Photoreduction of carbon dioxide with water over ... - Can Li Group

4 downloads 251397 Views 151KB Size Report
Received 9 October 2002; received in revised form 13 December 2002; accepted 22 February .... (9) NR-1000 data acquisition system; (10) laptop computer.
Applied Catalysis A: General 249 (2003) 11–18

Photoreduction of carbon dioxide with water over K2 Ti6O13 photocatalyst combined with Cu/ZnO catalyst under concentrated sunlight Guoqing Guan a,∗ , Tetsuya Kida a , Tomohiro Harada b , Munetoshi Isayama b , Akira Yoshida a a

Clarification Composite Materials Group, National Institute of Advanced Industrial Science and Technology, AIST Kyushu, Shuku-machi, 807-1 Tosu, Saga 841-0052, Japan b Fukuoka Industrial Technology Center, 3-2-1 Kamikoga, Chikushino City, Fukuoka 818-8540, Japan Received 9 October 2002; received in revised form 13 December 2002; accepted 22 February 2003

Abstract Photoreduction of CO2 with water into valuable organic compounds under concentrated sunlight as well as Xe- or Hg-lamp irradiation was investigated using a Pt-loaded potassium hexatitanate (K2 Ti6 O13 ) photocatalyst or a composite catalyst in which the Pt-K2 Ti6 O13 photocatalyst was combined with a CO2 hydrogenation catalyst of Cu/ZnO. When the Pt-K2 Ti6 O13 photocatalyst was used under Xe- or Hg-lamp irradiation, H2 , CH4 , HCHO and HCOOH were formed. On the other hand, when the composite catalyst was used under concentrated sunlight, CH3 OH was successfully formed in addition to the above products. For the composite catalyst, H2 resulting from the water decomposition over the photocatalyst serves as the reducing agent for the CO2 hydrogenation over the Cu/ZnO catalyst. The reaction temperature in this study exceeded 580 K due to concentrating the sunlight. Such a high reaction temperature is sufficient for the Cu/ZnO catalyst to reduce CO2 with H2 into CH3 OH. It should also be noted that product yields for the photocatalysts were much improved under concentrated sunlight. This study revealed that a simultaneous supply of photons and thermal energy could improve the activity of photocatalysts. © 2003 Elsevier Science B.V. All rights reserved. Keywords: Photocatalyst; Cu/ZnO; K2 Ti6 O13 ; Photoreduction; CO2 hydrogenation; Solar energy; Concentrated sunlight

1. Introduction The rising atmospheric concentrations of greenhouse gases, especially CO2 , are among the most pressing social issues today. As a potential measure for utilizing CO2 emitted largely from industries as a fuel resource, the hydrogenation of CO2 to methanol on a copper-based catalyst has been extensively stud∗ Corresponding author. Tel.: +81-942-81-3641; fax: +81-942-81-3690. E-mail address: [email protected] (G. Guan).

ied, because methanol can easily be converted into hydrogen or other organic compounds. Although Cu-based catalysts such as Cu/ZnO are well known to be active for methanol synthesis under typical reaction conditions (473–523 K, a total pressure of 5–10 MPa and H2 /CO2 = 3) [1–5], this process requires a huge amount of H2 as the reducing reagent. Up to now, approximately 95% of total H2 production is based on steam reforming technology at present; unfortunately, this leads to a large emission of CO2 . Thus, a highly effective synthesis of H2 without the by-production of CO2 is favored to realize the CO2

0926-860X/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0926-860X(03)00205-9

