3172
Langmuir 2005, 21, 3172-3178
Characterization of TiO2 Nanoparticles Surface Modified with Aluminum Stearate Terry A. Egerton,*,† Neil J. Everall,‡ and Ian R. Tooley§ School of Chemical Engineering & Advanced Materials, Bedson Building, University of Newcastle, Newcastle NE1 7RU, U.K., ICI Measurement Science Group, Wilton Research Centre, Wilton, Redcar TS10 4RF, U.K., and Uniqema R&D, Wilton Research Centre, Wilton, Redcar TS10 4RF, U.K. Received October 22, 2004. In Final Form: January 12, 2005 This paper uses measurements of adsorption and vibrational spectra (DRIFTS, ATR, and Raman) to characterize TiO2 (rutile) nanoparticles that have been surface treated with aluminum and stearate, “aluminum stearate”. From these measurements, we have developed a model of titania particles covered by patches of “alumina”. Vibrational spectra, particularly the spectra of the carboxylate headgroups, show that the stearate then adsorbs on both the titania and the alumina. Surprisingly, the distribution of the stearate between alumina and titania is sensitive to the presence of water. As the water content decreases, the relative amounts of stearate on titania, rather than alumina, increase, and this increase is accompanied by a less ordered structuring of the stearate tails, as evidenced by a shift of the C-H stretching bands to higher frequencies and a broadening of the 1296 cm-1 Raman band. This effect is consistent with earlier observations that the presence of water reduced the bonding of stearate headgroups to the surface of titania. We have also shown that the dispersion in C12-C15 alkyl benzoate of aluminum stearate coated titania is sensitive to the presence of small amounts, ∼4%, of water. Finally, we have demonstrated that surface stearate, like surface alumina, reduces the rate of phototocatalytic oxidation of 2-propanol. A 7% stearate coating reduces acetone formation by a factor of 4. There is no evidence from these studies that, during the oxidation experiment, 2-propanol displaces stearate from the titania surface.
Introduction The interaction of stearic acid with the surface of nanoparticulate titanium dioxide has received considerable attention because stearic acid and/or its metal salts have been widely used to modify the surface of titanium dioxide pigments and as lubricants for pigment processing in plastics. The adsorption of stearate on pigmentary sized titania has been studied by a number of authors.1 Other authors have shown that the dispersion in nonaqueous liquids of titania treated with carboxyl terminated surfactants is sensitive to the presence of small quantities of water,2 and Polunina and co-workers3 concluded that covalent bonding between the titania surface and the carboxyl group of stearic acid decreased in the presence of water but increased if cations such as aluminum were present at the surface. Most recently, Britcher and coworkers4 used thermogravimetric analysis (TGA) and diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) to probe the adsorption of sodium stearate on titania pigment with a nominal size of 0.38 µm and with a surface coating of hydrous aluminate. The IR spectra indicated that the stearate bridged across two aluminum atoms in this coating. In contrast, McBride, on the basis of electron spin resonance (ESR) results from * Corresponding author. Phone: +44 (0) 191 222 5618. E-mail:
[email protected]. † University of Newcastle. ‡ ICI Measurement Science Group, Wilton Research Centre. § Uniqema R&D, Wilton Research Centre. (1) Ottewill, R. H.; Tiffany, J. M. J. Oil. Colour Chem. Assoc. 1967, 50, 844. (2) Crowl, V. T.; Malati, M. A. Discuss. Faraday Soc. 1966, 42, 301. (3) Polunina, I. A.; Mikhailova, S. S.; Roldughina, T. V. Compos. Interfaces 1999, 6, 49. (4) Capelle, H. A.; Britcher L. G.; Morris, G. E. J. Colloid Interface Sci. 2003, 268, 293.
