Reactive Layer-by-Layer Deposition of Poly ... - ACS Publications

ARTICLE pubs.acs.org/Langmuir

Reactive Layer-by-Layer Deposition of Poly(ethylene imine) and a Precursor of TiO2: Influence of the Sodium Chloride Concentration on the Film Growth, Interaction with Hexacyanoferrate Anions, and Particle Distribution in the Film Nadia Ladhari,†,‡ Christian Ringwald,†,‡ Ovidiu Ersen,§ Ileana Florea,§ Joseph Hemmerl
e,†,‡ and Vincent Ball*,†,‡,|| †

Institut National de la Sant
e et de la Recherche M
edicale, Unit
e Mixte de Recherche 977, 11 rue Humann, 67085 Strasbourg, C
edex, France ‡ Universit
e de Strasbourg, Facult
e de Chirurgie Dentaire, 1 place de l’H^opital, 67000 Strasbourg, France § Institut de Physique et Chimie des Mat
eriaux (IPCMS), Unit
e Mixte de Recherche 7504 du CNRS, 23 rue du Loess, 67087 Strasbourg, France

bS Supporting Information ABSTRACT: Films prepared according to a layer-by-layer (LBL) manner find increasing importance in many applications such as coatings with dedicated optical or electronic properties, particularly when including nanomaterials. An alternative way to prepare such hybrid layer-by-layer coatings is to perform sol
gel chemistry in a layer-by-layer manner. In this article, we highlight the importance of the NaCl concentration as a parameter to control the growth as well as the properties of LBL films made from poly(ethylene imine) as the organic counterpart and titanium IV (bisammoniumlactato)dihydroxyde ([Ti(lac)2(OH)2]2
) as the precursor of TiO2. An increase in the sodium chloride concentration leads to the faster growth of the film and to a decrease in the number of hexacyanoferrate anions remaining in the film after a buffer rinse. This may be due to a progressive increase in the fraction of negatively charged TiO2 as suggested by transmission electron microscopy. In the presence of 0.5 M NaCl, the fraction of TiO2 is close to 60% in mass. As a surprising finding, the films produced from 0.15 M NaCl are not homogeneously filled with TiO2 even if the film is produced in an LBL fashion. The increased concentration of TiO2 at the film
solution interface could constitute a barrier for the incorporation of the negatively charged redox probe.

’ INTRODUCTION The functionalization of solid
liquid interfaces by the alternating adsorption of positively and negatively charged polymers or colloids, yielding polyelectrolyte multilayer films1 or hybrid organic
inorganic films,2
4 has proven to be a versatile tool for modifying the interactions of the coated materials with the ambient medium.5 The alternating deposition of poly(sodium4-styrenesulfonate) and poly(diallyldimethyl ammonium chloride) allows the production of anticorrosive coatings6 whose efficiency can even be increased by the addition of corrosion inhibitors inside the PEM films.7 These PEM films can have a sensing activity by themselves, for instance, in response to changes in humidity,8 and they can display permselectivity versus different solvents9 or ions.10
12 Their use as sensors can be improved by the addition of catalysts such as enzymes.13
16 The layer-by-layer incorporation of nanoparticles has allowed a plethora of properties to be conferred on this kind of films: electrochromic applications;17,18 control of surface roughness and wettability;19,20 modification of mechanical properties,21
24 film reflectivity,25,26 and magnetic properties;27 and reactivity to external stimuli such as r 2011 American Chemical Society

electromagnetic radiation28 and changes in temperature.29 Finally, there are some films that can display a combination of optical and mechanical properties.30 Instead of incorporating the desired nanoparticles in an LBL manner, the possibility of synthezising them after reduction of the incorporated metallic cations31
33 or by a reactive layer-by-layer deposition34
36 has recently been demonstrated. This method, inspired by the catalytic role of polyamines in the production of biosilica,37,38 relies on the alternating deposition of a polyamine and either silicic acid39 or a hydrosoluble precursor of titanium dioxide, Ti(IV) (bisammonium lactato) dihydroxyde (denoted [Ti(lac)2(OH)2]2
herein).34
36 Most of the articles describing the reactive layer-by-layer (r-PEM) deposition of polyamines and the precursors of silica were aimed at demonstrating the feasability of this concept as well as highlighting some of its applications, for instance, photoinduced superhydrophilicity. Indeed, the (PEI-[Ti(lac)2(OH)2]2
)n films exhibit Received: March 24, 2011 Revised: May 13, 2011 Published: May 27, 2011 7934

dx.doi.org/10.1021/la201106r | Langmuir 2011, 27, 7934–7943

Langmuir photoinduced superhydrophilicity when the last deposited layer consists of TiO2 produced from its precursor, [Ti(lac)2(OH)2]2
.36 These films contain anatase nanoparticles that are 4.9 ( 1.2 nm in diameter whereas the particles formed in solution are extremely polydisperse when a solution containing [Ti(lac)2(OH)2]2
is mixed with the polyamine-containing solution. The same finding of a narrow particle size distribution has been described when [Ti(lac)2(OH)2]2
reacts in contact with recombinant silaffin adsorbed on (PAH-PSS)2 PEM films.40 In the later case, the formation of rather monodisperse titania nanoparticles was explained by the fact that the protein forms clusters of around 20 nm in diameter, which restricts the growth of the TiO2 particles. However, to extend the concept of inorganic multilayer films to applications, mainly as antibacterial coatings, as materials for dye-sensitized solar cells,41 and as templates for photomagnetic materials,42,43 where the produced TiO2 has to be coupled with cyanometallates, more fundamental knowledge about the reaction mechanism leading to TiO2 nanoparticles is required. This will be the main aim of this article, in which we will investigate the influence of the sodium chloride concentration on the film thickness, optical properties, interaction with charged redox probes, and particle size of TiO2 incorporated in the film. In addition, we will investigate the TiO2 distribution along the direction perpendicular to the films build up in the presence of 0.15 M NaCl. We choose to focus on the NaCl concentration as a control parameter because it is known that it changes the persistence length of the polyelectrolyte, here PEI, and hence the availability of the primary, secondary, and tertiary ammonium groups to react with [Ti(lac)2(OH)2]2
and to induce its condensation in TiO2. The first step of this interaction should be of an electrostatic nature, with [Ti(lac)2(OH)2]2
being a possible counterion of the ammonium groups carried by PEI. Of particular interest will be the structural characterization of the TiO2 particles produced during the r-PEM deposition as a function of the ionic strength. We will hence work at the same concentrations of the reactants as in our previous study, namely, at 5 mM in [Ti(lac)2(OH)2]2
and at 1 mg/mL in PEI. Finally, we will demonstrate on the basis of the cyclic voltammetry measurements of Fe(CN)6 4
present in the films that it is possible to precipitate Prussian Blue in the (PEI-[Ti(lac)2(OH)2]2
)m films. Note that in this article the r-PEM films are denoted by (PEI-[Ti(lac)2(OH)2]2
)m owing to their preparation method even if the films contain TiO2 particles.

