J . Chem. SOC., Faraday Trans. I , 1982, 78, 3659-3669 Photosensitised Dissociation of Water using Dispersed Suspensions of n-Type Semiconductors BY ANDREW MILLS* AND GEORGE PORTER Davy Faraday Research Laboratory of the Royal Institution, 21 Albemarle Street, London WlX 4BS Received 6th April, 1982 n-Type semiconductor powders (TiO, and SrTiO,) were used to sensitise the photochemical cleavage of water using U.V. light (A < 400 nm) under conditions of room temperature and pressure. Although several methods of piatinising these powders were used, one method in particular (involving precipitation of a platinum sol, by addition of an inert electrolyte, in the presence of the semiconductor powder) was found to produce an efficient photocatalyst for water reduction. Many different photocatalysts were tested for water reduction activity, using EDTA as an electron donor, and for water oxidation activity, using Fe3+ as an electron acceptor.In the absence of EDTA and Fe3+, U.V. irradiation of these photocatalysts liberated H, but 0, evolution was not observed. Reasons for these observations are discussed. In recent years numerous studies have been made on the use of semiconductors in photosynthetic and photocatalytic reacti0ns.l In particular, since the work of Fujishima and Honda2 a great deal of attention has focused on the use of semiconductors to cleave water photo~hemically,~~ and of the many availabL semiconductors it is the n-type oxides such as strontium titanate (SrTiO,) and titanium dioxide (TiO,) which have been most widely used, because of their high stabilities towards photo- corrosion and their favourable band energie~.~ Initially water splitting was achieved photoelectrochemically, using these n-type oxide semiconductors in the form of electrodes;, however, it now appears possible to use semiconductors in a particle form4 to carry out many of the reactions previously associated with semiconductor electrodes. In a particle system a ' short-circuited' photoelectrochemical cell (p.e.c.) may be constructed6 by depositing some platinum (Pt) onto a particle of a semicorl- ductor, the overall desired reaction occurring by electron and hole transfer at the two sites (i.e.Pt and semiconductor) on the particle. Such particulate systems appear, in general, much simpler and less expensive to use and construct than their p.e.c.counterparts. Also, materials may be used which are not available as single-crystal (or even polycrystalline) semiconductor electrodes, for reasons of high resistivity or difficulty in fabrication. Platinised semiconductor powders, such as TiO,, have been used to photo-oxidise water,4 acetic acid,lU alcohol^,^ hydrocar60ns,8 carbohydrates,'" active carbonld and biomass,lc and simultaneously reduce water. In the photochemical cleavage of water with semiconductor powders both oxidative and reductive reactions are believed to occur on the same particle, resulting in the simultaneous production of H, and 0,.9 On a particle the spatial separation between oxidative and reductive sites will be very smalllo in contrast to a p.e.c., thereby increasing the likelihood of back reaction.In order to reduce this experimental conditions are usually chosen to favour desorption of H, and 0, from the particle In general these conditions take the form of elevated temperatures and/or reduced pres~ures.~~-f Indeed Lehn et working on metallised SrTiO, powders, found that the photochemical cleavage of water only took place at reduced pressures 36593660 PHOTOSENSITISED DISSOCIATION OF WATER ( I 5 mmHgT) and was enhanced at elevated temperatures. Absence of water photo- decomposition, under conditions of room temperature and pressure, has also been reported for platinised TiO, powders by Pichat et ~ 2 1 . ~ In contrast to this, several workers have claimed water photodecomposition using semiconductor powders, under ambient conditions of temperature and pressure.For example, Bulatov and Khideke14a have claimed a quantum yield for H, production [O(H,)] of 6% for this reaction, using platinised rutile