nitroxide-labeled stearic acid analogues, concluded that the carboxylic acid bonds to alumina as a monodentate species.5 Nanoparticulate titanium dioxide is widely used as an inorganic sunblock and for this application may be dispersed in either aqueous or nonaqueous formulations. The titanium dioxide is coated with “aluminum stearate” in order to control dispersion in both “oil”, for example, C12-C15 alkyl benzoate, and water. In addition, the presence of the stearate reduces aggregation during drying and in this way makes the finished formulations more transparent to visible radiation. This paper first uses adsorption measurements to confirm that the alumina coats the surface of the titania nanoparticles, rather than forming separate particles of hydrous aluminum oxide. Spectroscopy is then used to probe the nature of the bonding between the stearate and a titania surface that is covered with patches of alumina. We show direct evidence that the packing of the stearate is modified by the presence of trace amounts of water. Finally, we report the effect of the stearate coating on the titania photoactivity. Experimental Section Chemicals. The high-area titania used as the starting material for this study is shown in Figure 1 and was made available by ICI-Uniqema. It was prepared by the hydrolysis of TiCl4, and the surface area (BET, N2) was ∼130 m2 g-1. Powder X-ray diffraction showed no phases other than rutile. The transmission electron micrograph of sample A (Figure 1) shows that the particles are acicular. Inspection of micrographs recorded at a lower magnification, and therefore imaging a much larger number of particles than that shown in Figure 1, leads to an average particle size of ∼15 × 75 nm long. However, the size of the primary particles (determined via X-ray line broadening) (5) McBride, M. B. J. Colloid Interface Sci. 1980, 76, 393.
10.1021/la047390d CCC: $30.25 © 2005 American Chemical Society Published on Web 02/26/2005
TiO2 Surface Modified with Aluminum Stearate
Figure 1. High-magnification transmission electron micrograph (TEM) of high-area rutile sample A showing acicularity of the crystals. The bar represents 50 nm. Also shown is a schematic diagram of the crystals shown in the TEM, indicating regions of internal surface area. is ∼7 nm. This is broadly consistent with the size derived from the surface area measurements (∼10 nm equivalent sphere diameter) but much smaller than that inferred from lowmagnification transmission electron micrographs (TEMs) or from optical measurements. The reason for this apparent anomaly is that the XRD size is measured in the 110 direction, across the crystals, whose length is in a direction parallel to the c axis and because, as exemplified by the drawing in Figure 1, the primary crystallites assemble to form larger secondary particles with their c axes oriented parallel to one another. Thus, the particle width of ∼15 nm shown in the labeling of the drawing is, in fact, the width of several primary crystallites assembled alongside one another. Titania Coated with Aluminum Stearate. All the coated powders were prepared from the same batch of uncoated higharea rutile, sample A. The required quantity of caustic sodium aluminate was added dropwise over 30 min into a suspension of sample A at 50 °C. The resulting alkaline slurry was stirred for 30 min during which time the temperature was raised to 75 °C. A hot sodium stearate solution was added to the slurry over 10 min. After the addition of the sodium stearate solution, the slurry was allowed to equilibrate and the pH adjusted to 6.5-7.0 before the slurry was filtered and dried in air at 110 °C. As no macroscopic stearate crystals were found in the coated product by X-ray diffraction (XRD) and since neither inorganic analysis nor surface analysis (XPS) detected residual sodium, we conclude that the coating consists essentially of aluminum and stearate. The targeted aluminum contents (expressed as % alumina) of the coated samples ranged from 0 to 18%. For a coating in which the reported level of alumina is 18% (9% aluminum), the stearate content of a pure aluminum stearate, Al(C17H35COO)3, would be (9/(27/849) ) 300%) relative to the mass of the uncoated titania. Since the target levels of stearate are much lower than this, it is clear that none of the coatings correspond to a stoichiometric aluminum stearate compound.