’ MATERIALS AND METHODS Chemicals and Substrates. All solutions were made from Milli-Q ultrapure water (F = 18.2 MΩ cm). All polyelectrolyte- and [Ti(lac)2(OH)2]2
-containing solutions were prepared in Tris (Tris(hydroxymethyl)aminomethane, Gibco BRL) buffer at 5 or 50 mM whose pH was fixed at 7.50 with concentrated hydrochloric acid. The ionic strength was increased by the addition of NaCl (Euromedex, Souffelweyersheim, France, ref 1112-A) up to 1 M. Branched PEI was purchased from Sigma (viscosity-averaged molecular mass of 750  103 g mol
1, ref P3143, lot 093K0098), and titanium IV (bisammoniumlactato)dihydroxyde was from Aldrich (50% w/w, ref 388165). The [Ti(lac)2(OH)2]2
-containing solutions were prepared just before the beginning of each reactive LBL deposition experiments. We checked by means of UV
vis spectroscopy (mc2 spectrophotometer, Safas, Monaco) that [Ti(lac)2(OH)2]2
solutions do not undergo an important spontaneous polycondensation. Indeed, if polycondensation did occur, then the appearance of particles in the solutions

ARTICLE

would result in an increase in the solution turbidity. This was never observed at λ = 500 nm for [Ti(lac)2(OH)2]2
at 5 mM as long as the ionic strength was lower than 1 M and for durations of shorter than 4 h. However, the addition of PEI to the [Ti(lac)2(OH)2]2
-containing solution produced an instantaneous increase in turbidity. Hence, in our investigation we worked with a NaCl concentration of between 0 and 1 M, with the Tris buffer being 5 or 50 mM and the pH being constant at 7.50 ( 0.05. The adsorption substrates were P-doped silicon (100) wafers (Wafernet, Inc., San Jose, CA) or quartz plates (1  4  0.1 cm3, Thuet, Blodesheim, France) for the characterization of the films by ellipsometry and UV
vis spectroscopy, respectively. They were cleaned with absolute ethanol and distilled water and immersed in a hot (∼70 °C) Hellmanex solution (2% v/v, Hellma, GmBh, M€ullheim, Germany) for half an hour, followed by intensive water rinsing and immersion in a hot 1 M hydrochloric acid solution (for 10 min). Finally, the substrates were rinsed with distilled water and blown dry with a stream of nitrogen. For cyclic voltammetry, the substrates were polished amorphous carbon electrodes (see later), and for TEM characterization, we used 300-mesh copper grids coated with Formvar. Finally, for infrared spectroscopy in the attenuated reflection mode, we used a ZnSe crystal that was cleaned with optical paper immersed in ethanol, in Hellmanex, and again in ethanol. The crystal was finally washed with Milli-Q water. Deposition of the r-PEM Films by Dipping. Contrary to our previous investigation, where we showed the deposition of r-PEM film by spraying PEI and [Ti(lac)2(OH)2]2
,36 we favored film deposition by alternating dipping in a PEI- and a [Ti(lac)2(OH)2]2
-containing solution. This comes from the fact that it is difficult to apply spray deposition on amorphous carbon electrodes and Formvar-coated electron microscopy grids that we used for our cyclic voltammetry and structural characterization, respectively. The r-PEM films were deposited by starting with the adsorption of PEI owing to its positive charge at pH 7.50 and owing to the negative surface charge density of all of the used substrates. The substrate was left in contact with the solution PEI (1 mg/mL) for t minutes, placed in contact for 5 min with the same buffer as that used to dissolve the PEI and [Ti(lac)2(OH)2]2
anions, put in a solution containing [Ti(lac)2(OH)2]2
at 5 mM for t minutes, and rinsed again with buffer for 5 min. These four immersion steps lead to the deposition of one “layer pair” and were repeated the desired number of times, m. The optimal adsorption time t was determined by means of the in situ monitoring of the adsorption kinetics by means of the quartz crystal microbalance with dissipation technique.

Characterization Experiments of the Surfaces of the Films. Quartz Crystal Microbalance. The QCM-D technique can be used to calculate the adsorbed mass on a piezoelectric quartz crystal because of the adsorbent’s influence on the crystal’s oscillation. A Q-Sense D300 device (G€otenborg, Sweden) was used, and the oscillations of the quartz crystal were excited close to its resonance frequency at about 5 MHz. When the exciting signal was stopped, the decay of the shear wave was followed as a function of time at the fundamental frequency as well as at its third, fifth, and seventh overtones. From this decay curve, the pseudofrequency and the dissipation were calculated as a function of time. The change in the reduced resonance frequency with respect to the signal in the presence of pure Tris buffer will be denoted as Δfn/n, where n is the overtone number (n = 1, 3, 5, or 7). For ideally thin and rigid films, the reduced frequency changes for all overtones overlap (as will be shown in the Results section) and the adsorbed mass per unit area of the surface, Γ, can be calculated using the Sauerbrey equation44 Γ¼


Fc v Δfn Δfn ¼
C n 2f1 2 n

ð1Þ

The density of the crystal Fc, the propagation speed v, and the fundamental frequency f1 of the oscillation are grouped in the material 7935

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir

ARTICLE

constant, C, that has a value of 16.6 ng cm
2 Hz1
for the employed crystals (according to the furnisher). The thickness d of the deposit can be calculated by assuming a homogeneous density F by d = Γ/F. We assumed here that F = 1.3 g/cm3, which means that the thickness obtained by QCM-D is only semiquantitative. Any precise information about the film density is not available for the moment owing to its composite nature: it contains PEI, TiO2, water, and certainly ions from the electrolyte (as will be demonstrated by infrared spectroscopy). Streaming Potential Measurements. The streaming potentials of the (PEI-[Ti(lac)2(OH)2]2
)m films deposited on glass slides were measured with a ZetaCAD device (CAD Instrumentation, Les Essarts le Roi, France). Two glass slides were mounted parallel to each other in the plexiglass sample holder and were separated by a 500-μm-thick poly(tetrafluorethylene) (PTFE) spacer. For all measurements, Tris buffer with a concentration of 5 mM was circulated between the samples. The r-PEM film was deposited on the substrates ex situ using the optimal deposition time determined by means of QCM-D. The streaming potential was measured five times on the same substrate, before and after functionalization with (PEI-[Ti(lac)2(OH)2]2
)m films, and the obtained values were averaged. These experiments were aimed to define if the reactive LBL deposition followed the regular charge inversion usually observed during the alternating deposition of polycations and polyanions.45 All of these experiments were performed by depositing the LBL films in the presence of 5 mM Tris at pH 7.5, with the streaming potential measurements being performed under the same conditions. The ζ potential was calculated from the measured streaming potential using the Smoluchowski relationship46 ζ¼

