In Situ Ir Studies Mechanism Of Methanol

Applied Catalysis A: General 171 (1998) 301±308

In situ IR studies on the mechanism of methanol synthesis over an ultra®ne Cu/ZnO/Al2O3 catalyst Qi Suna, Chong-Wei Liub, Wei Panb, Qi-Ming Zhub, Jing-Fa Denga,* a

b

Department of Chemistry, Fudan University, Shanghai 200433, China State Key Laboratory of C1 Chemistry and Technology, Tsinghua University, Beijing 100084, China Received 22 October 1997; received in revised form 4 February 1998; accepted 26 February 1998

Abstract Methanol synthesis from CO2 and CO/CO2 hydrogenation was carried out under real reaction conditions over an ultra®ne Cu/ZnO/Al2O3 (Cu/Zn/Alˆ60/30/10, molar ratio) catalyst. The formation and variation of surface species were recorded by in situ FT-IR spectroscopy. The mechanisms of methanol synthesis and RWGS reaction were discussed. The result revealed that methanol was formed directly from CO2 hydrogenation for CO2/H2 or CO/CO2/H2 reaction systems. b-HCOOÿ s was the necessary intermediate for methanol synthesis. A scheme of methanol and RWGS reaction was proposed. # 1998 Elsevier Science B.V. All rights reserved. Keywords: Methanol synthesis; Mechanism; In situ; FT-IR

1. Introduction In industrial processes, methanol is synthesized from hydrogenation of mixtures of CO and CO2 over Cu-based catalyst. Addition of a small amount of CO2 to the mixture of CO and H2 can promote methanol yield remarkably. Because of this attractive promoting effect, methanol synthesis from CO2 and H2 has caused considerable attention in recent years [1±7]. However, there is still controversy over some important questions, such as: 1. the role of carbon dioxide in the process of methanol synthesis;

*Corresponding author. Fax: +86 2165 641740; e-mail: [email protected] 0926-860X/98/$19.00 # 1998 Elsevier Science B.V. All rights reserved. PII S0926-860X(98)00096-9

2. whether CO or CO2 serves as the primary carbon source for methanol synthesis; 3. whether the inter-conversion between CO and CO2 via the water-gas shift (WGS)/or reverse water-gas shift (RWGS) is an indispensable process or not; and 4. what are the reaction intermediates involved in methanol synthesis. By the use of the temperature-programmed-desorption (TPD) [8±10], IR spectroscopy [11±13], chemical trapping methods [14] and 14 C or 13 C isotopic tracer [15,16], various surface species were found on Cubased catalyst in the course of methanol synthesis and some reasonable mechanisms were suggested, but most of these experiments were carried out at highvacuum or atmospheric pressure conditions using clean Cu single crystal, supported Cu/SiO2 catalyst

302

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308

or Cu-based catalyst with low copper content. The results obtained under these conditions diverge from the results obtained under real reaction processes. In this work, a high activity ultrafine Cu/ZnO/Al2O3 catalyst (Cu/Zn/Alˆ60/30/10, molar ratio) prepared by ``oxalate gel co-precipitation method'' [17±19] was used as working catalyst to study the mechanism of methanol synthesis from CO2 and CO/CO2 hydrogenation by means of FT-IR spectroscopy under real reaction conditions (high pressure, high temperature and feed gas continuously flowing over the catalyst). A scheme of methanol synthesis formation was suggested. 2. Experimental Cu/ZnO/Al2O3 catalyst (Cu/Zn/Alˆ60/30/10, molar ratio) used in this study was prepared by ``the oxalate gel co-precipitation method'' similar to that described previously [17±19]. The BET surface area of the Cu/ZnO/Al2O3 catalyst is 49.7 m2/g, Cu surface area measured by N2O titration is 36.3 m2/g and the metallic copper particle size evaluated by XRD and TEM is 10.7 nm. Infrared experiments were conducted on a 30 mg wafer of the mixture of the catalyst and g-Al2O3 (catalyst/g-Al2O3ˆ1/6, weight ratio) crushed to a powder smaller than 500 mesh placed inside an in situ reaction cell (shown as in Fig. 1). Infrared spectra were recorded with a PerkinElmer system 2000 FT-IR spectrometer. A resolution of 4.0 cmÿ1 was used throughout the investigation and

