Journal of Sol-Gel Science and Technology

, Volume 64, Issue 2, pp 465–479

Thickness effects in naturally superhydrophilic TiO2–SiO2 nanocomposite films deposited via a multilayer sol–gel route

Authors

    • LMGP (Grenoble Institute of Technology)
    • SIMaP (Grenoble Institute of Technology)
  • L. Rapenne
    • LMGP (Grenoble Institute of Technology)
  • P. Chaudouët
    • LMGP (Grenoble Institute of Technology)
  • G. Berthomé
    • SIMaP (Grenoble Institute of Technology)
  • M. Langlet
    • LMGP (Grenoble Institute of Technology)
Article

DOI: 10.1007/s10971-012-2878-4

Cite this article as:
Holtzinger, C., Rapenne, L., Chaudouët, P. et al. J Sol-Gel Sci Technol (2012) 64: 465. doi:10.1007/s10971-012-2878-4

Abstract

TiO2–SiO2 nanocomposite films of various thicknesses have been deposited via a multilayer sol–gel route. These films exhibit a natural and persistent superhydrophilicity, which allows considering new applications for easy to clean surfaces. Atomic force microscopy, scanning and transmission electron microscopy, as well as ellipsometry and UV/visible spectrometry measurements, were performed to study how the multilayer procedure influenced the morphology and composition of composite films in relation to their thickness. The natural and photo-induced wettability of these films was studied and discussed in relation to morphology, composition, and thickness features. It is concluded that, while such features did not significantly influence the natural wettability of nanocomposite films, their photo-induced wettability was considerably enhanced when increasing their thickness, which favored a faster superhydrophilicity photo-regeneration when this natural property started to disappear after a long aging period in ambient atmosphere.

Keywords

SuperhydrophilicityTiO2–SiO2 compositesSol–gel coatingsSurfacesEnhanced cleanability

1 Introduction

Cleanabiliy of material surfaces is a cause for concern for many industrial and domestic applications. In this context, we have performed extensive studies on sol–gel thin films deposited from composite TiO2–SiO2 sols [15]. We have shown that derived nanocomposite TiO2–SiO2 films exhibit a natural superhydrophilicity, i.e., a surface showing a water contact angle of zero or near zero degree. Depending on the experimental conditions, the superhydrophilicity can persist over a long duration of several weeks or months. Since superhydrophilic surfaces have higher affinity toward water than toward carbon contamination, the natural superhydrophilicity of nanocomposite TiO2–SiO2 films introduces a new concept of easy to clean surfaces, i.e. coated surfaces that can easily be cleaned or degreased by water without the need for detergent. Static and dynamic oil wettability experiments enabled us to demonstrate the enhanced cleanability of such superhydrophilic surfaces [4]. We have also shown that, over a very long duration the superhydrophilicity and enhanced cleanability of TiO2–SiO2 composite films progressively disappear, which has been attributed to atmospheric carbon contamination occurring at the film surface [15]. However, owing to the photocatalytic activity of TiO2 nanocrystallites present in the nanocomposite films [6], the superhydrophilicity of TiO2–SiO2 composite films can easily be recovered, since a short exposure to UV light yields the photocatalytic decomposition of carbon contaminants [4, 7]. Thus, our new approach yields important insights for future developments of easy to clean surfaces. Indeed, TiO2-coated surfaces, which are traditionally studied for such applications, exhibit a photoinduced superhydrophilicity only when exposed to UV light [6, 8, 9], and can therefore essentially be exploited in outdoor atmospheres. In contrast, our approach allows us considering easy to clean surfaces whose performances have practically no limitation in the long term and which can be used in atmospheres where UV light is not permanently present, like indoor atmospheres.

It is well known that TiO2–SiO2 composites have a strong catalytic power, which is generally attributed to deprotonated TiOx and/or protonated SiOx+ units formed at TiO2–SiO2 granular interfaces [1016]. As proposed in our previous papers, these localized electric charges may favour a molecular or dissociative adsorption of atmospheric water at the surface of the nanocomposite films, which would enhance natural water wettability properties of the films [15]. Such a mechanism would be similar to what occurs at the surface of TiO2 coatings exposed to UV light, i.e. a photoinduced superhydrophilicity of TiO2 surfaces resulting from photogenerated charges. In the case of TiO2–SiO2 mixed films deposited from polymeric TiO2 and SiO2 sols, we previously showed that such enhanced properties are never observed when TiO2 is not present in the form of anatase crystallites, i.e. when titanium and silicon are mixed at the molecular scale [1]. In contrast, all recent works performed using anatase crystallites preformed in solution, which are the subject of the present paper, systematically yielded enhanced wettability composite films [15]. This feature illustrates the intrinsic effect of such crystallites and reinforces the assumption that enhanced wettability properties arise from the presence of TiO2–SiO2 granular interfaces. In our previous paper, relations between interface effects and the TiO2–SiO2 sol composition were investigated by X-Ray Photoelectron Spectroscopy (XPS) measurements [3]. The results showed that it was not possible to precisely quantify the amount of Ti–O–Si bonds in composite films since, as these bonds are only located at TiO2–SiO2 granular interfaces, their amount is necessarily very weak. However, XPS measurements clearly indicated shifts of the O1s, Si2p, and Ti2p peaks in relation to the TiO2–SiO2 sol composition, which gave evidence of Ti–O–Si bonds. These shifts are explained by electronic interactions at TiO2–SiO2 granular interfaces, which would be induced by an interfacial substitution of Si(Ti) atoms by Ti(Si) atoms less (more) electronegative and more (less) polarizable within a SiO2 (TiO2) host network.

Beside granular interfaces effects, other physico-chemical and structural properties can influence the wettability of TiO2–SiO2 composite films. For instance, according to De Gennes et al. [17], morphological properties such as surface roughness and open porosity can modify the water contact angle, emphasizing the water wettability on the surface of rough hydrophilic films, or enhancing water impregnation within the thickness of porous films. The effects of roughness and open porosity on the wettability enhancement of naturally hydrophilic surfaces have been described by theoretical models developed by Wenzel [18] and Bico [19], respectively. In particular, the surface open porosity can induce roughness effects described by the first model and can also promote water impregnation features described by the second one. In the latter case, the model assumes that the surface behaves like a composite one, one component being the dry solid itself, the other one being constituted of impregnated water regions. In this case, the enhanced wettability of such porous surfaces relies on a bi-regime mechanism, i.e. water firstly impregnates the surface pores and can then preferentially spread at the surface of impregnated regions. Little research has been done to evaluate to which extent morphology effects, in relation to composition and thickness features, may influence the water wetting of composite surfaces. Thus, a better understanding of the superhydrophilicity observed in TiO2–SiO2 nanocomposite films requires complementary studies. The objective of this work was to investigate how a multilayer deposition procedure may influence the morphology, composition, and thickness of TiO2–SiO2 nanocomposite films, and how these properties can in turn influence the natural and photoinduced wettability of these films.