ΔE ηλ ΔP εε0

ð2Þ

where ζ, η, λ, and εε0 are the ζ potential, the solution viscosity, the solution conductivity, and the dielectric permittivity of water. ΔE/ΔP is the streaming potential, namely, the slope of the potential difference versus pressure difference curve. The potential difference ΔE was measured between two Ag/AgCl reference electrodes located on both sides of the measurement cell. The pressure difference ΔP between the two electrolyte compartments was varied with compressed air in increments of 5 kPa between
30 and þ30 kPa. Because the viscosity and the dielectric permittivity are temperature-dependent, the temperature of the solution was regularly measured in situ and its value was used to calculate temperature corrected values of the dielectric permittivity and viscosity. The solution conductivity was also measured in situ. Cyclic Voltammetry. The cyclic voltammetry (CV) experiments were carried out in a conventional three-electrode setup (CH Instruments, Austin, TX, USA, ref 604B) in which the working, reference, and counter electrodes were amorphous carbon (ref 104), Ag/AgCl (ref 111), and a platinum wire (ref 115). The solutions were deoxygenated before each measurement by bubbling nitrogen for 10 min. A freshly polished and cleaned working electrode was used for each experiment. The capacitive and the faradic currents were measured by cycling the potential difference between the working electrode and the reference electrode between
0.1 and þ0.65 V at scanning rates between 5 and 500 mV/s. CV was first performed on the pristine electrode. The capacitive current was measured in a buffer solution containing 50 mM Tris and x M sodium chloride at pH 7.5. Subsequently, the faradic current was measured in the same buffer solution and in the presence of 1 mM potassium hexacyanoferrate (K4Fe(CN)6, Sigma-Aldrich, ref P9387). The experiment was continued only if the potential difference between the oxidation and reduction peaks of hexacyanoferrate was lower than 80 mV. (Theoretically, it should equal 59 mV for a reversible one-electron process taking place at a temperature of 298 K.) The working electrode was then modified with the (PEI-[Ti(lac)2(OH)2]2
)n films using the optimal immersion times determined by means of QCM-D. The aim of these experiments was to measure the interactions of Fe(CN)6 4
with

the (PEI-[Ti(lac)2(OH)2]2
)n films as a function of the ionic strength of the buffer used during film deposition. The films were placed in contact with either a 1 mM hexacyanoferrate or a 1 mM ruthenium hexaamine II (Ru(NH3)2Cl2, Sigma-Aldrich) containing buffer solution, and CV experiments were conducted after different incubation times. Preliminary experiments have shown that for all NaCl concentration values of the buffer, the oxidation and reduction currents reached maximal values after less than 10 min. Hence, the experiments aimed to quantify the incorporation of Fe(CN)64
ions in the film were performed by incubating the film with the 1 mM electroactive ions containing buffer for 10 min. The electrode was then rinsed with buffer, and cyclic voltammograms were regularly recorded at a scan rate of 50 mV/s until the oxidation and reduction currents reached steady values. From this time on, we conducted CV at different scan rates. The surface concentration Γ of Fe(CN)64
anions or Ru(NH3)62þ cations accessible to oxidation (which could be lower than the actual concentration of redox probes trapped in the film) was obtained from the variation of the oxidation current, ipa, versus the potential scan rate, υ, according to47,48 Γ¼

4RTip n2 F2 Aυ

ð3Þ

where F, A, and T are the Faraday constant, the electrode area, and the absolute temperature. We assume the electrode area to be equal to the geometric area of the amorphous carbon disk (3.14 mm2), hence neglecting its roughness. This leads to an overestimation of the surface coverage of Fe(CN)64
. Ellipsometry. The thickness of the hybrid film grown on the silicon substrate was measured by means of ellipsometry (Jobin Yvon model PZ2000, France) at a constant angle of incidence (70°) after film deposition, rinsing with distilled water to avoid the crystallization of Tris or NaCl, and drying in a stream of nitrogen. Additional details are given in the Supporting Information section. UV
Vis Spectroscopy. The details of the UV
vis spectroscopy experiments are given in the Supporting Information section. The UV
vis and ellipsometry data were combined to obtain information about the optical absorption coefficient of the (PEI-[Ti(lac)2(OH)2]2
)m films as a function of the NaCl concentration used to deposit the films. This coefficient was obtained by dividing the slope of the absorbance versus the number of layer pairs, SA, by the slope of the thickness versus the number of layer pairs, Sd, according to ε¼

sA sd

ð4Þ

SA and Sd were obtained from the UV
visible absorption and ellipsometry data, respectively. Atomic Force Microscopy. Atomic force microscopy (AFM) was performed on (PEI-[Ti(lac)2(OH)2]2
)10 r-PEM films deposited on the silicon wafers also used for the ellipsometry measurements. Topography images in contact mode and in air were acquired with a Nanoscope IV instrument (Veeco, Santa Barbara, CA, USA) at a scanning frequency of 2 Hz. The employed cantilevers (Veeco, ref MSCT-AUWH) had a nominative spring constant of 0.01 N/m and were terminated with silicon nitride (nominative radius of curvature of 10 nm). AFM topographical images were analyzed to obtain the rootmean-square (rms) roughness of the r-PEM films deposited from solutions at different concentrations in NaCl. Compositional and Structural Characterizations. Infrared Spectroscopy in Attenuated Total Reflection Mode (FTIR-ATR). These experiments, whose details are given elsewhere,49 were performed to show the presence of Fe(CN)64
anions in (PEI-[Ti(lac)2(OH)2]2
)6 films and the appearance of Prussian Blue when the whole architecture was placed in contact with 31 mM FeCl3. (FeCl3 was dissolved in D2O at pD 6 to avoid the formation of iron III hydroxide.) The 7936

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir (PEI-[Ti(lac)2(OH)2]2
)6 architecture was built up according to the information obtained from the QCM-D experiments in the presence of 50 mM Tris
150 mM NaCl, with D2O being the solvent. The pH was adjusted to 7.9 to account for the 0.4 pH unit difference between H2O and D2O. After each polyelectrolyte adsorption, the polyelectrolyte solution was replaced by Tris buffer and the infrared spectrum of the film was acquired by accumulating 512 scans at 2 cm
1 spectral resolution on an Equinox 55 spectrometer (Bruker, Wissembourg, France). The detector was a liquid-nitrogen-cooled MCT detector. The transmitted intensity was compared to that transmitted by the naked ZnSe crystal to calculate the absorption spectrum of each layer. After the film buildup was completed, the buffer was replaced by a 1 mM potassium hexacyanoferrate-containing solution that was allowed to circulate over the film for 10 min (as for the CV experiments). Finally, the film was placed in contact with an aqueous FeCl3 solution at 31 mM. Thermogravimetric Analysis. The thermogravimetric analysis experiments (TGA) were performed on a SETARAM thermoanalyzer. Each sample was placed in a platinum crucible and heated from room temperature to 750 °C at a heating rate of 10 °C/min, using a 20/80 (v/v) O2/N2 mixture at a flow stream of 50 cm3/min. The sample being analyzed was the powder obtained by scratching the (PEI-[Ti(lac)2(OH)2]2
)m films away from their silicon substrate. Transmission Electron Microscopy. The transmission electron microscopy analysis was performed on a Jeol 2100F electron microscope operating at 200 kV and with a point-to-point resolution of 0.21 nm. Typical plan-view TEM images were thus acquired in bright-field mode on different areas of the analyzed specimens. The selected-area electron diffraction (SAED) patterns were taken over a 500-nm-diameter circular area. The electron energy loss spectra (EELS) were recorded on the same microscope using a standard postcolumn GIF imaging filter with a collection angle set to 30 mrad and a dispersion of 0.3 eV per energy channel. The mean spectra were obtained by averaging the individual energy loss of the electrons coming from a relatively large area of the layers. Tomography Analysis. The 3D morphology of the film and the spatial distribution of the TiO2 particles within the layer were investigated by electron tomography with the acquisition of the tilt series in bright-field mode.50 This technique consists of the volume reconstruction of a nano-object from a series of 2D images and thus allows us to solve the characteristics of interest in three dimensions. For the tomographic experiment, the tilt series was acquired by tilting the specimen over a range of (60°, with an image recorded every 2° in Saxton mode.51 The acquisitions were carried out at low temperature (about 100 K) in order to reduce the irradiation damage in the organic part of the layer during the total duration of the acquisition process (1 h). After the acquisition, the projections of the tilt series were spatially aligned using the cross-correlation algorithm implemented in the IMOD software.54 The volume calculation was performed using the algebraic reconstruction techniques52 by using the TomoJ plugin implemented in the general mode Image software53 with 10 iterations.