50 scans taken over a 20 s interval were averaged to achieve a satisfactory signal-to-noise ratio. The ultrahigh purity (>99.999%) gases of H2, CO, N2 and chemical purity CO2 (>99.9%) were used. The gases Ê in the in¯ow were puri®ed by passing through a 5 A molecular sieve-trap. After reduction by H2 at 513 K for 4 h, the reaction cell packed with the catalyst was rapidly cooled to room temperature and ¯ushed with a high-purity N2 for 2 h. After switching N2 with reaction gas (CO2/H2ˆ1/3 or CO/CO2/H2ˆ20/5/75) and pressurizing slowly to 2.0 MPa, temperatureprogrammed-reaction (TPR) experiments (the ¯ow rate of reaction gas was 30 ml/min and the rate of temperature increase was 2 K/min) were performed and IR spectra were recorded synchronously. 3. Results and discussion Fig. 2 shows part of the IR spectra obtained in ¯owing CO2/H2 reaction gas at 2.0 MPa. It is found that a few adsorbed CO can be formed easily at room temperatures. The band at 2079 cmÿ1 is resolved only at a low temperature range and is attributed to CO adsorbed on a low Miller-index plane of reduced Cu [20]. The peaks at 2171 and 2118 cmÿ1 are characteristic of gaseous CO. With increase in the temperature, exceeding the level of 313 K, it is found that the intensity of bands of surface adsorbed CO (2079 cmÿ1) and gaseous CO (2118 and 2171 cmÿ1) decrease gradually, indicating the decrease of CO concentration over catalyst surface

Fig. 1. Schematic diagram of in situ IR reaction cell: (1) cell body; (2) cell core; (3) window frame; (4) NaCl crystal window; (5) O-ring; (6) sample fixing ring; (7) sample wafer.

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308

303

Fig. 2. IR spectra for methanol synthesis from CO2/H2 recorded during temperature-programmed-reaction (TPR) process at 2.0 MPa.

and gas phase, and the rato-vibrational spectrum of H2O (1600±1800 cmÿ1) begins to arise and increase. Meanwhile, monodentate formate species (mHCOOÿ) are observed at 1585 cmÿ1. The results indicate that CO species are formed from dissociative adsorption of CO2 over catalyst surface. With the temperature increasing further (higher than 323 K), during the increase of the intensity of peak at 1585 cmÿ1, the intensity of another peak at 1593 cmÿ1 due to bidentate carbonate species …b-CO2ÿ 3 † also increases more rapidly. Up to about 493 K, it is found that these two peaks weaken again. Fig. 3 shows the process of formation and change for CH3O and CH3OH species at various temperatures. It is found that several peaks at 2857 and 2940 cmÿ1, which are characteristic of b-HCOOÿ s , are present at lower temperatures, e.g. 313 K. As the temperature is raised further, the intensity of the peaks at 2857 and 2940 cmÿ1 is also increased and reaches a steady-state concentration above 413 K. However, the peaks at 2918 and 2965 cmÿ1, assigned to CH3Os and CH3OH species, respectively, occur and are ampli®ed gradually until the temperature is raised above 433 K. Such a result reveals that, although some adsorbed and gaseous CO are present at low temperature, no CH3Os (adsorbed methoxy species) or CH3OH were formed, and b-HCOOÿ s species also could be ignored. While the temperature is higher than 393 K, the amount of

Fig. 3. Formation and variation of methoxy group and methanol during temperature programmed reaction (TPR) process at 2.0 MPa for methanol synthesis from CO2‡H2.

CO species (COs and gaseous CO) and that of m-HCOOÿ s decrease, and CH3Os as well as product CH3OH increase due to the increase of b-HCOOÿ s . From these results, it is concluded that methanol is formed directly from hydrogenation of CO2, and that methanol formation from hydrogenation of carbon monoxide formed via RWGS is negligible. The key

304

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308

intermediate species for methanol synthesis from CO2 hydrogenation is bidentate formate …b-HCOOÿ s † instead of monodentate formate …m-HCOOÿ s †. Further, when we switch feed gas CO2/H2 with pure H2, all the band intensities of CO, CO2, m-CO2ÿ 3 , mHCOOÿ become weaker slowly, but CH3Os and CH3OH could still be clearly detected due to the existence of surface formate b-HCOOÿ at 2940 cmÿ1. When b-HCOOÿ (2940 cmÿ1) fades away, no methanol species could be observed by IR. The above experimental observation and analysis strongly indicate that the process of the hydrogenation of bidentate formate …b-HCOOÿ s † is the rate-limiting step for methanol synthesis from CO2 hydrogenation. This result is in agreement with previous reports [11,21]. TPR-IR experiments were also conducted with H2/ CO/CO2 feed gas. Fig. 4 shows the spectra obtained when the temperature was increased from 293 K with the rate of 2 K/min at a pressure of 2.0 MPa and gas ¯ow rate of 30 ml/min. It is found that a sharp peak at 2006 cmÿ1 ascribed to absorbed CO (carbonyl band)

