J . Chem. SOC., Faraday Trans. 1, 1984, 80, 1507-1515 Photocatalytic Hydrogen Evolution from Aqueous Hydrazine Solution over Precious-metal/Anatase Catalysts BY YOSHINAO OOSAWA National Chemical Laboratory for Industry, Higashi, Yatabe, Tsukuba, Ibaraki 305, Japan Received 23rd August, 1983 Aqueous hydrazine solution has been photocatalytically decomposed over precious- metal/anatase catalysts yielding only H,, N, and NH,. H,/N, (molar ratio of H, evolved to N, evolved) is usually ca. 1 and does not depend on the reaction conditions (pH, metal, reactant concentration). However, the value is > 1 when another hole scavenger (CH,OH) is present. N,/NH, is CQ. 0.5 and does not change in the presence of CH,OH. On the basis of these results, the reaction scheme is considered to be as follows: the electrons photoexcited into the conduction band of anatase are consumed through H, formation and the positive holes generated in the valence band are consumed through the simultaneous formation of N, and NH,. The dependence of the reaction rate on the concentration of N,H, and on the metal suggests that the rate-determining step is the formation of H, over the metal. In recent years, photocatalytic H, evolution from ~ a t e r l - ~ or organic- compound + water systems6* using precious-metal/semiconductor powders has been widely investigated from the standpoints of photo-to-chemical energy conversion assisted by light energy, reaction mechanism and so on.On the other hand, few reportssy have been published on photocatalytic H, evolution from systems with an inorganic electron donor other than water, although in some systems high photo- catalytic reaction rates and some simple products are expected, making it easier to determine the products and follow the reaction.Hydrazine, which is an inorganic electron donor, can reduce protons and produce H, thermodynamically. However, it has been reported that the thermal-catalytic decomposition of an acidic aqueous solution over platinum black proceeds according to the following reaction with no evolution of H,:l0 3N,H, + N, + 4NH,. In a previous communicationS photocatalytic evolution of H, from an aqueous solution of N2H,*2HC1 in the presence of Pt/TiO, has been reported: 2N,H4 + N, + H, + 2NH,. On the basis of preliminary experiments it has been proved that this reaction has the following features: (i) the rate of reaction is fast, (ii) the reaction produces only a few products (H,, N, and NH,) and so it is easy to obtain good mass balance and charge balance, which, despite their importance, have been reported for only a few cases of photocatalytic reactions in aqueous solution, and (iii) the successive reaction of primary products is minimal, which is rather different from organic electron donors.Therefore, this reaction system is considered to be suitable for an investigation of the features of photocatalytic reactions. This paper presents the results of a study of the photocatalytic decomposition of N2H4 in aqueous solution over precious-metal/anatase catalysts. 15071508 PHOTOCATALYTIC HYDROGEN EVOLUTION EXPERIMENTAL MATERIALS N2H4*H20, N2H4.2HC1, NH,C1 and CH,OH were all reagent grade.Deionized water was used without further purification. Purest-grade anatase (TP-2) was purchased from Fuji Titan. Each precious-metal/anatase was prepared by an impregnation H, reduction method (H,, 101 kPa, 200 "C, 4 h) from anatase and an aqueous solution of the precious-metal chloride (except for OsO,). X-ray powder diffraction measurements (Rigaku, Geigerflux, Cu Ka) confirmed that the crystal structure was not changed by loading of the precious metals. Average particle diameters of the anatase and the Pt/anatase (lop2 wt/wt) were measured using