Polyion Thin Films Prepared by

Aug 15, 2007 - The internal structure of porous TiO2 films prepared by electrostatic ... For a more comprehensive list of citations to this article, u...
0 downloads 0 Views 251KB Size
Download PDF
9860

Langmuir 2007, 23, 9860-9865

Internal Structure of Nanoporous TiO2/Polyion Thin Films Prepared by Layer-by-Layer Deposition R. Kniprath, S. Duhm, H. Glowatzki, N. Koch, S. Rogaschewski, J. P. Rabe, and S. Kirstein* Department of Physics, Humboldt UniVersity Berlin, Newtonstrasse 15, 12489 Berlin, Germany ReceiVed February 9, 2007. In Final Form: June 19, 2007 The internal structure of porous TiO2 films prepared by electrostatic layer-by-layer deposition was investigated. The films were prepared by alternate dipping of solid substrates into dispersions of TiO2 nanoparticles and polycations, polyanions, or pure buffer solution, respectively. The surface charge of the amphoteric TiO2 particles was controlled by the pH of the aqueous dispersions. The morphology of the film surface was investigated by means of scanning electron microscopy. It was found that the surface roughness strongly depends on the polymeric material used for the deposition process but is independent of the ionic strength of the solution or the molecular weight of the polyions. The samples with rough surfaces feature strong light scattering. The porosity and internal structure of the TiO2/ polyelectrolyte films were investigated by adsorption/desorption of dye molecules. A crude estimate yields an internal surface that is up to 160 times the plane surface of the substrate for a film thickness of 1 µm. The composition of the films was investigated by X-ray photoelectron spectroscopy (XPS). Detection of the XPS signal after each deposition step of the first three dipping cycles shows a significant increase of the relative surface coverage of Ti after the TiO2 deposition step and of PSS after the PSS deposition step. For later dipping cycles, such an increase was also detectable but less prominent.

Introduction Dye sensitized solar cells (DSSCs or Graetzel cells) have become a hot topic for research because they are considered to be a promising alternative to first generation silicon based technology, offering a low cost and high flexibility in the choice of components.1-5 The DSSCs have been attracting much attention6 since power conversion efficiencies exceeding 10% were demonstrated.3 This type of second generation photovoltaic device is based on a photoinduced charge separation at the surface of a TiO2 electrode, which is sensitized to the visible range of light by an adsorbed dye molecule. The photoexcited electron is supposed to move through the TiO2 electrode to the external metal contact, while the hole is transported to the counter electrode by an ionic current that is carried by an electrolyte. In general, the efficiency of solar cells depends on the photovoltage and photocurrent densities generated by incident light. While photovoltage is determined by the composition of materials that contribute to the charge separation, the photocurrent density is directly related to the density of absorbers within the active layer of the cell. In DSSCs, the active layer is usually composed of a TiO2 film coated with sensitizing dye molecules.1-3 A high density of absorbers is achieved by using a porous structure of TiO2 that generates a large internal surface area per cell surface area. A further means to raise the density of photogenerated charge carriers is to increase the path length of the incident light inside the active layer of the cell. However, since the charge carriers need to reach the conducting substrate to contribute to the photocurrent, the free charge carrier diffusion length sets a limit for the thickness of the absorbing layer. Another more * Corresponding author. E-mail: [email protected]. (1) Graetzel, M. Nature 2001, 414, 338. (2) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Mueller, E.; Liska, P.; Vlachopoulos, N.; Gra¨tzel, M. J. Am. Chem. Soc. 1993, 115, 6382. (3) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737. (4) Bisquert, J.; Cahen, D.; Hodes, G.; Ru¨hle, S.; Zaban, A. J. Phys. Chem. B 2003, 108, 8106. (5) Hagfeldt, A.; Gra¨tzel, M. Chem. ReV. 1995, 95, 49. (6) Nanu, M.; Schoonman, J.; Goossens, A. Nano Lett. 2005, 5, 1716.

