J . Chem. SOC., Faraday Trans. 1, 1984,80, 183-189 Photochemical Reduction of Oxygen Catalysed by Colloidal Cadmium Selenide BY JAMES R. DARWENT Department of Chemistry, Birkbeck College, University of London, Malet Street, London WClE 7HX Received 12th May, 1983 Colloidal CdSe sensitises the photoreduction of 0, and methyl viologen (MV2+) by cysteine. In the absence of cysteine the semiconductor is rapidly photo-oxidised. Rates of 0, photore- duction and yields of MV'+ tend towards a maximum value for MV2+ concentrations of 1 O-, mol dm-3. Flash-photolysis and steady-state measurements of 0, reduction are interpreted by a mechanism which shows that the low quantum yields are due to rapid recombination of photogenerated charge carriers. Only 2 of the photogenerated electrons are available to reduce 0,.Photosensitized electron-transfer reactions at the surface of semiconductor particles may provide a route for the conversion of solar These reactions are also important in photography and the photo-oxidation of pigments. Recent research has shown that dispersions of semiconductor particles can photodissociate water into H, and 0,.3-5 Most of this work has focussed on oxide semiconductors, which have large band gaps and absorb very little visible radiation, but a number of research groups have also shown that CdS can sensitise the photoreduction of water with visible light. Krasnovskii et aL6 were the first to demonstrate this reaction using CdS, methyl viologen (MV2+) and the enzyme hydrogenase. More recent work has shown that the enzyme can be replaced by metal catalysts (Pt, Rh or RuO,) deposited on the CdS particles, in which case the viologen relay is no longer Although CdS absorbs light up to 550 nm this represents only a small fraction of the solar spectrum, and ideally the absorption threshold should be in the region of 900 nm.12 Platinised semiconductor particles with band gaps in this region (Si, CdSe and copper phthalocyanine)s can sensitize the reduction of water, but the quantum efficiency is invariably orders of magnitude less than that obtained with CdS.The photosensitised reduction of 0, by colloidal CdSe has been investigated to provide a better understanding of the poor efficiencies of these low-band-gap semiconductors, and our results are presented below. EXPERIMENTAL Steady-state experiments were performed with an Applied Photophysics UV90 photoirrad- iation system using a 900 W Xe lamp and a grating monochromator.Light intensities were measured with a calibrated thermopile and absorption spectra were recorded with a Perkin-Elmer 200 spectrophotometer. Conventional flash-photolysis measurements were made with an Applied Photophysics K200 system using a nitrate filter to remove wavelengths < 400 nm as described previ0us1y.l~ Solutions for the flash-photolysis experiments were purged for 1 h with N,. Oxygen concentrations were measured with a Clark membrane oxygen electrode. A detailed 183184 PHOTOREDUCTION OF 0, USING COLLOIDAL CdSe CATALYST 0 I I 1 0 5 10 15 20 tlmin Fig. 1. Photosensitised reduction of 0, (MV2+, lo-, mol dm-,; cysteine, lo-, mol dm-3; CdSe, 2 x lo-, mol dm-3; NaLS, mol dm-3; phosphate buffer, lo-, mol dm-3; pH 6.8; A,, = 475 nm; IA = 3 x loA7 ein s-l).description of the experimental arrangement, sensitivity and calibration of this instrument is given in ref. (14). The solution (37 cm3) was stirred and thermostatted at 35 OC. Colloidal CdSe was prepared by passing H,Se gas into a solution of Cd(NO,), (5 x mol dm-,) in phosphate buffer (pH 7.0, lo-, mol drn-,) with lo-, mol dm-, sodium