a centrifugal automatic particle analyser (Horiba, CAPA-500) and were both 0.26 pm. Their specific surface areas were measured by the B.E.T.method (Shimadzu surface-area analyser 2205) with Ar as an adsorbate and were 18.3 m2 g-l and 17.8 m2 g-l, respectively. APPARATUS AND PROCEDURE The reaction was performed in a photocell (108 cm3) with a rectangular-parallelepiped lower part (35 x 35 x 60 mm) and a septum. Usually, the reaction mixture, consisting of precious- metal/anatase (10 mg) and an aqueous solution of reactant (30 cm3 of 0.1 mmol drn-,) in an argon atmosphere, was stirred and irradiated using a 500 W ultra-high-pressure mercury lamp (Ushio) at ca. 40 "C in the stationary state. The dark reaction was performed at 50 "C. ANALYSIS Gaseous products (H, and N,) were analysed quantitatively by gas chromatography (Yanaco G 180; MS-13X column). The liquid reaction mixture was filtered to remove solid photocatalyst, which was washed with HCl(l0 cm3 of 1 mol dm-3) in order to desorb reactant and products adsorbed on it.The filtrates and the washings were combined together and analysed." N,H, was analysed quantitatively by titration with KIO, so1ution.12 Concentrated NaOH was then added to the mixture and the resulting solution was distilled. NH, was analysed qualitatively by the indophenol method', and quantitatively by acid-base titration', of the distillate. U.v.-visible spectrophotometric analysis (Shimadzu U.V. 240) ascertained that no reaction mixture except for anatase had absorption in the wavelength region transmitted through the Pyrex glass ( A > ca. 300 nm). RESULTS REACTION OVER Pt/ANATASE Fig. 1 shows the time dependence of the volume of gases evolved when the aqueous solution of N2H4 or N,H4*2HC1 was irradiated in the presence of Pt/anatase wt/wt). The gas evolution rate was higher when N2H4 was used as a reactant than when N,H4*2HC1 was used.The volume of the gases is expressed in cm3 at s.t.p. (0 "C, 101 kPa) unless otherwise stated. No reaction occurred when a 420 nm cut-off filter was used or in the dark. When water (30 cm3) was irradiated in the presence of the Pt/anatase (lo-, wt/wt), a small amount of H, was evolved: 0.17 cm3 after a 19 h reaction. So H, evolution in the absence of N,H, (or N,H4*2HCl) may be due to photocatalytic decomposition of or photoassisted decomposition of water. l5 Products were analysed in a few reactions. Table 1 shows the results of the analysis. No compounds other than H,, N,, NH, and N2H4 were found by gas-chromatographic analysis with MS-13X (1 m, 50 "C) and chromosorb 103 (2.25 m, 170 "C).It is clear that the reactions presented in table 1 proceed photocatalytically, because they fulfil all of the following requisites: (i) dark reaction does not occur, (ii) the reaction stops when irradiated through a 420 nm cut-off filter (the band gap of anatase is 3.2 eV, therefore light of energy greater than that of the band gap is cut of€), (iii) mass balanceY. OOSAWA 1509 18 16 14 6 2 12 c: v) m 5 --. U 0 w v) w 10 - % I 8 Lr 0 w 5 6 9 4 2 / / i Y I I I I I I I I 1 2 3 4 5 6 7 8 time/h Fig. 1. Photocatalytic H, and N, evolution over Pt/anatase wt/wt). Reaction conditions: Pt/anatase, 10 mg; reactant, 30 cm3 of 0.1 mol dm-3 N2H4, A, H, and A, N,; reactant, 30 cm3 of 0.1 mol dm-3 N2H4.2HC1, 0, H, and 0, N,; (---) when the 420 nm cut-off filter was used.