TiO, in 0.5 mol dm-3 H,SO,. However, this appears unlikely, as electrochemical studies have shown that the rutile form of TiO, does not have a sufficiently negative flat-band potential ( Vfb) to reduce water5 and, in addition, Jaeger and Bardll using e.s.r. found no evidence for the production of radicals usually associated with water decomposition (such as OH' and OH;) on U.V. irradiation of rutile (platinised and unplatinised)/H,O suspensions. Other claims for the photo- chemical cleavage of water, under conditions of room temperature and pressure, have been made by Kawai and Sakata,lr using a mixture of RuO,, TiO, and Pt powders [@(H,) = 0.0273, and by Gratzel and coworkers,12 using colloids of TiO, coupled with RuO, and/or Pt [@(H,) = 30+ 10x1. In view of these varying reports it was decided to look more closely at the photochemical cleavage of water, under ambient conditions, using platinised TiO, and SrTiO,, with the overall aim of producing a reproducible and efficient photocatalyst for this reaction.EXPERIMENTAL MATERIALS Powdered anatase titanium dioxide (TiO,) was obtained from B.D.H. (9973, Aldrich (99.9+%) and Laporte (Tiona G, 98%; Nb,O,, 0.15%). Other powders used included rutile TiO, (Tiona 010,98%; Nb,O,, 0. IS%, Laporte), strontium titanate (SrTiO,, 2j4 99.5%, Alpha Inorganics) and cadmium sulphide (CdS, 99.999 %, Koch-Lite).Ethylenediaminetetra-acetic acid (EDTA), sodium citrate, ferric chloride (Fe C1;6H20) and chloroplatinic acid (5% w/v H,Pt Cl6.6H,O) were purchased from B.D.H. (A.R. grade). A RuO,/TiO, powder was prepared, using 50 mg of ruthenium tetroxide and 5 gm of TiO, (B.D.H., anatase) in 50 cm3 of water. The reaction mixture was placed in a 100 cm3 stoppered glass flask, stirred for 5 days, filtered and the grey powder residue dried in air. This powder has been found to catalyse the reduction of cerric ions in aqueous s01ution.l~ METHODS Methods used for platinisation of the semiconductor powders were as follows : (A) An N,-purged suspension was prepared containing 1 gm of the semiconductor powder, 15 mg of chloroplatinic acid and 10 cm3 of a 40% formaldehyde solution, stabilised by methanol (B.D.H.).This suspension was subsequently stirred continuously on irradiation for 8 h in a quartz vessel with a 250 W Hg medium-pressure lamp (this method is analogous to that reported by Gratzel et a1.).12 (B) A Pt/citrate sol was prepared by refluxing, for 4 h, a solution containing 30 mg of chloroplatinic acid, 30 cm3 of a 1 % sodium citrate solution and 120 cm3 of water. A third (50 cm3) of the resultant Pt sol was stirred with 1 gm of the semiconductor powder and 5.8 gm of sodium chloride added. The destabilisation of the sol, followed by Pt precipitation, appeared to be complete within seconds of the addition. (C) A Pt sol was prepared as described in (B), in the presence of 1 gm of the semiconductor. (D) A suspension was prepared, containing 1 gm of the semiconductor powder in 50 cm3 of a Pt sol [described in (B)] and the solvent subsequently removed by rotary evaporation.(E) A suspension was prepared containing 1 gm of the semiconductor powder and 15 mg of chloroplatinic acid in 10 cm3 of acetic acid (1 mol dm-,). Saturation of this suspension with H, brought about the reduction of the Pt salt to the metal. For all the methods described above, t 1 mmHg = 13.5951 x 980.665 x 10F Pa.A. MILLS AND G. PORTER 3661 the final platinised powder suspensions were filtered and repeatedly washed with distilled water before being dried in air. All steady-state experiments were performed with an Applied Photophysics clinical reactor, using a 900 W Xe lamp and a cold-water infrared filter.A cut-off filter was used to restrict the light to wavelengths > 400 nm for some measurements, and a high-radiance monochromator used for action-spectra measurements. The concentration of evolved oxygen or hydrogen was measured with a Clark membrane electrode, purchased from Rank Bros. A detailed description of the experimental arrangement, sensitivity and calibration of this instrument is given elsewhere.l4 