2 Experimental

2.1 Preparation of TiO2–SiO2 composite sols

Titania–silica composite sols were deposited from mixtures of TiO2 anatase crystalline suspensions (CS) and polymeric SiO2 sols. A polymeric TiO2 mother solution (MS) was first prepared by mixing tetraisopropyl orthotitanate (TIPT) with deionised water, hydrochloric acid, and absolute ethanol as solvent [20]. The TIPT concentration in the solution was 0.4 M, and the TIPT/H2O/HCl molar composition was 1/0.82/0.13. The solution was aged at room temperature for 2 days before use. Then, a CS composed of TiO2 nanocrystallites in absolute ethanol was prepared from the MS using a multi-step procedure that has been previously detailed [21]. Briefly, the MS was first diluted in an excess of deionised water (H2O/TIPT molar ratio of 90) and autoclaved at 130 °C for 6 h. Autoclaving yielded the crystallization of TiO2 particles diluted in an aqueous medium. An exchange procedure was then performed in order to remove water from the sol and form a CS diluted in absolute ethanol. The CS was composed of TiO2 primary nanoparticles of about 5–6 nm in diameter, crystallized in the anatase phase, and aggregated in polycrystalline secondary particles of 50–100 nm in size [21]. This procedure yielded a final TiO2 concentration of 0.24 M in the CS. The coatings resulting from this solution showed excellent optical qualities and the CSs exhibited a good stability over aging since no significant aggregation of TiO2 crystallites occurred in the short term [21]. Accordingly, the CSs could be stored at room temperature for days or weeks before being used in reproducible conditions for pure TiO2 film deposition or for mixing with a silica sol to form TiO2–SiO2 composite sols.

In our previous papers, we studied how the sol–gel reactivity of composite sols can influence the superhydrophilicity of derived films [2, 3]. We have shown that the best wettability properties were obtained with composite sols formulated from very weak reactivity silica sols where hydrolysis/polycondensation sol–gel reactions yielded short Si–O–Si oligomeric chains. This favoured a greater amount of TiO2–SiO2 granular interfaces and promoted an enhanced film superhydrophilicity. In the present work, according to previously optimized formulations [2, 3], silica sols were elaborated from mixtures of tetraethoxysilane (TEOS), deionised water, hydrochloric acid, and absolute ethanol as solvent. In a first step, the TEOS concentration was fixed at 2.4 M, and the TEOS/H2O/HCl molar composition was 1/0.5/0.004. The silica sol was then heated in oven for 48 h at 60 °C and then diluted in ethanol at a concentration of 0.24 M. Finally, CSs and silica sols were mixed in 20, 40, 60, and 80 Si/(Si + Ti) molar ratios. So-obtained composite sols were magnetically stirred for 30 min at room temperature, before being used for the deposition of composite films.

Pure TiO2 and SiO2 films were also deposited from the CS and from a highly reactive silica sol, respectively. As explained before, silica sols used for the preparation of composite ones exhibited very weak sol–gel reactivity. This weak reactivity can promote a liquid dewetting on the substrate surface because of insufficiently developed polymeric Si–O–Si chains that are not able to counteract surface tension effects during the post-deposition liquid–solid (sol–gel) transformation [2, 3]. Dewetting effects are all the more prone to occur when the silica component becomes preponderant in the composite sols [5]. Accordingly, total dewetting was observed when a pure silica sol of poor reactivity was used. For this reason, pure SiO2 films of reference were deposited from a stronger reactivity sol adapted from our previous works [2, 3]. In this sol, the TEOS concentration was fixed at 1.5 M, and the TEOS/H2O/HCl molar composition was 1/2.2/0.00013. Here again, the silica sol was heated at 60 °C during 48 h and then magnetically stirred for 30 min at room temperature before deposition.

2.2 Film deposition

Thin film deposition was performed on 3 × 3 cm² (100) silicon wafers by spin-coating (300 μL of sol, spin-speed of 3,000 rpm) at room temperature. Prior to the deposition, the wafers were pre-heated in air at 500 °C for 2 h in order to pre-oxidize the silicon surface and eliminate any organic dust at the surface. After cooling, the substrates were simply cleaned with ethanol and then dried with air spray. As will be illustrated below, this overall procedure generally yielded pure and composite films of good homogeneity and optical quality. Soda lime glasses were also used as substrates to performed UV/visible transmission measurements. Films of various thicknesses were deposited using a multilayer procedure. After each deposition operation, the film was stabilized by annealing in air at 500 °C for 5 min before a new deposition operation. After single-layer or multilayer depositions, pure and composite films were finally annealed in air at 500 °C for 2 h. In this paper, the SX-n nomenclature stands for films consisting of n single-layers (n = 1, 3, or 7) deposited from a composite sol with a Si/(Si + Ti) ratio of X mol %.

2.3 Characterizations

UV/visible transmission spectra were collected using a Perkin Elmer Lambda 35 spectrophotometer in the 190–1,100 nm spectral range on films deposited on soda lime glass substrates. All other characterizations were performed on films deposited on silicon wafers. The film morphology was characterized by surface imaging and RMS roughness measurements using an Atomic Force Microscope (AFM, Digital Instrument Nanoscope Multimode) in tapping mode. RMS roughness values were estimated from AFM images by spectral analysis on 1 μm2 areas. Surface imaging was also performed by Field Emission Gun-Scanning Electron Microscopy (FEG-SEM) using a FEI QUANTA 250 ESEM FEG operated at 10 or 15 keV. Cross-section high-resolution Transmission Electron Microscopy (TEM) observations were performed using a JEOL-2010 LaB6 instrument operated at 200 keV. Energy Dispersive X-ray (EDX) analyses were performed during TEM observations to determine the Si/Ti atomic ratio from intensities of the Si (1.74 keV) and Ti (4.51 keV) peaks. Probed areas had a typical size of around 50 nm. XPS measurements were performed using a XR3E2 apparatus from Vacuum Generator employing an Mg Kα source (1,253.6 eV). The X-ray source was operated at 15 kV for a current of 20 mA. Before collecting data, the samples were put in equilibrium for 12 h in an ultra high vacuum chamber (10−10 mbar) to control desorption of the studied samples. Photoelectrons were collected by a hemispherical analyzer at a constant take-off angle of 90°. The film thicknesses and refractive indices were measured with a Gaertner L116B ellipsometer operated at 633 nm wavelength. According to our previous works, the volume porosity of the films was derived from their refractive index using the Lorentz-Lorenz relationship and taking into account the refractive index of the bulk material [3]. For composite films, the bulk material refractive index has been derived from a Lorentz-Lorenz mixing law taking into account the molar fraction of TiO2 and SiO2 in the films [3].