(N and H) and charge balance hold well and (iv) the turnover number (moles of N2H4 consumed/moles of photocatalyst)* is > 1. Table 2 shows the average evolution rate of H, and N, for the initial few hours over various precious-metal/anatase catalysts ( lo-, wt/wt). Neither the reaction with the cut-off filter nor the dark reaction at 50 "C proceeded. Therefore, the gas evolutions presented in table 2 are also assumed to proceed photocatalytically. In almost all cases, H,/N, was ca. 1. The gas-evolution rate decreased in the following order: Pt > Pd > Ir > Rh > 0 s > Ru. This order seems to reflect the magnitude of hydrogen overvoltage required to obtain a current density of 2 mA cm-, in 0.05 mol dm-3 H2S04 (Pt < Ir < 0 s and Pd < Rh < Ru).16 REACTION IN THE PRESENCE OF CH,OH In order to obtain information on the photocatalytic decomposition of N,H, in aque- ous solution, the reaction was performed in the presence of another electron donor, i.e.CH,OH. Neither photocatalytic H, evolution from CH,OH + H,O(v/v = 1 / 1) in the presence of anatase nor the dark reaction in the presence of Pt/anatase (lo-, wt/wt) proceeded. The results of the analysis are shown in table 1. The mass balance of nitrogen held fairly well in both cases. Note that N,/NH3 is close to 0.5 * Turnover number (moles of reactant consumed/moles of photocatalyst used) is used as a guide as to whether the reaction is stoichiometric or catalytic.1510 PHOTOCATALYTIC HYDROGEN EVOLUTION Table 1.Analysis data reactant N,H, 2HC1 N2H4. 2HC1 N2H4 * 2HCP N2Hdd + CH,OHey f + CH30He,Q reaction time/h N2H, consumed/mmol product/mmol H2 N, NH3 N reactant N product H reactant H product e- consumeda pf consumedb turnover numberc mass balance/mequiv. charge balance/mequiv. 8 0.84 0.42 0.46 0.79 1.68 1.71 3.36 3.21 1.63 1.84 6.7 6.5 0.95 0.56 0.50 0.83 1.90 1.83 3.80 3.64 1.95 2.00 7.6 15 0.66 0.47 0.31 0.58 1.32 1.20 - 5.3 ~ 15 0.23 0.68 0.14 0.24 0.46 0.52 - 1.8 Reaction conditions: 0.1 mol dm-3 N2H, or NZH,*2HC1; amount of reactant, 30 cm3; a Calculated as 2H,+ NH,. Calculated as [moles of N2H, consumed 500 W ultra-high-pressure Hg lamp (Ushio) was photocatalyst, Pt/anatase ( lo-, wt/wt), 10 mg.per rnol of Pt/anatase ( lo-, wt/wt) used]. used. Calculated as 4N2. 450 W Xe lamp (Oriel) was used. 1 0.1 rnol dmV3 CH,OH. 1 rnol dm-, CH,OH. Table 2. H, and N, evolution rates (cm3 h-l) over precious-metal/anatase catalysts pho tocatalyst H2 N2 P t / ana tase 2.06 2.07 Ir/anatase 0.80 0.8 1 Rh/ana tase 0.70 0.74 Os/anatase 0.061 0.068 Ru/anatase 0 0 anatase 0 0 Pd / an at ase 1.58 I .55 Reaction conditions: reactant, 0.1 mol dm-, N,H4*2HC1; photocatalyst, 10 mg. even in the presence of a large excess of CH,OH. On the other hand, H,/N, > 1 in the absence of CH,OH. DEPENDENCE OF THE REACTION ON N,H, CONCENTRATION The dependence of the H, evolution rate on the concentration of N,H, is shown in fig. 2. In these experiments, a small amount of the Pt/anatase (lo-, wt/wt, 2 mg) was used in order to keep the conversion and the decrease in the concentration of reactant small even in the most dilute solution.In this concentration region, the rate was higher at lower concentrations and became constant at concentrations > 0.1 mol drn-,. HJN, was 1 even at a concentration of 10 rnol dmP3 : reaction time, 14 h; H, evolved, 21.9 cm3; N, evolved, 20.8 cm3; H,/N, = 1.05.Y. OOSAWA 2.0; - I s - 1.6- 5 1 - e, Y 2 1 . 2 - 3 0 . 8 - c - .& I ? - '0 151 1 I I 0.01 0.03 0.1 0.3 1 3 10 20 concentration/mol dm-3 Fig. 2. Dependence of H2 evolution rate on NzH4 concentration. Reaction conditions: Pt/anatase wt/wt), 2 mg. Table 3. Photocatalytic decomposition of NH, 0.4 Pt Pd Rh Ir, Ru, 0 s p hotoca tal y s t /anatase /anatase /anatase /anatase reaction time/h 5 6 23 4-1 5 volume of gas/cm3 0.52 0.29 0.46 < 0.01 0.28 < 0.01 0.21 < 0.01 H2 N2 Reaction conditions: reactant 0.1 mol dm-, NH4C1; photocatalyst, 10 mg.DECOMPOSITION OF NH, NH, was formed in the photocatalytic reaction of aqueous N,H, solution over Pt/anatase as shown in table 2. NH, formed photocatalytically is also able to decompose phot~catalytically.