Langmuir, Vol. 21, No. 7, 2005 3173 Adsorption Measurements. Prior to measurement of adsorption, samples were evacuated at ∼1 Pa for 12 h at 110 °C and then allowed to cool naturally. Adsorptions were measured at -195 °C using a McBain balance. Nitrogen (1.3 kPa) was introduced to the vacuum system and allowed to equilibrate for 60 min. The spring extensions due to the uptake of nitrogen were then measured and the nitrogen pressure recorded, before repeating the process for increasing pressures of nitrogen (1.380 kPa). Brunauer-Emmett-Teller (BET) surface areas assumed σN2 ) 0.162 nm2. Infrared Spectroscopy. DRIFT spectra were recorded using a Spectra-Tech DRIFTS accessory mounted in a Bio-Rad FTS40 infrared spectrometer. All samples were dried in situ at 110 ( 1 °C. Spectra were referenced to an initial KBr reference spectrum and were recorded between 1000 and 4000 cm-1 using 100 scans at a 4 cm-1 resolution. Attenuated total internal reflection (ATR) infrared spectra were obtained using a single bounce diamond-prism reflection accessory (“Golden Gate” Graseby-Specac Ltd). A background spectrum, at a 4 cm-1 resolution, was obtained using the clean diamond crystal with no sample in contact. Spectra were then recorded (typically averaging 400 scans) by placing a sample on top of the diamond crystal and applying light pressure with an anvil to ensure intimate contact. The evanescent field generated by reflecting an infrared beam inside the prism penetrates several microns into the sample. Therefore, it fully penetrates bundles of the nanophase TiO2 crystals and the measured spectrum samples both coating and the “bulk” of the TiO2 crystal, not just the surface. Where appropriate, ATR measurements were complemented by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS). DRIFT spectra could be measured at elevated temperatures but were susceptible to artifacts due to packing and particle size differences. Raman Spectroscopy. Raman spectra were recorded using a Kaiser Holoprobe spectrometer (Kaiser Optical Systems, Inc, Ann Arbor, MI) with a 785 nm laser generating ∼80 mW of power at the sample. The laser was focused onto a lightly compacted portion of the powder and the backscattered light collected to produce the Raman spectrum. The spectral resolution varied with wavelength but was of the order of 5 cm-1 at the center of the spectrum. Typically, 20 scans of ∼12 s duration were averaged to record a spectrum. Measurement of 2-Propanol Oxidation. Photocatalytic activity was determined by measuring, at 30 °C, the photogeneration of acetone by hand-stirred dispersions of 0.4 g of TiO2 in 50 mL of 2-propanol. The reaction was carried out in a cylindrical Pyrex vessel illuminated, from the base, by two Philips PL-L 36W 09 actinic lamps, as previously described.6 Radiation was passed through a Pyrex heat filter, and cell temperature was controlled at 30 °C by a closely fitting cylindrical heater. The reaction mixture was sampled by a hypodermic syringe through a port fitted with a septum cap and filtered to remove titanium dioxide. Samples were analyzed for acetone by gas chromatography (Cambridge GC94, Chromosorb wax 60/80 mesh at 70 °C) calibrated with mixtures of acetone and 2-propanol using a diethyl ether internal standard. Straight-line calibration plots (R2 ) 0.997) were obtained.
Results and Discussion Coating of Titania by Aluminum Stearate. The challenge of demonstrating that a coating has formed a skin on the original particles, rather than simply forming fresh discrete particles, is particularly demanding for nanoparticles because of the difficulty of detecting surface layers by transmission electron microscopy. For the titania nanoparticles of the type used in this study, although highmagnification transmission micrographs of coated samples show no evidence of alumina as a separate precipitate, it has not been possible to detect inorganic coatings by highmagnification transmission electron microscopy. Even when a heavy (20%) cerium oxide coating, chosen because its high electron density facilitates detection, was exam(6) Egerton, T. A.; Tooley, I. R. J. Mater. Chem. 2002, 12, 1111.
3174
Langmuir, Vol. 21, No. 7, 2005
Egerton et al.
Table 1. Surface Area of Titania Samples Coated with Aluminum Stearatea % alumina
% stearate
BET area after outgassing at 110 °C ( 5 m2 g-1
% alumina
% stearate
BET area after outgassing at 110 °C ( 5 m2 g-1
0 18 18 18 18
0 0 8 12 16
132 130 110 95 85
18 12 10 7
12 12 12 12
95 65 64 61
a
The quoted compositions are those targeted in the preparation experiments. Table 2. Effect of Energetic Heterogeneity on the BET Areas Derived from the Calculated Isothermsa
a
C constant of alumina or titania accounting for 50% of the total surface
C constant of low-energy surface accounting for 50% of the total surface
relative BET area derived from computed composite isotherms
BET C constants derived from computed isotherms
200 200 200 200 200 200 200
200 100 50 20 10 5 1
1 1.000 0.998 0.988 0.960 0.899 0.674
200 135 85 46 32 26 35
The homotattic surface is assumed to have an area of 1.