The hydrophilicity of the films surfaces was quantified from measurements of the water contact angle using the sessile drop method. Experiments were performed at 20 °C within an environmental chamber using a KRUSS G 10 goniometer connected to a video camera. Several deionised water droplets (pH = 5.7) of 0.5 μL volume were spread at different points on the samples surfaces where water contact angles were measured for statistical purpose. Then, samples were stored in the dark under ambient condition in open atmosphere. During aging, water contact angles were periodically measured using the previously mentioned statistical procedure in order to study the effects of natural aging on the film wettability. Superhydrophilicity photo-regeneration experiments were performed using a 100 W black light mercury lamp (PAR38, Sylvania), principally emitting at a 365 nm wavelength, placed on the top of the environmental chamber and illuminating the samples through a quartz window. Before UV exposure, the films were copiously rinsed with water and then dried with air spray. As previously shown, this pre-treatment does not modify the natural film wettability but favours a faster superhydrophilicity photo-regeneration [4].

3 Results and discussion

3.1 Morphology features

FEG-SEM and AFM images of SX-n films are illustrated in Figs. 1 and 2, for single-layer and 7-layers films, respectively. Previous studies showed that single-layer and multilayer TiO2 (S0-n) films deposited from a pure CS exhibited a good optical quality (not illustrated here), which relies on a fine and homogeneous granular morphology [21]. This morphology is illustrated in Figs. 1a, 2a and 1b, 2b, where FEG-SEM and AFM images show that the multilayer deposition procedure does not modify morphological features of pure TiO2 films. In particular, a RMS roughness of 3.5 ± 1 nm was deduced from AFM analyses for all these films, irrespectively of the number of deposited layers. According to the AFM analyses performed in this work, a roughness variation of ±1 nm was considered to depict the experimental error, i.e. the roughness uniformity for a given sample and the repeatability between samples elaborated in similar conditions. Figure 3 and its insert show the UV/visible transmission spectra of S20-1, S20-7, S60-1, and S60-7 composite films deposited on glass substrates. Above a 600 nm wavelength, the spectra of all coated glass substrates fairly fit the bare substrate spectrum. This observation firstly depicts the good homogeneity and transparency of the films, which shows that the deposition of TiO2 particles capped with oligomeric silica chains does not particularly alter the optical quality of composite films whatever the SiO2 content and the number of deposited layers. It also means that no significant interference fringes are observed in the coated glass spectra, as it would be the case if the mismatch between the substrate and film refractive indices induced multi-reflexions at the film/air and film/substrate interfaces [21]. This second assertion suggests that the composite films exhibit a refractive index close to that of the glass substrate (1.52). Below 600 nm, two spectral regions can be appreciated, as illustrated in the insert of Fig. 3. Below 400 nm, the spectra are dominated by absorption features arising from the energy band gap of TiO2 crystallites present in the composite films (the band gap is 3.2 eV for bulk anatase, which corresponds to a wavelength of 380 nm). Accordingly, the absorption edge of S20-n and S60-n films is similarly red-shifted when the number of deposited layers increases from 1 to 7, which depicts the stronger UV light absorption of multilayer films caused by a greater number of deposited TiO2 crystallites. Between 400 and 600 nm, the insert of Fig. 3 illustrates some additional spectral features, which can no longer be attributed to absorption by TiO2 crystallites. On the one hand, the transmission level of glass substrates coated with S20-n films is rather close to that of the bare substrate, whatever the number of deposited layers. On the other hand, compared to the bare substrate spectrum, the transmission level of substrates coated with S60-n films exhibits a certain decrease with decreasing wavelengths, and this effect is all the more marked as the number of deposited layers increases. This observation depicts that, for Si-rich composite films, light diffusion phenomena cause some slight optical losses at short wavelengths in the visible spectral range, which can in turn be attributed to morphological features.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig1_HTML.jpg
Fig. 1

FEG-SEM images for S0-1 (a), S0-7 (b), S20-1 (c), S20-7 (d), S60-1 (e), and S60-7 (f) films. The scale bar depicted in a also accounts for the other images. Arrows shown in e, f are pointing at surface cavities

https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig2_HTML.jpg
Fig. 2

AFM images for S0-1 (a), S0-7 (b), S20-1 (c), S20-7 (d), S60-1 (e), and S60-7 (f) films. The scale bar depicted in a also accounts for the other images. Arrows shown in e, f are pointing at surface cavities

https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig3_HTML.gif
Fig. 3

UV/visible transmission spectra for a bare glass substrate (filled circle) and for glass substrates coated with SX-1 (dotted lines) and SX-7 (full lines) composite films deposited from a S20 (open triangle) and a S60 (open square) sol. The inset shows a magnified view of these spectra in the short wavelengths region