efficient means to increase the path length of light within the absorbing layer is to induce backscattering of the incident light. This has been achieved with an additional scattering layer on top of the active layer and with scattering centers inside the layer.7 Other important factors for the functionality of DSSCs are photochemical stability of the sensitizer and conductivity of the transport layer for charge carriers to the counter electrode. Recently, new types of TiO2 based cell designs have been proposed that use more stable counter electrodes and sensitizer materials.8 For example, solid-state electrolytes9,10 or inorganic6 or polymeric hole conductors11-13 were used to replace the liquid electrolyte as a counter electrode, and semiconductor quantum dots14,15 were introduced as sensitizers instead of organic dyes. However, the most crucial component of all these photovoltaic systems is the nanostructured TiO2 film.16 For the production of these thin films, different methods are being used, most of which involve either high temperatures,17 high pressures,18 or complex chemistry.19 A simple process for preparing nano(7) Hore, S.; Vetter, C.; Kern, R. Sol. Energy Mater. Sol. Cells 2006, 90, 1176. (8) Gregg, B. A. J. Phys. Chem. B 2003, 107, 4688. (9) Tokuhisa, H.; Hammond, P. T. AdV. Funct. Mater. 2003, 13, 831. (10) Lowman, G. M.; Tokuhisa, H.; Lutkenhaus, J. L.; Hammond, P. T. Langmuir 2004, 20, 9791. (11) Qiao, Q. Q.; McLeskey, J. T. Appl. Phys. Lett. 2005, 86, 153501. (12) Qiao, Q. Q.; Su, L. Y.; Beck, J.; McLeskey, J. T. J. Appl. Phys. 2005, 98, 94906. (13) Da¨ubler, T. K.; Glowacki, I.; Scherf, U.; Ulanski, J.; Ho¨rhold, H.-H.; Neher, D. J. Appl. Phys. 1999, 86, 6915. (14) Robel, I.; Subramanian, V.; Kuno, M.; Kamat, P. V. J. Am. Chem. Soc. 2006, 128, 2385. (15) Yu, P.; Zhu, K.; Norman, A. G.; Ferrere, S.; Frank, A. J.; Nozik, A. J. J. Am. Chem. Soc. 2006, 110, 25451. (16) Benkstein, K. D.; Kopidakis, N.; van de Lagemaat, J.; Frank, A. J. J. Phys. Chem. B 2003, 107, 7759. (17) Park, N.-G.; Schlichtho¨rl, G.; van de Lagemaat, J.; Cheong, H. M.; Mascarenhas, A.; Frank, A. J. J. Phys. Chem. B 1999, 103, 3308. (18) Lindstro¨m, H.; Holmberg, A.; Magnusson, E.; Malmquist, L.; Hagfeldt, A. J. Photochem. Photobiol., A 2001, 145, 107. (19) Saito, Y.; Kambe, S.; Kitamura, T.; Wada, Y.; Yanagida, S. Sol. Energy Mater. Sol. Cells 2004, 83, 1. (20) He, J.-A.; Mosurkal, R.; Samuelson, L. A.; Li, L.; Kumar, J. Langmuir 2003, 19, 2169. (21) Schulze, K.; Kirstein, S. Appl. Surf. Sci. 2005, 246, 415.

