoparticle monolayers of arbitrary coveragebetween 0.1 and 0.41.

I [M]10-4 10-3 10-2 10-1

θ

0.0

0.1

0.2

0.3

0.4

0.5

Fig. 12. The dependence of the maximum coverage of hematite particles on micaon the ionic strength. The points denote experimental data obtained by SEM orAFM, and the solid line shows the theoretical results calculated from Eq. (7) byadopting the effective hard particle concept.

58 M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59

Another issue that was studied in our work was the stability ofhematite monolayers, which was determined by performing con-trolled desorption experiments. Thus, in the first stage of theseexperiments, a hematite monolayer of a defined coverage was pro-duced as described above. Afterward, the hematite suspension wasreplaced by pure electrolyte of a defined pH and ionic strength andthe particles were allowed to desorb under diffusion-controlledtransport for a prescribed period of time. Their final coverage ofparticles remaining on the surface was determined ex situ by directSEM and AFM enumeration.

A typical kinetic desorption run obtained for the initial coverageof hematite particles Ho = 0.25, pH = 3.5, I = 10�2 M, is presented inFig. 13. As can be seen for the desorption time of 3600 min (60 h),the decrease in the initial coverage of the monolayer was only 4%.Analogous results were obtained in other desorption runs per-formed for an ionic strength range of 10�4 to 3 � 10�2 M and apH range of 3.5–7.4. In all cases, desorption of hematite particleswas negligible for times up to 60 h, which suggests hematite mon-olayers were indeed stable under these conditions.

More extensive studies on the electrokinetic properties ofhematite monolayers and their stability under dynamic conditions(flow) are presented in the forthcoming works.

t1/2 [min1/2]

0 20 40 60

θ/θ 0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Fig. 13. Kinetics of hematite particle desorption determined by SEM, pH = 3.5,I = 10�2 M, T = 293 K and initial coverage of particles Ho = 0.25. The solid linedenotes a linear fit of experimental results.

4. Conclusions

The nanosized hematite particle suspension, synthesized by anacidic hydrolysis of ferric chloride was stable for a broad range ofionic strength and bulk concentration up to 2800 ppm. ForpH < 8.9 (isoelectric point), the hematite particles were positivelycharged, which facilitated their deposition driven by electrostaticinteractions on negatively charged substrates.

This suspension was used to produce hematite monolayers onmica via a controlled self-assembly process carried out under dif-fusion controlled transport. The absolute coverage of depositedparticles was quantitatively determined by AFM and SEM imaging.It was confirmed that the deposition was diffusion controlled, withthe initial rate proportional to the bulk concentration of particles.On the other hand, for long adsorption times, the saturation cover-age was attained, increasing systematically with ionic strength.The kinetic runs were adequately reflected by the RSA model.

Additionally, particle desorption measurements confirmed thatfor a higher coverage range, hematite particle desorption was prac-tically negligible within the time period of 60 h, indicating that themonolayers were stable.

Our experimental data proved, therefore, that it is feasible toproduce uniform and stable hematite nanoparticle monolayers(films) of desired coverage in self-assembly processes controlledby the bulk suspension concentration and the ionic strength. Suchmonolayers may find practical applications as substrates for selec-tive protein and nanoparticle deposition, sensors, and photoelec-trodes in various electrocatalytic applications.

Acknowledgments

This work was financially supported by the Research Grant:POIG 01.01.02-12-028/09-00.

The authors are grateful to Katarzyna Luberta – Durnas for herhelp in performing the XRD analysis.