As evidenced by brightness and colour contrasts in FEG-SEM and AFM images in Figs. 1c, 2c and in Fig. 1d, 2d for S20-1 and S20-7, respectively, the number of deposited layers neither significantly influenced the morphology of Si-poor composite films nor their RMS roughness. However, compared to pure TiO2 ones, the composite films exhibited a rougher surface characterized by a RMS roughness of 7 nm ± 1 nm. In contrast to pure TiO2 films and Si-poor composite ones, Si-rich composite films exhibited thickness dependent morphology features, as illustrated in Figs. 1e, 2e and Figs. 1f, 2f for S60-1 and S60-7 films, respectively. The S60-1 film exhibited a very specific sponge-like morphology characterized by large surface cavities inducing local variations of the film thickness. These surface cavities are illustrated by brightness or colour contrasts in the FEG-SEM and AFM images of Figs. 1e, 2e (see arrows). The surface cavities may also be present in the S20-1 film (Figs. 1c, 2c), but to a weaker extent and with much smaller diameter and depth. We previously discussed the presence of surface cavities in relation to the silica content for single-layer (SX-1) composite films [5]. The formation of these films arises from hetero-condensation reactions occurring between short oligomeric Si–O–Si chains and TiO2 nanocrystallites. These reactions yield Ti–O–Si bridging bonds resulting in the deposition of nanocomposite films constituted of TiO2 particles capped with oligomeric Si–O–Si chain. Yet, since the silica component arises from a poorly reactive sol, silica chains are not sufficiently developed to counteract surface tension effects taking place during the liquid–solid transformation. Thus, dewetting effects can occur during the post-deposition solvent evaporation, and may in turn influence the morphology of SX-1 films. We previously showed that these dewetting effects are all the more pronounced as the Si amount increases and the Si + Ti total concentration decreases in composite sols [5]. In extreme cases, dewetting promoted a partial denuding of the substrate, or even a total denuding. For SX-1 films studied in the present work, dewetting effects only promoted the formation of cavities at the surface of Si-rich composite films, but did not yield any partial denuding of the substrate even for S80-1 films [5]. EDX analyses performed in different places of Si-rich composite films during TEM cross section observations also indicated a very homogeneous Si–Ti atomic distribution, which indicated that, despite the formation of surface cavities, Si and Ti species were uniformly distributed over the entire substrate surface [5]. Increasing the number of layers deposited from a Si-rich composite sol promoted significant enhancement in surface cavity features. Indeed, compared to FEG-SEM and AFM images of the S60-1 film (Figs. 1e, 2e), those of S60-7 film (Figs. 1f, 2f) exhibited much larger cavities. These morphology modifications also promoted thickness dependent roughness evolutions. Indeed, while the multilayer deposition procedure did not significantly influence the RMS roughness of S20-n composite films (7 nm ± 1 nm), we measured a RMS roughness of 9 nm ± 1 nm on S60-1 composite films, but this value jumped up to 14 ± 1 nm as soon as the number of layers was superior to 3. All these observations can in turn explain that S20-n films are almost not affected by optical losses, whatever their thickness, while surface cavities can cause some slight diffusion losses at short wavelengths in the spectra of S60-n films and these losses increase with increasing film thicknesses.

3.2 Cross-section TEM/EDX studies

TEM studies were performed on different SX-7 composite films. Dark field images (Figs. 4a, 4b for a S20-7 film and a S60-7 film, respectively) indicated the uniform distribution of TiO2 crystallites through the film thickness for all TiO2–SiO2 compositions studied in this work (see white spots). The mean diameter of TiO2 particles fairly agreed with that of primary crystallites formed in the CS (5–6 nm). These observations show that: (1) the silica sol is able to impregnate TiO2 aggregates formed in the CS and efficiently cap primary TiO2 single crystals, which probably reinforces the formation of TiO2–SiO2 granular interfaces, (2) capping of TiO2 nanocrystallites with a silica shell prevents any thermal growth of these crystallites during the post-deposition treatment at 500 °C, and (3) the multilayer deposition procedure does not disturb the uniform distribution of TiO2 particles through the film thickness irrespectively of the TiO2–SiO2 sol composition.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig4_HTML.jpg
Fig. 4

Cross-section dark field TEM image of a S20-7 (a) and a S60-7 (b) film

In contrast, it appeared that the sol composition significantly influenced the distribution of Si species through the thickness of multilayer films. It is illustrated in Figs. 5a, 5b by cross-section EDX measurements realized during the TEM observation of S20-7 and S60-7 films, respectively. Figure 5a shows that the Si/Ti atomic ratio deduced from EDX measurements remains fairly constant through the whole thickness of the S20-7 film. This observation depicts an homogeneous distribution of both Ti and Si species, even if the Si/Ti atomic ratio of 0.40 ± 0.04 determined from these measurements is quite greater than the value expected from a S20 composite sol (Si/Ti = 0.25). This indicates a certain excess of Si species and the actual formation of a 70TiO2–30SiO2 composition film instead of the 80TiO2–20SiO2 expected composition. For comparison, the EDX measurements performed in different places of a S60-1 (single-layer) composite film indicated a composition that fairly agreed to that of the composite sol [5]. These features suppose a certain lack of control in the synthesis conditions of Si-poor composite films, and will be analyzed in next sections. In contrast, EDX measurements performed on different places of a S60-7 film evidenced a marked Si/Ti atomic ratio gradient through the film thickness (Fig. 5b). This gradient points out a Si/Ti atomic ratio of about 1.75 (35TiO2–65SiO2 composition) in the deepest layers of the film and about 1 (50TiO2–50SiO2 composition) in the surface layers, which have to be compared to a 1.5 ratio (40TiO2–60SiO2 composition) expected from a S60 composite sol.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig5_HTML.gif
Fig. 5

Evolution of the Si/Ti atomic ratio, determined from EDX measurements performed during cross-section TEM observations, as a function of the distance from the substrate for a S20-7 (a) and a S60-7 (b) composite film. The dotted and continuous arrows represent the theoretical Si/Ti ratio expected from a S20 and a S60 sol composition, respectively. The lines are drawn to guide the eye

The differences between S20-7 and S60-7 Si/Ti evolutions were repeatedly observed on different samples. They suggest impregnation features during the multilayer deposition procedure. As it has been previously shown, S60-1 films exhibit rather large surface cavities. It is thus possible that, during the subsequent deposition operations, the composite sol impregnates and partially fills in these cavities. However, since the composite sol is constituted of solid TiO2 particles on the one hand, and short oligomeric Si–O–Si chains on the other hand, these latter are likely to preferentially impregnate the cavities, while TiO2 particles would be retained at the surface of the last deposited layer. Such a process is expected to act cumulatively at each deposition of a new layer. This description can explain previous observations on S60-n films: (1) a marked Si/Ti gradient through the thickness of multilayer S60-n films, yielding an excess (deficit) of Si species in deeper (surface) layers, and (2) the formation of larger surface cavities when the number of deposited layers increases, which would arise from cumulative effects of the Si species impregnation. In contrast to S60-1 composite films, S20-1 ones did not exhibit any marked surface cavities. This suggests that impregnation feature may only account when large cavities are present at single-layer surfaces. Thus, it would explain that: (1) multilayer S20-n films do not exhibit any Si/Ti gradient through their thickness, and (2) the morphology of such composite films does not significant evolve during the multilayer procedure.