’ RESULTS AND DISCUSSION We first investigated the kinetics of the deposition of both PEI, followed by a buffer rinse, and the reaction of [Ti(lac)2(OH)2]2
anions with already deposited PEI by means of QCM-D. These experiments were performed at constant pH, constant reactant concentration, and constant temperature (25 °C); the only parameter that was changed was the concentration of NaCl. A typical experiment is displayed in Figure 1, with the buffer being 50 mM Tris þ 200 mM NaCl, corresponding to a total ionic strength of 242 mM (the pKa of Tris is equal to 8.2). It appears clearly that the changes in reduced frequency are the same for the third, fifth, and seventh harmonics of the quartz crystal, implying

ARTICLE

Figure 1. (A) Evolution of the reduced frequency change as a function of time at the third, fifth, and seventh harmonics of the silica-coated quartz crystal during the deposition of a (PEI-[Ti(lac)2(OH)2]2
)m r-PEM film. The deposition was performed in the presence of 50 mM Tris þ 200 mM NaCl at pH 7.50. The vertical black lines correspond to the beginning of the [Ti(lac)2(OH)2]2
adsorption. For simplicity, the injections of PEI are not labeled but are clearly visible in between two successive injections of [Ti(lac)2(OH)2]2
. (B) Film thickness, calculated from the data in part A, at the end of each layer pair as a function of the number of deposited layer pairs. The calculation was made by assuming a density of 1.3 g cm
3 for the film. The full line corresponds to the linear regression to the data. The dashed line corresponds to the limit of the 95% confidence interval.

that the use of the Sauerbrey equation is justified to calculate the surface coverage and the film thickness directly from the changes in the reduced frequency.53 The evolution of the film thickness as a function of the number of layer pairs is displayed in Figure 1B: the growth is linear with a slope of around 19.5 nm per layer pair at this NaCl concentration. It appears also from Figure 1A that the deposition kinetics of both PEI and [Ti(lac)2(OH)2]2
was fast, being achieved in less than 2 min. (We consider the deposition kinetics to be achieved when the change in reduced frequency is less than 1 Hz/min.) When the concentration in NaCl was increased, the deposition kinetics of the [Ti(lac)2(OH)2]2
anions was slowed down but it nevertheless reached a steady state in less than 5 min (Figure 1 of the Supporting Information). Hence, all of the r-PEM films in the forthcoming part of this study were deposited by using a deposition time of t = 5 min for both the PEI and [Ti(lac)2(OH)2]2
deposition steps. Note that this reactive layer-by-layer deposition obeys all of the characteristics of electrostatically driven deposition with charge overcompensation after every deposition, as shown by zeta potential measurements (Figure 2). Nevertheless, we found that the regular alternation between a positive zeta potential ((12.6 ( 2.2) mV 7937

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir

Figure 2. Zeta potential of r-PEM films made from the alternating deposition of PEI and [Ti(lac)2(OH)2]2
(each deposition step lasting over 5 min) in the presence of 5 mM Tris buffer at pH 7.50. As explained in the text, all of these points were obtained on independently prepared films. Odd integers correspond to the deposition of PEI, and even integers correspond to the deposition of [Ti(lac)2(OH)2]2
anions (leading to the formation of TiO2). The horizontal dashed line corresponds to ζ = 0.

for PEI) and a negative zeta potential ((
23.2 ( 7.8) mV after the deposition of [Ti(lac)2(OH)2]2
) was found only when the [Ti(lac)2(OH)2]2
anions were allowed to react with the PEI ending film immediately after the PEI adsorption had reached a steady state. Indeed, when we measured the zeta potential of a PEI ending film as a function of time, we found that its absolute value decreased to zero before charge inversion and finally decreased to the zeta potential obtained after the deposition of the last TiO2 layer. However, the zeta potential of a TiO2 ending film (obtained after the reaction of [Ti(lac)2(OH)2]2
with the PEI ending film) remained constant for at least 48 h (Figure 2 of the Supporting Information). This shows that PEI was intrinsically unstable on such films and the whole architecture was stable after the reactive deposition of [Ti(lac)2(OH)2]2
. Another possible interpretation would be the progressive exposure of TiO2 at the film
solution interface and a concomitant penetration of PEI to the inside parts of the film through a diffusion process. We will address this point later on in this article. We showed in our previous investigation that the (PEI[Ti(lac)2(OH)2]2
)m films prepared in the presence of 50 mM Tris at pH 7.50 remained stable for weeks when stored in this buffer.36 Hence, all of the zeta potentials given in Figure 2 were measured on a freshly prepared film ending with the deposition of PEI or [Ti(lac)2(OH)2]2
. Because the films were stable, from the point of view of their surface potential, when TiO2 was the last deposited layer, all of the film properties investigated in this study were measured on (PEI-[Ti(lac)2(OH)2]2
)n films, not on (PEI-[Ti(lac)2(OH)2]2
)n-PEI films. We then investigated the deposition of (PEI-[Ti(lac)2(OH)2]2
)m films as a function of the NaCl concentration by means of ellipsometry and UV
vis spectroscopy. A typical buildup experiment followed by UV
vis and ellipsometry at an ionic strength of 191 mM is displayed in Figure 3 of the Supporting Information. For the ellipsometry experiments, we measured the thickness of the deposit (see the experimental part concerning the refractive

ARTICLE

Figure 3. Thickness increment per layer pair of (PEI-[Ti(lac)2(OH)2]2
)m films as a function of the NaCl concentration of the buffer solution during the film buildup as measured by QCM-D (O) and ellipsometry (b) Each point in this curve is the result of an independent buildup experiment as shown in Figure 1 for QCM-D and in Figure 4 of the Supporting Information for ellipsometry. The dotted lines correspond to the limits of the 95% confidence interval of the linear regressions. It has to be noted that the ellipsometry data depend on the relative humidity of the ambient air in which the measurements were made.