Figure 2. Diagrammatic representation of the suggested blocking of pores by stearate and the influence of alumina on this process.
ined using a Philips CM200ST FEG-STEM microscope, coating could not be clearly detected. Therefore, we first report results of adsorption measurements which were used to investigate whether the aluminum stearate forms separate particles or whether it is intimately associated with the surface of the titania nanoparticles. Table 1 shows the nitrogen surface areas for titania particles coated with aluminum stearate and outgassed at 110 °C. The results in Table 1 show that coating the parent rutile with 18% alumina, by the same method as that used for the aluminum stearate coatings, but without stearate addition, caused a negligible decrease, 10 the BET area of the energetically heterogeneous surface is within 5% of the energetically homogeneous surface, we consider that energetic heterogeneity does not significantly affect the measured areas of our stearate modified titania. (The measured C value of 40 for the 18% alumina 16% stearate coating is also consistent with a C value between 10 and 20 for the stearate patches.) Therefore, our model must attempt to explain the changes in BET area. The next stage in this study was to use vibrational spectroscopy to probe the nature of the stearate/titania and stearate/alumina interactions. (7) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity, 2nd ed.; Academic Press: London, 1982. (8) Day, R. E.; Parfitt, G. D.; Peacock, J. Discuss. Faraday Soc. 1971, 52, 215. (9) Amati, D.; Kovats, E. S. Langmuir 1988, 4, 329.
TiO2 Surface Modified with Aluminum Stearate
Figure 3. DRIFT spectra of (a) aluminum tristearate and (b) stearic acid.
Figure 4. ATR spectra of coated titanias, with spectra of pure aluminum stearate and sodium stearate for reference.
The Basis of the Spectral Assignments. Figure 3 shows the DRIFT spectra of stearic acid and aluminum tristearate (Aldrich). The band at 3675 cm-1 in the spectrum of aluminum tristearate is assigned to free aluminum hydroxide groups.4 The strong peaks at 2850 and 2920 cm-1 are assigned to the symmetric and asymmetric vibrations of sCH2 and sCH3 groups, respectively. The strong band at 1705 cm-1 due to stretching vibrations of the CdO group of the undissociated carboxylic group in stearic acid is replaced, in the spectrum of aluminum stearate, by strong absorptions at 1468 and 1585 cm-1, assigned to the antisymmetric and symmetric stretches of bidentate carboxylate. The 1468 cm-1 band is actually a mixed mode, since the symmetric stretch will be coupled with the motions of the extended alkyl chain. It is also overlapped by the CH2 scissoring mode that occurs near this frequency. The large (>100 cm-1) separation of the bands indicates that the carboxylate is bridging rather than chelating to a single aluminum ion but is probably too small to indicate monodentate coordination to a single aluminum ion, as was proposed by Oberg et al. for stearate adsorbed onto sputtered aluminum.10 Figure 4 contrasts the ATR spectra of aluminum tristearate powder with spectra of (aluminum free) stearate-capped titania. The stearate coated titania shows bands at ∼1558 and 1420 cm-1, consistent with bridging coordination, and 1510 and 1470 cm-1, which can be (10) Oberg, K.; Persson, P.; Shchukarev, A.; Eliasson. Thin Solid Films 2001, 397, 102.