12

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

hydrogenation process in practice. One of the ideal approaches for solving this problem would be H2 production from water over photocatalysts under solar radiation, and a simultaneous supply of reaction heat to CO2 hydrogenation catalysts by concentrating the solar radiation. It has been shown in many studies that H2 is generated from water over photocatalysts under light irradiation [6]. This could be the most promising way for H2 production in the future, if inexpensive, efficient, and corrosion-resistant photocatalysts become available. It is also emphasized that photocatalysts should preferably work under sunlight when this process is utilized in practice, but very little is known about their activity under sunlight. Thus, we previously investigated the photocatalytic production of H2 from water over various metal oxides under sunlight and found that Pt-loaded potassium hexatitanate (K2 Ti6 O13 ), which has a perovskite-type tunnel structure, showed the highest activity among the catalysts tested [7]. Photoreduction of CO2 with water over photocatalysts such as TiO2 , SiC, CdS and ZnS has also been studied for the direct conversion of CO2 into usable organic compounds such as CH4 , HCHO, HCOOH and CH3 OH [8]. Noticeably, CH3 OH is formed when TiO2 photocatalysts loaded with Cu components were used [9–11]. Mizuno et al. reported that CO2 was photocatalytically reduced with water to HCOOH and CH3 OH in supercritical CO2 containing TiO2 and Cu particles [9]. However, when Cu was directly loaded on TiO2 , CH4 and C2 H6 were mainly formed rather than HCOOH and CH3 OH [10]. On the other hand, Tseng et al. reported that CH3 OH was produced over a Cu-loaded TiO2 catalyst prepared by a sol–gel method under UV irradiation. Moreover, the CH3 OH yield was significantly improved by adding NaOH into the reaction solution [11]. This means that a caustic solution increases the dissolved CO2 concentration and that OH− can act as an effective electron donor, thereby improving the CO2 reduction rate. These findings suggest that Cu components loaded on photocatalysts play a crucial role in the reduction of CO2 into CH3 OH, but the formation of CH3 OH is partly dependent on reaction conditions and catalyst preparation methods. In the present study, we designed a process for producing CH3 OH from CO2 and water under concentrated sunlight, in which H2 evolved from water over photocatalysts was expected to serve as the reducing

agent for the CO2 hydrogenation over a Cu-based catalyst. This is in contrast to the direct conversion of CO2 over photocatalysts. Concentrated sunlight was found to provide a sufficient reaction temperature up to 580 K for the Cu-based catalyst to work properly. For this process, we choose K2 Ti6 O13 as the photocatalyst and combined Pt-loaded K2 Ti6 O13 with a CO2 hydrogenation catalyst of Cu/ZnO. 2. Experimental 2.1. Catalyst preparation The chemicals used in the present study were of reagent grades. They were purchased from Wako Pure Chemical Industry and were used without further purification. The Cu/ZnO catalyst (Cu:Zn = 1:1 in molar ratio) was prepared by a co-precipitation method where 40 ml of a mixed solution containing copper and zinc nitrates (1.0 M each) was added dropwise to 100 ml of a Na2 CO3 solution (1.1 M) at 343 K for 0.5 h under vigorous stirring. After aging for 2 h, precipitates were filtered, thoroughly washed with distilled water, and dried at 373 K overnight. The obtained powders were calcined in air at 623 K for 2 h and reduced at 573 K under H2 flow (20 ml/min) for 3 h. The K2 Ti6 O13 photocatalyst was synthesized as follows: 5.0 g of TiO2 (P-25) powders was dispersed in distilled water containing 1.4 g of K2 CO3 at room temperature. The resulting samples were dried and then calcined at 1213 K for 20 h in air. The Cu/ZnO-loaded K2 Ti6 O13 photocatalyst (hereafter referred to as Cu/ZnO/K2 Ti6 O13 ) with the composition of 0.05Cu:0.05Zn:0.9K2 Ti6 O13 (w:w:w) was prepared by impregnation. K2 Ti6 O13 powders were impregnated in a mixed solution of metal nitrites at a designated ratio under stirring. The resulting powder samples were dried overnight at 373 K, followed by calcinations at 873 K for 2 h in air and subsequently reduced at 573 K for 6 h under H2 flow (20 ml/min). Pt fine particles were loaded on either Cu/ZnO/K2 Ti6 O13 or K2 Ti6 O13 by a photochemical deposition method where 3.0 g of Cu/ZnO/K2 Ti6 O13 or K2 Ti6 O13 powders were dispersed in 60 ml of Na2 CO3 solution (2 M) containing 0.019 g of H2 PtCl6 . The above suspension was irradiated with a 150 W Hg-lamp to deposit Pt on Cu/ZnO/K2 Ti6 O13 or K2 Ti6 O13 particles. Finally,