index of the film) after 10 layer pairs or every 2 layer pairs in order to investigate if the drying
rehydration step would influence the film growth. It appeared that the regular drying
rehydration cycles did not affect the deposition process: the thickness of (PEI-[Ti(lac)2(OH)2]2
)10 films was the same (within experimental errors of about 5%) whether the film had never been dried during its deposition or dried five times (i.e., every two layer pairs (data not shown)). As expected, the thickness per layer pair (i.e., the slope of the buildup curves obtained from the QCM-D or from the ellipsometry curves increased with the NaCl concentration of the solutions, Figure 3). Both thickness increments per layer pair appeared almost linear as a function of the salt concentration up to 1.0 M. The experiments could not be pursued above 1 M in NaCl owing to the appearance of spontaneous turbidity in the [Ti(lac)2(OH)2]2
-containing solutions in less than 2 h. It also appears that the thickness increment was higher when investigated by QCM-D then by ellipsometry by a factor of close to 7. This reflects an important salt-induced swelling of the (PEI-[Ti(lac)2(OH)2]2
)n r-PEM films. Indeed, the films were characterized after drying but in the presence of ambient humidity (always lower than 100%) when ellipsometry is used, whereas the frequency shifts were recorded in the presence of buffer solution in the QCM-D experiments. It is well known that QCM-D monitors the whole bound mass at the interface, with the water of hydration included.55 The UV
vis spectra of the films were consistent with those of deposits containing TiO2 in the form of anatase over the whole range of investigated salt concentration (Figure 3 of the Supporting Information). It appears that the absorbance increment per layer pair obtained from the UV
vis experiments also increases with the ionic strength (Figure 4 of the Supporting Information). This finding, in combination with an almost linear increase in the thickness increment per layer pair as a function of the ionic strength (Figure 3), suggests that the optical absorption coefficient 7938

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir

ARTICLE

Table 1. Evolution of the Surface Concentration in Hexacyanoferrate Retained in the Film as a Function of the NaCl Concentration in the Buffer Solution Used during Film Buildup salt concentration (mM)

a

Γ (10
13 mol/cm2)a

d (nm)b

C (mol/L)c

41.7

9.76

43.1

2.26  10
4

191.7

9.83

59.6

1.65  10
4

341.7

5.05

58.0

0.87  10
4

541.7

2.48

68.5

0.36  10
4

1041.7

0.67

88.6

0.08  10
4 b

Surface concentration as calculated from eq 3. Film thickness obtained from ellipsometry data (Figure 3. c Average concentration of hexacyanoferrate in the film calculated according to C = Γ/d.

Figure 4. Change in the oxidation current peak as a function of the scan rate for (PEI-[Ti(lac)2(OH)2]2
)m=6 films prepared in the presence of 50 mM Tris buffer with different concentrations of added NaCl: (O) 0, (b) 150, (blue `) 300, (9) 500, and (red () 1000 mM. The r-PEMfilm-coated electrode was placed in contact with a 1 mM potassium hexacyanoferrate-containing solution (for 10 min) and washed with buffer until the release rate of hexacyanoferrate became very slow. From that point on, the CV experiments were performed at different scan rates. The lines correspond to linear regressions to the data. The inset represents the slope of the current vs scan rate as a function of the total ionic strength of the solution.

of the (PEI-[Ti(lac)2(OH)2]2
)m films as defined by eq 4 should be almost independent of the ionic strength. This indeed appears to be the case: the absorption coefficient remains constant at 0.007 ( 0.002 over the whole range of the investigated NaCl concentrations (from 0 to 1 M) This finding deserves a discussion from a more structural point of view: the amount of TiO2 deposited per layer pair and the film thickness increased when the ionic strength increased, with the net result that ε remained approximately constant and independent of the NaCl concentration. This does not mean that the particle size of TiO2 remains constant; it might well be that the observed effect is due either to an increase in the number of particles, which is proportional to the film thickness (and hence to the ionic strength) at constant particle size, or to a change in the particle size and number with the salt concentration. On the micrometer scale, the films were homogeneous without a significant increase in the rms roughness (Figure 5 of the Supporting Information) when the ionic strength increased. The rms roughness values were 28.4, 28.4, and 36.1 nm when the film deposition was performed in the presence of NaCl at 0, 200, and 1000 mM, respectively. Before going to structural considerations related to the particle size distribution in the film, we investigated how the affinity of the film for hexacyanoferate anions changes with the NaCl concentration used during the buildup. A typical cyclic voltammogram of a (PEI-[Ti(lac)2(OH)2]2
)m=6 film prepared in the presence of 50 mM Tris, 150 mM NaCl, and a 1 mM potassium hexacyanoferrate solution for 10 min (which sufficient to reach a steady state in the CV signal) and rinsed with buffer for different times is displayed in Figure 7 of the Supporting Information. It appears that some of the initially incorporated hexacyanoferrate anions are leached out of the film but steady oxidation and reduction currents were obtained after 10
20 h (Figure 8 of the

Supporting Information). The number of hexacyanoferrate anions remaining in the film after the buffer rinse did not allow us to calculate the true number of noncompensated charges in the polyelectrolyte multilayer film. It is found that the number of hexacyanoferrate anions retained in the film decreased with the NaCl concentration (inset of Figure 4). Equation 3 allows us to calculate the surface concentrations in hexacyanoferrate anions from the data in the inset of Figure 4 (Table 3). The peak width at half-maximum of the oxidation and reduction peaks was close to 90 mV (Figure 7 of the Supporting Information shows a typical example for the films deposited in the presence of 50 mM Tris buffer with 150 mM NaCl), as expected for equilibrium electrochemistry in a thin film configuration justifying the use of eq 3 to calculate the surface concentration from the measured peak height. Knowing the film thickness from the data in Figure 3, it was then possible to calculate the average concentration of the anionic redox probe in the film (Table 1). The trend is the same as that displayed in the inset of Figure 4 but is amplified by the fact that the film thickness increased when the salt concentration increased during the buildup (Figure 3). Note that in all cases the hexacyanoferrate concentration in the film was lower than in the bulk solution (1 mM). This was not the case when purely organic multilayer films made from poly(allylamine) and poly(-L-glutamic acid) were placed in contact with hexacyanoferrate-containing buffer solutions:49 the concentration of hexacyanoferrate anions in such films can be 3 orders of magnitude higher than in the solution used to load the films. In our case, the number of negative charges due to TiO2 may be very high owing to the fact that TGA analysis showed that the layer consists of 60 wt % of the TiO2 compound and 40 wt % of the PEI polymer when the film is deposited from a 50 mM Tris buffer containing 500 mM NaCl (Figure 9 of the Supporting Information). The (PEI-[Ti(lac)2(OH)2]2
)m films do not retain ruthenium hexamine (not shown), which suggests, as in our previous investigations, that the negatively charged redox probe is retained in the r-PEM film through its interactions with the positively charged groups of PEI.56 The anions are able to diffuse in the multilayer film owing to the fact that their electroneutrality is ensured by extrinsic charge compensation.57 This means that in the present case the positive charges of PEI are not quantitatively matched by the negative charges of TiO2 (its isoelectric point is at around 4.7
6.258) and that the film remains electroneutral through the incorporation of counteranions from the electrolyte 7939