Langmuir, Vol. 21, No. 7, 2005 3175
reasonably assigned to chelated carboxylate.11 There is also a weak band at ∼1650 cm-1 due to adsorbed water. The separation of the asymmetric and symmetric carboxylate vibrations forms the basis of these assignments.12 Earlier work by Wu et al. on stearate-capped titania nanoparticles found similar spectra and reported two bands at 1520 and 1456 cm-1 assigned to bidentate chelating species, instead of bands indicative of bridging carboxylate.11 These differences probably reflect differences of titania surface structure associated with Wu’s TiO2 precipitation method (hydrolysis of TiCl4 in the presence of stearic acid). When interpreting the spectrum of stearate adsorbed onto alumina coated titania, it is fortunate that, although the diagnostic value of the symmetric aluminum stearate band is limited because it coincides with titanium stearate bands, the asymmetric stretch of aluminum stearate (∼1585 cm-1) does not overlap any of the titanium stearate bands. It is not possible to decompose any given spectrum of a coated sample into a simple linear combination of the titanium stearate and aluminum stearate reference spectra shown in Figure 3. The reference spectrum of aluminum stearate was obtained from a sample of microcrystalline aluminum tristearate, in which the stearate will experience a somewhat different local environment compared with it being adsorbed onto (patchy) nanophase alumina coatings. Also, the spectrum of nanophase titanium stearate could well be modified in the presence of a dispersed, patchy covering of alumina. Thus, the relative band intensities and breadths of the stearate bands would be expected to be different for the coated samples. The aluminum stearate was used solely to indicate the position of the asymmetric carboxylate stretch when coordinated to aluminum. Work is underway to prepare samples of pure nanophase alumina that has been coated with stearate. These samples will be used to provide a better fingerprint IR spectrum of stearate adsorbed onto small alumina regions and hence allow spectral decomposition. Interaction of Carboxylate Headgroups with the Alumina/Titania Surface. The spectrum of stearate adsorbed onto 10.5% alumina coated TiO2 is also shown in Figure 4 and is recognizable as a superposition of the spectra of aluminum stearate and stearate coated titania, albeit with varying proportions of the chelating and bridging forms of titanium stearate. (The possibility that the 1560 cm-1 band in the coated samples arises from residual sodium stearate was excluded because of the lack of detectable sodium.) Since, in separate experiments on alumina (only) precipitation and stearate (only) precipitation, the alumina deposits at a higher pH than stearate, we assume that first the alumina deposits on the titania surface and that subsequently stearate deposits on the alumina coated titania. On the basis of the results of our earlier study,6 referred to above, 10.5% alumina is insufficient to completely cover the titania surface. The particle surface must therefore have patches of alumina coating and areas of uncoated titania. The presence of IR bands of stearate bonded to both Ti and Al (Figure 6) is fully consistent with the expected heterogeneous surface. We now consider how the distribution of stearate between the two types of patches depends on the levels of stearate (loading) or water (drying). Packing of the Stearate Alkyl Chains. Vibrational spectra of stearates are sensitive to the ordering of the alkyl chains because the frequencies and the bandwidths (11) Wu, X.; Wang, D.; Yang, S. J. Colloid Interface Sci. 2000, 222, 37. (12) Nakamoto, K. Infrared and Raman Spectra of Inorganic Coordination Compounds; Wiley: New York, 1978; p 232.
3176
Langmuir, Vol. 21, No. 7, 2005
Figure 5. ATR spectra showing the effect of loading (6.5, 13.5, and 27% stearate) on stearate chain conformation for a fixed level of 10.5% alumina.
Figure 6. Changes in ATR spectra when wet aluminum stearate coated titania is dried in an oven held at 110 °C for 12 h.