3.3 Ellipsometric measurements

The refractive indices and thicknesses of SX-n films were deduced from ellipsometric measurements. A refractive index of 1.90 ± 0.05 was measured for all single and multilayer TiO2 films deposited from a pure CS. This value is weaker than that of bulk anatase TiO2 (2.50) [22], which is attributed to a significant volume porosity of the CS-derived TiO2 films [3, 23]. Accordingly, the Lorentz-Lorenz relationship indicated a volume porosity of around 25 % for such films. This porosity probably arises from nanopores formed at grain boundaries, which suggests that TiO2 crystallites pre-formed in the CS do not undergo an optimal compaction during their deposition in liquid phase. For comparison, a refractive index of 1.45 was measured on a reference SiO2 single-layer film deposited from a reactive silica sol. This value closely agrees with that of bulk vitreous silica [22], which depicts the strong density, i.e. a volume porosity close to 0 %, of a glass-like SiO2 film deposited through a sol–gel route. Generally speaking, all single-layer and multilayer S20-n, S40-n, and S60-n films exhibited a refractive index in the range of 1.50 ± 0.1, which confirms conclusions drawn from UV/visible transmission spectra suggesting that the composite films refractive indices are close to that of a glass substrate. Thus, contrary to what would be expected from a mixing law between the high index TiO2 and low index SiO2 components, refractive indices of TiO2–SiO2 films do not follow a continuous decrease when increasing the Si/(Si + Ti) ratio in composite sols. As discussed in our previous articles for single-layer films, it is likely that, while increasing the Si/(Si + Ti) ratio in composite sols, the decrease in the film refractive index is modulated by porosity effects [3, 5]. Firstly, the Lorentz-Lorenz relationship yields a volume porosity of around 55 % for single and multilayer S20-n films, which is considerably greater than the value derived for pure TiO2 films. In our previous papers dealing with single-layer films, we explained that the presence of oligomeric silica chains in Si-poor composite films increases the porosity of such films compared to pure TiO2 films [3, 5]. This can probably be extrapolated to multilayer films since, as we previously mentioned, SEM, AFM, and TEM analyses showed that the multilayer procedure did not influence the morphology of Si-poor composite films, and did not lead to any impregnation features nor structural modifications. According to the Lorentz-Lorenz relationship, further increase in the SiO2 molar fraction of SX-n films yields a continuous decrease of their volume porosity, i.e. a volume porosity of around 45 and 30 % for S40-n and S60-n films, respectively. In the case of single-layer films, we explained that, above a Si/(Si + Ti) threshold ratio of around 20 %, the decreasing film refractive index induced by an increasing SiO2 molar fraction is counteracted by enhanced density effects arising from the partial filling of inter-granular pores by Si–O–Si chains in excess [3, 5]. Here again, it appears that such features can be extrapolated to multilayer composite films.

Figure 6 and its insert illustrate the influence of the number of deposited layers on the thickness of films deposited from sols of various compositions. The insert shows that, for a fixed sol composition, the film thickness linearly increased with the number of deposited layers. Besides, for a fixed number of deposited layers, the film thickness was also strongly influenced by the Si/(Si + Ti) ratio in the sol. Figure 6 indicates that variations of this ratio induced similar changes in the thickness of single-layer and multilayer films. This influence can firstly depict some modifications in the sol rheology (liquid viscosity or surface tension), which would influence the amount of impregnated matter during the liquid phase deposition in relation to the Si/(Si + Ti) ratio, but they can also be discussed in relation to the composite nature of TiO2–SiO2 films. On the one hand, it appears that thickness variations illustrated in Fig. 6 fairly correlate previously mentioned volume porosity variations, i.e. for single-layer and multilayer films, the thickness firstly increases when increasing the Si/(Si + Ti) ratio up to 20 %, and then it undergoes continuous decrease with further increase of the Si/(Si + Ti) ratio. Increased porosity effects can probably explain that, for a fixed number of deposited layers, the thickness of S20-n films was systematically greater than that of corresponding pure TiO2 (S0-n) films. On the other hand, in the case of single-layer films, we explained that, above a Si/(Si + Ti) threshold ratio of around 20 %, the film thickness principally depends on the amount of deposited TiO2 crystallites [5]. It is because Si species arising from the poorly reactive silica sol component are essentially expected to form oligomeric and thus very short Si–O–Si chains capping TiO2 crystallites. Here again, this observation can probably be extrapolated to multilayer films. A decreasing amount of deposited TiO2 crystallites would in turn explain data of Fig. 6 showing that increasing the Si/(Si + Ti) ratio above 20 % in the composite sols promoted a continuous decrease in the thickness of SX-n films. The null thickness plotted in Fig. 6 for pure silica sols illustrates total dewetting effects arising from the very poor reactivity of these sols. Previous explanations are partly supported by TEM studies. We have shown in Sect. 3.2 that the multilayer deposition of S60-n films is governed by an impregnation process, i.e. newly deposited Si species partially impregnate pores of the previously deposited layers while newly deposited TiO2 particles are retained at the surface. If Si species influenced the thickness of S60-n films, the film thickness would not linearly increase with the number of deposited layers as illustrated in the insert of Fig. 6. These observations reinforce the conclusion that the thickness of composite films essentially depends on the amount of deposited TiO2 crystallites and confirm that the decreasing film thickness illustrated in Fig. 6, when increasing the Si/(Si + Ti) ratio above 20 %, arises from a decreasing amount of deposited TiO2 crystallites.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig6_HTML.gif
Fig. 6

Evolution of the film thickness with the Si/(Si + Ti) molar ratio for SX-1 (a), SX-3 (b), and SX-7 (c) composite films. The inset represents the evolution of the film thickness with the number of deposited layers for S0-n (a), S20-n (b), S40-n (c), and S60-n (d) composite films. The lines are drawn to guide the eye

3.4 Wettability properties: aging and statistical features

The evolution of the water contact angle during natural aging in ambient atmosphere, i.e. in the absence of UV light, has preliminarily been studied for single-layer films issued from sols of various formulations. For each formulation, different series of sols have been synthesized and derived samples have been aged during several weeks or months spreading over a whole period of more than 1 year. Aging data are compiled in Fig. 7 and its insert. Immediately after thermal treatment at 500 °C, all the films, independently on their formulation, exhibited a superhydrophilic character. TiO2 films deposited from pure CSs rapidly lost their initial superhydrophilicity over aging, since the water contact measured on those surfaces reached a value around 40° after only few days of aging. A longer aging yielded a slower but continuous contact angle increase (Fig. 7a). This loss of superhydrophilicity is attributed to carbon contamination occurring at the film surface over aging in natural atmosphere [15]. Compared to TiO2 films, SiO2 ones deposited from reactive silica sols also underwent a superhydrophilicity loss but it occurred more gradually and slowly (Fig. 7b). For these films, a water contact angle of 30° was reached after aging of about 8 weeks. In contrast to pure TiO2 and SiO2 films, S40-1, S60-1, and S80-1 composite films exhibited a much slower water contact angle increase over aging (Fig. 7c). Such composite films naturally maintained their superhydrophilicity, i.e. a water contact angle of 5° or less, for at least 8 weeks. Longer aging periods promoted further contact angle increase and then a progressive loss of superhydrophilicity. S20-1 composite films constitute a particular case that will be analyzed in the following.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig7_HTML.gif
Fig. 7