The solutions were stirred and thermostatted at 25 0.5 "C and H, and 0, formation confirmed by gas chromatography. Before each experiment the solutions were sonicated for ca. 1 min in a Dawe ultrasonic bath to disperse the powder. The solutions were then transferred to the photochemical cell and purged with oxygen-free nitrogen prior to illumination.The initial rate of H, (or 0,) evolution [or R(02J) was determined from the concentration of H, (or 0,) evolved over the first 4 min of irradiation. Photoacoustic spectra were recorded at the City University with the help of Mr C . Morrison. RESULTS AND DISCUSSION SACRIFICIAL SYSTEMS The photoreduction of water to H,, using just semiconductor powders, has been achieved in several systems but with low efficiencie~.*~?l~ This may be improved by depositing Pt onto the semiconductor powder particles, thus lowering the overpotential for water reduction.1° There are two major methods used to deposit Pt onto semiconductor powders. Developed by Gratzel et a1.l29 l6 the first method involves the preparation of a Pt/citrate sol, followed by removal of the protective citrate (using an ion-exchange resin), addition of the semiconductor powder and sonication of the resulting mixture.However, using this procedure we have encountered problems, including Pt adhering to the resin and incomplete precipitation of the non-protected Pt sol onto the semiconductor powder. Preferred12b over the citrate reduction method (because it is easier to perform and gives results of excellent reproducibility) is the second method of platinisation, i.e. photoplatinisation. Although there are several variations of this method, they all involve the U.V. irradiation of a semiconductor powder suspended in a solution containing a Pt salt and, usually, an electron donor (e.g. acetic acid,17 forrnaldehydel2 or ethanolle). In general this procedure is suitable only for photostable semiconductors because long, high-intensity irradiations are involved.With this in mind, several alternative [(B)-(E)], as well as established (A), platinisation techniques were used for producing photocatalysts, in order to determine a facile routine method of depositing Pt onto a support, resulting in reproducible batches of catalyst without elaborate techniques. The photocatalysts produced by these methods were compared1* for water reduction activity using a test system incorporating an electron donor (such as EDTA) to scavenge, efficiently and irreversibly, the ' hole' of the 'electron-hole pair' produced on irradiation of the semiconductor with band-gap light. The remaining conductance- band electron should then be able to migrate to a Pt site, where water reduction can occur.In a typical experiment a sonicated suspension of a platinised semiconductor powder (37 mg) in EDTA solution (37 cm3, lo-, mol dm-3) was placed into the Pyrex irradiation cell of a H,-detecting Clark membrane electrode, purged with N,, and the initial rate of H, production [R(H,)] determined on irradiation. The results of this work are shown in table 1 , from which it appears that only method (B) produces photocatalysts of similar (if not greater) activity to those produced by a photo- platinisation technique [i.e. method (A)]. Using method (A), other metals (Rh, Ru,3662 PHOTOSENSITISED DISSOCIATION OF WATER TABLE 1 .-COMPARISON OF THE ACTIVITIES OF DIFFERENT PHOTOCATALYSTS TOWARDS WATER REDUCTION AND OXIDATION R ( H 2 ) a in R(oz)a in type of semiconductor method of lo-, mol dm-, R(H2)a in lo-, mol dmP3 support platinisation EDTA (%) H,O (%) FeC1, (%) SrTiO, SrTiO, SrTiO, SrTiO, TiO, (anatase, B.D.H.) TiO, (anatase, B.D.H.) TiO, (anatase, B.D.H.) TiO, (anatase, B.D.H.) TiO, (anatase, B.D.H.) TiO, (anatase, B.D.H.) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Aldrich) TiO, (anatase, Laporte) TiO, (anatase, Laporte) TiO, (rutile, Laporte) TiO, (rutile, Laporte) TiO, (anatase, B.D.H.)/RuO, TiO, (anatase, B.D. H .)/RuO, TiO, (anatase, B.D.H.)