Ind. Eng.
50
Chem. Prod. Res. Dev. 1985,2 4 , 50-55
Synthesis of Methanol from Carbon Monoxide and Hydrogen over a Copper-Zinc Oxide-Alumina Catalyst Rajesh M. Agnyt and Chrlstos G. Takoudis' School of Chemical Engineering, Purdue University, West La fayette, Indiana 47907
The kinetics of methanol synthesis from carbon monoxide and hydrogen over a commercial Cu/ZnO/A120, has been investigated at temperatures between 250 and 290 OC and pressures between 3 and 15 atm. At the reaction conditions studied, methanol is the main product. A two-parameter kinetic model that predicts observed patterns quantitatively is presented. Based on this study, it is proposed that the formyl species, CHO-, appears to be the most abundant surface intermediate, and the ratedetermining step seems to be the surface reaction between a methoxy intermediate and an adsorbed hydrogen atom. A mechanism for the methanol synthesis is also proposed. The main feature of this mechanism is the shift from a carbon-surface bond to an oxygen-surface bond during
the reaction.
Introduction Among the main industrial organic reactions, the synthesis of methanol is an outstanding example of the practical importance of catalytic processes. It stands out with its high selectivity, long term performance of the catalysts, and lack of catalyst poisoning by oxidizing gases (Herman et al., 1979). The modern low-pressure methanol synthesis catalysts are usually based on Cu/ZnO/Cr203 or Cu/ZnO/Al,03. These mixed oxide catalysts, produced by coprecipitation methods, have been found to be considerably more active than the individual components (Herman et al., 1979; Henrici-Oliv6 and Oliv6, 1984). Most kinetic studies of methanol synthesis over Cu/ZnO catalysts have been carried out using mixtures of carbon monoxide and hydrogen, but in some cases (Klier et al., 1982) the synthesis gas has also contained carbon dioxide. Power rate laws as well as complicated rate expressions have been proposed over the years. The detailed kinetics of the surface processes is virtually unknown. With the exception, perhaps, of the near-atmospheric work by Saida and Ozaki (1964), investigations of the kinetics of methanol synthesis over Cu/ZnO catalysts have been done a t pressures on the order of 75 atm (Herman et al., 1979; Klier et al., 1982; Shimomura et al., 1978; Leonov et al., 1973; Rozovskii et al., 1975; Kagan et al., 1975). A major drawback of several such kinetic studies is the use of integral reactors. Due to the exponential effect of temperature on reaction rate as described by the Arrhenius law, small. deviations from isothermality can lead to large errors in measured rates for reactions with large heat effects. In the present work, we study the kinetics of methanol synthesis over a commercial Cu/ZnO/A1203 catalyst in a differential reactor a t pressures in the range from 3 to 15 atm and temperatures between 250 and 290 "C (Agny, 1984). Experimental Section A high-pressure unit is used for the kinetic experiments (Figures 1 and 2). Hydrogen (UHP, 99.999% pure) and carbon monoxide (CP, 99.5% pure) were supplied by Matheson Gas Products. Hydrogen is led through a Matheson oxygen purifier OR-25 followed by a gas drier and a Matheson moisture indicator 465. A gas drier filled with molecular sieves and a moisture indicator are also 'Harris Semiconductor, P.O. Box 883, MS 06-OO0, Melbourne,
FL 32901.
