J . Chem. Soc., Faraday Trans. 1, 1985, 81, 61 -68 Correlation of the Crystal Structure of Titanium Dioxide Prepared from Titanium Tetra-2-propoxide with the Photocatalytic Activity for Redox Reactions in Aqueous Propan-2-01 and Silver Salt Solutions BY SEI-ICHI NISHIMOTO, BUNSHO OHTANI, HIROSHI KAJIWARA AND TSUTOMU KAGIYA* Department of Hydrocarbon Chemistry, Faculty of Engineering, Kyoto University, Sakyo-ku, Kyoto 606, Japan Received 8th March, 1984 Titanium dioxide (TiO,) has been prepared by the hydrolysis of titanium tetra-2-propoxide, followed by calcination at various temperatures (T,) up to 1000 "C. The content and crystallite size of anatase in the TiO, powders increased upon increasing T, up to 550 "C. In the T, range 550-600 "C a mixture of anatase and rutile was obtained.A further increase in T, resulted in TiO, of rutile structure only. The photocatalytic activities of these TiO, powders for redox reactions were evaluated in the following systems: (1) aqueous propan-2-01 solution, (2) aqueous Ag,SO, solution and (3) aqueous Ag,SO, solution containing propan-2-01. The anatase TiO, showed photocatalytic activity in all these systems, the activity increasing with crystal growth. In aqueous propan-2-01 solution the activity is dramatically enhanced by partial coverage of the TiO, with platinum black. The photocatalytic activity of the rutile TiO, powder was comparable to or even greater than that of anatase when the reaction system included the silver salt, but was negligibly small for aqueous propan-2-01 solution regardless of the partial Pt coverage.Photocatalysed reactions such as the dehydrogenation of alcohol~l-~ or the photo-Kolbe reaction of carboxylic using dispersed TiO, particles have been widely investigated. In many cases various kinds of commercially available TiO, powders have been used as photostable catalysts with sufficient oxidizing and reducing abilities.1°-13 Although the correlation between the physical properties, e.g. bulk crystal and surface structures, and the photocatalytic activity for oxidizing water has been reported for the polycrystalline TiO, electrode systems,14 TiO, suspension systems are still a subject of investigation. We have recently characterized the photocatalytic activity of TiO, powders prepared from Ti(SO,), by hydrolysis and calcination a t various temperatures up to 1000 "C.l5 The activity of such TiO, powders, when mixed with platinum black, for the dehydrogenation of propan-2-01 in aqueous solution depends on the crystal structure; i.e. the activity of anatase TiO, was adequate whereas that of rutile was negligible. The superior photocatalytic activity of anatase compared with rutile has previously been observed for the oxidation of liquid propan-2-01 in the presence of 0,.l6 Unfortunately, a small amount of residual sulphate ions caused considerable lowering of the photocatalytic activity of the TiO, powders obtained from Ti(SO,), because they decreased the pH of the aqueous s u ~ p e n s i o n . ~ ~ In order to characterize the intrinsic photocatalytic activities it is therefore desirable that TiO, powders free from such contamination are prepared by an alternative method and are subjected to further investigation.This paper describes the effects of the physical properties of anion-free TiO, powders prepared from titanium tetra-2-propoxide [Ti(OPr),]l on the photocatalytic activity for redox reactions in aqueous solution. 