lauryl sulphate (NaLS) as a supporting surfactant. H,Se was generated by adding solid CdSe to dilute hydrochloric acid. Both the acid and the cadmium solution were continuously purged with N, to avoid reaction of 0, with H,Se. The solution was then purged overnight to remove excess H,Se.The average radius of the particles was determined by transmission electron microscopy. Cd(NO,), (AnalaR) and NaLS (specially purified for biochemical work) were purchased from B.D.H. Solid CdSe (99.999%) was from Koch-Light. Water was doubly distilled. RESULTS In the last two years a number of research groups have investigated electron-transfer reactions in colloids of semiconductor particles of either Ti0215-17 or CdS.,? 159 These colloids are transparent and stable for a period of weeks and so provide a suitable medium in which to study photoredox reactions at the semiconductor/liquid interface. No previous work has been reported on the photochemical reactions of colloidal CdSe, which has a band gap of 1.7 eV and represents an attractive semiconductor for the conversion of solar energ~.~~q 26 Colloidal CdSe was prepared by bubbling H,Se into an oxygen-free solution ofcadmium nitrate ( 5 x lod3 mol dm-3) and sodium lauryl sulphate (NaLS) ( lo-, mol dm-3).This resulted in a relatively clear brown colloid containing particles with an average radius of 100 nm. Optical-absorp- tion studies showed that these colloids were stable for several weeks. Photosensitised reduction of 0, provides a simple model system with which to study electron-transfer reactions of semiconductor particle^.^^-^^ The reaction can be3 d I v) 3 2 E E m . n 0 7 S ' 0 0 J. R. DARWENT 5 10 [MV2+]/10-3 mol dm-3 185 5.0 T E a - ii! I- = 5 . - t 2.5 > 0 Fig. 2. Yield of MV'+ (a) and rate of 0, photoreduction (0). (Unless otherwise stated conditions as in fig.1.) monitored continuously with a membrane polarographic detector, and unlike the photoreduction of water no additional catalysts are needed. When colloidal CdSe is illuminated in the presence of 0, the semiconductor is photo-oxidised and a red precipitate of Se forms in a period of minutes. This reaction can be prevented by the presence of a sacrificial electron donor, such as cysteine, which is preferentially oxidised. This approach has been particularly successful with conventional semiconductor e l e c t r ~ d e s . ~ ~ ~ 26, 30 Colloidal CdSe will then sensitise the photoreduction of 0, by cysteine with no appreciable loss of CdSe. The rate of this reaction is dramatically increased by low concentration of methyl viologen. Fig. 1 shows a typical example of the concentration-time profile for the photoreduction of 0,.The rate is essentially independent of 0, concentration for the first two half-lives. In the dark, the reduction of 0, by cysteine is insignificantly slow at this pH and temperature. The effect of MV2+ concentration on the rate of 0, reduction is presented in fig. 2. This shows that the rate tends to a maximum value for [MV2+] > lo-, mol dm-3, at which point the rate is 7.5 times faster than in the absence of MV2+. Microsecond flash photolysis was used to provide a better understanding of the mechanism of this system. When CdSe colloids were illuminated with a short flash of visible light (3, > 400 nm) in the presence of cysteine ( lo-, mol dmP3) and MV2+ ( 10-4-10-2 mol dmP3) the absorption spectrum of reduced viologen (MV'+) (Amax 395 and 605 nm) was observed.In the absence of 0,, MV'+ was formed as a permanent product. All the MV'+ was produced during the 10 ps photoflash. The yield of MV'+ was proportional