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir

ARTICLE

Figure 5. ATR-FTIR spectra of the (PEI-[Ti(lac)2(OH)2]2
)6 film (solid black line), the same film placed in contact with 1 mM potassium hexacyanoferrate for 10 min (þ) and subsequently with 31 mM iron III chloride for 30 min (solid blue line).

or from the lactate anion obtained through the decomplexation of [Ti(lac)2(OH)2]2
. To check the possibility of the presence of lactic acid, the chelating anion of Ti(IV) in the [Ti(lac)2(OH)2]2
complex anion in the r-PEM films, we performed some ATR
FTIR experiments (Figure 5). It appeared that in addition to the peaks attributed to TiO2 (at 1123 and 1058 cm
1 59) there was an intense peak centered at 1620 cm
1 as a result of lactate anions. Unfortunately, we were not able to quantify the concentration of lactic acid in these films owing to the evanescent nature of the wave penetrating from the ZnSe crystal in the film. When the film was placed in contact with a 1 mM hexacyanoferrate-containing buffer for 10 min, a strong peak attributed to the
CN bond was apparent at 2036 cm
1 whereas the band at 1620 cm
1 has decreased in intensity. This suggest, as in other polyelectrolyte multilayer films,49 that the electroactive anions are trapped in the architecture through an ion-exchange mechanism. The decrease in retained hexacyanoferrate anions upon an increase in the NaCl concentration (inset of Figure 4) suggests that there are fewer available amino sites for anion exchange or that the film structure and porosity have changed in such a way that the electroactive anions can no longer have access to these sites. Indeed, observation of the film morphology by means of TEM suggests that the density of the inorganic component was increased when the sodium chloride concentration of the buffer used to deposit the film was increased (Figure 6). This means also that an increasing number of negatively charged TiO
groups is available to compensate for the positive charge of PEI and that the amount of intrinsic charge compensation increases with the salt concentration, decreasing the ability of the films to act as anion-exchange membranes. Owing to the fact that an increase in ionic strength during the deposition of r-PEM films simultaneously induced an increase in film thickness (Figure 3) and a decrease in the amount of bound hexacyanoferrate (Figure 4), we will focus on the films prepared at an intermediate NaCl concentration, namely, at 150 mM where their retention of hexacyanoferrate is almost unchanged with respect to the maximal value (inset of Figure 4). For such optimal conditions, we investigated the ability of the films to act

Figure 6. TEM micrographs of (PEI-[Ti(lac)2(OH)2]2
)6 films prepared in the presence of 50 mM Tris buffer (pH 7.50) in the presence of NaCl at different concentrations. The scale bar is the same, 500 nm, for all micrographs.

as a reactor to allow us to produce Prussian Blue when hexacyanoferrate is already present. We also wanted to investigate the distribution of TiO2 in these films. In addition, the composite films containing TiO2, PEI polymer, and Fe(CN)64
anions were placed in contact with D2O containing a large excess of FeCl3 (31 mM). After 30 min of incubation with this solution, the film was rinsed with 50 mM Tris buffer (containing 150 mM NaCl) and its IR spectrum was acquired. Figure 5 shows that the elongation band attributed to
CN groups is shifted to higher wavenumbers, which is usually associated with the formation of a CN-Fe-CN- network and hence with Prussian Blue.60 This will be investigated in more detail in a forthcoming report and will open a route to the easy production of thin films displaying photomagnetic behavior. Let us now consider the nature of the TiO2 phase and the distribution of the inorganic material along the film thickness. Such knowledge is of fundamental importance to future applications of the films, such as in photomagnetic coatings and 7940

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir

ARTICLE

Table 2. Interplanar Distances for the TiO2 Nanoparticles Deduced from SAED Patterns, Compared to the Expected Ones for an Ideal Anatase Phase experimental distances from SAED (Å)

3.58

2.38

1.91

1.66

theoretical distances (in Å) for the anatase phase and corresponding Miller indices

3.51 (101)

2.33 (112)

1.89 (200)

1.66 (211)

Figure 7. (A) Typical selected-area electron diffraction (SAED) pattern taken on the area illustrated in the TEM image (B) corresponding to the film prepared in the presence of 0.15 M NaCl. (C) EELS spectrum recorded in the energy range containing the Ti-L23 and O-K edges. (Inset) EELS signal after the background extraction.

photovoltaic devices. Because the electron density of the TiO2 compound is much larger than that of the PEI polymer, the dark areas in the bright-field TEM images presented in Figure 6 (for the three analyzed layers) are assigned to TiO2 . By comparing the TEM images taken on the samples with different NaCl concentrations, we can observe that the relative amount of the inorganic component increases with the ionic strength of the buffer used to prepare the film. In addition, a direct inspection of the TEM images suggested that the TiO2 nanoparticles had a roughly rounded shape and were grouped in aggregates. This typical aggregation was confirmed by the analysis of tomographic reconstructions obtained on one of the films, as one can see hereafter. The electron diffraction measurements performed on the three analyzed films demonstrated the random crystallographic orientation of the individual TiO2 nanoparticles and the preferential formation of the anatase phase. Indeed, by comparing the interplanar distances deduced experimentally from the SAED patterns to the theoretical ones, very good agreement can be observed (Table 2). A typical SAED pattern made of polycrystalline rings corresponding to the most intense diffraction spots is shown in Figure 7.

In addition, the EELS spectrum recorded in the energy range containing the O-K and Ti-L23 edges, shown in Figure 7c, is characteristic of the anatase phase. More precisely, the determination of the O/Ti ratio and the fine structure of the spectrum at the Ti-L23 edge emphasized the presence of TiO2 anatase, thus confirming the results obtained by SAED analysis. Note that, as in our previous study,36 the anatase phase was obtained without any thermal treatment. To obtain information on the spatial distribution of the TiO2 nanoparticles within the film, a tomographic analysis was performed on the film prepared in the presence of 0.15 M NaCl. The results are summarized in Figure 8, which presents one of the typical TEM images from the tilt series used to reconstruct the volume of the analyzed area and several slices extracted from the reconstruction. Note that these cross-sectional slices are extracted from the reconstruction perpendicular to the surface of the film. This allowed us to visualize the film in cross section and to extract some particular characteristics, such as the mean film thickness, about 100 nm, which does not fluctuate much laterally along the film, confirming its homogeneity. In addition, by analyzing in more detail the volume reconstruction, we can 7941

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir

ARTICLE

would have expected a regular distribution of TiO2 along the thickness of the films. These experimental findings are in contradiction to this expectation. They strongly suggest a migration of TiO2 to the film
solution and film
substrate interfaces during the storage of the film. Indeed, the films were stored for several days in an ambient atmosphere between the end of their preparation and the TEM analysis. Note that this explanation is consistent with the decrease in the ζ potential of the PEI ending films as function of time (Figure 2 of the Supporting Information). Even if PEI was the last deposited layer of the (PEI-[Ti(lac)2(OH)2]2
)6-PEI films, we found a charge reversal of the film and a steady-state potential identical to that of a (PEI-[Ti(lac)2(OH)2]2
)6 film, which is possibly explained by a progressive change in the film composition at the solid
liquid interface. Inward diffusion of PEI in the film and outward diffusion of TiO2 nanoparticles could explain the origins of both the ζ-potential change with time for PEI-ending films and the inhomogeneous particle distribution along the direction perpendicular to the film surface. More investigations are needed to understand such migration phenomena in r-PEM films. The present observation is of the highest interest for the preparation of films presenting gradients in nanoparticles along the film thickness.