6162 PHOTOCATALYTIC ACTIVITY OF TITANIUM DIOXIDE EXPERIMENTAL PREPARATION OF TiO, POWDER Titanium tetra-2-propoxide [Ti(OPr),] was supplied by Wako Pure Chemicals and distilled before use (b.p. 92.0-94.0 "C, 0.47 kPa). A mixture of Ti(OPr), (100 cm3, 0.34 mol) and propan-2-01 (1 80 cm3) was added dropwise to an ice-cooled mixture of propan-2-01 (450 cm3) and distilled water (150 cm3) with vigorous stirring. The resulting white precipitate of titanic acid was filtered off, washed repeatedly with distilled water and precalcined for 24 h at 120 "C in air.The TiO, powder thus obtained was subjected to further calcination in air at various temperatures with an electric furnace equipped with a programmed controller (Ohkura EC 53/2 PB). The heating was first performed for 4 h at a constant rate of ca. 1-4 "C min-l until a specified temperature (T,) was attained and continued for 5 h at T,, after which cooling was performed at the same constant rate as in the case of increasing temperature. X-RAY DIFFRACTION ANALYSIS The crystal structures of the TiO, powders were determined by an X-ray diffraction method, using a Rigaku Geigerflex 2013 diffractometer (target, Cu; filter, Ni; 35 kV; 20 mA; scanning speed, 1" min-l). The contents of anatase and rutile in the TiO, were evaluated by integration of the most intense peaks 28 = 25.4" [d = 0.352 nm, the (011) plane of anatase] and 27.3" [d = 0.325 nm, the (1 10) plane of rutile], respectively, by reference to CaCO, as an internal standard.18 The calibration curves for the anatase and rutile were obtained using commercially available anatase (Merck) and rutile (prepared by heating the Merck TiO, powder at 1200 "C for 10 h in aiP), respectively.The mean crystallite size (L) was determined from the broadening (8) of the most intense line in the X-ray diffraction pattern, after corrections for the Ka doublet and instrumental broadening based on the Scherrer equation20 (L = kl/@cos 8, where A is the radiation wavelength, 8 is the Bragg angle and k = 0.90).PHOTOREACTION A finely ground TiO, powder (50 mg), with or without platinum black (typically 5 wt % , Nakarai Chemicals), was suspended in distilled water (5.0 cm3) or aqueous Ag,SO, solution (0.025 mol drn-,, 5.0 cm3) in a glass tube (18 mm dia. x 180 mm, transparent for exciting-light wavelengths > 300 nm). The suspension was purged with Ar for at least 30 min and sealed off with a rubber cap. Propan-2-01 (38 mm3, 0.50 mmol) was injected through the cap by a syringe. The Ar-purged TiO, suspension was irradiated under magnetic stirring at room temperature with a merry-go-round apparatus equipped with a 400 W high-pressure mercury arc (Eiko-sha 400). PRODUCT ANALYSIS A portion (0.2 cm3) was withdrawn from the gas phase (30.0 cm3) of the sealed sample and subjected to analysis for volatile products such as H, and 0,, using a Shimadzu GC 4A gas chromatograph equipped with t.c.d.and a 5A molecular-sieve column (3 mm diameter x 3 m) with Ar carrier at 100 "C. Propan-2-01 and acetone were analysed with a Shimadzu GC 6A gas chromatograph equipped with f.i.d. and polyethylene glycol 20M on a Celite 545 column (3 mm diameter x 2 m) with N, carrier at 90 "C. The procedure and apparatus for the determination of the amount of deposited Ag have been described e1~ewhere.l~ RESULTS AND DISCUSSION PHYSICAL PROPERTIES OF THE TiO, POWDERS PREPARED FROM Ti(OPr), The weight of TiO, powder obtained on calcination, relative to the weight before calcination at 120 "C, decreased with increasing c, attaining a constant value of ca.50%. The phase transition from anatase to rutile was observed in the range 600-650 "C. This transition temperature is considerably lower than that (750-800 "C) for TiO, prepared from Ti(S0,),.l5 The mean crystallite size of the anatase (LA) is plotted against T, in fig. 1. The value of LA did not appreciably change at T, d 550 "C,S-I. NISHIMOTO, B. OHTANI, H. KAJIWARA AND T. KAGIYA 200 150 { 100 4 50 01 I 0 200 400 600 800 1000 T, 1°C Fig. 1. Variation in the mean crystallite size of anatase (LA) as a function of calcination temperature (T,). 