varied from weak to intense in the temperature range 293±343 K and then to weak again with the increase of temperature. It is coincident with the variation of the peak at 1634 cmÿ1 due to the bending mode of H2O physisorbed on Cu. However, no formate or carbonate species appear under these conditions. As the temperature increases to above 353 K, the peak at 2006 cmÿ1 disappears and a broader peak at 1972 cmÿ1 grows rapidly and then decreases gradually. Moreover, the bands at 1591 and 1601 cmÿ1, which are characteristic of m-HCOOÿ and b-CO2ÿ 3 species, respectively, increase gradually and reach the maximum at 493 K. Furthermore, when the temperature is increased again, two bands of 1591 and 1601 cmÿ1 diminish. This variation is similar to the change of CO2‡H2 in the TPR-IR experiment. However, it is also found that, besides the bands of 2855 and 2940 cmÿ1 (ascribed to bidentate formate b-HCOOÿ s and CH3Os, respectively), product CH3OH is also detected upon the temperature of 443 K. This indicates that the necessary intermediate is still

Fig. 4. IR spectra for methanol synthesis from CO/CO2/H2 recorded during temperature-programmed-reaction (TPR) process at 2.0 MPa.

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308 ÿ b-HCOOÿ s species instead of m-HCOOs or other species. It is noteworthy that the intensity of bands of the rato-vibrational spectrum of H2O at 1600± 1800 cmÿ1 is obviously weaker for CO/CO2‡H2 system than that for CO2/H2. So the addition of CO could inhibit the reaction of RWGS. Namely, the formation of water species could be suppressed and the formation rate of methanol could be enhanced by the addition of CO into CO2/H2 feed gas. Fig. 5 shows the methanol formation rates for feed gases containing only CO2 and H2, containing only CO and H2 as well as containing CO/CO2 mixture and H2. It is found that both the methanol formation rates for the individual CO and CO2 hydrogenation are much lower than that for CO/CO2 mixture hydrogenation, and the methanol formation rate for CO hydrogenation is lower than for CO2 at 499 K. It is clear that the addition of CO into CO2/H2 feed gas promotes the formation rate of methanol signi®cantly. Especially, we found that the variation of the amount of water is very signi®cant for CO2/H2 and CO/CO2/H2 reaction systems. After addition of CO into CO2/H2, only a trace of water was detected and the amount is three orders of magnitude lower than that of CO2/H2 reaction system. This result was also found in this in situ IR±TPRS experiment and in previous kinetic and catalytic activity testing [19]. According to the Arrhenius equation, ln(TOF) of methanol formation is proportional to the inverse temperature. By plotting the formation rate from kinetic and activity testing versus 1/T (Fig. 6), an apparent activation energy could be determined from the slope. The slope is measured at the temperature ranging from 433 to 453 K, where the rates of both methanol synthesis and RWGS start to increase rapidly before the conversion signi®cantly alters the gas-phase composition. The activation energy for methanol synthesis is 20.7 kcal/mol and that for RWGS reaction is 22.73 kcal/mol. After the addition of CO into CO2/H2 feed gas, however, the activation energy for methanol synthesis is decreased to 6.52 kcal/mol, although there is no signi®cant change in the activation energy for RWGS reaction. It is clear that the addition of CO into CO2/H2 feed gas could lead to the decrease of the activation energy for methanol synthesis. As a result, the rate of methanol formation is increased and the methanol selectivity is promoted to a large extent. This is in agreement with the results from this in situ IR experiment.

305

Fig. 5. Methanol synthesis rate for CO/H2, CO2/H2 and CO/CO2/ H2 feeds over the ultrafine Cu/ZnO/Al2O3 catalyst. Pˆ2.0 MPa, space velocityˆ4500 hÿ1.

According to the above analysis, some suggestions for methanol synthesis from CO/CO2/H2 could be proposed: 1. The procedures of dissociative adsorption and hydrogenation of CO2 are two parallel competitive reactions. At low temperatures, the formation of

Fig. 6. Arrhenius plot ln(TOF) for methanol formation and RWGS reaction versus the inverse reaction temperature (1/T) over Cu/ ZnO/Al2O3 catalyst for CO2‡H2 (&,*) and CO/CO2‡H2 (~).