associated with the methylene groups depend on whether the alkyl chains are fully extended or contain gauche rotamers. For example, CH2 asymmetric and symmetric stretching frequencies are lower for fully extended chains than for the folded chains and ordered extended chains usually give rise to narrower vibrational bands, due to a reduction in the range of molecular conformations present (each with a slightly different vibrational frequency).13 Figure 5 demonstrates that, for coated samples with fixed alumina (10.5%) and increasing (from 6.5 to 27%) stearate levels, the C-H stretching frequencies decrease as the stearate loading increases. (We calculate, assuming a 0.2 nm2 footprint, that for stearate levels below 30% adsorption would not exceed the monolayer capacity of a 140 m2 g-1 adsorbent.) This shift is interpreted as a consequence of the stearate chains packing more closely as the loading is increased, leading to a denser structure with fewer gauche defects. The trend may be compared with the reported14 increase from 2849 to 2851 cm-1 in the symmetric CH2 stretch of sodium stearate as the chains become conformationally disordered upon raising the temperature. These authors also noted that in the ordered stearate the asymmetric methylene stretch occurred at 2919 cm-1 but did not quote its upshift as the temperature was raised. Effect of Drying Titania Coated with Aluminum Stearate. All of the above spectra were recorded on samples that had been dried in an oven held at 110 °C for 12 h. We now turn our attention to the spectroscopic evidence of changes in the coating structure when the washed, wet aluminum stearate coated product is dried. Coordination of Carboxylate to the Surface. Figure 6 shows the significant changes in the ATR-IR spectra (13) Snyder, R. G. J. Chem. Phys. 1967, 47, 1316. (14) Gavrilko, T.; Drozd, M.; Puchkovskaya, G.; Naumovets, A.; Tkachenko, Z.; Viduta, L.; Baran, J.; Ratajczak, H. J. Mol. Struct. 1998, 450, 136.
Egerton et al.
Figure 7. Changes in the ATR-FTIR C-H stretching bands when wet aluminum stearate coated titania was dried in an oven held at 110 °C for 12 h.
when wet aluminum stearate coated titania filter cake (60-70% water) is “dried” for 12 h at 110 °C. (The water content of the wet filter cake was estimated from the weight loss upon drying at 110 °C for 12 h.) For nominally dried samples, the weight loss after a further 12 h was ∼4% water. (Control experiments showed that stearate was not lost upon drying.) In the wet filter cake, bands of bidentate-bridged aluminum stearate dominate, although there are also weaker bands consistent with the presence of bidentate-bridged titanium stearate. In the dried sample, the aluminum stearate bands are relatively weak in comparison with the titanium stearate, which now shows bands consistent with chelating and bridged stearate (although chelated titanium stearate (1510 cm-1 band) is still less prominent than in the stearate-capped titania sample (Figure 4)). Therefore, the main conclusion is that drying increases the proportion of stearate bound to titanium rather than aluminum. This could arise either by “free” stearate (or stearic acid) binding to bare patches of titania or by stearate recoordinating from aluminum sites to titanium sites. Because these spectra provide a relative not an absolute quantitation, they do not differentiate between these possibilities, but the second alternative seems more probable. Figure 7 shows the increases in the ATR-FTIR C-H stretching frequencies of a nanoparticulate TiO2 with a 10.5% alumina 13.5% stearate coating when the washed, still-damp sample (60-70% water) was dried at 110 °C overnight (∼4% water). Upon drying the sample, the C-H stretching frequencies increased. One interpretation is that in the wet sample the stearate is a densely packed, ordered alkyl chain structure in which the chains are fully extended and that during drying the packing decreases and the chains become less extended. The view that drying causes conformational disordering is supported by the broadening of the band, due to in-phase CH2 twisting, at 1296 cm-1, in the Raman spectrum (Figure 8). This is attributed to a decrease in the degree of order in the stearate coating, due to the introduction of chain defects. An overall interpretation of the spectra is that on the wet samples the stearate groups are mainly on the alumina surface patches but after drying they occupy both titania and alumina sites. The redistribution decreases their packing density, as shown by the changes in conformation of the stearate tail. Since water molecules are known to adsorb more strongly on rutile than on alumina, we speculate that the removal of the residual few % of water during drying removes the water molecules, which, in the wet cake, disproportionately block the titania surface sites. This interpretation is consistent with the earlier conclu-
TiO2 Surface Modified with Aluminum Stearate
Figure 8. Change in Raman spectra upon drying wet (6070% water) aluminum stearate coated rutile to ∼4% water.
Figure 9. Effect of powder moisture content on viscosity for dispersions of aluminum stearate coated sample A in C12-C15 alkylbenzoate.