to the flash intensity and showed a similar dependence on [MV2+] to that found for the rate of 0, reduction (fig. 2).186 PHOTOREDUCTION OF 0, USING COLLOIDAL CdSe CATALYST DISCUSSION The results described above can be described by the mechanism in scheme 1. hv + CdSe Cd Se( h +e -) CdSe(h+e-) + cysSH CdSe(e-) + MV2+ MV'++O, CdSe(e-) (cysSH (light absorption) (1) -+ CdSe(h+e-) -+ CdSe (charge recombination) (2) -+ CdSe(e-) + $(cysS), + H+ (oxidation of cysteine) (3) (4) -+ CdSe + MV'+ -+ MV2++0i- (oxygen reduction) ( 5 ) + CdSe (electron loss) (6) (electron trapping) = cysteine). Scheme 1.Absorption of light by the CdSe particles leads to the formation of an electron-hole pair (e- h+). In the absence of a sacrificial electron donor h+ will recombine with e- or oxidise selenide, leading to the formation of selenium and the destruction of the semiconductor. Cysteine can act as a suitable electron donor and prevent photo- oxidation of the particle^.^^ 8 y 2 5 1 2 6 q 30 A fraction of the photogenerated electrons are then trapped at a high negative potential in the conduction band of CdSe (ECB = - 1.2 V us NHE)31 and are available to reduce 0, to 0,- (E, = - 0.16 V).279 32 This is a slow reaction, presumably because of the low solubility of 0, in water (ca. 2 x 1 0-4 mol dm-3 under these experimental conditions) and alternative reactions, such as the reduction of cystine and Cd2+, will result in electron loss.The rate of 0, reduction is increased by low concentrations of MV2+ ( 10-3-10-2 mol dm-3) which will bind to the anionic surfactant close to the surface of CdSe and accept electrons from the conduction band of the semiconductor. Subsequent electron transfer from MV'+ to 0, is known to be a diffusion-controlled reaction.33 The energetics for these reactions are summarised in fig. 3. Under these experimental conditions, 0;- is known to disproportionate, so that the overall reaction is oxidation of cysteine and reduction of oxygen to H,0,.32 Applying the steady-state approximation to the concentration of e- and MV'+ leads to the following equation for the rate of oxygen reduction: 1 k6 -+ - 1 R - R, aI, [MV' +]k,al,' (7) R and R, (mol s-l) are the rates of 0, reduction in the presence and absence of MV", a is the fraction of photogenerated electrons which are trapped [this will depend on the relative rates of reactions (2) and ( 3 ) ] , I , (ein s-l) is the rate of absorption of photons, k,(s-l) is the rate of electron loss, and k , is the rate of electron transfer to MVS2+.This equation assumes that the rates of reactions (2) and (3) are essentially constant (the concentration of cysteine will only drop by 2% during the reduction of 0,) and that reaction (2) is fast compared with the rate of MV'+ generation. The equation predicts correctly that the rate will be independent of the 0, concentration. A plot of eqn (7) is shown in fig.4, from which cc = (2.0f0.1) x lo-, and k 6 / k , = (1.4k0.2) x mol dmP3. This shows that the low efficiency of colloidal CdSe results from the small value of a, i.e. the high rate of charge recombination compared with the reaction of h+ with cysteine.J . R. DARWENT 187 Fig. 3. Redox diagrams for the photosensitised reduction of 0,. 