10.1021/la700385v CCC: $37.00 © 2007 American Chemical Society Published on Web 08/15/2007

Internal Structure of Nanoporous TiO2/Polyion Thin Films

structured thin films at room-temperature conditions provides the recently developed layer-by-layer deposition technique,22 which is based on electrostatic attraction between oppositely charged polyelectrolytes. However, this technique was also demonstrated to work well with inorganic charged species such as metal colloids,23,24 or nanoparticles of semiconductors,25 or metal oxides such as TiO2.26 Recently, it was demonstrated20,21,27 that this simple dip coating process can be applied for the fabrication of nanoporous TiO2 films that serve as electrodes in DSSCs. The process combines charged polymers (i.e., polyelectrolytes) and oppositely charged TiO2 nanoparticles in a sequential film built up under ambient conditions, and it offers thickness control on a sub-micrometer scale.20 However, such TiO2 films are of interest for a broad readership because they are used in various other applications such as, for example, hydrophilic coatings28 for biological applications.26,29 In this work, we investigated the role of the polyions used in the layer-by-layer deposition process on the structure and composition of TiO2 films. The surface morphology, the internal structure such as the inner surface per outer surface area, and the material composition of these TiO2 films were investigated by scanning electron microscopy (SEM), dye coating experiments, and X-ray photoelectron spectroscopy (XPS), respectively. Experimental Procedures Materials and Methods. TiO2 nanoparticles (99.9% anatase) with a mean diameter of 40 nm (surface of 38 m2/g) were obtained from Alfa Aesar as a white powder and used as received. Aqueous suspensions with particle concentrations of 2 g/L were obtained by adjusting the pH to 2.4 by titration with hydrochloric acid (25%, Merck) or to a value of 10 by using 0.5 mM solutions of Na2CO3and NaHCO3- (both from Aldrich). Acidic buffer solutions were prepared from HCl and small amounts of citrate (received from Dechant). Poly(styrene sulfonate sodium salt) (PSS, Mw ) 70 000 g/mol), poly(dimethylammonium chloride) (PDAC, Mw ) 400 000500 000 g/mol), and poly(ethylene imine) (PEI, Mw ) 55 000) were obtained from Aldrich and used as received. If not stated otherwise, the polyions were dissolved in 10-2 M aqueous solutions without additional salt. The dye cis-bis(isothiocyanato)bis(2,2′-bipyridyl4,4′-dicarboxylato)-ruthenium(II) (Ru535) was obtained from Solaronix, and rhodamine 6G (R6G) and fluorescein were obtained from Aldrich and used as received. For all aqueous solutions, ultrapure Millipore water was used (18 MΩ resistivity). Absorption spectra were taken using a Shimadzu UV-2101 PC spectral photometer. The AFM measurements were performed using a Nanoscope III (Veeco Instruments). All images were recorded in tapping mode. For XPS measurements, samples were introduced in a custom ultrahigh vacuum (UHV) system (base pressure 2 × 10-11 mbar). Spectra were recorded with a hemispherical analyzer (PHOIBOS 100) at a 20 eV pass energy using Al KR1/2 radiation. No charging or beam damage of samples could be observed. SEM was performed using a JSM 840A microscope with a 30 kV electrode voltage. Film Preparation. As substrates, either plain glass or silicon wafers (covered with a native layer of SiO2) or glass slides covered by a thin layer of indium tin oxide (ITO) (PGO) were used. The (22) Decher, G. Science 1997, 277, 1232. (23) Kotov, N.; Dekany, I.; Fendler, J. H. J. Phys. Chem. 1995, 99, 13065. (24) Schmitt, J.; Decher, G.; Dressick, W. J.; Brandow, S. L.; Geer, R. E.; Shashidhar, S.; Calvert, J. M. AdV. Mater. 1991, 9, 61. (25) Gao, M. Y.; Kirstein, S.; Mo¨hwald, H.; Rogach, A. L.; Weller, H. J. Appl. Phys. 2000, 87, 2297. (26) Lvov, Y.; Ariga, K.; Onda, M.; Ichinose, I.; Kunitake, T. Langmuir 1997, 13, 6195. (27) Agrios, A. G.; Cesar, I.; Comte, P.; Nazeeruddin, M. K.; Gra¨tzel, M. Chem. Mater. 2006, 18, 5395. (28) Kommireddy, D. S.; Patel, A. A.; Shutava, T. G.; Mills, D. K.; Lvov, Y. M. J. Nanosci. Nanotechnol. 2005, 5, 1081. (29) Kommireddy, D. S.; Lvov, Y. M.; Mills, D. K. J. Biomed. Nanotechnol. 2005, 1, 1.