Figure 10. (9) uncoated rutile sample A; (0) Degussa P25; (2) sample A plus 7% stearate (ex sodium stearate); (×) sample A plus 7% stearate (ex sodium stearate) and 7% alumina.
sion3 that water decreased the bonding between carboxylic headgroups and titania. We also note that as for stearate treated titania pigment this change in the water content of stearate/alumina/titania nanoparticles affects the rheology of the derived dispersions, as shown in Figure 9. A fuller study is currently underway. Photocatalytic Activity of the Surface Modified Particles. Titanium dioxide is a well-known photocata-
Langmuir, Vol. 21, No. 7, 2005 3177
lyst,15 and since this photocatalysis often involves hydroxyl radical intermediates, the effect of the aluminum stearate layer on the photocatalytic activity of the TiO2 particles is of significant interest. We now consider the effect of aluminum stearate coatings on the photoactivity of the rutile nanoparticles. We have monitored this activity by comparing the activity of different coatings for the photocatalytic oxidation of 2-propanol to acetone. The oxidation of 2-propanol by UV irradiated titanium dioxide has been widely studied,6,16-19 and it has been shown that the catalytically active intermediates are hydroxyl radicals formed by the capture of UV generated positive holes by adsorbed or surface hydroxyl ions. Under the conditions used in this study, no acetone was generated either by TiO2 in the absence of UV or by UV radiation in the absence of TiO2. Thus, the rate of acetone formation is a convenient measure of the UV initiated generation of hydroxyl radicals at the surface of TiO2. The effect of catalyst loading on the rate of acetone formation has been reported previously by us,19 and on the basis of these results, a catalyst loading of 8 g dm-3 was selected for comparisons of catalytic activity. An earlier study,6 although the details of the deposition were different, demonstrated that increasing the alumina coating level from 7.5 to 14% approximately halved the rate of acetone production. In the present experiments, stearate by itself (sample A plus 7% stearate ex sodium stearate) significantly reduced the measured rate of acetone formation. Separate gas-phase experiments have shown that adsorbed stearate is photocatalytically oxidized to carbon dioxide. It is therefore probable that the decrease in the 2-propanol oxidation rate is at least partially due to reaction of the surface generated radicals with the stearate. The rate of acetone production was reduced further by the combination of 7% stearate with 7% alumina. Therefore, the conclusion from the photoactivity studies is that both alumina and stearate reduce 2-propanol oxidation. This study has not sought to determine the mechanism by which the alumina reduces photoactivity. However, alumina coatings are widely used to reduce the photocatalytic activity of pigment-grade titania21 and it is probable that their role is to act as a physical barrier between the surface radicals and the alcohol molecules. (Because of the wide band gap of alumina, transport of holes across the alumina layer is unlikely.) There is no evidence, for example, increasing acetone formation rate with time, that 2-propanol, which is less polar than water, displaces stearate from the surface sites responsible for hydroxyl radical generation. (Even the deliberate addition of water to the 2-propanol did not increase the rate of acetone formation.) We have shown earlier19 that the state of dispersion of a TiO2 catalyst modifies the measured photocatalytic oxidation of 2-propanol, and it is probable that the comparative activities of our samples are modified by dispersion effects. Despite this, there is a general trend of decreasing photoactivity with increasing total coating (15) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem Rev. 1995, 95, 69. (16) Cundall, R. B.; Rudham, R.; Salim, M. S. J. Chem. Soc., Faraday Trans 1 1976, 72, 1642. (17) Egerton, T. A.; King, C. J. J. Oil. Colour Chem. Assoc. 1979, 62, 386. (18) Fraser, I. M.; MacCallum, J. R. J. Chem. Soc., Faraday Trans. 1 1986, 82, 607. (19) Egerton, T. A.; Tooley, I. R. J. Phys. Chem. B 2004, 108, 5066. (20) Jin, C. FTIR Studies of TiO2 Pigmented Polymer Photodegradation. Ph.D. Thesis, University of Newcastle, 2004. (21) Wiseman, T. J.; Howard, P. B. U.S. Patent 3859115, Sept 5, 1974.
3178
Langmuir, Vol. 21, No. 7, 2005
Egerton et al.