References

[1] A.S. Teja, P.Y. Koh, Prog. Cryst. Growth Charact. Mater. 55 (2009) 22–45.[2] S.M. Ahmed, J. Leduc, S.F. Haller, J. Phys. Chem. 92 (1988) 6655–6660.[3] J.A. Glasscock, P.R.F. Barnes, I.C. Plumb, A. Bendavid, P.J. Martin, Thin Solid

Films 516 (2008) 1716–1724.[4] Ch. Li, Y. Shen, M. Jia, S. Sheng, M.O. Adebajo, H. Zhu, Catal. Commun. 9 (2008)

355–361.[5] Y. Wang, B.H. Davis, Appl. Catal., A 180 (1999) 277–285.[6] Ch. Wu, P. Yin, X. Zhu, Ch. OuYang, Y. Xie, J. Phys. Chem. B 110 (2006) 17806–

17812.[7] A. Esteban-Cubillo, J.M. Tulliani, C. Pecharroman, J.S. Moya, J. Eur. Ceram. Soc.

27 (2007) 1983–1989.[8] E.L. Miller, D. Paluselli, B. Marsen, R.E. Rocheleau, Thin Solid Films 466 (2004)

307–313.[9] F. Zhou, S. Kotru, R.K. Pandey, Thin Solid Films 408 (2002) 33–36.

[10] Y.S. Hu, A. Kleiman-Shwarsctein, A.J. Forman, D. Hazen, J.N. Park, E.W.McFarland, Chem. Mater. 20 (2008) 3803–3805.

[11] A. Duret, M. Grätzel, J. Phys. Chem. B 109 (2005) 17184–17191.[12] F. Le Formal, M.M. Grätzel, K. Sivula, Adv. Funct. Mater. 20 (2010) 1099–1107.[13] L. Yuberto, M. Ocaña, A. Justo, L. Conttreras, A.R. González-Elipe, J. Vac. Sci.

Technol., A 18 (2000) 2244–2248.[14] B.J. Kim, E.T. Lee, G.E. Jang, Thin Solid Films 341 (1999) 79–83.[15] A.A. Tahir, K.G. Upul Wijayantha, S. Saremi-Yarahamadi, M. Mazhar, V. McKee,

Chem. Mater. 21 (2009) 3763–3772.[16] E. Matijevic, P. Scheiner, J. Colloid Interface Sci. 63 (1978) 509–524.[17] E. Matijevic, R.S. Sapieszko, J.B. Melville, J. Colloid Interface Sci. 50 (1975) 567–

581.[18] R.S. Sapieszko, R. Patel, E. Matijevic, J. Phys. Chem. 81 (1977) 1061–1074.[19] T. Sugimoto, K. Sakata, J. Colloid Interface Sci. 152 (1992) 587–590.[20] T. Sugimoto, M.M. Khan, A. Muramatsu, Colloids Surf., A 70 (1993) 167–169.[21] T. Sugimoto, M.M. Khan, A. Muramatsu, H. Itoh, Colloids Surf., A 79 (1993)

233–247.[22] T. Sugimoto, K. Sakata, A. Muramatsu, J. Colloid Interface Sci. 159 (1993) 372–

382.[23] T. Sugimoto, Y. Wang, H. Itoh, A. Muramatsu, Colloids Surf., A: Physicochem.

Eng. Aspects 134 (1998) 265–279.

M. Ocwieja et al. / Journal of Colloid and Interface Science 386 (2012) 51–59 59

[24] X. Wang, X. Chen, L. Gao, H. Zheng, M. Ji, Ch. Tang, T. Shen, Z. Zhang, J. Mater.Chem. 14 (2004) 905–907.

[25] J. Lian, X. Duan, J. Ma, P. Peng, T. Kim, W. Zheng, ACS Nano 3 (2009) 3749–3761.

[26] J. Zhang, J. Buffle, J. Colloid Interface Sci. 174 (1995) 500–509.[27] U. Björkstèn, J. Moser, M. Grätzel, Chem. Mater. 6 (1994) 858–863.[28] W. Wang, L. Liang, A. Josh, B. Gu, J. Mater. Chem. 18 (2008) 5770–5775.[29] L. Huo, W. Li, L. Lu, H. Cui, S. Xi, J. Wang, B. Zhao, Y. Shen, Z. Lu, Chem. Mater. 12

(2000) 790–794.[30] X.G. Peng, Y. Zhang, J. Yang, B.S. Zou, L.Z. Xiao, T.J. Li, J. Phys. Chem. 96 (1992)

3412–3415.[31] F. Kuhnen, K. Barmettler, S. Bhattacharejee, M. Elimelech, R. Kretzschmar, J.