Influence of the aging time under ambient atmosphere on the water contact angle for pure TiO2 (S0-1) (a) and pure SiO2 (S100-0) single-layer films (b), and for S40-1 (open triangle), S60-1 (multi symbol), and S80-1 (plus) single-layer composite films (c). The lines are drawn to guide the eye. The inset represents the same evolutions for pure SiO2 (open square) and S20-1 composite films (filled triangle). Continuous and dotted lines shown in this inset correspond to two specific series of S20-1 samples

Even if Fig. 7 clearly depicts the ability of TiO2–SiO2 composite films to naturally maintain their superhydrophilicity over a long period in the absence of UV light, additional insights in the wettability properties of aged films may be provided. Indeed, Fig. 7a shows that wettability measurements performed over aging on different series of pure TiO2 films yielded a very broad dispersion of contact angle values. Water contact angles measured over aging on pure SiO2 films also exhibited certain dispersion, but much less marked than for TiO2 films (Fig. 7b). In contrast, contact angles measured on S40-1, S60-1, and S80-1 composite films all exhibited a very narrow dispersion as long as the films maintained their superhydrophilicity (up to 8 weeks aging) (Fig. 7c). Accordingly, over this period, it was not possible to reliably classify films of different Si/(Si + Ti) ratio in terms of superhydrophilicity persistence. However, it should also be noted that when these composite films started to lose their superhydrophilicity, the water contact angles measured on different series of samples exhibited a greater dispersion.

As we previously mentioned, S20-1 composite films constituted a particular case, which is illustrated in the insert of Fig. 7. The water contact angles measured on S20-1 films over aging appeared globally weaker than those measured on aged SiO2 films, similarly to what occurred for other composite films. However, the dispersion observed for S20-1 films was much broader, similarly to what occurred for pure TiO2 films. On the one hand, some series of samples exhibited a very slow contact angle increase over aging (dotted line in the insert of Fig. 7), which was comparable to that observed for other SX-1 compositions. On the other hand, for some other series, the contact angle increase was much more rapid and rather close to that observed for pure SiO2 films (continuous line in the insert of Fig. 7). As mentioned in the experimental section, aging experiments have been performed under ambient conditions in open atmosphere. Such conditions were chosen to study our films in real indoor pollution conditions. Besides, these experiments were performed over long aging durations and on several series of samples studied at different periods of the year. It is likely that indoor pollution can vary over such a long duration owing to seasonal climatic variations, e.g., indoor relative humidity and/or temperature variations. This observation may explain experimental dispersions illustrated in Fig. 7 and its insert for different series of samples. For reasons that are not clearly elucidated yet, and according to experimental dispersions illustrated in Fig. 7 and its insert, Ti–rich films (TiO2 and S20-1 films) appear to be very sensitive to the pollution conditions, whereas it seems that Si-rich films (SiO2 films and S40-1, S60-1, and S80-1 composite ones) exhibit an enhanced robustness toward those variations. Accordingly, Si-rich composite films not only exhibit a naturally persistent superhydrophilicity but also a remarkable reproducibility over an aging period of at least 8 weeks. The particular case of S20-1 composite films can be discussed as follows. As shown by EDX measurements, and contrary to Si-rich composite films, S20-1 films do not necessarily exhibit a perfectly controlled TiO2–SiO2 stoechiometry, which may suppose a certain lack of control in the synthesis conditions of such Si-poor composite films. This feature has not been investigated in more details, since such films are not of primary interest for our studies. However, when considering very strong differences between the wettability behaviours of pure TiO2 films (Fig. 7a) and S40-1 composite films (Fig. 7c), it is very likely that weak variations in the TiO2–SiO2 stoechiometry of S20-1 films can critically influence their wettability properties, which would explain strong dispersions illustrated in the insert of Fig. 7.

3.5 Influence of thickness on the natural wettability properties

Previously exposed water wettability features illustrate the difficulty to precisely study and classify composite films in terms of superhydrophilicity persistence. Indeed, owing to the enhanced persistence of their superhydrophilicity, a classification between best composite films requires long aging durations. Experimental conditions over such long aging durations can in turn be perturbed by fluctuations in the polluting atmosphere, which can bother a reliable comparison between aged composite films studied in separate series. We have attempted to set up accelerated aging protocols, which would have minimized long time experimental uncertainties. At this moment, these attempts remain unsuccessful. In the present work, in order to progress in the understanding of the composite films wettability, we have thus decided to study single series of samples aged together over the same aging period. This choice obviously reduces the possibility to draw statistical conclusions, but it minimizes uncertainties arising from variations in pollution conditions over a long period of time. In this section and the following, two series of samples have been studied. Figure 8 illustrates the water contact angle reached after 14 weeks aging for a first series of single-layer and multilayer films of various compositions. While the water contact angle measured on pure TiO2 and SiO2 films exhibited a very elevated value of 70° and 40°, respectively, all the contact angles measured on S20-n, S40-n, and S60-n composite films lied in a 10 ± 4° range (note that, in this series, S20-n films exhibited good performances). Even when the composite films can no longer be considered to be superhydrophilic, the weak contact angle measured on these aged films provides new evidence that compared to pure films, composite ones are able to naturally maintain an enhanced hydrophilicity over a very long aging duration. Besides, contact angle values illustrated in Fig. 8 for composite films do not depict any significant influence of the film thickness on their wettability. Indeed, the 10 ± 4° range can reasonably be considered to lie within the experimental error, and whatever their thickness, composite films aged for a very long duration never exhibit the superhydrophilicity observed for shorter aging durations.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig8_HTML.gif
Fig. 8

Influence of the Si/(Si + Ti) molar ratio on the water contact angle measured after 100 days of aging under ambient atmosphere for SX-1 (open circle), SX-3 (open square) and SX-7 (open triangle) composite films. The line is drawn to guide the eye