~~ A few experiments were performed to confirm this point. NH,Cl and precious-metal/anatase ( lop2 wt/wt) were used as reactant and photocatalyst. The results are shown in the table 3. Neither the reaction with the cut-off filter nor the dark reaction proceeded. Therefore the gas evolution in table 3 is considered to proceed photocatalytically. The gas evolution rates with NH4C1 were much lower than those with N,H,.Therefore the contribution from the photocatalytic decomposition of NH, is negligibly small compared with that of N,H4 under the present reaction conditions. DEPENDENCE OF H, EVOLUTION RATE ON THE Pt LOADING RATIO Eight kinds of Pt/anatase with different Pt loading ratios were prepared and H, evolution over them was investigated. The results are presented in fig. 3. H, evolution was detected even when Pt/anatase (lo+ wt/wt) was used. The H, evolution rate increased with increasing Pt loading ratio and reached saturation above a loading ratio of lo-, wt/wt. It was observed that the H, evolution rate is almost constant over a wide range of Pt loading ratios. The Pt coverage in Pt/anatase ( 5 x wt/wt) was estimated to be ca.4%: the Pt surface area estimated from the half-width of the X-ray diffraction peak was 0.71 m2 g-l and the B.E.T. surface area was 17.5 m2 g-l. Therefore the cause of the saturation in the H, evolution rate at low Pt loading ratios1512 PHOTOCATALYTIC HYDROGEN EVOLUTION Pt/anatase (wt/wt) Fig. 3. Dependence of H, evolution rate on Pt loading ratio. Reaction conditions: Pt/anatase, 10 mg; 30 cm3 of 0.1 mol dm-3 N,H,. is not caused by a decrease in the number of photons striking the anatase surface. Note that even Pt/anatase with a very low Pt loading ratio showed high activity: e.g. Pt/anatase ( wt/wt) showed an activity of approximately one-fourth that of the Pt/anatase (lo-, wt/wt). Further study is required to understand these points.DISCUSSION REACTION SCHEME A few standard redox potentialsls for the present reaction system are presented in fig. 4, together with the bottom of the conduction band (Gb)l9 and the top of the valance band (Kb)” of anatase at pH 7. Their relative location changes little with a change of pH. Reactions (1)-(3) correspond to the three redox pairs in fig. 4. Aque- ous hydrazine solution is thermodynamically unstable and can be decomposed thermodynamically according to reaction (4) or (5) by combination of reactions (1) and (2) or reactions (1) and (3), respectively. As has been stated already, however, these reactions did not take place thermocatalytically under the reaction conditions used in the present study. N, + 5H+ + 4e- N,H; (1) 2H+ + 2e- + H, (2) N,H: + 3H+ + 2e- + 2NHi (3) N,H4 + N2+2H, AGO = -22 kcal mol-1 (4) N,H4 --+ 4N2 + QNH, AGO = - 46 kcal mol-l.( 5 ) When the anatase is irradiated by light of energy greater than that of the band gap, the electrons photoexcited into the conduction band can reduce protons producing H, [reaction (2)] or reduce N2H4 producing NH, [reaction (3)] thermodynamically. On the other hand, the positive holes generated in the valence band can oxidize N,H, producing N, [reaction (l)] thermodynamically. Following three features of the reaction can be used to elucidate the photocatalytic reaction scheme : (i) H,/N, is ca. 1 over a wide range of reaction conditions (pH, loaded metal, N2H4Y. OOSAWA 1513 -1 0 w x z ;> b. 2 1 --- 2 3 anatase Fig. 4. Energetic correlation between a few standard redox potentials,18 the bottom of the conduction band (&b)l@ and the top of the valence band (&b)'@ of anatase at pH 7.concentration), (ii) N,/NH, is ca. 0.5 and the value does not change even in the presence of another electron donor (CH,OH) and (iii) the photocatalytic decom- position