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

the Pt-loaded Cu/ZnO/K2 Ti6 O13 and the Pt-loaded K2 Ti6 O13 photocatalyst (0.3 wt.% Pt each; hereafter referred to as Pt-Cu/ZnO/K2 Ti6 O13 and Pt-K2 Ti6 O13 , respectively) were thoroughly rinsed with distilled water. The prepared catalysts were analyzed on an X-ray diffractometer (Rigaku, RINT-1400) with Cu K␣ radiation. Diffuse reflectance spectra were recorded with a UV-Vis spectrometer (Jasco V-550). The BET surface areas of the sample powders were measured by a nitrogen adsorption method. 2.2. Photocatalytic reactions Photoreduction of CO2 with water under Xe- or Hg-lamp irradiation at room temperature was carried out using a photoreactor (173 cm3 ) equipped with a flat quartz window (63.3 cm2 ). The catalysts were dispersed on a piece of filter paper fixed on the bottom of the reactor, into which a designated amount of distilled water was introduced. The filter paper acts as a wet bed for the catalyst layer. After degassing by a vacuum system, CO2 was introduced into the reactor, which was externally irradiated through the window with 300 W Xe- or 150 W Hg-lamps. Fig. 1 shows a schematic drawing of an experimental system for the photoreduction of CO2 with water

13

under concentrated sunlight. As illustrated in Fig. 1, the prepared catalysts were placed on wet quartz wool in an optical quartz tube cell (reaction cell, 24.7 cm3 ). A designated amount of distilled water was introduced into the reaction cell. In this case, the quartz wool was used as a wet bed for the catalyst layer. After the reaction cell was deaerated by a vacuum system, CO2 was introduced into the cell. The reaction cell was set on a support bar fixed on an equatorial mount (GPD-type, Vixen, Japan) driven by a sky sensor, which can track the sun automatically. In order to focus the sunlight on the catalyst layer, a concave mirror (diameter = 200 mm) was also fixed on the support bar so that the catalyst layer was positioned at the focal point of the mirror. The reaction temperature was monitored with a thermocouple, which was connected to an NR-1000 data acquisition system (Keyence, Japan) and a laptop computer. Experiments were done between 9:30 and 15:30 h on sunny days. Solar insolation upon the reaction cell was measured with a pyronometer (Ishikawa Sangyou Co. Ltd., Japan). H2 and CH4 produced in the gas phase were determined using a gas chromatograph (GC) (Shimadzu, GC-14B) equipped with a molecular sieve 5A column. The gases were detected with a thermal conductivity detector (TCD) and a flame ionization detector (FID), respectively. Ar was used as a carrier gas in this case. HCHO produced in the liquid phase was determined using a GC (Shimadzu, GC-6A) equipped with a Porapak T column and a TCD. CH3 OH was determined using a GC (Shimadzu, GC-17A) equipped with a capillary column and an FID. He was used as a carrier gas for the analysis of products in the liquid phase. HCOOH was analyzed using a high-pressure liquid chromatograph (HPLC) (JASCO, LC-2000) with a silica-based column. The amount of samples taken from the liquid phase for the analysis was 5 ␮l. 3. Results and discussion 3.1. Characterization of catalysts

Fig. 1. Schematic drawing of experimental set-up for CO2 reduction under concentrated sunlight: (1) concave mirror; (2) reaction cell; (3) sunlight; (4) GPD-type equatorial mount; (5) thermocouple; (6) catalysts; (7) quartz wool; (8) temperature compensator; (9) NR-1000 data acquisition system; (10) laptop computer.