Langmuir, Vol. 23, No. 19, 2007 9861 glass substrates were first cleaned and hydrophilized by immersion for 12 h in a solution of sulfuric acid and hydrogen peroxide (volume ratio 1:1, Piranha solution; Caution! Piranhia solution is extremely acidic) and then immersed for 10 min in a solution of water, ammonia, and hydrogen peroxide (ratio 5:1:1, at 75 °C), and finally, they were rinsed with pure water. On all substrates, one layer of PEI was deposited by immersion in a 10 mM solution for 5 min. After washing with plain water, the samples were prepared according to the following three different protocols. Type A: the sample was alternately dipped into a solution of TiO2 particles (pH 9.9) and a pure buffer solution (pH 2).21 The immersion time was different at the first and the later dipping cycles. For the first cycle, the immersion time was 5 min in the TiO2 and buffer solution, respectively, and at the later cycles, it was 1 min and 20 s. Type B: alternate dipping in the TiO2 solution (pH 2.4) (immersion time 5 min for the first dip and 1 min for the following dips) and the solution of the polyanion PSS (immersion time 5 min for the first dip and 20 s for the following dips). Type C: alternate dipping into the TiO2 solution (pH 9.9) (immersion time 5 min for the first dip and 1 min for the following dips) and the solution of the polycation PDAC (immersion time 5 min for the first dip and 20 s for the following dips). Between every dip, the sample was washed with pure water by immersion into cascade rinse tanks and dried under a flow of nitrogen. A dipping cycle always was a full sequence of dipping as described previously including washing and drying. Dye Loading. For the dye adsorption/desorption experiments, the following protocol was used: the samples were first immersed into the dye solution for several hours to allow for diffusion of the dye into the film and adsorption at the surface. The dye concentration was 0.24 mM in ethanol for Ru535 and 1 mM in isopropanol for R6G and fluorescein, respectively. According to the recipe of Solaronix, the samples were heated initially to 70 °C and the solution to 80 °C, then the samples were immersed and heated for another 2-3 h and cooled overnight. Afterward, the samples were taken out, the remaining solution was removed, and the samples were dried with nitrogen. The dyes were extracted from the film by immersion of the sample in clean solvent (ethanol for Ru535 and isopropyl alcohol for R6G) for longer times, up to 3 days for R6G. The amount of dye that was released was determined by measuring the optical absorption of the solution using a molar extinction of 46 000 L mol-1 cm-1 for Ru535 and 10 500 L mol-1 cm-1 for R6G.

Results and Discussion Film Growth and Surface Morphology. The formation of porous TiO2 films using layer-by-layer deposition of amphoteric TiO2 nanoparticles and polyions was first described by He et al.20 Because of the amphoteric nature of the TiO2 nanoparticles, the films can be prepared using polycations (mostly PDAC or PAH) or polyanions (PSS). Later, it was shown that films can be prepared without using polyelectrolytes as counterions.21 In all investigations, it was seen that the structure and properties of the films depend on the chemical nature of the polyelectrolytes used as counterions in the layer-by-layer deposition process. To clarify the influence of the counterions in a more systematic way, three types of films were prepared using different anionic and cationic materials for the alternating adsorption processes. Type A used negatively charged TiO2 particles without any polyelectrolytes; in this case, the surface charge of the adsorbed layers was altered by immersion into a buffer solution of pH 2.21 Type B used positively charged TiO2 particles combined with polyanionic PSS. Type C used negatively charged TiO2 particles combined with polycationic PDAC. In the first step, the preparation protocol for the films was optimized with respect to the deposition times. It was found that the first deposition cycle (adsorption of TiO2 nanoparticles, washing, drying, adsorption of polyelectrolyte, washing, drying) was most important for the growth of films with a high lateral

9862 Langmuir, Vol. 23, No. 19, 2007

Kniprath et al.