63 0 200 400 600 800 1000 T , 1°C Fig. 2. T,-dependent yields of H, (YH2, 0) and acetone (I;CH3)2C0, 0 ) on the irradiation (10 h) of TiO, (50 mg) suspended in aqueous propan-2-01 (38 mm3, 500 pmol) solution (5.0 cm3) under Ar. but rapidly increased in the narrow T, range 55&610 "C.The mean crystallite size of the rutile obtained at T, 2 600 "C was estimated to be > 200 nm, although its exact value could not be determined using the Scherrer equation. PHOTOCATALYTIC ACTIVITY OF TiO, AND Ti0,-Pt IN AQUEOUS PROPAN-2-OL SOLUTION The TiO, powders, prepared as above, when suspended in aqueous propan-2-01 solution and irradiated at Aex > 300 nm under Ar produced H, and acetone. Fig. 2 shows that yields of both H, ( YH2) and acetone ( Y(CH3)2CO) over an irradiation period of 10 h were strongly dependent on the calcination temperature T,. Although the reproducibility of the data in fig. 2 was relatively poor (ca. 30%), a trend is evident64 PHOTOCATALYTIC ACTIVITY OF TITANIUM DIOXIDE 240 200 160 -.1 g 120 5;- 80 40 0 t I ) 0 200 400 600 800 1000 T, 1°C Fig.3. T,-dependent yields of acetone (qCH3)2CH, 0 ) and H2(YH2, 0) on the irradia- tion (1 h) of 5 wt % Pt loaded Ti02 (50 mg) suspended in aqueous propan-2-01 (38 mm3, 500 pmol) solution (5.0 cm3) under Ar. that YH, and Y(CH3),C0 increased with increasing T, until the maximum values ( YH2 w 7 pmol and Y(cH3)2C0 w 10 pmol) were attained at T, = 600 "C. For the treatment at T, > 610 "C, YH2 and qCH,),CO decreased to a greater extent relative to the maximum values. TiO, for T, = 1000 "C, which consists only of rutile, was virtually ineffective for the formation of H, and acetone. As has been well doc~mented,l-~? 21 partial coverage of TiO, ( T , = 610 "C) with a small amount of Pt (up to 5 wt%) enhances the rate of H, formation (RH2).A saturation limit of RH2 was observed in the range of Pt coverage from 2.5 to 5.0 wt% , which was 100-fold greater than without Pt. Fig. 3 illustrates the T, dependences of YH2 and qCH3)2C0 for 1 h photoirradiation of the TiO, powders covered with 5 wt% Pt (Ti0,-Pt). The profiles of these T, dependences are seen to be essentially identical to those in fig. 2, suggesting that the photocatalytic activity of a given TiO, particle is intrinsically determined by the solid properties of the particle but enhanced by the surface Pt. It is evident from fig. 3 that the yields of H, and acetone from Ti0,-Pt are equal within experimental error (reproducibility & 5% ), regardless of q.The equivalence of H, and acetone produced by the photodecomposition of propan-2-01 catalysed by anatase-Pt has been also demonstrated by Teratani et aL3 The net photoreaction in the present system is therefore represented as follows : hv > 800 nm (CH,),CHOH - (CH,),CO + H,. Ti0,-Pt In absence of Pt, YHz was less than I;CH3)2C0. The lower yield of H, is accounted for by the reduction of TirV on the illuminated TiO, catalyst to form Ti11J22 as a non-catalytic side reaction, because the white suspension of TiO, was observed to turn grey during the photoirradiation : hv > 300 nm i(CH,),CHOH +TiIV i(CH,),CO + Ti"' + H+. (3)S-I. NISHIMOTO, B. OHTANI, H. KAJIWARA AND T. KAGIYA 65 Thus, in the absence of partial Pt coverage, the two photoreactions (2) and (3) proceed in competition.It is also plausible that H,, as a photoproduct, reduces TiIV to TiIII. Fig. 2 provides an estimate that at most 60% of the photogenerated reducing species, i.e. electrons photoexcited to the conduction band of TiO,, would be consumed for the self-reduction of TiO, and