present in the carbon monoxide line. The flow rates of the gases are controlled with Hoke metering valves (Operator, 0121L2D; Valve, 1315G4Y). Immediately downstream are Matheson mass flow transducers 8142 whose output is displayed on a coupled Matheson mass flow indicator 8142. The gases are then led through Nupro F series in-line filters and Matheson check valves 61841'4 to a Hoke three-way ball valve (Operator, OllY6F; Valve, 7673G4Y). The three-way ball valve directs the gaseous mixture either to the reactor or to the bypass. The reactor system (Figure 2) consists of a U-tube made of stainless steel 304,0.95 cm 0.d. and 0.09 cm thick. The U-tube is suspended vertically in a Techne fluidized sand bath SBS-4. In the direction of flow, the U-tube contains a length of glass wool, a preheating section of Potters silica glass spheres P-047, a second length of glass wool to isolate the catalyst bed from the glass spheres, and a final glass wool plug above the catalyst bed to prevent carry-out of the catalyst (Agny, 1984). An Omega iron-constantan thermocouple probe TT-J-30 positioned in the bath near the catalyst bed in the U-tube measures the bath temperature. Independent experiments reveal that (i) the difference between the bath temperature and the temperature inside the fixed bed reactor is less than 0.5 "C, and (ii) the catalytic fixed bed we use is isothermal (Agny, 1984). The reactor system is pressurized by a Grove back pressure regulator 91W. The reactor effluent is led to a Carle sampling valve 7707. During analysis, the effluent sample is carried by the carrier gas (hydrogen) to a Varian Aerograph Gas Chromatograph 202-B. A Chromosorb 102 column permits separation of CO, C02, and methanol. The gas chromatograph is coupled with a Houston Instrument chart recorder A521X-14K to enable qualitative and quantitative analysis of the effluent. Based on gas chromatographic analysis of CO bubbled through methanol at different temperatures, it is concluded that the weight percent of the components in the reactor effluent can be estimated by using Dietz's thermal conductivity weight factors (1967). The retention times of COz,HzO, CO, and CH30H were established separately for each component in the gas chromatograph. Typically, a 5-g Cu/ZnO/A1203catalyst sample (C792-01 of United Catalysts, Inc., Kentucky) of average size equal to 0.077 cm is reduced with hydrogen at 1atm and 290 "C for 15 h (Agny, 1984). No further pretreatment is done during the rest of the experimental span. A reaction interval of 1 h is found to be sufficient to produce repro-
0196-4321/85/ 1224-OO5O$01.50/0 0 1985 American Chemical Society
Ind. Eng. Chem. Prod. Res. Dev., Vol. 24, No. 1, 1985 51
Two Stage Regula t o r
Check Valve
Moisture Indicator
- --I
--o--w-t
:' ?d To By ass
fJ
Mass Flow Indicator Indicator
URn i d i r e c t i o n a l On/off Valve
N2
1
I
3-wa
Valve
I
L--
To
1 1.
Needle '1. Met. Valve1
Oxygen Oxygen Purifier
HZ
Reactor System
S c = > m M ---I; -u-+--o--cl-o -1 Single
K?--+
Stage Regulator
F
Transducer Transducer
Figure 1. R e a c t a n t s d e l i v e r y system.
\
Sys tern
fl
3-way Valve
"
Sand Bath
A B
U tube Reactor __
To Sampling Valve and Gas Chromatograph
E
D Flow C o n t r o l
C
Contaminant
On-Off House Air
\
W
L i n e Regulator
Figure 2. R e a c t o r system. Table I. Composition of Cu/ZnO/A120, Catalyst (39-2-01 of United Catalysts, Inc., on an Oxygen-Free Basis element
cu Zn
A1
wt
70
40.6 50.3 9.1
atom %
36.5 44.1 19.4
ducible steady-state product distributions. Steady-state reaction data are collected at temperatures between 250 and 290 "C, pressures in the range of 3 to 15 atm, H2/C0 = 2.1-2.4, and total flow rates between 75 and 115 cm3/min (STP). The composition of the Cu/ZnO/A1203 catalyst measured with a Dispersive Analyzer is presented in Table I. The specific surface area of this catalyst, determined by BET with N2, is 37.8 m2/g. Results The Cu/ZnO/A1203catalyst is tested at three temperatures, 250,270, and 290 "C. Steady-state conversions are reached within 1h after the pressure, temperature, gas flow rate, and feed gas composition are established. Typical measured steady-state conversions are summarized in Table I1 along with the conditions used. No deactivation