Table 3. Rates of 2-Propanol Oxidation by 8 g dm-3 Aluminum Stearate Coated Rutile sample number
% alumina
% stearate
106 × 2-propanol oxidation rate (mol m-3 min-1)
A B C D E F G H I J
0 0 7 7 7 7 7 10 12 12
0 7 1 3 7 10 12 12 10 12
154 36 20 8 6 3 9 9 12.4 8
level, from 0 to ∼18% coating. Table 3 shows that, for a fixed stearate level of 10 or 12%, increasing alumina from 7 to 12% causes an insignificant reduction in photoactivity (samples G and J) or a small increase in photoactivity (samples F and I), perhaps because the reduction in photoactivity due to stearate is partially negated by the redistribution of stearate to the extra alumina. This redistribution may explain why samples with >10% alumina are more photoactive than predicted. Conclusions Five clear conclusions emerge from this work. First, stearate reduces the measured surface area of the titania particles and simulated BET isotherms demonstrate that this is not an artifact of creating an energetically heterogeneous surface. Second, vibrational spectroscopy demonstrates that stearate adsorbs on both the titania and alumina surfaces and the distribution between these two surfaces depends on the moisture content. Third, the conformation (and probably packing) of the stearate chains depends on the moisture content. Fourth, the effect of moisture on the stearate is mirrored by reversible changes in the rheology of derived suspensions in C12-C15 alkyl benzoate. Last, both alumina and stearate reduce the measured photoactivity of the titania. The simplest explanation of the measured decrease in surface area with increasing amount of stearate is that the stearate blocks the pores between the titania primary particles. If the diameter of the primary particles is of the order of 8 nm, the pores between them would be of the order of 1 nm, that is, of a size to be blocked by stearate molecules with a length of ∼2.7 nm. This model assumes that the stearate tails are perpendicular to the surface, as is implied by the spectroscopic results from which we inferred increasing ordering of vertically oriented stearate molecules as the stearate levels increased. For a fixed
amount of stearate (12%), the surface area decreases as the alumina level decreases from 18 to 7%, even though alumina by itself does not significantly alter the BET area. Since the spectroscopic results suggest that on dried samples the stearate adsorbs on both alumina and titania sites, we speculate that the alumina may tend to adsorb on the outside of a secondary particle with stearate then being distributed between the alumina and the titania surfaces. Increasing the amount of alumina would decrease the chance of stearate adsorbing on the titania surface at sites at which it can block pores between the titania primary units. The spectra show that the distribution of titania between the alumina and titania surfaces depends on moisture content. Water adsorbs strongly on TiO2, and this unexpected result may indicate that water displaces stearate from the rutile to the alumina surface. This conclusion is consistent with earlier reports2,3 that the dispersion of stearate treated titania in nonaqueous liquids is sensitive to the presence of traces of water. Further studies of this effect are in hand. Detailed spectroscopic analysis also confirmed that the C-H stretching absorption moved to lower frequencies as the stearate coverage increased. This observation confirms those of Britcher and co-workers who, for 200 nm particles of titania pigment, used the point at which the bands cease to shift to determine the stearate loading at which the closest packed structure has been obtained.4 Finally, the results clearly demonstrate that stearate reduces the photo-oxidation of 2-propanol and that this decrease is comparable with that obtained with alumina. The probable reason for this reduction is that the hydroxyl radicals which are formed upon irradiation may oxidize either stearate in the coating or 2-propanol molecules in the surrounding media. We have demonstrated this oxidation of the stearate in independent experiments in which a physical mixture of uncoated titania and stearate has been exposed to UV and carbon dioxide formation was monitored by IR absorbance 2350 cm-1.20 We therefore conclude that sacrificial oxidation of the stearate which coats the titania particles contributes significantly to the decrease in migration of hydroxyl radicals from the titania surface. Acknowledgment. This work was carried out as part of an EPSRC, CASE award with Solaveil business of Uniqema. The authors thank Uniqema for permission to publish this paper. It is a pleasure to acknowledge Dr. L. Kessell (Solaveil) for helpful discussion. LA047390D