0 1 2 3 l/[MV2']/10-3 dm3 mol-' Fig. 4. Reciprocal plot of MV'+ yield (0) and rate of 0, photoreduction (0) against 1/[MV2+] (variable concentrations of MV2+, otherwise conditions as in fig. 1).188 PHOTOREDUCTION OF 0, USING COLLOIDAL CdSe CATALYST This kinetic scheme can also describe the flash-photolysis results, from which [MV'+]- [e30+k,[e],[MV2+] -- 1 1 k6 where [el, is the concentration of electrons trapped during the photoflash by reaction (3) and [MV'+] is the total concentration of MV'+ formed for a given concentration of MVe2+.The results of the flash-photolysis experiments are also shown in fig. 4, where there is good agreement between the two sets of data. From this figure k,/k, is again (1.4 If: 0.2) x mol dm-3 and [el,, is (4.8 -f 0.5) x lo-' mol dm-3. CONCLUSION Poor quantum yields are observed for electron-transfer reactions sensitised by colloidal CdSe suppported by NaLS, since charge recombination occurs more rapidly than oxidation of cysteine in this system. This problem may be overcome by using a surfactant support which could facilitate oxidation of cysteine. Alternatively it might be possible to reduce the rate of charge recombination by adsorbing metal ions onto the particles, an approach which has been very successful with liquid-junction photovoltaic cells.25* 2 5 y 34 This work was supported by the S.E.R.C.and London University Central Research Fund. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 A. J. Bard, J. Phys. Chem., 1982, 86, 172. M. Gratzel, Acc. Chem. Res., 1982, 15, 376. A. V. Bulatov and M. L. Khidekel, Zzv. Akad. Nauk SSSR, Ser. Khim. Nauk, 1976, 1902. J. M. Lehn, J. P. Sauvage and R. Ziessel, Nouv. J. Chim., 1980, 4, 623. E. Borgarello, J. Kiwi, E. Pelizzetti, M. Visca and M. Gratzel, J. Am. Chem. SOC., 1981, 103, 6324. A. A. Krasnovskii, G. P. Brin, A. N. Luganskaya and V. V. Nikandrov, Dokl. Akad. Nauk SSSR, 1979, 249, 896. J. R.Darwent and G. Porter, J. Chem. SOC., Chem. Commun., 1981, 145. J. R. Darwent, J. Chem. SOC., Faraday Trans. 2, 1981, 77, 1703. K. Kalyanasundaram, E. Borgarello and M. Gratzel, Helv. Chim. Acta, 1981, 64, 362. E. Borgarello, K. Kalyanasundaram, E. Pelizzetti and M. Gratzel, Helv. Chim. Acta, 1982, 65, 243. J. R. Harbour, R. Wolkow and M. L. Hair, J . Phys. Chem., 1981, 85,4026. G. Porter and M. D. Archer, Interdisc. Sci. Rev., 1976, 1, 119. M. A. West, Creation and Detection of Excited States, ed. W. R. Ware (Marcel Dekker, New York, 1976), vol. 4. A. Mills, A. Harriman and G. Porter, Anal. Chem., 1981, 53, 1254. D. Duonghong, J. Ramsden and M. Gratzel, J . Am. Chem. SOC., 1982,104,2977. A. Henglein, Ber. Bunsenges. Phys. Chem., 1982, 86, 241. M. A. Fox, B. Lindig and C. Chen, J. Am. Chem. SOC., 1982, 104, 5828. J. Kuczynski and J. K. Thomas, Chem. Phys. Lett., 1982,88,445. A. Henglein, Ber. Bunsenges. Phys. Chem., 1982, 86, 301. A. Henglein, J . Phys. Chem., 1982, 86, 2291. Z. Alfassi, D. Bahnemann and A. Henglein, J. Phys. Chem., 1982, 86, 4656. R. Rossetti and L. Brus, J . Phys. Chem., 1982, 86, 4470. K. Kalyanasundaram, E. Borgarello, D. Duonghong and M. Gratzel, Agnew. Chem., Int. Ed. Engl., 1981, 20, 987. K. Metcalfe and R. E. Hester, J. Chem. SOC., Chem. Commun., 1983, 133. A. Heller, Acc. Chem. Res., 1981, 14, 154. M. S. Wrighton, Acc. Chem. Res., 1979, 12, 303. J. R. Harbour and M. L. Hair, J . Phys. Chem., 1977,81, 1791.J. R. DARWENT 189 2R J. D. Spikes, Photochem. Photobiol., 1981, 34, 549. 29 R. E. Stephens, B. Ke and D. Tnvich, J . Phys. Chem., 1955, 59, 966. 3o A. Kirsch-de Mesmaeker, A. M. Decoster and J. Nasielski, Solar Energy Materials, 1981, 4, 203. 31 A. B. Ellis, S. W. Kaiser, J. M. Bolts and M. S. Wrighton, J . Am. Chem. SOC., 1977, 99, 2839. 32 D. T. Sawyer and J. S. Valentine, Acc. Chem. Res., 1981, 14, 393. 33 J. R. Darwent and K. Kalyanasundaram, J . Chem. SOC., Faraday Trans. 2, 1981,77, 373. 34 A. Heller, K-C. Chang and B. Miller, J. Am. Chem. SOC., 1978, 100, 684. , (PAPER 3/761)