formation of TiIII. In contrast, the presence of Pt on the TiO, surface could prevent electron trapping of TiIV into Ti111. Moreover, the Pt is responsible for the efficient charge separation of the photogenerated electron-hole pairs in the Ti02,10 thereby promoting the reduction of protons to H, and the oxidation of propan-2-01 to acetone: (TiO,) + hv -+ e- + h+ (4) Pt e-+ H+ -+ Pt-H iH2 (5) h++$(CH,),CHOH -+ ~(CH,),CO+H+. ( 6 ) Reactions (4)-(6) are in accord with the stoichiometry of reaction (2).By reference to the structural evidence, it is clear from fig. 2 and 3 that the TiO, powders containing anatase (T, = 120-620 "C) show sufficient photocatalytic activity to produce H, and acetone, while those containing rutile only (T, b 650 "C) show negligible activity even when covered with Pt. The activity of anatase-containing TiO, powders increased with increasing q. For the TiO, powders (T, = 600-650 "C) containing both anatase and rutile, the activity decreases with decreasing content of anatase. The negligible activity of rutile in aqueous propan-2-01 solution is attributed to the disadvantageous energetics for the reduction of protons to H, compared with anatase : because a rutile electrode exhibits a more positive flat-band potential than an anatase electrode, the energy of photoexcited electrons in the conduction band of rutile is expected to be lower than that of those in anata~e.~~,, PHOTOCATALYTIC ACTIVITY OF TiO, IN AQUEOUS SOLUTIONS OF SILVER SALT Photoirradiation (Aex > 300 nm) of aqueous suspensions of these TiO, powders containing Ag,SO, led to the formation of 0, and the deposition of Ag metal on the TiO, parti~1es.l~.l9 The T, dependences of the yields of O,( YO,) and the Ag deposit ( YAg) over an irradiation period of 1 h are shown in fig. 4. A linear relationship with a slope of 0.23 was obtained between the YO, and YAg, in accord with the following net photoreaction: hv > 300 nm 4Ag++2H20 - 4Ag+02+4H+. TiO, (7) Support for the release of H+ in this scheme was obtained by the observation of a rapid decrease in the pH (from ca.4 to 2) of the suspension during photoirradiation. A slightly larger amount of Ag deposit compared with the stoichiometry in reaction (7) would be attributed to the partial photoadsorption of another product, O,, on the TiO, s ~ r f a c e . ~ ~ - ~ ~ Fig. 4 shows that both Y,, and YO, are relatively small and nearly independent of T, when TiO, powders prepared in the lower T, region (120-550 "C) are used. Upon raising T, from 550 to 600 "C the photocatalytic reaction became more rapid to give eventually maximum values of YAg z 110 pmol and YO, z 25 pmol. Y,, and Yo, then decreased and approached constant values of 80 and 20 pmol, respectively, in the T, range 800-1000 "C. The surface area of rutile probably decreases but activity per unit area is constant in this T, range.Note that in the presence of Ag+ ions the66 PHOTOCATALYTIC ACTIVITY OF TITANIUM DIOXIDE 100 80 3 60 x 40 20 0 0 200 400 600 800 1000 T, 1°C Fig. 4. T,-dependent yields of Ag metal (YAP, 0) and 0, (Yo,, a) on irradiation (1 h) of an aqueous TiO, suspension in Ag,SO, solution (250 pmol Ag+, 5.0 cm3) under Ar. TiO, powders containing the rutile structure only, which were practically inactive in aqueous propan-2-01 solution without Ag+, showed a larger activity than the anatase TiO, (T, < 600 "C). In particular, the mixed anatase-rutile powder (600 6 &/"C < 620) showed the highest activity for the formation of Ag and 0,, in contrast to the activity for propan-2-01 dehydrogenation (fig.2 and 3). These facts clearly demonstrate that the photocatalytic activity of rutile is essentially comparable to that of anatase in certain