/RuO, CdS CdS A1203 A1203 none B none B none 5 3 4 0 28 46 11 22 14 0 96 100 13 63 2 5 8 2 0 48 0 21 0 40 24 4 136 3 0 0 0 0 0 0 2.0 2.6 0 1.6 0.7 0 6.0 10.8 1.6 4.1 0 0 0 0 0 3.7 0 0 0 0 0.3 0 0.5 0 0 0 8.0 5.9 8.8 9.6 9.8 5.7 1 .o 5.2 10.3 21 .o 15.5 13.7 13.4 16.8 15.8 17.0 25.8 23.2 8.5 5.4 4.0 0 0 a A relative rate of H, production [R(,,J or oxygen production [R(,,,] of 100% corresponds to 1.2 x irradiation of a platinised [method (B)] TiO, powder (Aldrich, 37 mg), in the presence of a lo-, mol dm-, EDTA solution (37 cm3), with light of 360 k 20 nm and intensity 10l6 photon s-l (as determined using a calibrated thermopile) resulted in H, production with a formal quantum yield15 of 3 & 1 %.mol (gas) min-l; Ir, Co) were deposited on TiO, from their respective salts4h but only Rh showed any appreciable photocatalytic activity (see table 1).Surprisingly, SrTiO, appears a less active photocatalyst for water reduction than anatase TiO,, even though its flat-band potential (&) is more red~cing.~ Also, R(H2) values for anatase TiO, appear dependent on the commercial source of this material. One possible explanation of these results may lie in the surface chemistry dependence of n-type oxides on their method of preparation and subsequent chemical history.lg (This may also be one of the reasonsA. MILLS AND G . PORTER 3663 for the varying reports concerning photochemical water cleavage using these semi- conductors.) Interestingly, water reduction was also achieved in EDTA solution using a platinised rutile TiO, powder, even though this reaction is thermodynamically unfavourable (AG > 0), for reasons stated previously. However, niobium oxide (Nb,O,) is known to shift V,, for rutile cathodically,12b720 and this oxide is present as an impurity (0.15 %) in the rutile TiO, powder used.Accumulation of negative charge (or conductance-band electrons) at a Pt site owing to the irreversible nature of the EDTA oxidation by valence-band holes may also contribute towards the shift in Gb, increasing the likelihood of water reduction. This latter effect would not be expected in the absence of EDTA and, indeed, no H, evolution was observed with rutile under such conditions. It is known from electrochemical studies that RuO, electrodes exhibit overvoltages for H, evolution similar to those for Pt,,l and recently Amouyal et ~ 1 . ~ ~ reported RuO, to be an effective redox catalyst for H, generation from a Ru(bipy)t+/methyl viologen/EDTA sacrificial system.However, Gratzel and have reported evidence that, when bound to TiO,, RuO, and Pt specifically catalyse water oxidation and reduction, respectively. We have found RuOJTiO, powders on irradiation with U.V. light, in either the presence or absence of EDTA, showed little ability to reduce water. In agreement with the results of Darwent and P ~ r t e r , ~ ~ ~ ~ irradiation (2 > 400 nm) of a CdS powder (1 mg ~ m - ~ ) , suspended in a EDTA solution (37 cm3), resulted in H, evolution, the rate of which was enhanced greatly (ca. x 45) by platinising the CdS powder using method (B). Harbour et al.15" have recently reported no H, production from CdS powders except in the presence of both Pt and EDTA, but it is worth noting that method (B) represents a quick, easy and efficient method of platinisation of powders which, in contrast to established photoplatinisation techniques, is amenable to platinisation of powders not photostable to U.V.light (e.g. CdS, CdSe and dye-coated semiconductors) or which do not absorb light (e.g. Al,03).17a7 24 The rate of water reduction, sensitised by the photocatalysts, in the presence of X/nm FIG. 1.