Figure 1. Scanning electron micrographs for samples of type A (TiO2 particles only), type B (TiO2 particles and PSS), and type C (TiO2 particles and PDAC). Films were prepared by 60 dipping cycles. Images of three different magnifications are shown with a scan area of 120 µm × 80 µm (left), 40 µm × 26 µm (middle), and 6 µm × 4 µm (right).

homogeneity. Therefore, during this first cycle, rather long immersion times of 5 min were selected. For the subsequent dipping cycles, the immersion times could be reduced to less than 1 min without a significant change in the deposited material during one cycle and also without changes in the internal film structure. Another important process step was the drying procedure. Blowing the samples with a gentle stream of nitrogen resulted in homogeneous films on a macroscopic scale of squared centimeters. Using these optimized parameters, it was possible to grow films with hundreds of layers and final thicknesses of several micrometers within a few hours. This means that the time needed for film production could be reduced by a factor of 5-10 with respect to the times reported in previous publications.20,21,30 To determine the sample thickness, we removed a narrow stripe of adsorbed material from the substrate by scratching with a knife and measured the height of the edges by AFM. By measuring the film thickness after several succeeding deposition cycles, the thickness increment per deposition cycle could be evaluated. These measurements were performed for several samples at different numbers of deposition cycles. For samples B and C, a mean increase in thickness of 24.0 and 24.4 nm per deposition cycle was measured, respectively. Similar growth rates were reported in previous publications,20,30 but a larger (30) Kommireddy, D. S.; Shashikanth, M. S.; Lvov, Y. M.; Mills, D. K. Biomaterials 2006, 27, 4296.

increase in the thickness for negative TiO2 particles (analogous to type C samples) was reported in ref 20. Here, the thickness increase per dipping cycle was on the order of 50% of the mean diameter of the TiO2 particles, which indicates that at each dipping cycle, a submonolayer of particles was adsorbed. For type A samples, the growth of the films was less regular, and no reproducible growth rate per dipping cycle was observed. The typical overall thickness of the samples after 50 deposition cycles was on the order of 500 nm, which would give a mean increment of 10 nm per cycle. However, this value was scattered by more than 50% for different samples. The growth of type A samples is described in ref 21 in detail, and it was assumed there that this type of film formation was favored by a sol-gel process (i.e., chemical linkage between TiO2 particles by hydrogen bonding). The surface morphology of all samples was investigated using SEM. In Figure 1, micrographs of the three types of samples are presented at different magnifications. All samples show rough surfaces, and agglomerates of nanoparticles can be identified, which have sizes ranging from a few hundred nanometers to micrometers. However, while type A and B samples exhibit a rather homogeneous distribution of TiO2 particles with very few agglomerates, type C samples show a very craggy surface with a heterogeneous distribution of particles coagulated into large scale agglomerates. The most striking feature of this surface roughness is the relation to light scattering. For samples of equal thicknesses, type C samples typically show a much higher opacity than samples of type A or B. This leads us to assume very strong