Fig. 2 shows XRD diffraction patterns of the Cu/ZnO catalyst (a) before and (b) after reduction at 573 K for 6 h. Before the reduction, peaks corresponding to CuO and ZnO are clearly observed. However, after the reduction, the intensity of the peaks

14

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

Fig. 2. XRD patterns of the Cu/ZnO catalyst before (a) and after (b) hydrogen reduction at 573 K.

corresponding to CuO decreased considerably, whereas those for ZnO remained almost unchanged. On the other hand, peaks corresponding to Cu appeared in the XRD pattern after the reduction. Metallic Cu is the active site for the CH3 OH synthesis over Cu/ZnO-based catalysts [2–5]; ZnO is claimed to act not only as the support for Cu but also as a promoter for the CO2 hydrogenation [3,12,13]. The BET specific surface area of the prepared Cu/ZnO catalyst was 23.5 m2 /g. Fig. 3(a) and (b) shows XRD diffraction patterns of K2 Ti6 O13 and Cu/ZnO/K2 Ti6 O13 , respectively. It can be seen that Cu and ZnO were dispersed on the K2 Ti6 O13 photocatalyst. In this case, K2 Ti6 O13 can act as a support for the Cu/ZnO catalyst. This could promote transportation of H2 evolved from water on the K2 Ti6 O13 photocatalyst to the active sites of the Cu/ZnO catalyst and could thus improve the CO2 hydrogenation rate, compared to a simple mixture of the above two catalysts. The BET specific surface areas of K2 Ti6 O13 and Cu/ZnO/K2 Ti6 O13 were 1.08 and 3.06 m2 /g, respectively. Fig. 4 shows the diffuse reflectance spectra of K2 Ti6 O13 (a), Pt-K2 Ti6 O13 (b), Cu/ZnO/K2 Ti6 O13 (c) and Pt-Cu/ZnO/K2 Ti6 O13 (d). These results indicate that Pt-K2 Ti6 O13 and Pt-Cu/ZnO/K2 Ti6 O13 can respond to light of wavelength shorter than 420 nm. The absorption edges around 340 nm are not shifted in any sample, indicating that loading Pt and Cu/ZnO

Fig. 3. XRD patterns of the prepared catalysts. (a) K2 Ti6 O13 , (b) Cu/ZnO/K2 Ti6 O13 .

on K2 Ti6 O13 has no significant effect on the light absorption behavior of K2 Ti6 O13 , despite the attachment of ZnO, which is a wide band-gap semiconductor. This is probably due to the small loading amount of Cu/ZnO. The upward shift of light absorbance for Cu/ZnO/K2 Ti6 O13 in the wavelength region from 420 to 600 nm can be explained in terms of the light absorption of black Cu/ZnO dispersed on the K2 Ti6 O13 .

Fig. 4. Diffuse reflectance absorption spectra of (a) K2 Ti6 O13 , (b) Pt-K2 Ti6 O13 , (c) Cu/ZnO/K2 Ti6 O13 and (d) Pt-Cu/ZnO/K2 Ti6 O13 .

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

15

Table 1 Reaction products resulting from the reduction of CO2 with H2 O under Xe- or Hg-lamp irradiation at ambient temperaturea Catalyst

Lamp

Yields of products (␮mol/g catalyst) H2

CH4

HCHO

HCOOH

CH3 OH

Cu/ZnO Cu/ZnO

Xe Hg

0 Trace

0 0

0 0

0 0

0 0

Pt-K2 Ti6 O13 Pt-K2 Ti6 O13

Xe Hg

93.0 408.2

1.04 1.22

0 2.86

0 1.38

0 0

Cu/ZnO/Pt-K2 Ti6 O13 b Cu/ZnO/Pt-K2 Ti6 O13 b

Xe Hg

35.4 76.9

0.96 1.08

0 1.58

0 0.94

0 0

Pt-Cu/ZnO/K2 Ti6 O13 Pt-Cu/ZnO/K2 Ti6 O13

Xe Hg

49.6 102.1

1.32 1.64

1.24 3.42

7.70 18.26

0 0

a

Each experiment was carried out using 0.3 g of catalyst with 77 kPa of CO2 and 4.0 ml of H2 O in a photoreactor equipped with a flat quartz window of ca. 63.6 cm2 . b 0.15 g Cu/ZnO catalyst and 0.15 g Pt-K Ti O 2 6 13 catalyst were physically mixed.