-Action spectrum of platinised [method (B)] anatase TiO, (B.D.H., 37 mg) in EDTA solution (0.1 mol drnp3, 37 cm3).3664 PHOTOSENSITISED DISSOCIATION OF WATER EDTA showed a strong wavelength dependence, as illustrated for platinised anatase TiO, (fig. 1). Hydrogen was only produced by light of 3, -c 400 nm since, above this wavelength, no light is absorbed by anatase TiO,, which has a band-gap of 3.2 eVll (corresponding to 388 nm).This action spectrum (fig. 1) matched the photoacoustic spectrum recorded for the photocatalyst which, in turn, was identified from the literaturez5 as anatase TiO,. Work with Al,O, (Ebg > 7 eV24) indicated that the oxide support must be light-absorbing for H, evolution to occur, and that light (A > 300 nm) absorbed by EDTA does not lead to detectable H, production. The photocatalysts were also tested for water oxidation activity by using an electron acceptor (FeC1,) to scavenge the ‘electron’ of the ‘electron-hole pair’ which is produced on irradiation of the semiconductor with band-gap light. This allows the remaining valence-band ‘hole’ to migrate to the semiconductor surface where water oxidation can occur.Using a procedure similar to that described for the determination of R(H,) [repiacing EDTA with FeCl, (37 cm3, mol dm-, H,SO,)] the initial rate of 0, production [R(o,,] was determined for each catalyst (37 mg) with an 0, Clark membrane electrode, and the results are also shown in table 1. From these results it would appear that platinisation reduces the abilities of the semiconductors to oxidise water. This may partly be the result of decreased light absorption by the semiconductor, due to the Pt deposited onto its surface. Also, thermodynamically, there is the possibility of Pt competing with water for oxidation. However, although Pt oxidation is claimed by Kawai and Sakata26 to occur in a system containing Pt, TiO, and active carbon, Sat0 and Whiteld found no evidence for this but rather that platinum oxide is reduced by TiO, on U.V.irradiation. Although the valence-band ‘ hole ’ is an extremely oxidising species, having an overpotential for water oxidation of ca. 1.9 eV, some workers have claimedlc. improved yields by coupling or mixing the TiO, particles with RuO, (one of the best electrode materials for water oxidation).lc! 27 However, a RuO,/TiO, powder (see table l), upon illumination in the presence of FeCl,, showed no enhancement of R(OP) compared with the TiO, support (similar results were found for W03).,* From table 1, Co deposited onto TiO, appears to enhance R(02), although Sat0 and WhiteLd claim that Co (along with Ni) is readily photo-oxidised by TiO,.Interestingly, Co2+ ions have r e ~ e n t l y ~ ~ ~ ~ ~ been used to catalyse water oxidation by Ru3+, Os3+ and Fe3+ tris-bipyridyls. Work with Al,O, indicated that the oxide support must be light absorbing for 0, evolution to occur and, although Fe3+ ions are known31 to oxidise water on irradiation with U.V. light, no evidence for this was found under the experimental conditions of pH and concentration used. mol dm-, in 5 x NON-SACRIFICIAL SYSTEMS From the above work it is clear that platinised TiO, and SrTiO, can photochemically reduce or oxidise water in the presence, respectively, of an electron donor or acceptor. However, in the photochemical cleavage of water both processes would have to occur simultaneously, i.e.(1) hv semiconductor .+ e- + h+ electron-hole pair 2e-+ 2H+ 2 H, 4h+ + 2H,O -, 0, + 4H+A. MILLS AND G . PORTER 3665 4 I 0 time/min FIG. 2.-Typical H, concentration, measured on a Clark membrane electrode, against time profiles on U.V. irradiation of several aqueous anatase TiO, (B.D.H.) suspensions (37 mg of semiconductor dispersed in 37 cm3 of H,O), platinised by the following methods: (a) B; (b) A; (c) E and ( d ) none. The arrow denotes the start of U.V. irradiation. leading to H, and 0, evolution from the same particle. The rates of H, evolution in water (i.e. no EDTA) for each of the photocatalysts were determined (see table 1) and, although reduced by ca. 10 times, appeared to follow similar trends to those found for water reduction in EDTA (an example of some typical H, concentration against time profiles is given in fig.2). By varying the amount of chloroplatinic acid used in method (A), or the amount of Pt colloid used in method (B), the estimated Pt concentration (expressed as a percentage of Pt) deposited onto the semiconductor powder may be varied. The results of this work, using platinised TiO,/water suspensions, are shown in fig. 3 and indicate, for both procedures, an optimum Pt concentration of