306

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308

CO (corresponding to the bands of 2007 cmÿ1) along with the adsorbed Os by dissociative adsorption is a prior process. These adsorbed oxygen species (Os) combine with nascent hydrogen (Hs) to form hydroxy species (OHs) and then to H2O, so the peak at 1634 cmÿ1 is observed. As temperature increases, the rate of CO2 adsorption/ hydrogenation increases and the process of dissociative adsorption of CO2 is suppressed, so the bands of 1634 and 2007 cmÿ1 diminish gradually and then disappear. 2. With the increase of temperature, the bands at 1591 and and 1601 cmÿ1 (ascribed to m-HCOOÿ s m-CO2ÿ 3 species, respectively) increase gradually, and then diminish after reaching the maximum, while the bands at 2857 …b-HCOOÿ s †, 2918 and 2940 cmÿ1 (CH3Os) increase gradually along with the variation of the intensity of the band at 2965 cmÿ1 (CH3OH). This indicates that the formation of CH3Os and CH3OH species are accompanied by the formation of bidentate formate ÿ …b-HCOOÿ s †, and b-HCOOs species is a key intermediate, which undergoes stepwise hydrogenation to form CH3Os then to CH3OH. The rate-limiting step for methanol synthesis over the ultrafine Cu/ ZnO/Al2O3 catalyst is the hydrogenation of b-HCOOÿ s . 3. The effects of addition of CO to the CO2/H2 feed gas are ascribed to the inhibition of the RWGS reaction (or enhancement of the WGS reaction). Water produced via methanol synthesis from CO2 hydrogenation is consumed by the fast water-gas shift reaction that simultaneously provides CO2, which is the feed gas for methanol synthesis. As a result, the limitation of thermodynamic equilibrium for CO2 hydrogenation to CH3OH is removed and leads to a higher methanol yield (methanol synthesis formation rate) and selectivity. This result has been discussed in kinetic terms previously [19]. From Fig. 6, it is also found that with the addition of CO into CO2/H2 feed gas, apparent activation energy for methanol synthesis is decreased. It led to the promotion of methanol formation rate and selectivity. One explanation is the following: in the absence of CO, H2O is formed earlier via RWGS reaction and it prohibits the reaction of methanol synthesis; when CO is introduced, relative high coverage of CO leads to the

reduction of dissociative adsorption of CO2 on the catalyst surface and the inhibition of RWGS reaction. Therefore, the reaction of methanol synthesis can take place quickly. Moreover, the Os species formed via b-HCOOÿ s hydrogenation to H2COs is consumed by the adsorbed COs to form [CO2]s rapidly. Thus, a direct result is that methanol formation rate and selectivity is increased. From such an analysis, it is easy to illustrate the pathway of methanol synthesis and RWGS reaction by the scheme shown in Fig. 7: In the absence of CO, the process of dissociative adsorption for CO2 could take place easily over the catalyst surface [22] and there is a strong tendency for pathway (I) to happen; even if the reaction occurs via pathway (II), this surface (O=C±OH)s could dissociate to form COs and OHs easily due to the absence of CO in gas phase. However, when CO is introduced into CO2/H2 feed gas, the process of dissociative adsorption for CO2 is inhibited due to the high coverage of CO. The reaction tends to take place via pathway (II) to form O=C±OHs, which then isomerizes to form other intermediate species b-HCOOÿ s , but could not dissociate to form COs and OH easily. The direct result is that IR band absorption intensity at 1600±1800 cmÿ1 ascribed to the rato-vibrational spectrum of H2O in CO/CO2/H2 system is much weaker than that in CO2‡H2 system, and methanol formation rate and selectivity are promoted significantly. On the other hand, if the conclusions that b-HCOOÿ s species are a key intermediate and hydrogenation of b-HCOOÿ s is the rate-limiting step for methanol synthesis are true, it seems to be ambiguous to explain the fact of the lowering of apparent activation energy after introducing CO gas into CO2/H2 mixture. Although we did not make further study on the adsorption/ desorption species and hydrogenation activity of those adsorbed species on this ultra®ne Cu/ZnO/Al2O3 catalyst, some previous research works have given some powerful hints which could be helpful to explain the above divergency. Fujita et al. [23], found that two types of adsorbed formate species, HCOO±Cu and HCOO±Zn, can be formed and hydrogenate to form CH3O±Zn and then methanol, but the rate constant of the hydrogenation of HCOO±Cu is about 10 times greater than that of the hydrogenation of HCOO±Zn and the activation energy of the former is lower than that of the latter. Based on this result, we can easily explain the mechanism of methanol synthesis and the