3.2. Photoreduction of CO2 under Xe- or Hg-lamp irradiation Table 1 shows the amounts of products resulting from the photoreduction of CO2 with water under Xe- or Hg-lamp irradiation. Each experiment was carried out using 0.3 g of catalyst with 77 kPa of CO2 and 4.0 ml of distilled water for 6 h in the photoreactor described above at room temperature. As shown in Table 1, no reaction products were detected even when the Cu/ZnO catalyst was irradiated with an Hg-lamp, indicating that the Cu/ZnO catalyst has no photocatalytic activity for the reduction of CO2 with water. On the contrary, H2 and CH4 were evolved for the Pt-K2 Ti6 O13 photocatalyst under Xe-lamp irradiation, and HCHO and HCOOH were also obtained under Hg-lamp irradiation. This shows that the Pt-K2 Ti6 O13 photocatalyst has activity for H2 evolution and CO2 reduction. However, in this study, the amount of O2 evolution was negligible, suggesting that water is oxidized mainly into OH radicals or H2 O2 over the Pt-K2 Ti6 O13 photocatalyst [14,15]. On the other hand, when the Cu/ZnO and Pt-K2 Ti6 O13 powders were physically mixed through grinding, the yields of products decreased under both Xe- and Hg-lamp irradiation. The decrease in activity observed here might be due to the light shadowing effect of the Cu/ZnO powders on the Pt-K2 Ti6 O13 when they were mixed, because the Cu/ZnO catalyst has no photocatalytic activity and the light incident

on Cu/ZnO is not utilized for the photocatalytic reaction, but lost as heat. For the Cu/ZnO catalyst, a relatively high temperature is necessary for the CO2 hydrogenation [1–5]. Thus, at room temperature, the Cu/ZnO catalyst shows no effects on the reduction of CO2 . However, the organic compound yields were increased when the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst was used, despite a decrease in the H2 yield. In particular, HCHO and HCOOH were formed even under Xe-lamp irradiation. In general, photogenerated electrons in semiconductors tend to move to the loaded metal particles, which facilitates the separation of electron-hole pairs and improves the photocatalytic reaction rate. Thus, for Pt-Cu/ZnO/K2 Ti6 O13 , the formation of HCHO and HCOOH is attributable to the Pt-K2 Ti6 O13 and Cu-K2 Ti6 O13 junctions present in the composite catalyst; the metallic Cu loaded on photocatalysts can act as active sites for CO2 reduction, as in the case for Cu/TiO2 [9–11]. It has been reported that CO2 can photocatalytically be reduced into CH3 OH with H2 O over Cu-loaded TiO2 catalysts under UV irradiation [9–11]. However, no CH3 OH was detected when the prepared catalysts were exposed to Xe- or Hg-lamp irradiation at room temperature in the present study. Thermodynamically, transforming 1 mol of CO2 with water into CH3 OH requires 228 kJ of energy. Thus, the CO2 hydrogenation over the Cu/ZnO catalyst is always performed at relatively high temperatures and pressures. This is probably the reason why no CH3 OH was formed over

16

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

the prepared catalysts when used at room temperature. We next tried to reduce CO2 into CH3 OH at relatively high temperatures over the prepared catalysts, to which photon and thermal energy was provided by concentrating the solar radiation. 3.3. Photoreduction of CO2 under concentrated sunlight The reduction of CO2 with water over the prepared catalysts under concentrated sunlight was performed using 0.3 g of catalysts with 202 kPa of CO2 and 4.0 ml of H2 O in the photoreactor shown in Fig. 1. Table 2 summarizes the yields of obtained reaction products over various catalysts under concentrated or non-concentrated sunlight. It was found that CH3 OH was formed over the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst under concentrated sunlight, whereas no CH3 OH was formed under non-concentrated sunlight. The average reaction temperatures in the photoreaction cell increased up to 583 K, which was in the range of the normal operation temperature for the Cu/ZnO catalyst. Obviously, such high reaction temperatures are responsible for the CH3 OH formation over the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst. Although the reaction pressure also affects the CH3 OH yield to some extent, the reaction temperature is a prerequisite for the CH3 OH formation in this system. On