Internal Structure of Nanoporous TiO2/Polyion Thin Films

light scattering that we attribute to the microstructure observed for type C samples. It is well-known that in the case of layer-by-layer deposition of linear polyelectrolytes, the thickness growth of the film strongly depends on the ionic strength of the polyelectrolyte solutions.31,32 It is assumed that the chain conformation in solution is changed from a stretched to a coiled state due to screening of the charges with an increasing salt concentration. Adsorption of polyelectrolytes in a coiled state then results in thicker films than adsorption of extended chains.31 Therefore, it is an interesting question if the conformation of the polyelectrolyte in solution may also influence the structure and morphology of the TiO2 films. To investigate this effect, similar samples were prepared using polyelectrolyte solutions that contained NaCl salt at concentrations of 1 mol L-1. Although this concentration provides screening lengths of a few angstroms, no difference in the surface structure of the films could be detected. The same result was found when the molecular weight of the polyelectrolytes was changed. Since the resulting SEM images are quite identical to those of Figure 1, they are provided as Supporting Information. It seems that the morphology and structure of the films is not controlled via the physical conformation of the polyelectrolyte chains in solution but via their chemical structure. Dye Loading and Internal Structure Analysis. The different surface roughness of the films could be related to their internal structure. An important quantity that characterizes the internal structure of such porous materials is the internal surface per substrate area. This internal surface was determined by adsorption/ desorption of dye molecules as described in the Experimental Procedures. Three different dye molecules were used: R6G, a cationic molecule, fluorescein in its form as a free base, and the photosensitizer Ru535 (also called N3) that is known to bind specifically to TiO2 surfaces.1-3 It was found that the adsorption/ desorption experiment worked for all samples with the Ru535 dye, but a very different behavior was found for other dyes. Using R6G, only samples of type B could be loaded with a significant amount of dye. In this case, an even larger amount of dye was adsorbed by the film than in the case of using Ru535, and a bright red color was obtained. The other samples, type A and C, virtually did not pick up any color during a 45 min immersion into a solution of R6G. Likewise, all attempts to coat type A-C thin films with fluorescein were not successful. Only spurious amounts of the dye were absorbed by the films. This diverse adsorption behavior of the different dyes may be explained either by electrostatic properties or by specific adsorption. Electrostatic forces could influence the adsorption of the dyes due to a distinct surface potential caused by the outermost layer. However, such electrostatic forces can be excluded because in this case, one would have expected complementary behavior for R6G and fluorescein due to their opposite charges. Within the films, the charges were compensated on a very short length scale that provided electroneutrality and hence did not influence the loading with dye. From these considerations, we conclude that the different adsorption behavior of the dyes is due to specific binding that is most pronounced for Ru535 at TiO2. In the following discussion, only the Ru dye was used to investigate the internal surface of the films. It was a general finding that the amount of dye absorbed per area increased with the film thickness (cf. Figure 2). The increase was linear for type B and C samples, while the relationship was less clear for type A samples, which showed lower thicknesses and smaller amounts of deposited dye. (31) Steitz, R.; Jaeger, W.; von Klitzing, R. Langmuir 2001, 17, 4471. (32) Shiratori, S. S.; Rubner, M. F. Macromolecules 2000, 33, 4213. (33) Yeh, J. J.; Lindau, I. Atom. Data Nucl. Data Tables 1985, 32, 1.