Colloid Interface Sci. 231 (2000) 32–41.[32] M. Ocwieja, Z. Adamczyk, M. Morga, A. Michna, J. Colloid Interface Sci. 364

(2011) 39–48.[33] M. Zembala, Z. Adamczyk, Langmuir 16 (2000) 1593–1601.[34] M. Zembala, Z. Adamczyk, P. Warszynski, Colloids Surf., A: Physicochem. Eng.

Aspects 195 (2001) 3–15.[35] M. Schudel, S.H. Behrens, H. Holthoff, R. Kretzschmar, M. Borkovec, J. Colloid

Interface Sci. 196 (1997) 241–253.[36] Y.T. He, J. Wan, T. Tokunaga, J. Nanopart. Res. 10 (2008) 321–332.[37] A.V. Delgado, F.G. Gonzalez-Cabállero, Croat. Chem. Acta 71 (1998) 1087–

1104.[38] A. Bentaleb, P. Vera, A.V. Delgado, Mater. Chem. Res. 37 (1994) 68–75.[39] Z. Adamczyk, M. Nattich, M. Wasilewska, M. Zaucha, Adv. Colloid Interface Sci.

168 (2011) 3–28.[40] M. Dabkowska, Z. Adamczyk, J. Colloid Interface Sci. 366 (2012) 105–113.

[41] Z. Adamczyk, J. Colloid Interface Sci. 229 (2000) 477–489.[42] Z. Adamczyk, Particles at Interfaces: Interactions, Deposition, Structure, first

ed., Elsevier, Academic Press, Amsterdam, 2006.[43] M. Ocwieja, Z. Adamczyk, K. Kubiak, J. Colloid Interface Sci. 376 (2012) 1–11.[44] Z. Adamczyk, Curr. Opin. Colloid Interface Sci. (2012), http://dx.doi.org/

10.1016/j.cocis.2011.12.02.[45] Z. Adamczyk, L. Szyk, Langmuir 16 (2000) 5730–5737.[46] J. Kleimann, G. Lecoultre, G. Papastavrou, S. Jeanneret, P. Galletto, G.J.M. Koper,

M. Borkovec, J. Colloid Interface Sci. 303 (2006) 460–471.[47] R. Pericet-Camora, G. Papastavrou, M. Borkovec, Langmuir 20 (2004) 3264–

3270.[48] I. Popa, B.P. Cahill, P. Maroni, G. Papastavrou, M. Borkovec, J. Colloid Interface

Sci. 309 (2007) 28–35.[49] E.A.M. Brouwer, E.S. Kooij, H. Wormeester, B. Poelsema, Langmuir 19 (2003)

8102–8108.[50] E.S. Kooij, E.A.M. Brouwer, H. Wormeester, B. Poelsema, Colloids Surf., A:

Physicochem. Eng. Aspects 222 (2003) 103–111.[51] E.A.M. Brouwer, E.S. Kooij, M. Hakbijl, H. Wormeester, B. Poelsema, Colloids

Surf., A: Physicochem. Eng. Aspects 267 (2005) 133–138.[52] Y. Yuan, M.R. Oberholzer, A.M. Lenhoff, Colloids Surf., A: Physicochem. Eng.

Aspects 165 (2000) 125–141.[53] M.R. Oberholzer, J.M. Stankovich, S.L. Carine, D.Y.C. Chan, A.M. Lenhoff, J.

Colloid Interface Sci. 194 (1997) 138–153.[54] M.R. Böhmer, E.A. van der Zeeuw, G.J.M. Koper, J. Colloid Interface Sci. 197

(1998) 242–250.[55] M. Semmler, E.K. Mann, J. Ricka, M. Borkovec, Langmuir 14 (1998) 5127–5132.