the other hand, no products were detected over the Cu/ZnO catalyst in the absence of the Pt-K2 Ti6 O13 photocatalyst even at 568 K, as shown in Table 2. This suggests that H2 is indispensable for the CO2 hydrogenation over the Cu/ZnO catalyst. The product yields for the Pt-K2 Ti6 O13 photocatalyst were increased compared to those obtained under non-concentrated sunlight. A similar phenomenon was reported by Fujii and Masuda [16], who found that heating the reaction solution led to a remarkable improvement in the photocatalytic activity of a photocatalyst for water decomposition. It is therefore possible that both the thermal and the photon energy provided by concentrated sunlight facilitate CO2 reduction and water decomposition over photocatalysts more efficiently than normal sunlight does. However, no CH3 OH was detected over the Pt-K2 Ti6 O13 photocatalyst even at such high temperatures; Pt loaded on photocatalysts has no activity for CH3 OH synthesis. In order to ascertain whether CO2 is hydrogenated into CH3 OH over the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst by H2 resulting from the water decomposition over the Pt-K2 Ti6 O13 photocatalyst, the two catalysts, Cu/ZnO and Pt-K2 Ti6 O13 , were packed in series inside the reactor in a 1:1 (w:w) and were irradiated with concentrated sunlight. CH3 OH was obtained even in this case, as shown in Table 2.

Table 2 Reaction products resulting from the reduction of CO2 with H2 O under concentrated sunlighta Catalyst

Average reaction temperature (K)

Solar insolation (kWh/m2 )

Maximum atmospheric temperature (K)

Yields of products (␮mol/g catalyst) H2

CH4

Cu/ZnO

568

5.71

288

Trace

0

Pt-K2 Ti6 O13

547

5.32

283

196.93 (120.37)c

1.37 (1.08)

33.56 (2.92)

Cu/ZnO/Pt-K2 Ti6 O13 d Cu/ZnO/Pt-K2 Ti6 O13 e Cu/ZnO/Pt-K2 Ti6 O13 f

571 583 553

6.02 6.34 5.41

300 306 284

89.8 110.2 174.37

0.69 0.78 0.86

Pt-Cu/ZnO/K2 Ti6 O13

556

5.57

286

93.31 (51.74)

1.73 (1.42)

a

HCHO

0

HCOOH

0

TONb CH3 OH

0



124.31 (4.87)

0 (0)



4.31 9.83 10.97

14.52 19.59 43.47

13.60 15.98 18.93

0.0035 0.012 0.0045

13.72 (2.77)

57.27 (10.08)

32.03 (0)

0.041

Each experiment was carried out using 0.3 g of catalyst with 202 kPa of CO2 and 4.0 ml of H2 O in the photoreactor as shown in Fig. 1. TON is defined as the ratio of the amount of CH3 OH produced to that of the active site Cu in the catalyst. c Data in the parentheses were obtained under non-concentrated sunlight. d 0.15 g Cu/ZnO catalyst and 0.15 g Pt-K Ti O 2 6 13 catalyst were physically mixed. e 0.05 g Cu/ZnO catalyst and 0.25 g Pt-K Ti O 2 6 13 catalyst were physically mixed. f 0.15 g Cu/ZnO catalyst and 0.15 g Pt-K Ti O 2 6 13 catalyst were packed in series inside the photoreactor. b