of the catalyst is observed at the experimental conditions studied, based on periodic tests conducted at "standard conditions" which were chosen to be 11.2 atm, 250 "C and H2/C0 ratio equal to 2.3. As a consequence of a rigorous design of our differential reactor, heat and mass transfer effects are negligible (Agny, 1984). Methanol is the main product at the conditions studied. Carbon dioxide and water (in trace amounts) are formed as side products. No other byproducts are observed at the conditions used. The effect of temperature on the reaction rate is shown in Figure 3. Error bars are shown at each data point to illustrate the range of possible reaction rates due to the maximum relative experimental error of 15%. It is observed that the reaction rate increases as the temperature increases in the temperature range 250 to 290 "C. Data reported in Table I1 demonstrate that the reaction rate increases as the total pressure increases at constant temperature and H2/C0 ratio. Table I11 facilitates comparison of the measured CO conversions to methanol with the equilibrium ones. In contrast to measurements at 250 and 270 "C, conversions obtained at 290 "C and in the pressure range of 3 to 15
Ind. Eng. Chem. Prod. Res. Dev., Vol. 24, No. 1, 1985
52
Table 11. Steady-State Data of Methanol Synthesis over Cu/ZnO/Ala03Catalyst C79-2-01 of United Catalysts, Inc.
run no. 1 2 3
reactor press., atm 14.3 12.6 11.2 9.4 7.7 6.2 14.2 11.0 9.5 7.8 6.2 3.4 14.3 11.2 9.5 7.7 6.2 3.5
reactor temp, "C 250 250 250 250 250 250 270 270 270 270 270 270 290 290 290 290 290 290
4
5 6 7 8 9 10 11 12 13 14 15 16 17 18
flow rate at
STP, H2/C0 2.4 2.4 2.2 2.2 2.1 2.2 2.3 2.2 2.4 2.2 2.2 2.1 2.4 2.3 2.2 2.1 2.2 2.2
conv of CO to methanol,
cm3/min 102 100 77 79 83 79 104 112 101 100 103 106 101 116 100 103 97 99
%
co
0.27 0.24 0.22 0.19 0.15 0.13 0.91 0.61 0.54 0.42 0.32 0.16 1.35 0.96 0.81 0.63 0.48 0.22
29.4 28.9 31.1 31.6 32.5 31.6 29.4 30.9 29.5 30.9 30.9 32.0 29.0 29.8 30.6 31.4 30.7 31.3
reactor effluent compn, % CHSOH CO2 70.5 0.08 a 0.07 a 70.9 0.07 a 68.8 68.3 0.06 a 67.4 0.05 a 68.3 0.04 a 70.0 0.26 0.27 68.8 0.18 0.09 70.2 0.16 0.07 68.9 0.13 0.07 68.9 0.11 0.06 67.9 0.05 a 70.0 0.40 0.51 69.6 0.30 0.26 68.8 0.27 0.27 68.1 0.20 0.24 69.0 0.15 0.14 68.6 0.07 a
H2
H20 a a a
a a a a a
a a a
a a a a a
a a
Trace amounts.
Table 111. Measured and Equilibrium CO Conversions to Methanol (YO) 250 "C 270 "C run no. 1 2
3 4
5 6
measd 0.27 0.24 0.22 0.19 0.15 0.13
equil 18.0 15.0 12.0 9.0 6.0 4.0
---
--
run no. 7 8 9 10 11 12 7
I
measd 0.91 0.61 0.54 0.42 0.32 0.16
equil 9.0 5.0 4.0 3.0 2.0 0.6
run no. 13 14 15 16 17 18
,
01 1.70
I .EO
: I - , iii+
I
I
1.90
x
2.00
2 .IC
equil 4.6 2.8 2.0 1.3 0.9 0.3
followed. A statistical package involving linear and nonlinear regression was used to compare the rates predicted by an assumed model with the measured reaction rates (Agny, 1984). Since evidence for the formyl (CHO-) and the methoxy (CHBO-)species has been reported for Cu/ZnO catalysts during methanol synthesis (Sawsey et al., 1982; Griffin and Roberts, 1982), two-parameter and three-parameter kinetic models which can be derived from Langmuir-Hinshelwood type rate expressions were tried. These contained driving force groups representing a bimolecular surface reaction between an adsorbed hydrogen atom and either a methoxy or a formyl adsorbed species. The form of such rate expressions examined is
r = k ( PcOph2! .60
290 "C measd 1.35 0.96 0.81 0.63 0.48 0.22
e/(.)"