does not proceed over anatase without loaded metal, which is different from the case of NH,OH.,O At first sight, feature (i) can be explained by the combination of reactions (l), (2) and (3); that is (2)+(3)-(1). However, it is difficult to explain features (ii) and (iii) using this reaction scheme. At first, it is reasonable to assume that N, is formed through the hole-consumption pathway, because N, is the only oxidation product in the present reaction system.Seemingly, there are two kinds of reduction product: H, and NH,. It may also be assumed that H, is formed through the electron-consumption pathway, because H, evolution was only possible by loading precious metals which have a low H, overpotential, as shown in table 2. On the other hand, problems arise if it is assumed that NH, is formed in the electron-consumption pathway. Feature (ii) is very significant. When CH,OH is added to the N2H4 system, some of the positive holes generated by irradiation will react with CH,OH. Then some electrons generated by irradiation will not recombine with positive holes. If NH, were formed in the electron-consumption pathway, some of the electrons not recombining with positive holes would be used in the formation of both H, and NH,.As a result, N,/NH, would be < 0.5, the value obtained in the absence of CH,OH. On the other hand, H,/NH, would be 2, the value obtained in the absence of CH,OH. However, the results shown in table 1 contradict these expectations. Therefore feature (ii) shows that NH, is not formed in the electron- consump tion pathway . Concerning feature (iii), it is useful to compare it to the case where NH,OH is a reactant.,O When NH,OH was used as the reactant, the photocatalytic reaction proceeded even over anatase without any loaded metal, yielding N,, N,O and NH,. In that case, NH, was formed in the electron-consumption pathway, because NH, is the only reduction product. If NH, could have been formed in the electron- consumption pathway for N2H, system also, the photocatalytic reaction [reactions (1) and (3)] over the anatase should have occurred.In practice, however, the photocatalytic reaction over anatase does not proceed. Therefore, feature (iii) also supports the idea that NH, formation over anatase cannot be an electron-consumption pathway in the N2H4 system. 50 FAR 11514 PHOTOCATALYTIC HYDROGEN EVOLUTION As NH, formation cannot be an electron-consumption pathway, H, formation should be the only electron-consumption pathway. Therefore, on the basis of the above discussion, it is reasonable to assume that N, and NH, are formed together in the hole-consumption pathway. It is tentatively postulated that tetrazane (N,H,) is formed as a reaction intermediate. In several reaction systems it has been assumed that positive holes are consumed through the formation of hydroxy radical and through direct reaction with reactant.,l This may also be the case in the present reaction system.Therefore the main reaction will be represented by the reaction scheme shown below. anatase ec(e1ectron) + p+(hole) hv (electron-consumption pathway) H++e-+H' 2H' + H, (over precious metal) (hole-consumption pathway) H,O+p+ -+ H++OH' N,H, + p+ + N,Hi + H+ N,H, +OH' + N2Hi + H20 2N2Hi + N,H,[ = NH,(NH),NH,] N,H, -+ N, + 2NH,. This reaction scheme can explain the three features of the reaction described above. In some cases: H2/N2 < 1. It can be explained by assuming that the following reaction scheme is a side reaction. over precious metal. 