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

Thus, we conclude that H2 for the CO2 hydrogenation over the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst was provided by the water decomposition over the Pt/K2 Ti6 O13 photocatalyst. However, the CH3 OH yield was lower than that for the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst. It is well known that active atomic H is formed on Pt that is loaded on catalysts by a H2 spillover effect [17]. The Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst may have relatively intimate contact between the Pt-K2 Ti6 O13 and Cu/ZnO catalysts. This could assist in the transport of H2 formed over Pt-K2 Ti6 O13 to the neighboring surface of Cu/ZnO, thereby increasing the CH3 OH yield. This assumption is supported by the fact that, when the Cu/ZnO and Pt-K2 Ti6 O13 catalysts were physically mixed, the CH3 OH yield was decreased to some extent, as shown in Table 2. In addition, other compound yields for the physical mixture of the two catalysts were lower than those for the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst. This probably means that the photocatalytic active site of Pt on K2 Ti6 O13 is partly masked by mixing the Cu/ZnO and Pt-K2 Ti6 O13 catalysts, resulting in low yields for H2 , CH4 , HCHO and HCOOH. Thus, the catalyst preparation method employed here, in which Cu/ZnO was loaded on K2 Ti6 O13 with the subsequent loading of Pt, is useful for improving product yields. It is reported that CH4 can be partially oxidized into CH3 OH with the active oxygen species such as OH radicals or O2 2− over metal-loaded catalysts at high temperatures [18]. We thus speculate on a possibility that a part of the active oxygen species generated on the photocatalyst may migrate to a neighboring Cu/ZnO site and convert CH4 to CH3 OH under concentrated sunlight. Furthermore, other published results also suggest that HCHO and HCOOH can be reduced into CH3 OH with H2 resulting from water splitting over the Cu/ZnO-based catalysts [5,19–21]. The obtained results show the effectiveness of the combination of the Cu/ZnO catalyst with the K2 Ti6 O13 photocatalyst for synthesizing CH3 OH from CO2 and water under concentrated sunlight. However, as shown in Table 2, the apparent turnover number (TON), defined as the ratio of the amount of CH3 OH produced to that of the active site Cu in the catalyst, was low for all the cases. Such a low CH3 OH yield can often been seen in the literature [8–11], implying the difficulty of the photocatalytic reduction of CO2 . On the other hand, this study revealed that

17

CH3 OH was catalytically formed over the Cu/ZnO catalyst by the H2 resulting from the water decomposition over the K2 Ti6 O13 photocatalyst when the Pt-Cu/ZnO/K2 Ti6 O13 composite catalyst was used. Thus, the observed low TON values must be due to the insufficient amount of H2 produced over the K2 Ti6 O13 photocatalyst, because the Cu/ZnO catalyst is reported to be active for the CO2 hydrogenation under the reaction conditions of 101 kPa and 483 K [22], which are similar to those used in this study (202 kPa, 556 K). The obtained TON values suggest that more active photocatalysts, which can supply a sufficient amount of H2 for the CO2 hydrogenation over the Cu/ZnO catalyst to proceed efficiently, should be developed to improve the CH3 OH yields in this reaction system. 3.4. Reaction scheme for the reduction of CO2 under concentrated sunlight From the above results, we propose a possible reaction scheme for the reduction of CO2 with water over the Pt-K2 Ti6 O13 photocatalyst combined with CO2 hydrogenation catalyst of Cu/ZnO under concentrated sunlight as follows. (i) Incident photons are absorbed by K2 Ti6 O13 , and the photoexcited electrons and holes are produced. (ii) The electrons move to the loaded Pt on K2 Ti6 O13 , where the water decomposition and CO2 reduction proceed competitively. Water molecules adsorbed on Pt are photodecomposed into H2 and OH radicals or O2 2− . On the other hand, CO2 adsorbed on Pt is activated to become CO2 − or C radicals and is reduced into CH4 , HCOOH and HCHO. Moreover, a part of photocatalycally produced CH4 is oxidized into CH3 OH on Cu/ZnO by the OH radicals or O2 2− . (iii) H2 produced over Pt loaded on K2 Ti6 O13 moves to the neighboring Cu/ZnO surface by a H2 spillover effect. CO2 chemisorbed and activated on Cu/ZnO is finally reduced into CH3 OH by the resulting H2 . Parts of the photocatalycally produced HCHO and HCOOH are also reduced into CH3 OH by H2 over Cu/ZnO. 4. Conclusions We performed the photoreduction of CO2 with water using the composite catalyst, for which the Pt-K2 Ti6 O13 photocatalyst was combined with the