and
103
Figure 3. Temperature dependence of measured methanol formation rates.
atm are one-third to one-half of their equilibrium counterparts. Therefore, at 290 OC, the reverse reaction becomes important. Since no specific experiments were conducted from the standpoint of mechanistic considerations, no additional information about surface intermediates was obtained. Starting with power rate laws, several rate expressions were tried to fit our experimental data. Emphasis was laid, during this data fitting, on rate expressions containing a small number of parameters. The reaction rate was found to be approximately proportional to the total pressure. This enabled us to eliminate several rate expressions if the observed dependence of rate on total pressure was not
respectively, where X = Pco, pi:,P c p f i 5 ,PcOpH2, and PcOph:. These forms of X correspond to the assumption that each of the following species may be the most abundant surface intermediate: molecularly adsorbed CO, dissociatively adsorbed hydrogen, and surface intermediates with H2/C0 ratio in the adsorbate of 0.5, 1.0, and 1.5, respectively. For the three-parameter models, two or more of these species are adsorbed in significant quantities. It is found that the best possible fit is obtained with the rate expression (Agny, 1984)
Ind. Eng. Chem. Prod. Res. Dev., Vol. 24, No. 1, 1985 53 17.60
17.26
18.32
18.68
19.0q
19.VO
-iB.SO
-18.50
X
/
I
I
/
-19.50 7
= I
El
/A
Figure 4. Arrhenius plot for the methanol synthesis (eq 3). 0
Table IV. Preexponential Factors and Overall Activation Energies for Methanol Synthesis over Cu/ZnO/AI2Oa Catalyst C79-2-01 of United Catalysts, Inc.
temp, OC 250 270 290
preexp factor, g-mol/(g of cat. s atm) 13600 1890 48
overall act. energy, cal/g-mol
34000 f lo00 29500 f 1000 25500 i 1000
where k = k , exp(-E/RT) and n = -1.30 f 0.03. The Arrhenius plot of the temperature dependence of k is presented in Figure 4. With a single preexponential factor and a single overall activation energy obtained from linear regression of the rate constant data, large deviations (30-40%) are observed between predicted and measured reaction rates. Therefore, it is propesed that there is a decrease in the overall activation wfergy with temperature (Figure 4). The overall activatiton energies and the corresponding preexponential facPors computed at each temperature are presented in TabPe IV (Agny, 1984). The error accompanying the estimation of the activation energy is about f l O O O cal/g-mol based on the error involved in calculating the activation anergy from the Arrhenius plot and the 15% maximum relative error in observed reaction rates. Equation 3 fits all OW data with a maximum relative error of less than 15% (Figure 5). Discussion Features of the Proposed Kinetic Model. From the structure of rate expressibn 3 we note the following. (a) The kinetics of the reverse reaction is accounted for. A necessary constraint on a kinetic expression intended to cover a range of methanol concentrations is that the net rate should become zero at equilibrium. (b) This rate expression coritains only two parameters, the rate constant and an empirical constant, n. Recall that most of the kinetic models proposed for methanol synthesis (Klier et al., 1982; Saida and Ozaki, 1964; Leonov et al., 1973) invariably contain a large number of parameters; thus, experimental data can usually be fitted by more than one model. (c) It is evident that the reaction rate depends on a positive power of partial pressure of hydrogen and on a negative power of the partial pressure of carbon monoxide. This type of dependence of the reaction rate on the partial pressure of CO and H2is also evident from the rate expression proposed by Klier et al. (1982) for methanol
1
a
2
4
7
E
6
MEASURED REACTION RATE. [CMOLIG CRT/SECI X 108 Figure 5. Comparison between predicted reaction rates (eq 3) and measured reaction rates.