Langmuir, Vol. 23, No. 19, 2007 9863

The linear increase of the amount of adsorbed dye proves that the whole internal surface of type B and C films is accessible for the dye molecules, independent of film thickness; therefore, we call this film structure highly porous. The deviations from a linear relationship observed for type A samples may indicate an obstructed diffusion for thicker samples (>500 nm). From the slope of the graphs in Figure 2, the dye loading per unit sample area and per thickness increment can be determined. These data are listed in column 4 of Table 1. Assuming a mean area of 1 nm2 per molecule occupied by one dye molecule adsorbed at the TiO2 surface, the so-called roughness factor (R.F.) is calculated. This factor describes the surface accessible by dye molecules divided by the substrate area.3 For 1 µm thick films coated with Ru535 dye, the measured roughness factors were 163, 45, and 81 for films prepared with PDAC, PSS, and a pH 2 buffer solution, respectively (cf. Table 1). The highest roughness factors were observed for films of type C, which also exhibit the highest surface roughness in SEM images. The value of 163 observed in this case is approximately 2 times the value reported for TiO2 films that were prepared by the conventional sintering method.3,19 For comparison, the amount of R6G absorbed into films with PSS was significantly higher, yielding a roughness factor of 1240. This value heavily exceeds the value of 200 that can be theoretically predicted for a film built up by monodisperse particles of 15 nm in diameter.3 Therefore, it is likely that the R6G dye was absorbed by the porous film in an amount that exceeds the monolayer coverage (i.e., the pores are filled with a condensed phase of dyes). Surface Composition. The striking differences in surface morphology and dye adsorption for samples prepared using different polyelectrolytes raised the question about the role that the chemical composition of the polyelectrolytes is playing for the extent of polymer incorporation into the thin films. XPS was selected as a method that provides information about the amount of different elements in the film. To enable specific detection of the polymer, mostly samples of type B were investigated because this polymer contains sulfur. The signals of carbon, nitrogen, and oxygen are influenced by contamination from ambient air and are therefore not suitable for composition analysis. The signal from sulfur can be ascribed reliably to PSS since the element is typically not present in surface contaminants. The samples were prepared on ITO covered glass substrates to obtain conductive substrates. Because XPS is sensitive to very thin layers at the surface of a sample, the films were analyzed after each of the first six coating steps of the first three dipping cycles. Additionally, we analyzed the surface composition of a sample after coating with 60 dipping cycles and after one additional PSS coating step. From the XPS spectra, the signals of sulfur, titanium, and indium were used to determine the relative surface concentrations of these elements. We obtained relative atomic concentrations by division of the count rates of photoelectrons with atomic sensitivity factors.33 To calculate the surface composition, we assumed that the surface consisted exclusively of PSS, TiO2, and uncoated ITO substrate. The relative concentrations of these materials were converted into relative area fractions occupied by the respective elements. Taking into account the density of the materials that contain the respective elements (PSS for S, TiO2 for Ti, and ITO for In), a specific volume for each element is evaluated, and one obtains volume ratios of the three types of materials. Assuming that the maximum depth for photoelectron emission without inelastic scattering inside the material is identical

9864 Langmuir, Vol. 23, No. 19, 2007

Kniprath et al.

Figure 2. Amount of dye molecules (Ru535 or R6G) adsorbed per substrate area of the three different films of type A-C with increasing film thickness. The amount of dye loading by the films was determined by redissolving the dyes in organic solvents by immersion of the films in the respective solvent (see text for further details). Table 1. Summary of Amount of Dye Adsorbed at Internal Surface of Nanoporous TiO2 Films Prepared in Three Different Ways and Calculated R.F.a

Figure 2 a b c d

polyelectrolyte (sample type) PDAC (C) PSS (B) pH2 (A) PSS (B)

dye Ru535 Ru535 Ru535 rhodamine 6G

R.F. at adsorbed film amount thickness of dye (mmol/m2/µm) of 1 µm 0.270 0.075 0.134 2.053

163 45 81 1240

a Data were evaluated from graphs of Figure 2 using a mean area per molecule of 1 nm2. See text for more details.

for the three types of materials, one can interpret the volume ratio as a ratio of surface coverage. Figure 3 displays the results for three dipping cycles of a bare substrate (first six data points) and of one dipping cycle of a thick sample pre-coated with 60 dipping cycles (last two data points). The PSS surface coverage was 39% after the first dip into the PSS solution. It decreased by a few percent during the following TiO2 dips and increased again after the PSS dips, so that it fluctuated around ∼40% coverage during the first three dipping cycles. The first three TiO2 dips all yielded increases of roughly 15% TiO2 surface coverage, which was only slightly decreased by the subsequent PSS dips, to reach 44% after the third dipping cycle. The apparent bare ITO substrate decreased with all dipping steps and was finally decreased to 24% after three cycles. For the thick films, the ITO signal had vanished. The PSS and TiO2 fractions reached 40 and 60%, respectively, with a fluctuation of (1.5%. We assume the surface composition of