18

G. Guan et al. / Applied Catalysis A: General 249 (2003) 11–18

Cu/ZnO catalyst, under concentrated sunlight. It was found that CO2 was reduced into CH3 OH over the composite catalyst, depending largely on the reaction temperature. By concentrating the sunlight, the reaction temperature increased to 580 K, which was sufficiently high for the Cu/ZnO catalyst to reduce CO2 with H2 into CH3 OH. In this system, H2 resulting from the water decomposition over the Pt-K2 Ti6 O13 photocatalyst serves as the reducing agent for the CO2 hydrogenation over the Cu/ZnO catalyst. This study demonstrates that combining photocatalysts with CO2 hydrogenation catalysts is effective for synthesizing CH3 OH from CO2 and water without addition of H2 under concentrated sunlight. However, the CH3 OH yield is quite low at the present. The development of more efficient photocatalysts for H2 production holds the key to this reaction system.

Acknowledgements The financial support provided for this work from Japan Science and Technology Corporation (JST) is gratefully acknowledged. References [1] H.J. Freund, M.W. Roberts, Surf. Sci. Rep. 25 (1996) 225. [2] Y. Borodko, G.A. Somorjai, Appl. Catal. A: Gen. 186 (1999) 355. [3] Y. Choi, K. Futagami, T. Fujitani, J. Nakamura, Appl. Catal. A: Gen. 208 (2001) 163.

[4] J. Toyir, P.R. Piscina, J.L.G. Fierro, N. Homs, Appl. Catal. B: Environ. 29 (2001) 207. [5] Z.M. Hu, K. Takahashi, H. Nakatsuji, Surf. Sci. 442 (1999) 90. [6] T. Bak, J. Nowotny, M. Rekas, C.C. Sorrell, Int. J. Hydrogen Energy 27 (2002) 991. [7] W. Shangguan, A. Yoshida, Int. J. Hydrogen Energy 24 (1999) 425. [8] H. Yoneyama, Catal. Today 39 (1997) 169. [9] T. Mizuno, H. Tsutsumi, K. Ohta, A. Sagi, H. Noda, Chem. Lett. 46 (1994) 1553. [10] K. Adachi, K. Ohta, T. Mizuno, Sol. Energy 53 (1994) 187. [11] I.H. Tseng, W.C. Chang, J.C.S. Wu, Appl. Catal. B: Environ. 37 (2002) 37. [12] R. Burch, S.E. Golunski, M.S. Spencer, J. Chem. Soc., Faraday Trans. 86 (1990) 2683. [13] M.S. Spencer, Catal. Lett. 50 (1998) 37. [14] G. Munuera, J.P. Espinos, A. Fernandez, P. Malet, A.R. Gonzalez-Elipe, J. Chem. Soc., Faraday Trans. 86 (1990) 3441. [15] M. Shibata, A. Kudo, A. Tanaka, K. Domen, K. Maruya, T. Onishi, Chem. Lett. (1987) 1017. [16] T. Fujii, T. Masuda, in: Proceedings of the Abstract of 67th Conference of Chemical Society of Japan, Tokyo, 3M108, 1994. [17] U. Roland, T. Braunschweig, F. Roessner, J. Mol. Catal. A 127 (1997) 61. [18] K. Otsuka, Y. Wang, Appl. Catal. A: Gen. 222 (2001) 145. [19] K. Tabata, Y. Teng, Y. Yamaguchi, H. Sakurai, E. Suzuki, J. Phys. Chem. A 104 (2000) 2648. [20] T. Takemoto, D. He, Y. Teng, K. Tabata, E. Suzuki, J. Mol. Catal. A 179 (2002) 279. [21] T. Takemoto, D. He, Y. Teng, K. Tabata, E. Suzuki, Appl. Catal. A: Gen. 225 (2002) 177. [22] S. Fujita, M. Usui, H. Ito, N. Takezawa, J. Catal. 157 (1995) 403.