As explained in the introduction, electric charges are supposed to be localized at TiO2–SiO2 granular interfaces of the composite films. Such charges, which are also present at the surface of composite films, may in turn induce a molecular or dissociative water adsorption at the film surface and contribute to the superhydrophilicity of the films. Beside charge effects, thickness effects leading to composition or morphology (roughness, porosity) variations can also influence the wettability of composite films. However, in the present work, such thickness effects do not seem playing a major role. For instance, for Si-rich S60-n composite films, this study firstly showed that the impregnation features arising from a multilayer deposition procedure can induce some modifications in the TiO2–SiO2 composition of the surface layers, which primarily govern wettability properties. However, these weak composition modifications have no practical consequence since, according to Fig. 8, the film composition is not a critical parameter and similar wettability features were obtained in a large S40-n to S80-n composition range. Moreover, we have also shown that increasing the thickness of Si-rich composite films induced some RMS roughness increase. However, this increase is not significant enough to influence their wettability properties. Accordingly, it has been reported that the roughness does not influence the wettability of a given surface below a threshold value around 100 nm [24], a value by far much larger than all RMS roughnesses reported in the present work. Finally, an increased surface porosity, for instance increased cavity dimensions arising from the multilayer deposition procedure of a Si-rich sol (Figs. 1, 2), can also influence the water contact angle measurement. In such conditions and during wettability measurements, deposition of a water drop can impregnate surface pores, yielding a composite water/solid surface showing an enhanced affinity for water, i.e. an enhanced hydrophilicity as described by Bico [19]. However, the data illustrated in Fig. 8 show that, whatever the eventual influence of a water impregnation mechanism, this latter does not significantly influence the wettability properties of composite films. To conclude, it seems that roughness or porosity effects do not play a major role in the natural wettability of TiO2–SiO2 composite films, which reinforces the assumption that the superhydrophlicity of these films is intrinsically governed by granular interfaces effects, i.e. electrical charges effects at TiO2–SiO2 interfaces. Electrical measurements are currently in progress in our group in order to better assess such effects.

3.6 Influence of thickness on the photo-induced wettability properties

As explained before, the progressive loss of the superhydrophilicity of composite films over aging, illustrated in Fig. 7, is not an irreversible process, since this property can easily be regenerated through a short UV exposure as far as TiO2 crystallites are present in sufficient amount. This photo-regeneration results from: (1) the degradation of organic contaminants present at the film surface after aging, which is induced by the photocatalytic activity of TiO2 crystallites present in the composite films, and (2) a saturation of the surface with OH groups arising from atmospheric humidity, similarly to mechanisms yielding a super-hydrophilic surface when pure titania is exposed to UV light [6, 8, 9]. This photo-regeneration mechanism has previously been studied in our recent paper for single-layer composite films [4]. On the one hand, we showed that, similarly to what happened for pure SiO2 films, no photo-regeneration of the super-hydrophilicity was possible for composite films with an excess of silica [4]. Indeed, it is very likely that the silica network which embeds the TiO2 crystallites totally shields interaction between these crystallites and organic species present at the film surface, and thus cancels any photo-reaction at the surface. On the other hand, composite films with a reduced silica amount successfully underwent a superhydrophilicity photo-regeneration. This photo-regeneration could even be performed for a rather important Si/(Si + Ti) ratio of 60 % (S60-1 film) [4]. However, this latter observation has to be tempered by the fact that in that study, before UV exposure, the aged S60-1 film still exhibited a very weak water contact angle of 6°.

In the present work, superhydrophilicity regeneration experiments were performed on a new series of S60-n films of various thicknesses all aged for more than 2 months in similar conditions. After aging, all of them exhibited a very comparable water contact angle of 18 ± 1° (Fig. 9) which is rather high as compared to those plotted on Fig. 7c. At first, this observation corroborates that the thickness of composite films does not influence their wettability over aging. Then, it clearly illustrates the experimental dispersion occurring when different series of samples are not aged at the same time over a long duration, pointing out the necessity to study simultaneously aged samples in order to draw fine conclusions. The evolution of the water contact angle measured for this series of aged S60-n films during their UV light exposition is illustrated in Fig. 9. For comparison, the same evolution measured on an aged single-layer TiO2 film is illustrated in the insert of Fig. 9. This latter graphic shows that the photo-induced superhydrophilicity of a pure TiO2 film relies on a bi-regime mechanism. During the first illumination stage of around 30 min, the contact angle does not appreciably vary, after what it suddenly drops meaning that the surface has become superhydrophilic. In a previous paper dealing with the photo-induced properties of pure TiO2 films deposited from a CS, we explained that these properties rely: (1) on the photo-generation of charge carriers through the film thickness, and (2) on the migration of these carriers from the deeper layers of the film toward the outer surface, this migration being essentially governed by the diffusion of charge carriers at granular interfaces [23]. When the carriers reach the surface, they can in turn promote a photocatalytic decomposition of organic matter and a photo-induced OH surface saturation, yielding a superhydrophilic surface [6, 8, 9]. The bi-regime mechanism illustrated in the insert of Fig. 9 traduces the time separation of these two phenomena. Indeed, the photocatalytic decomposition of the organic pollutant present at the film surface may occur during the first minutes of the UV exposition without any noticeable water contact angle decrease, after what, once the surface has been sufficiently decontaminated, the photo-reactions leading to the super-hydrophilicity can take place [23].
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig9_HTML.gif
Fig. 9

Evolutions of the water contact angle for aged S60-1 (a), S60-3 (b), and S60-7 (c) composite films exposed to UV light. The inset shows the same evolutions for a single-layer pure TiO2 film. The lines are drawn to guide the eye

These photo-induced mechanisms can then be extrapolated to the photo-regeneration of the superhydrophilicity of composite films. As depicted in Fig. 9a, a single-layer S60-1 film exposed to UV light progressively recovered its superhydrophilicity but, even after a long UV exposure of 100 min, it never rigorously exhibited a contact angle of 0°. Besides, a bi-regime superhydrophilicity photo-regeneration could hardly be appreciated for the S60-1 film. It suggests that, in contrast to pure TiO2 films, the photocatalytic decomposition of organic pollution and photo-induced OH saturation of the surface occur more or less simultaneously at the surface of Si-rich composite films. This feature is probably due to the fact that, owing to the stronger affinity of silica species for water compared to that of a titania ones, a Si-rich composite film can undergo faster OH saturation. However, it also appears that the superhydrophilicity photo-regeneration observed for a single-layer S60-1 film (Fig. 9a) was very slow compared to that observed for a pure TiO2 film (insert of Fig. 9). In contrast to a S60-1 film, a 0° contact angle was rigorously reached after only 40 min of UV exposure for a S60-3 film (Fig. 9b) and the superhydrophilicity photo-regeneration was again faster for a S60-7 film, since the 0° value was reached after only 25 min of UV exposure (Fig. 9c). These data evidence that the thickness strongly influences the superhydrophilicity photo-regeneration duration of composite films. Such influence can hardly be attributed to roughness or water impregnation features since, as previously shown, such features do not seem to influence the wettability of composite films.