ca. 0.5% (ca. 20 times lower than those used by other workers).l29 1 7 b 9 24 Also, from fig. 3(a), at Pt levels above the optimum, R,H2, gradually falls (presumably due to the Pt screening the TiO, surface, thereby reducing the actual light intensity ‘seen’ by the semiconductor). This would indicate that Pt deposit levels > 0.50,; are not only unnecessary but may also be detrimental towards the rate of H, production.The rate of water reduction using a semiconductor photocatalyst was found also to be a function of semiconductor concentration ([SC]), as shown in fig. 4. The against [SC] profile is very similar to that reported by Rao et aL3, for ZnO and may be explained as follows. Initially, by increasing [SC] the amount of light absorbed by the semiconductor increases and therefore increases (A-B). Eventually a point is reached at which all the incident light is absorbed and R(H,) can increase no further (B). A further increase in [SC] only reduces the penetration depth of the incident light. However, this may well increase the likelihood of losing scattered light to the exterior and, in turn, account for the reduction of R(H,) with further increases in [SC]. Reduction of the penetration depth to such a level that the light loss due to scattering is almost constant may then account for the levelling-off of [i.e. (C-D)] at higher with pH for a platinised TiO, powder [SCI - Fig.5 illustrates the variation of3666 PHOTOSENSITISED DISSOCIATION OF WATER I I I 2 Pt (%) ( b ) 100- * 1 I 0. I 0.2 0.3 Pt (%) FIG. 3.-Relative rate of H, production [R!H?)] on U.V. irradiation of aqueous suspensions (1 mg ~ m - ~ , 37 cm3) of anatase TiO, (Aldrich) platinised by (a) method (A) [R(,2, = 100 corresponds to 1.3 x mol dm-3 (H,) min-l] or (b) method (B) [R&) = 100 corresponds to 9 x lo-' mol dm-3 (H,) min-'1. dispersed in water.From electrochemical the conductance and valence bands of TiOz electrodes remain constant relative to the reduction and oxidation potentials of water regardless of pH, and therefore (assuming TiO, particles also show this behaviour) should not produce any variation of with pH. In contrast, the surface nature of TiO, is pH-dependent, and for anatase TiO, two acid-base equilibria are known:34 -Ti-OH + H++ -Ti-OHi pKa,-4.98 -Ti-0-+ H++ -Ti-OH p Ka2- 7 .a (4) resulting in a point of zero zeta potential35 (P.z.z.P.) of 6.39. Work by Bard et aZ.3s on anatase TiO, particulate systems indicates that platinisation shifts the p.z.z.p. toA. MILLS AND G . PORTER 3667 I 6 FIG. 0 5 0 100 I50 [ SC I Img 4.-Relative rate of H, production [R(&)] on U.V. irradiation of varying amounts of semiconductor (in this case, platinised [method (B)] anatase TiO, (B.D.H.)} dispersed in 37 cm3 of H,O.I 1 I 0 I 2 3 4 5 6 7 8 9 I011 PH FIG. 5.-Relative rate of H, production [R(H2)] as a function of pH, for a platinised [method (B)] anatase TiO, (B.D.H.) powder (37 mg) dispersed in 37 cm3 of solution. Variation of pH was achieved by adding H,SO, or NaOH. lower pH values (ca. pH 4), probably due to anion (Cl-) adsorption. This p.z.z.p. (i.e. pH 4) corresponds roughly to the position of maximum R(H2) (see fig. 5), indicating that water reduction occurs best on a neutrally charged surface. A positively charged surface and, to a lesser extent, a negatively charged surface (produced by lowering or raising the pH, respectively) appear less favourable for water reduction.However, other factors may also be involved, for example TiO, shows ageing effects in H,SO, 37 and is soluble in this and alkaline In addition, it was noticed that the ease of dispersion of the photocatalyst decreased with decreasing pH.3668 PHOTOSENSITISED DISSOCIATION OF WATER So far we have described only the evolution of H, on U.V. irradiation of our photocatalysts in the presence of water. However, under identical irradiation conditions as those used above, no 0, evolution was observed. Although measurement of small concentrations of 0, in the presence of H, with a Clark electrode did prove difficult,14 this result was confirmed by gas chromatography. Results