photoreaction systems, as in this case where the more reducible Ag+ ions but not protons can react with the photogenerated electrons of rutile. Furthermore, the observed highest activity of the anatase-rutile mixture demonstrates that the crystal structure alone cannot explain the activity. It is likely that the surface area25 and the porosity29 also have significant effects on the photocatalytic activity of TiO, powders. PHOTOCATALYTIC ACTIVITY OF TiO, IN AQUEOUS SOLUTIONS OF SILVER SALT AND Addition of propan-2-01 to the aqueous suspension of TiO, containing Ag,SO, resulted in the oxidation of propan-2-01 into acetone together with the formation of 0, and the deposition of Ag metal.Fig. 5 shows variations of the product yields YAg, Yo, and qCHs)2C0 over the 1 h photoirradiation as a function of T,. Formation of 0, by the TiO, powders at T, < 550 "C was negligible, while a small amount of 0, (ca. 5 pmol) was obtained for T, 600 "C. YAg and ~ C H , ) , C O increased with increasing T, to attain their maxima of ca. 125 and 50 pmol at T, = 650 "C, and then decreased toward constant values on further increase in The total yield of oxidation reduction products satisfied a stoichiometry given by 2 Y;CH3)2C0 + 4YO2 = YAg. Clearly, the oxidation of both water [reaction (7)] and PROP AN-2-OL (> 650 "C). propan-2-01 hv > 300 nm (CH3),CHOH + 2Ag+ (CH,),CO + 2Ag+ 2H+ (8) TiOl is involved in the present system, although the proportion of water oxidation is much smaller.It is seen from fig. 5 that the TiO, powders at T, >, 550 "C, which consist onlyS-I. NISHIMOTO, B. OHTANI, H. KAJIWARA AND T. KAGIYA 67 140 120 100 80 5 1 x 60 40 20 0 0 200 400 600 800 1000 T, 1°C Fig. 5. T,-dependent yields of Ag metal (YAP, O), acetone ( Z;CH3)2C0, 0 ) and 0, (Yo,, 0 ) on irradiation (1 h) of an aqueous TiO, (50 mg) suspension in Ag,SO, solution (250 pmol Ag+, 5.0 cm3) containing propan-2-01 (38 mm3, 500 pmol) under Ar. of anatase, give rise to propan-2-01 oxidation almost exclusively according to reaction For T, > 600 "C the profile of the T,-dependent activity in this system (fig. 5 ) is similar to that without propan-2-01 (fig.4), although propan-2-01 predominantly undergoes oxidation instead of water. Since the TiO, powders in this T, range contained an increasing proportion of rutile as T, increased (fig. I), this similarity seems to originate largely from the action of the rutile. In contrast, the apparent photocatalytic activity of the TiO, powders at & < 550 "C, which contain only anatase, is at least three-fold enhanced by the addition of propan-2-01. This is clearly a result of the oxidation of the added propan-2-01 that would occur at the illuminated anatase TiO, more readily than that of water. Compared with the results for TiO, prepared from Ti(SO,),, the activity of TiO, powders from Ti(OPr), seems to be greater in this photoreaction system: YAg values for the TiO, powders from Ti(OPr), ( T , = 600 "C, LA = 53 nm) and Ti(SO,), ( T , =700 "C, LA = 37 nm)15 were 73 pmol for 1 h irradiation and 87 pmol for 2 h irradiation, respectively.ENHANCEMENT OF THE PHOTOCATALYTIC ACTIVITY OF TiO, POWDERS BY REDUCIBLE (3). OR OXIDIZABLE SPECIES As described above, the photocatalytic activity of TiO, powders suspended in aqueous solution depends on both the crystal structure and the solution species to be oxidized or reduced by the photogenerated hole (h+) or electron (e-), respectively. The latter effect was clearly demonstrated by the different photocatalytic activities in the three photoreaction systems. Almost no reaction occurred on the photoirradiation of an aqueous TiO, suspension in the absence of propan-2-01 or Ag+. However, the addition of