1 H' + N2H4 + N2Hj N2Hi + H' + 2NH,, In some cases, on the other hand, H,/N, > 1.It can be explained by assuming that the following reaction scheme is a side reaction or by assuming the decomposition of the NH, formed photocatalytically : N,H, + 2p+ -+ N2H; + H+ N2H; + N,H, + H+ N2H2 + 2p+ + N, + 2H'. RATE-DETERMINING STEP The dependence of the reaction rate on reactant concentration often provides useful information about the rate-determining step of the reaction. Nevertheless, the dependence of the rate of photocatalytic H, evolution on the reactant concentration is reported only for a few cases. In the case of N2H4 solution, the rate decreased with increasing concentration above 0.1 mol dm-3. It is clear that N,H4 depresses H, evolution in this concentration region in the present reaction system. This means that the N,H,-consumption step is not rate-determining.Moreover, as described already, the order of the H, evolution rate over precious-metal/anatase corresponded well to the order of magnitude of the hydrogen overvoltage of these precious metals. ThisY. OOSAWA 1515 suggests that the rate-determining step is H, formation on the surface of the precious metal. Electrolytic H, formation over precious metals is considered to be a combination of hydrogen atoms adsorbed on the surface of the precious metal. On the surface of these metals loaded on anatase, N,H, may also be adsorbed. The dependence of the H, evolution rate on the N,H, concentration depicted in fig. 2 may be understood through the inhibition effect of N,H, on the combination of hydrogen atoms on the Pt surface.1 H. Yoneyama, M. Koizumi and H. Tamura, Bull. Chem. SOC. Jpn, 1979, 52, 3449. J-M. Lehn, J-P. Sauvage and R. Ziessel, Nouv. J. Chim., 1980, 4, 623. S. Sat0 and J. M. White, Chem. Phys. Lett., 1980, 72, 83. E. Borgarello, J. Kiwi, E. Pelizetti and M. Gratzel, J. Am. Chem. SOC., 1981, 103, 6324. K. Domen, S. Naito, T. Ohnishi, K. Tamaru and M. Soma, J. Phys. Chem., 1982, 86, 3657. T. Kawai and T. Sakata, Nature (London), 1980, 286,474. Y. Oosawa, J. Chem. SOC., Chem. Commun., 1982, 221. E. Borgarello, K. Kalyanasundaram and M. Gratzel, Helu. Chim. Acta, 1982, 65, 243. lo L. F. Audrieth and B. A. Ogg, The Chemistry of Hydrazine (Wiley, New York, 1951). l1 C. A. Streuli and P. R. Averell, The Analytical Chemistry of Nitrogen and Its Compounds (Wiley, New 'I P. Pichat, J-M. Herrman, J. Disdier, H. Courbon and M-N. Mozzanega, Nouu. J . Chim., 198 1,5627. York, 1970), part I. R. A. Penneman and L. F. Audrieth, Anal. Chem., 1948,20, 1058. l3 W. T. Bolleter, C. J. Bushman and P. W. Tidwell, Anal. Chem., 1961, 33, 592. l4 J. E. Devries and E. St Clair Ganz, Anal. Chem., 1953, 25, 973. l5 S. Sat0 and J. M. White, J. Phys. Chem., 1981, 85, 592. l6 S. Srinivasan and F. J. Salzano, Int. J. Hydrogen Energy, 1977,2,53. The order of H,-selectivity shown in ref. (20) (Pd > Rh > Ru and Pt > Ir > 0 s ) also corresponded well to the magnitude of their hydrogen overpotential. Encyclopedia of Electrochemistry of the Elements, ed. A. J. Bard (Marcel Dekker, New York, 1978), vol. viii. l7 Q-S. Li, K. Domen, S. Naito, T. Onishi and K. Tamaru, Chem. Lett., 1983, 321. l9 M. V. Rao, K. Rajeshwar, V. R. P. Vernekker and J. Dubow, J. Phys. Chem., 1980, 84, 1987. 2o Y. Oosawa, J. Phys. Chem., in press. When NH,OH-HCl(aq) (30 of 0.1 mol dm-3) was irradiated by a 500 W ultra-high-pressure Hg lamp over a photocatalyst [precious-metal/anatase (lop2 wt/wt), 10 mg], H,, N,, N,O and NH, were formed. Relative reaction rate (normalized to anatase) and H, selectivity [(amount of electron consumed for H, evolution)/(amount of electron consumed for the reduction)] were as follows: anatase, 1, 0; Ru/anatase, 0.55, 0.002; Os/anatase, 0.18,O; Rh/anatase, 0.44,0.17; Ir/anatase, 0.59,0.12; Pd/anatase, 1.02,0.69; Pt/anatase, 0.91,0.37. 21 K. Jones, Comprehensive Inorganic Chemistry, ed. A. F. Trotman-Dickenson et al. (Pergamon Press, Oxford, 1973), vol. 2. (PAPER 3/ 1490) 50-2