Figure 8. Cross-section slices taken perpendicular to the surface of the film, showing that the TiO2 nanoparticles with a size larger than 2 nm are located mainly on the two surfaces, namely, at the substrate film and substrate gas-phase interfaces. The slices were reconstructed from images at different positions along the film. The upper and lower arrows indicate the location of the film
air and film
substrate interfaces, respectively. The (PEI-[Ti(lac)2(OH)2]2
)10 films were deposited from a 50 mM Tris buffer containing 150 mM NaCl.

observe that the most of TiO2 nanoparticles are located on the two surfaces of the layers. A statistical analysis performed on all of the nanoparticles contained inside the analyzed area allowed us to determine their mean size, about 6 to 7 nm. However, small, dark areas can also be observed within the layer, with a size lower than 2 nm, which could be associated with some very small TiO2 nanoparticles The presence of negatively charged TiO2 inside the film is essential to ensuring its cohesion. Indeed, because PEI is positively charged, the polymer chains have to be surrounded by negatively charged particles. The inhomogeneous distribution of anatase along the direction perpendicular to the film surface is a fascinating and highly unexpected result because the film was produced by a regular step-by-step deposition consisting of the alternating deposition of the polycation and the [Ti(lac)2(OH)2]2
anion. Hence, one

’ CONCLUSIONS In this investigation, we studied the influence of the NaCl concentration on the growth of reactive PEM films. It appears that the thickness increment per layer pair increased in a linear manner when the NaCl concentration was increased up to 1 M. The number of retained hexacyanoferrate anions in the films suddenly decreased when they were deposited from solutions above 200 mM in NaCl. We attribute this finding to a higher incorporation of TiO2 as suggested from TEM micrographs. The higher mass fraction in TiO2, up to 60% for films built at 0.5 M in NaCl, leads to an increase in the negative charge density of the film and a concomitant decrease in its anionic exchange capacity. The optimal ionic strength for the synthesis of Prussian Blue by postincubation with a solution containing Fe3þ cations was 0.15 M. Under these conditions, TiO2 was present in the form of anatase and its distribution across the film was not homogeneous as found by means 3D tomography. This last finding is of the highest interest because the r-PEM films presented a gradient in their nanoparticle distribution even if they were produced in a step-by-step manner with a constant thickness increment for each deposited layer pair. ’ ASSOCIATED CONTENT

bS

Supporting Information. Buildup of a (PEI-[Ti(lac)2(OH)2]2
)n r-PEM film in the presence of Tris and NaCl. Evolution of the zeta potential with time for a (PEI-[Ti(lac)2(OH)2]2
)6-PEI7 and of a (PEI-[Ti(lac)2(OH)2]2
)6 film in the presence of Tris buffer. Thickness change of a (PEI-[Ti(lac)2(OH)2]2
)m film as measured by ellipsometry and absorbance change at 240 nm of a (PEI-[Ti(lac)2(OH)2]2
)m film as measured by UV
vis spectroscopy. Evolution of the slope of the absorbance of the (PEI-[Ti(lac)2(OH)2]2
)m films as a function of the NaCl concentration. Surface topography of (PEI-[Ti(lac)2(OH)2]2
)10 films as measured by AFM. Cyclic voltamograms performed on (PEI-[Ti(lac)2(OH)2]2
)7 films prepared in the presence of Tris-NaCl to follow the release of trapped hexacyanoferrate 7942

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943

Langmuir anions. Release kinetics of hexacyanoferrate anions from a (PEI-[Ti(lac)2(OH)2]2
)6 film. Thermogravimetric analysis. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. )

Present Addresses

Current address: Department for Advanced Materials and Structures, Centre de Recherche Public Henri Tudor, 66 rue de Luxembourg, L-4002 Esch-sur-Alzette, Luxembourg.

’ ACKNOWLEDGMENT We acknowledge Cuong Pham-Huu (LMSPC, Universit
e de Strasbourg) for ATG measurements and Dr. Martin Brinkmann (Institut Charles Sadron, CNRS Unit
e Propre 22) for the SAED experiments. ’ REFERENCES (1) Decher, G.; Hong, J. D.; Schmitt, J. Thin Solid Films 1992, 210
211, 831–835. (2) Keller, S. W.; Kim, H.-N.; Mallouk, T. E. J. Am. Chem. Soc. 1994, 116, 8817. (3) Kotov, N. A.; Dekany, I.; Fendler, J. H. J. Phys. Chem. 1995, 99, 13065. (4) Caruso, F.; Lichtenfeld, H.; Giersig, M.; M€ohwald, H. J. Am. Chem. Soc. 1998, 120, 8523–8524. (5) Ariga, K.; Hill, J. P.; Ji, Q. Phys. Chem. Chem. Phys. 2007, 9, 2319–2340. (6) Fahrat, T. R.; Schlenoff, J. B. Electrochem. Solid State Lett. 2002, 5, B13–B15. (7) GRigoriev, D. O.; K€ohler, K.; Skorb, E.; Shchukin, D. G.; M€ohwald, H. Soft Matter 2009, 5, 1426–1432. (8) Kleinfeld, E. R.; Ferguson, G. S. Chem. Mater. 1995, 7, 2327–2331. (9) Toutianoush, A.; Krasemann, L.; Tieke, B. Colloids Surf., A 2002, 198
200, 881–889. (10) Toutianoush, A.; Tieke, B. Mater. Sci. Eng., C 2002, 22, 135–139. (11) Stair, J. L.; Harris, J. J.; Bruening, M. L. Chem. Mater. 2001, 13, 2641–2648. (12) Adusumili, M.; Bruening, M. L. Langmuir 2009, 25, 7478–7485. (13) Lvov, Y. M.; Lu, Z.; Schenkman, J. B.; Zu, X.; Rusling, J. F. J. Am. Chem. Soc. 1998, 120, 4073–4080. (14) Liu, H.; Hu, N. J. Phys. Chem. B 2006, 110, 14494–14502. (15) Weidinger, I. M.; Murgida, D. H.; Dong, W-f.; M€ohwald, H.; Hildebrandt, P. J. Phys. Chem. B 2006, 110, 522–529. (16) Rusling, J. F.; Hvastkovs, E. G.; Hull, D. O.; Shenkman, J. B. Chem. Commun. 2008, 141–154. (17) Moriguchi, I.; Fendler, J. H. Chem. Mater. 1998, 10, 2205–2211. (18) Liu, S.; M€ohwald, H.; Volkmer, D.; Kurth, D. G. Langmuir 2006, 22, 1949–1951. (19) Zhai, L.; Cebeci, F. C.; Cohen, R. E.; Rubner, M. F. Nano Lett. 2004, 4, 1349. (20) Zhai, L.; Berg, M. C.; Cebeci, F. C.; Kim, Y.; Milwid, J. M.; Rubner, M. F.; Cohen, R. E. Nano Lett. 2006, 6, 1213–1217. (21) Mamedov, A. A.; Kotov, N. A.; Prato, M.; Guldi, D. M.; Wicksted, J. P.; Hirsch, A. Nat. Mater. 2002, 1, 190. (22) Tang, Z.; Kotov, N. A.; Magonov, S.; Ozturk, B. Nat. Mater. 2003, 2, 413. (23) Podsiadlo, P.; Michel, M.; Lee, J.; Verploegen, E.; Wong Shi Kam, N.; Ball, V.; Lee, J.; Qi, Y; Hart, A. J.; Hammond, P. T.; Kotov, N. A. Nano Lett. 2008, 8, 1762–1770. (24) Lu, C.; D€onch, I.; Nolte, M.; Fery, A. Chem. Mater. 2006, 18, 6204–6210.