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308

307

Fig. 7. Scheme for mechanism of methanol synthesis and RGWS/GWS reaction from CO2/H2 and CO/CO2/H2 over ultrafine Cu/ZnO/Al2O3 catalyst.

fact of the lowering of apparent activation energy after introducing CO into CO2/H2 mixture. For CO/H2 reaction system, a mount of Os or OHs is formed with CH3O±Zn being formed from HCOO±Cu hydrogenation, as shown in Fig. 7. The adsorbed Os or OHs could inhibit the hydrogenation of HCOO±Cu, so, under steady-state condition, methanol formation rate is low and apparent activation energy is high. However, after introducing CO into CO2/H2 system, CO reacts quickly with Os to form CO2 and remove the adsorbed Os from the equilibrium system. As a direct result, the hydrogenation of HCOO±Cu is promoted and methanol formation rate is accelerated as well as the apparent activation energy of methanol synthesis is lowered. 4. Conclusion The mechanism of methanol synthesis and RGWS reaction was studied by using in situ IR technique. The experimental results indicate that methanol was

formed directly from CO2 hydrogenation both for CO2 and CO/CO2 hydrogenation. b-HCOOÿ s is the key intermediate for methanol synthesis and hydrogenation of b-HCOOÿ s is the rate-limiting step for methanol synthesis. The addition of CO to CO2/H2 system not only affects the reaction pathway but also lowers the apparent activation energy of methanol synthesis. References [1] B. Denise, R.P.A. Sneeden, J. Mol. Catal. 17 (1982) 359. [2] B. Denise, O. Cherifi, M.M. Bettahar, R.P.A. Sneeden, Appl. Catal. 48 (1989) 365. [3] Y. Amenomiya, Appl. Catal. 30 (1987) 57. [4] J.A.B. Bourzutschky, N. Homs, A.T. Bell, J. Catal. 124 (1990) 73. [5] Y. Nitta, O. Suwata, Y. Ikeda, Y. Okamoto, Catal. Lett. 26 (1994) 345. [6] R.A. Koeppel, A. Baiker, Appl. Catal. 84 (1992) 77. [7] J.S. Lee, K.H. Lee, S.Y. Lee, Y.G. Kim, J. Catal. 144 (1993) 414.

308

Q. Sun et al. / Applied Catalysis A: General 171 (1998) 301±308

[8] D.L. Roberts, G.L. Griffin, J. Catal. 110 (1988) 117. [9] D.L. Roberts, G.L. Griffin, Appl. Surf. Sci. 19 (1984) 298. [10] D. Bianchi, J.L. Gass, M. Khalfauah, S.J. Teicher, Appl. Catal. A 123 (1995) 89. [11] D.B. Clarke, A.T. Bell, J. Catal. 154 (1995) 314. [12] G.N. Stelios, A.J. Marchi, G.F. Froment, Appl. Catal. A 86 (1992) 45. [13] J.F. Edwards, G.L. Schrader, J. Catal. 94 (1985) 175. [14] G.A. Vedage, R. Pitchai, R.G. Herman, K. Klier, Amer. Chem. Soc. Div. Fuel Chem. Prepr. Paper 29 (1984) 196. [15] G.C. Chinchen, K. Mansfield, M.S. Spencer, CHEMTECH 20 (1990) 692. [16] K.C. Waugh, Catal. Today 15 (1992) 51.

[17] J.F. Deng, Q. Sun, Y.L. Zhang, S.Y. Chen, D. Wu, Appl. Catal. A 139 (1996) 75. [18] Q. Sun, Y.L. Zhang, H.Y. Chen, J.F. Deng, D. Wu, S.Y. Chen, J. Catal. 167 (1997) 92. [19] Y.L. Zhang, Q. Sun, J.F. Deng, D. Wu, S.Y. Chen, Appl. Catal. A 158 (1997) 105. [20] N. Sheppard, T.T. Nguyen, in: R.J.H. Clark, R.E. Hester (Eds.), Advances in Infrared and Raman Spectroscopy, Heydon, London, 1978. [21] M. Bowker, R.A. Hadden, H. Houghton, J.N.K. Hyland, K.C. Waugh, J. Catal. 109 (1988) 263. [22] R.A. Hadden, H.D. Vandervell, K.C. Waugh, G. Webb, Catal. Lett. 1 (1988) 27. [23] S. Fujita, M. Usui, H. Ito, N. Takezawa, J. Catal. 157 (1995) 403.