synthesis over Cu/ZnO catalysts at 75 atm and 225 to 250 “C. (d) The activation energy, E , which represents an algebraic sum of heats of adsorption of reactants, heats of surface reactions, and the activation energy of the ratedetermining step, decreases with increasing temperature (Table IV). Such a behavior is expected for a typical Langmuir-Hinshelwood rate expression (Table IV). Note that experiments including methanol in the feed stream were not carried out (Agny 1984). Mechanistic Considerations. The term P C & ! rep~; resents a formyl species, CHO-, formed by a reactive collision between CO and an adsorbed hydrogen atom. Mechanisms A and B (Herman et al., 1979; Kung, 1980) shown below, involve the formyl species as one of the surface intermediates
!
O H
C
II O I1 x
~
H
\/
O
C c
*I
1~
H
H
\c/
I, \
O H H
H
7
II*
H
/
C-OH
*1
CH,OH
*
”I e t
*
(A)
b
I
I
V
where * and v represent a surface site and an oxygen vacancy, respectively. Assuming the formyl species to be the most abundant surface intermediate and the last reaction step of both mechanisms to be rate determining, a rate expression of the following form can be derived (Agny, 1984)
r=
\
(1
+ KPcopRy
(4)
A comparison of eq 3 and 4 suggests that eq 3 may be
54
Ind. Eng. Chem. Prod. Res. Dev., Vol. 24, No. 1, 1985
derived from eq 4 by using the approximation (Vannice, 1975) (1 KX) ?I! axn (0 In I1)
+
where a is only temperature dependent. In mechanism A, the intermediates are proposed to be bonded to the surface cation via the carbon atom (Herman et al., 1979). Such a proposal is feasible for noble-metal catalysts because relatively strong metal carbon bonds can be formed, and there is ample evidence that CO adsorbs molecularly on these metals with the carbon end pointing toward the metal (Kung, 1980). However, it is proposed that the large dipole moments and the relatively uneven electron density distribution on oxide surface favor bonding of the surface to the oxygen atom of the intermediate (Kung, 1980). Thus a methoxy type derivative should be more likely. In mechanism B (Kung, 1980), the formation of the methoxy intermediate from the formyl species involves rupture of two bonds and formation of three bonds. An elementary step is expected not to involve multiple bond ruptures and formations. This proposed elementary step of mechanism B must therefore be looked upon cautiously in the absence of independent evidence that supports such a step. The indication of the formyl species from the kinetics observed and the examination of the shortcomings of mechanisms A and B allow us to suggest the mechanism (Agny, 1984)
i I H
CO
+ CI /
/I
$ 'C'
H
H
I 8
Q I 9
I
e
I
p H
9
0 ;t CH30H
f 8
I I
V
for the supported Cu/ZnO catalyst we used. Here the symbols *, 0, and v represent a metal cation, a zinc or an oxygen site, and an oxygen vacancy, respectively. I represents a likely intermediate between the formyl and methoxy species (Agny, 1984). Its possible structure needs to be investigated. Formaldehyde bonded to the surface as shown H
O&H
* is a possible substitute for I. Rate equation 4 can be derived from this mechanism by assuming that the formyl species is the most abundant surface intermediate and #at the surface reaction between the methoxy species and an adsorbed hydrogen atom is the rate-determining step (Agny, 1984). The main difference between this mechanism and mechanism B is the addition of an intermediate I between the formyl and methoxy intermediates. The general belief regarding CO and H2 adsorption on Cu/ZnO catalysts is that CO is molecularly adsorbed on a metal cation (predominantly Cu(1)) and hydrogen is dissociatively adsorbed at the Zn-0 pair sites. It is suggested that the presence of Cu(1) in ZnO also greatly increases the density of oxygen vacancies as is required by charge balance when Cu(1) exists as a substitutional impurity for Zn(I1) (Kung, 1980). Recently, coadsorption of CO and H2 on ZnO and Cu/ZnO catalysts has been investigated at 270 K and subatmospheric pressures by Saussey et al. (1982). The infrared spectrum of a mixture of CO and H2adsorbed on ZnO or Cu/ZnO reveals a pair of weak bands which have been assigned to vibrations of a surface formyl species.