ARTICLE

(25) Zhai, L.; Nolte, A. J.; Cohen, R. E.; Rubner, M. F. Macromolecules 2004, 37, 6113. (26) Wu, Z.; Lee, D.; Rubner, M. F.; Cohen, R. E. Small 2007, 3, 1445. (27) Suda, M.; Einaga, Y. Angew. Chem., Int. Ed. 2009, 48, 1754–1757. (28) Dai, Zh.; D€ahne, L.; Donath, E.; M€ohwald, H. J. Phys. Chem. B 2002, 106, 11501. (29) Quinn, J.; Caruso, F. Langmuir 2004, 20, 20–22. (30) Shim, B. S.; Tang, Z.; Morabito, M. P.; Agarwal, A.; Hong, H.; Kotov, N. A. Chem. Mater. 2007, 19, 5467–5474. (31) Wang, T. C.; Rubner, M. F.; Cohen, R. E. Chem. Mater. 2003, 15, 299–304. (32) Chia, K. K.; Cohen, R. E.; Rubner, M. F. Chem. Mater. 2008, 20, 6756–6763. (33) Nam, K. T.; Kim, D.-W.; Yoo, P. J.; Chiang, C.-Y.; Meethong, N.; Hammond, P. T.; Chiang, Y.-M.; Belcher, A. M. Science 2006, 312, 885–888. (34) Shi, X.; Cassagneau, T.; Caruso, F. Langmuir 2002, 18, 904. (35) Caruso, F.; Shi, X.; Caruso, R. A.; Susha, A. Adv. Mater. 2001, 13, 740. (36) Laugel, N.; Hemmerl
e, J.; Ladhari, N.; Arntz, Y.; Gonthier, E.; Haikel, Y.; Voegel, J.-C.; Schaaf, P.; Ball, V. J. Colloid Interface Sci. 2008, 324, 127–133. (37) Coradin, T.; Durupthy, O.; Livage, J. Langmuir 2002, 18, 2331. (38) Sumper, M.; Brunner, E. Adv. Funct. Mater 2006, 16, 17. (39) Laugel, N.; Hemmerl
e, J.; Porcel, C.; Voegel, J.-C.; Schaaf, P.; Ball, V. Langmuir 2007, 23, 3706–3711. (40) Kharlampieva, E.; Slocik, J. M.; Singamaneni, S.; Poulsen, N.; Kr€oger, N.; Naik, R. R.; Tsukruk, V. V. Adv. Funct. Mater. 2009, 19, 2303–2311. (41) O’Regan, M.; Gr€atzel, M. Nature 1991, 353, 737–740. (42) Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K. Science 1996, 272, 704. (43) Yamamoto, T.; Saso, N.; Umemura, Y.; Einaga, Y. J. Am. Chem. Soc. 2009, 131, 13196–13197. (44) Sauerbrey, G. Z. Phys. 1959, 155, 206–222. (45) Ladam, G.; Schaad, P.; Voegel, J. C.; Schaaf, P.; Decher, G.; Cuisinier, F. J. G. Langmuir 2000, 16, 1249. (46) Butt, H.-J; Graf, K.; Kappl, M. Physics and Chemistry of Interfaces, 2nd ed.,Wiley-VCH: Weinheim, Germany, 2006; Chapter 5. (47) Brown, A. P.; Anson, F. C. Anal. Chem. 1977, 49, 1589–1595. (48) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.: John Wiley & Sons: New York, 2001; Chapter 14, p 590. (49) Grieshaber, D.; V€or€os, J.; Zambelli, T.; Ball, V.; Schaaf, P.; Voegel, J.-C.; Boulmedais, F. Langmuir 2008, 24, 13668–13676. (50) Ersen, O.; Hirlimann, C.; Drillon, M.; Werckmann, J.; F. Tihay, F.; Pham-Huu, C.; Crucifix, C.; Schultz, P. Solid State Sci. 2007, 9, 1088–1098. (51) Saxton, W. O.; Baumeister, W.; Hahn, M. Ultramicroscopy 1984, 13, 57–70. (52) Gordon, R.; Bender, R.; Herman, G. T. J. Theor. Biol. 1970, 24, 471–481. (53) Messaoudii, C.; Boudier, T.; Schancez Sorzano, C. O.; Marco, S. BMC Bioinf. 2007, 6, 288. (54) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391. (55) H€o€ok, F.; V€ or€os, J.; Rodahl, M.; Kurrat, R.; B€oni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. Colloids Surf., B 2002, 24, 155. (56) H€ubsch, E.; Fleith, G.; Fatisson, J.; Labb
e, P.; Voegel, J.-C.; Schaaf, P.; Ball, V. Langmuir 2005, 21, 3664–3669. (57) Farhat, T. R.; Schlenoff, J. B. J. Am. Chem. Soc. 2003, 125, 4627–4636. (58) Parks, G. A. Chem. Rev. 1965, 65, 177–198. (59) Ladhari, N.; Hemmerl
e, J.; Ringwald, C.; Haikel, Y.; Voegel, J.-C.; Schaaf, P.; Ball, V. Colloids Surf., A 2008, 322, 142–147. (60) Jaiswal, A.; Colins, J.; Agricole, B.; Delhaes, P.; Ravaine, S. J. Colloid Interface Sci. 2003, 261, 330. 7943

dx.doi.org/10.1021/la201106r |Langmuir 2011, 27, 7934–7943