similar to these have been reported by Gratzel et al.,g712 where several hours of irradiation lapse before 0, evolution is observed on U.V.irradiation of Pt/RuO,/TiO, colloids. This effect has been attributed12b to the time taken for O,, which is readily photoadsorbed by TiO, as Oi-,39 to occupy all the vacant adsorption sites and has been suggestedg as a means of separating H, and 0, produced from such systems. Blank experiments involving U.V. irradiation of air-saturated aqueous suspensions of TiO, (and Pt/TiO,) confirmed that oxygen is readily photoadsorbed onto these powders (ca. 2 x lop6 mol 0, for 40 mg semiconductor dispersed in 40 cm3 water). The platinised forms of TiO, appeared to photoadsorb 0, faster ( t = 10 min) than TiO, ( t = 20 min). Attempts to displace the photoadsorbed 0, (produced on irradiation of a Pt/TiO, powder suspended in water) by addition of sodium phosphateg (Na,PO,, 5 x lo-, mol dm-3) proved ineffective.Also, irradiation of a Pt/TiO, powder suspen- sion (40 cm3, 1 mg ern-,) in the presence of 0.1 mol dm-, Na,PO, solution not only failed to liberate 0, but prevented water reduction as well. Similar inhibiting effects for Na,PO, have been found by Malati and Seager40 in the photo-oxidation of primary alcohols by TiO, powders and were attributed to the adsorption of phosphate ions41 onto the surface sites, rendering them less reactive. Although most of the irradiations performed on the photocatalysts were of short duration (ca. 30 min), initial work on prolonged irradiations ( t > 3 h) of Pt/TiO, (anatase) suspensions indicated that no 0, is evolved and H, evolution ceased under such conditions.This may well be the result of 0, (and Oi-) accumulation onto the TiO, particle’s surface, to such an extent that reduction of 0, and oxidation of Oi-- become themajor photoprocesses involved. However, regeneration ofthe photocatalyst appears possible by saturating the suspension with N,, although the reasons for this remain unclear. The adsorption of one or more of the photochemically produced species on to a semiconductor particle’s surface may prove an important limiting factor when considering such materials for use in practical solar-energy devices. few methods exist for probing the details of the reaction mechanisms or the catalyst properties in situ. Photoacoustic or reflectance spectroscopic techniques allow us to determine the energy of the band of the powders themselves, and spin trapping has been employed to identify intermediate radicals formed in these processes.ll The fact that the particles scatter and reflect as well as absorb light makes evaluation of anything but formal quantum yield^^^^,^ difficult. In a recent development Bard and employed photoelectrophoretic and electrochemical techniques to characterise a particulate TiO, photocatalyst.Work is currently in progress in this laboratory using techniques such as these to achieve a better understanding of the semiconductor particle systems described above. As pointed out by Bard and We thank the S.E.R.C., the E.E.C. and G.E. (Schenectady) for financial support of this work. We are indebted to Miss M. L. Zeeman for her assistance.We thank Mr C. Morrison and the City University for the photoacoustic spectroscopy and Dr A. Harriman and Dr J. Darwent for many helpful discussions.A. MILLS AND G. PORTER 3669 1 (a) B. Kraeulter and A. J. Bard, J. Am. Chem. SOC., 1978, 100, 2239, 5985; 1977, 99, 7729. (b) S. N. Frank and A. J. Bard, J. Phys. Chem., 1977, 81, 1484. (c) T. Sakata and T. Kawai, Nouv. J. Chim., 1981, 5, 279; Nature (London), 1980, 286, 474; 1979, 282, 283. ( d ) S. Sat0 and F. M. White, J. Phys. Chem., 1981,85, 336. (e) J. C. Hemminger, R. Carr and G. A. Somarjai, Chem. Phys. Lett., 1978, 57, 100. cf) S. Sato, J. Chem. SOC., Chem. Commun., 1982, 26. A. Fujishima and K. Honda, Nature (London), 1972, 238, 37; Bull. Chem. SOC. Jpn, 1971, 44, 1148. (a) M.S. Wrighton, D. S. Ginley, P. T. Wolczanski, A. B. Ellis, D. L. Morse and A. Lintz, Proc. Natl Acad. Sci. USA, 1975, 72, 