This is, to our knowledge, the first report presenting direct evidence for a formyl species resulting from CO and H2 interaction on Cu/ZnO catalysts. Deluzarche et al. (1978) claim to have identified, by reactive scavenging, the formyl species on nickel. Kung (1980) presents certain organometallic chemistry analogues to substantiate the formyl and methoxy species as intermediates in the mechanism B for methanol synthesis on ZnO and Cu/ZnO catalysts. These findings support our suggestion that adsorbed formyl species can be the most abundant surface intermediate (Agny, 1984), although further studies are necessary for such a proposal. Formaldehyde has never been observed as an intermediate in methanol synthesis over ZnO or Cu/ZnO catalysts. A plausible explanation, perhaps, is that formaldehyde being a highly reactive intermediate undergoes rapid conversion in some species like the methoxy. Saida and Ozaki (1964), who have studied the kinetics of methanol synthesis over a Cu/ZnO/Cr203 catalyst at near atmospheric pressures, conclude that quantitative interpretation of their data is possible only by assuming a reaction sequence involving formaldehyde. In a very recent work, Tawarah and Hansen (1984) have investigated the kinetics of methanol decomposition over ZnO in the temperature range 563 to 613 K. Based on their observations, they proposed the following reaction sequence CH3OH + CH30H* CH30H* + * F= CH30* + H* CH30* + H* + HCHO* + H2 * HCHO* + * CHO* H* CHO* + H* F= H2 CO + 2*
+
+
+
The symbol * represents a surface site. They confirm the presence of formaldehyde in the reactor effluent with a mass spectrometer and with chemical methods. Invoking the principle of microscopic reversibility, it may be concluded that some of these intermediates participate in the methanol synthesis reaction. Ample evidence exists for the methoxy species on ZnO and ZnO/Cr203catalyst (Deluzarche et al., 1977; Nagarjunan et al., 1963; Borowitz, 1969; Ueno et al., 1971). The zinc cation is reported to play the active role in ZnO catalysts (Saida and Ozaki, 1964; Kortuem and Knehr, 1974). In Cu/ZnO catalysts, the active site has been suggested to be Cu(1) dissolved in ZnO (Herman et al., 1979). Since the active forms appear to be the metal cations for both ZnO and Cu/ZnO catalysts, it seems reasonable to assume the formation of methoxy species on the Cu/ZnO catalysts. Recently, the methoxy species has also been observed on a Cu/ZnO catalyst during methanol synthesis by Griffin and Roberts (1982). Tsuchiya and Shiba (1965) have studied methanol synthesis over Cu/ ZnO/Cr203catalysts at atmospheric pressure and 250 "C. They report enhanced adsorption of CO and Hz in the presence of each other and also find the average composition of the adsorbate, H2(a)/CO(a)= 1.5. This indicates the most likely presence of the methoxy species, CH30 on the surface. For Cu/ZnO catalysts it is proposed that the intermediates, in the course of their progress toward the product methanol, orient themselves to form the more favorable oxygen-surface bond (Kung, 1980). These arguments lend support to the inclusion of a methoxy intermediate in the mechanism proposed for methanol synthesis over a Cu/ZnO/A1203catalyst. Formation of COP There can be considered to be three possible ways of C 0 2 formation 2co === cos + c (1)
Ind. Eng. Chem. Prod. Res. Dev., Vol. 24, No. 1, 1985 55
CO
+ H2O + C02 + H2
(11)
and