We believe that the data illustrated in Fig. 9 can be discussed in relation to the inter-granular diffusion mechanism of photo-generated charge carriers during their migration toward the film surface. In the case of pure TiO2 films, we previously indicated that the photo-induced superhydrophilicity of such films weakly depended on the film thickness (not shown here) [23]. In other words, we showed that only the outer film layers were necessary to provide a sufficient tank of charge carriers in order to promote optimal photo-induced mechanisms at the film surface. This can be related to an efficient inter-granular diffusion of the charge carriers in pure TiO2 films exposed to UV light, i.e. in photo-conducting films [6]. In composite films, even if the silica amount is weak enough to allow an inter-granular diffusion of the charge carriers, it is likely that this diffusion is partially hindered by the presence of (electrically isolating) silica chains at the surface of TiO2 crystallites. Charge carriers may temporarily be blocked at the surface of TiO2 crystallites and can in turn undergo a recombination process, which is known to be the most influencing mechanism limiting the photocatalytic activity [6]. In that case, the probability that charge carriers reach and react at the film surface is reduced, and the charge carrier tank can no longer be restrained to surface layers as it is the case for pure TiO2 films. Thus, increasing the film thickness seems to provide an additional charge carriers tank which would explain why, compared to pure TiO2 films, composite ones exhibit a strong thickness dependence of their photo-induced wettability properties.

The influence of the thickness on surface mechanisms inducing a superhydrophilicity photo-regeneration in composite films has been partially confirmed by XPS measurements. In principle, the O1s XPS peak might provide information on specific bonds arising from the photo-induced OH saturation of the film surface. However, in our case, since this peak accounts for the presence of multiple bonds (Si–O, Si–OH, Ti–O, Ti–OH), drawing conclusions on OH groups adsorbed at the film surface would require a very speculative multi-peak deconvolution [1]. Thus, in the present study, we only focused on photocatalytic mechanisms yielding the decomposition of organic contaminants. Fig. 10 illustrates the C1s peak (284.6 eV) measured on aged S60-1 and S60-7 composite films before and after UV exposure for 45 min. As explained in our previous papers, the C1s peak principally accounts for the carbon contamination arising from aging under ambient atmosphere [1, 3]. In order to enhance the effect of carbon contamination and allow an easier detection of this peak, the composite films illustrated in Fig. 10 were preliminarily aged for about 10 months. It is also worthwhile mentioning that the C1s peak does not necessarily account for the actual surface chemistry of aged films. Indeed, the films were put in equilibrium for 12 h in an ultra high vacuum chamber before collecting XPS data which may promote partial desorption of species adsorbed at the film surface. For this reason, XPS measurements cannot be used to quantitatively analyze adsorbed species, but only to illustrate phenomenological trends. After 45 min of UV exposure, the aged S60-7 film totally recovered its superhydrophilicity while the S60-1 film only underwent a partial water contact angle diminution, similarly to what is illustrated in Fig. 9. These trends are correlated to XPS data illustrated in Fig. 10. On the one hand, Fig. 10b shows that the UV exposure of S60-7 film promotes a significant decrease of the C1s peak intensity, which illustrates the photocatalytic decomposition of organic contaminants. However, since this peak is still intense after UV exposure, it suggests that the superhydrophilicity photo-regeneration does not require total carbon decontamination as far as sites available for OH adsorption are present in sufficient amount. On the other hand, Fig. 10a shows that the 45 min UV exposure is not sufficient to promote any significant carbon decontamination for the S60-1 film, which is in agreement with the fact that this film did not recover its superhydrophilicity after UV exposure. To conclude, the data of Fig. 10 illustrate that a greater amount of photo-induced charge carriers reach the surface of thicker composite films, which can in turn be correlated to the data of Fig. 9 showing that a thicker S60-n film undergoes a faster superhydrophilicity photo-regeneration.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-012-2878-4/MediaObjects/10971_2012_2878_Fig10_HTML.gif
Fig. 10

C1s peak measured by XPS on aged S60-1 (a) and S60-7 (b) composite films before (dotted lines) and after (continuous lines) exposition to UV light for 45 min. The films were preliminarily aged for 10 months under ambient atmosphere. The C1s peak has been normalized to the sum of the Si2p (103.6 eV) and Ti2p (458.5 eV) peak intensities and to the intensity values measured before UV exposure

4 Conclusion

TiO2–SiO2 nanocomposite films of various thicknesses have been deposited via a multilayer sol–gel route. AFM, FEG-SEM, TEM, ellipsometry, and UV/visible measurements show that, depending on the TiO2–SiO2 sol composition, the multilayer procedure induces some changes in the morphology and composition of composite films. TiO2–SiO2 nanocomposite films exhibit a natural and persistent superhydrophilicity, which allows considering new applications for easy to clean surfaces. This natural superhydrophilicity has been studied in relation to morphology, composition, and thickness features arising from a multilayer deposition procedure. It is concluded that such features do not significantly influence the natural wettability of TiO2–SiO2 composite films, which particularly suggests that surface roughness effects or water impregnation through the pores of more or less thick films do not play a major role in the enhanced wettability of these films. These conclusions reinforce the assumption that the natural superhydrophlicity of composite films is intrinsically governed by granular interface effects, i.e. electrical charge effects at TiO2–SiO2 interfaces. Electrical measurements are presently in progress in our group in order to better assess such effects. In contrast to their natural wettability, the photo-induced wettability of TiO2–SiO2 composite films is considerably enhanced for thicker films. Photo-induced mechanisms allow a superhydrophilicity photo-regeneration when this natural property starts to disappear after a long aging period in ambient atmosphere owing to surface carbon contamination. This photo-regeneration is attributed to the photocatalytic decomposition of carbon species at the film surface, which arises from TiO2 crystallites present in the films, and a concomitant OH saturation of the composite film surface. As suggested by XPS measurements, the faster superhydrophilicity photo-regeneration observed for thicker films can in turn be attributed to a greater amount of photo-generated charge carriers reaching the film surface, which in turn promotes more efficient photo-induced surface mechanisms.

Copyright information

© Springer Science+Business Media New York 2012