1518. (b) M. S. Wrighton, A. B. Ellis, P. T. Wolczanski, D. L. Morse, H. B. Abrahamson and D. S. Ginley, J. Am. Chem. SOC., 1976,98,2774. (c)T. Watanabe, A. Fujishima and K. Honda, Bull. Chem. SOC. Jpn, 1976, 49, 355. * (a) A. V. Bulatov and M. L. Khidekel, Izu. Akad. Nauk SSSR, Ser. Khim., 1976, 1902. (6) G. N. Schrauzer and T. D. Guth, J. Am. Chem. Soc., 1977, 99, 7189. (c) H. van Damme and W. K. Hall, J. Am. Chem. SOC., 1979, 101, 4373. ( d ) S. Sat0 and J. M. White, Chem. Phys. Lett., 1980, 72, 83. (P) T. Kawai and T. Sakata, Chem. Phys. Lett., 1980, 72, 87. (f) K. Domen, S. Naito, M. Soma, T. Onishi and K. Tamaru, J. Chem. SOC., Chem. Commun., 1980, 543.Cg) F. T. Wagner and G. A. Somorjai, Nature (London), 1980,285, 559. (h) J. M. Lehn, J. P. Sauvage and R. Ziessel, Nouu. J. Chim., 1980, 4, 623. H. P. Maruska and A. K. Ghosh, Solar Energy, 1978, 20, 443. P. Pichat, J-M. Herrmann, J. Disdier. H. Courbon and M-N. Mozzanega, Nouv. J. Chim., 1981, 5, 627. S. Sato and J. M. White, Chem. Phys. Lett., 1980, 70, 131. M. Gratzel, Acc. Chem. Rex, 1981, 14, 376. P. C. Jaeger and A. J. Bard, J. Phys. Chem., 1979, 83, 3146. J. Kiwi, E. Pelizetti, M. Visca and M. Gratzel, J. Am. Chem. SOC., 1981, 103, 6324. Soc., Dalton Trans., 1982, 1213. ti A. J. Bard, J. Protochem., 1979, 10, 50. lo A. J. Bard, J. Phys. Chem., 1982, 86, 172. l 2 (a)D. Duonghong, E. Borgarelloand M. Gratzel, J . Am. Chem. Soc., 1981,103,4685.(b) E. Borgarello, l3 (a) A. Mills and M. L. Zeeman, J. Chem. SOC., Chem. Commun., 1981, 948. (b) A. Mills, J . Chem. l4 A. Mills, A. Harriman and G. Porter, Anal. Chem., 1981, 53, 1254. l5 (a) J. R. Darwent and G. Porter, J . Chem. Soc., Chem. Commun., 1981, 145. (b) J. R. Darwent, J. Chem. SOC., Faraday Truns. 2, 1981, 77, 1703. (c) J. R. Harbour, R. Wolkow and M. L. Hair, J. Phys. Chem., 1981, 85, 4026. P. A. Brugger, P. Cuendet and M. Gratzel, J. Am. Chem. SOC., 1981, 103, 2923. A. J. Bard, J. Am. Chem. SOC., 1978, 100, 4317. A. Mills, J. Chem. SOC., Chem. Commun., 1982, 367. l9 G. D. Parfitt, Prog. Surf. Membr. Sci., 1976, 11, 181. 2o P. Salvador, Sol. Energy Mater., 1980, 2, 413. 21 D. Galizzioli, F. Tantardini and S. Trasatti, J. Appl. Electrochem., 1974, 4, 57. 22 E. Amouyal, P. Keller, A. Moradpour, J. Chem. Soc., Chem. Commun., 1980, 1019. 23 R. Humphry-Baker, J. Lilie and M. Gratzel, J. Am. Chem. Soc., 1982, 104, 422. 24 W. W. Dunn and A. J. Bard, Nouv. J. Chim., 1981, 5, 651. 25 M. J. Adams, B. C. Beadle, A. A. King and G. F. Kirkbright, Analyst (London), 1976, 101, 553. 26 T. Kawai and T. Sakata, J. Chem. SOC., Chem. Commun., 1979, 1047. *: M. Morita, C. Iwakura and H. Tamura, Electrochim. Acta, 1978, 23, 331. 28 J. R. Darwent and A. Mills, J. Chem. Soc., Faraday Trans. 2, 1982, 78, 359. 29 V. Ya. Shafirovich, N. K. Khannov and V. V. Strelets, Nouu. J. Chim., 1980, 4, 81. 30 A. Harriman, G. Porter and P. Walters, J. Chem. Soc., Faraday Trans. 2, 1981, 77, 2373. 31 A. A. Krasnovsky and G. P. Brin, Dokl. Akad. Nauk SSSR, 1962, 147, 656. 32 M. V. Rao, K. Rajeshwar, V. R. Pal Verneker and J. Du Bow, J. Phys. Chem., 1980, 84, 1987. 33 E. C. Dutoit, F. Cardon and W. P. Gomes, Ber. Bunsenges. Phys. Chem., 1976,80, 475. 34 P. W. Schindler and H. Gamsjager, Discuss. Faraday SOC., 1971, 52, 286. 3s B. Reichman and C. E. Byvik, J. Electrochem. SOC., 1981, 128, 2601. 37 L. A. Harris and R. H. Wilson, J. Electrochem. SOC., 1976, 123, 1010. 3H Handbook of Chemistry and Physics, ed. R. C. Weast (CRC Press, Cleveland, Ohio, 61st edn, 1980). 3y (a) A. H . Boonstra and C. A. H. A. Mutsaers, J. Phys. Chem., 1975, 79, 1694. (6) G. Munuera, V. Rives-Amau and A. Saucedo, J. Chem. SOC., Faraday Trans. 1, 1979,75,736. (c) A. R. Gonzalez- Elipe, G. Munuera and J. Sona, J. Chem. SOC., Faraday Trans. 1, 1979, 75, 748. l7 (a) A. J. Bard, W. W. Dunn and B. Kraeulter, U S . Patent, 1981, no. 4 264 421. (6) B. Kraeutler and W. W. Dunn, Y. Aikawa and A. J. Bard, J . Am. Chem. SOC., 1981, 103, 3456. 40 M. A. Malati and N. J. Seager, J . Oil Colour Chem. Assoc., 1981, 64, 231. 41 H. P. Boehm, Discuss. Faraday SOC., 1971, 52, 264. (PAPER 2/585)