Figure 3. Summary of XPS of a film of TiO2 nanoparticles and PSS measured after each of the first six deposition steps. Surface area fractions covered by PSS and TiO2 and area fraction of uncovered substrate (ITO) determined by photoelectron count rates from the core levels of S2p, Ti2p, and In3d are drawn vs respective deposition steps. Last two values (right of vertical line) were obtained from last two deposition steps after 60 deposition cycles.

thick films to be independent of thickness. Therefore, we expect the measured surface composition averages and fluctuations to be similar for all thick type B samples. For type C samples, we used the nitrogen concentration to determine the surface coverage of PDAC. Concerning the polymer surface coverage, the results we obtained point in the same direction as the ones for type B samples, but unfortunately, we could not fully separate contributions from PDAC and from

Internal Structure of Nanoporous TiO2/Polyion Thin Films

ambient nitrogen within the nitrogen signature. This obstructed a detailed quantitative analysis of the results, which are therefore not given in detail. The outer surface is the only part accessible for XPS characterization. But, due to the sequential deposition process and the porous film structure, we do not expect any qualitative difference in the chemical composition of the internal surface on the one hand and the outer surface on the other. Therefore, the significant presence of polymer material on the outer surface of the samples leads us to infer that the internal surface has a similar composition.

Conclusion We investigated the structural properties of TiO2 thin films built up with a layer-by-layer self-assembly technique. We alternately dipped glass samples into suspensions of anatase TiO2 nanoparticles and either into solutions of strong polyions or into an acidic buffer solution containing no polymer. The time needed for sample preparation could be drastically reduced by optimization of the preparation conditions, which allowed the fabrication of films with a micrometer thickness. XPS measurements showed that, at least for the case of PSS used as polyanions, the polyions cover a significant fraction of the thin film surface after each production step. Dye adsorption and desorption measurements gave us a quantitative measurement of the internal surface of the films. We found that the adsorbed amount of molecules increased with film thickness. This result indicates that a connected network of nanopores provides for a large internal surface and allows molecules in solution to penetrate deep into the films by diffusion. This process, however, was very sensitive to the dye used for probing and the polyion type used for film production. It is assumed that the polyion mostly determines the internal structure, which causes different internal surfaces. The different loading of diverse dyes is assumed to be due to specific adsorption of the dyes at one of the materials present in the films.

Langmuir, Vol. 23, No. 19, 2007 9865

Scanning electron micrographs revealed that also the surface structure of the films depends sensitively on the choice of the polyions. Interestingly, films prepared using PDAC exhibit micrometer sized agglomerates on their surfaces that we do not observe on other sample types. They also exhibit less transparency, a feature that we ascribe to strong light scattering at the microstructures. From XPS measurements, it was concluded that the polyelectrolytes are incorporated into the films in significant amounts, which is in contradiction to our previous conclusion that was drawn from absorption spectroscopy using dye labeled polyelectrolytes.21 Those measurements were probably falsified by charge-transfer mechanisms or photodegradation of the dyes. The porous structure of the TiO2 films prepared by the layerby-layer deposition technique makes them promising candidates for sensor or photovoltaic applications. They have a large internal surface area, their system of interconnected pores allows for post-production sensitization, and their microstructured surfaces are expected to effectively trap light. All these findings are significant for the production of efficient photovoltaic systems. Acknowledgment. This work was supported by the Deutsche Forschungsgemeinschaft by the Research Training Group 1025 “Fundamentals and Functionality of Size and Interface Controlled Materials: Spin- and Optoelectronics” (R.K. and J.P.R.) and by the Emmy Noether-Program (N.K., H.G., and S.D.), which we gratefully acknowledge. We thank E. Poblenz for her dedicated support in sample preparation. Supporting Information Available: Scanning electron micrographs for samples of type B (TiO2 particles and PSS) and type C (TiO2 particles and PDAC) showing films under different salt conditions of polyelectrolyte solution and for different molecular weights of polyelectrolytes. This material is available free of charge via the Internet at http://pubs.acs.org. LA700385V