propan-2-01 to this system led to the formation of a small amount of H,68 PHOTOCATALYTIC ACTIVITY OF TITANIUM DIOXIDE and acetone, some of which was produced by a non-catalytic process.When Pt black was loaded on the TiO, powders, the activity increased to yield stoichiometric amounts of H, and acetone. These facts show that an easily oxidizable species such as propan-2-01 is more effective for the trapping of the photogenerated hole, which has little ability to oxidize water. In addition, the Pt loading enhanced the electron trapping by H+ and depressed the self-reduction of TiIV to T F , especially in the case of anatase. The effect of Ag+ was also evident: TiO, could oxidize water to 0, with the aid of the easy reduction of the Ag+. A further enhancement of the catalytic activity of TiO, was observed on the addition of both propan-2-01 and Ag,SO,, particularly in presence of anatase (see fig.4 and 5 ) . We thank Prof. Satohiro Yoshida (Kyoto University) for his valuable advice on X-ray diffraction measurements. We also thank the Instrumental Analyses Research Centre of Kyoto University for permission to use atomic absorption spectrometers. Both referees are thanked for their careful and constructive reports. T. Kawai and T. Sakata, J. Chem. SOC., Chem. Commun., 1980,694; Nature (London), 1980,286,474; Chem. Lett., 1981, 81. P. Pichat, J-M. Herrmann, J. Disdier, H. Courbon and M-N. Mossanega, Nouv. J. Chim., 1981, 5, 627. S. Teratani, J. Nakamichi, K. Taya and K. Tanaka, Bull. Chem.SOC. Jpn, 1982, 55, 1688. K. Domen, S. Naito, T. Onishi and K. Tamaru, Chem. Lett., 1982, 555. B. Kraeutler and A. J. Bard, J. Am. Chem. SOC., 1978, 100, 2239; 5985. H. Reiche and A. J. Bard, J. Am. Chem. Soc., 1979, 101, 3127. I. Izumi, F-R. F. Fan and A. J. Bard, J. Phys. Chem., 1981,85, 218. H. Yoneyama, Y. Takao, H. Tamura and A. J. Bard, J. Phys. Chem., 1983,87, 1417. 1982, 86, 172. 'I H. Reiche, W. W. Dunn, K. Wilbourn, F-R. F. Fan and A. J. Bard, J. Phys. Chem., 1980,84, 3207. lo A. J. Bard, J. Photochem., 1979, 10, 59; Science (Washington D.C.), 1980, 207, 139; J. Phys. Chem., l1 A. J. Nozik, Annu. Rev. Phys. Chem., 1978, 29, 189. l 2 M. S. Wrighton, P. T. Wolczanski and A. B. Ellis, J. Solid State Chem., 1977, 22, 17. l 3 M. A. Fox, Acc. Chem. Res., 1983, 16, 314. l4 C. Stalder and J. Augustynski, J. Electrochem. Soc., 1979, 126, 2007. I5 S. Nishimoto, B. Ohtani, A. Sakamoto and T. Kagiya, Nippon Kagaku Kaishi, 1984, 246. l6 R. B. Cundall, R. Rudham and M. Salim, J. Chem. SOC., Faraday Trans. I , 1976,72, 1642. l8 R. A. Spurr and H. Myers, Anal. Chem., 1957,29, 760. l 9 S. Nishimoto, B. Ohtani, H. Kajiwara and T. Kagiya, J. Chem. SOC., Faraday Trans. I , 1983,79,2685. *O H. Klug and L. E. Alexander, X-Ray Diyraction Procedures (Wiley, New York, 2nd edn, 1974), 21 J. Disdier, J-M. Herrmann and P. Pichat, J. Chem. SOC., Faraday Trans. I , 1983, 79, 651. z2 A. D. Buss, M. A. Malati and R. Atkinson, J. Oil Colour Chem. Assoc., 1976, 59, 369. 23 M. V. Rao, K. Rajeshwar, V. R. Pai Verneker and J. DuBow, J. Phys. Chem., 1980, 84, 1987. 24 A. Mills and G. Porter, J. Chem. SOC., Faraday Trans. I , 1982, 78, 3659. 28 E. Borgarello, J. Kiwi, M. Gratzel, E. Pelizzetti and M. Visca, J. Am. Chem. Soc., 1982, 104, 2996. 26 A. H. Boonstra and C. A. H. A. Mutsaers, J. Phys. Chem., 1975, 79, 1694. 27 G. Munuera, V. Rives-Arnau and A. Saucedo, J . Chem. SOC., Faraday Trans. I , 1979,75, 736. 28 A. R. Gonzalez-Elipe, G. Munuera and J. Soria, J. Chem. SOC., Faraday Trans. I , 1979, 75, 748. 29 L. Kruczynski, H. D. Gesser, C. W. Turner and E. A. Speers, Nature (London), 1981,291, 399. M. R. Harris and G. Whitaker, J. Appl. Chem., 1962, 12, 490. p. 618. (PAPER 4/384)