(111) A , , + CO + Ared, + COZ where A,, and Ared.represent the oxidized and reduced forms of the catalyst, respectively. In our reaction system, the carbon and oxygen balances based on CO in the feed and CO, C02,and CH30H in the reactor effluent are always closed. This enables us to suggest that reaction I can be eliminated as a source of CO, (Agny, 1984). Van Herwijnen and Dejong (1980) have investigated the kinetics and mechanism of reaction I1 on a Cu/ZnO catalyst in the range of 172 to 230 OC and pressures ranging from 1to 6 atm. With a feed mixture containing CO, N2, and small amounts of H20,they find their Cu/ZnO catalyst to be active in the forward direction of reaction 11. This finding allows us to suggest that at our experimental conditions (250-290 "C, 3-15 atm), reaction I1 is likely. Herman et al. (1979) have tested a Cu/ZnO/A1203 catalyst for a H2/C0 (76/24) mixture without C02. Although this catalyst is found to be selective to methanol, it is rapidly and irreversibly deactivated. Their examination of the catalyst by optical spectroscopy reveals only the pink color that is characteristic of copper. They attribute this to reaction 111, where A,=, represents their suggested active site-Cu(1) dissolved in ZnO. In view of these arguments and lack of any spectroscopical analysis of our catalyst, both reactions I1 and I11 are probably responsible for the formation of CO, at the conditions of our experiments. Conclusion The kinetics of methanol synthesis from carbon monoxide and hydrogen over a commercial Cu/ZnO/A1203 catalyst has been investigated at temperatures between 250 and 290 OC and pressures between 3 and 15 atm. The highest catalytic activity is observed at 290 "C. A twoparameter kinetic model is presented. It quantitatively describes the observed patterns for the entire range of the experimental conditions studied. Furthermore, it is proposed that the formyl species, CHO-, appears to be the most abundant surface intermediate, and the rate-determining step seems to be the surface reaction between the methoxy intermediate and an adsorbed hydrogen atom. Based on this study and current knowledge of supported Cu/ZnO catalysts, a mechanism for the methanol synthesis is proposed. The main feature of this mechanism is the shift from a carbon-surface bond to an oxygen-surface bond during the course of reaction. Finally, it is suggested that
carbon dioxide is formed by two reactions, the water gas shift reaction and a redox equilibrium between the oxidized and reduced forms of the catalyst. Acknowledgment Support was provided by NALCO and CONOCO Companies t h r o u p Coal Research Center, Purdue University. The- thors are grateful to R. Diffenbach of the Department of Energy for providing the catalyst. Nomenclature E = overall activation energy, cal/g-mol K = reaction rate constant, g-mol/(g of cat. s atm) K O = preexponential factor, g-mol/(g of cat. s atm) Kq = equilibrium constant of the methanol synthesis reaction n = empirical constant r = reaction rate, g-mol/(g of cat. e ) R = universal gas constant, cal/(g-mol K) T = temperature of reaction, K Registry No. CO, 630-08-0; Cu, 7440-50-8; ZnO, 1314-13-2; methanol, 67-56-1. Literature Cited Agny, R. M. M.S. Thesis, Purdue University, 1984. Borowitz, J. L. J . Catal. 1969, 73, 106. Deiuzarche, A.; Kieffer, R.; Muth, A. Tetrahedron Lett. 1977, 3 8 , 3357. Deluzarche, A.; Hinderman, J. P.; Kieffer, R. Tetrahedron Lett. 1978, 39, 2787. Dietz, W. A. J . Gas Chromatogr. 1967, 5 , 68. Griffin, G. L.; Roberts, D.L. "Muiti-component Adsorbate Interactions of Cu/ ZnO Methanol Synthesis Catalysts", paper presented in the 75th Annual Meeting AIChE, Los Angeles, CA, Nov 14-18, 1982. Henrici-Oiive, G.; Olive, S. "Catalyzed Hydrogenation of Carbon Monoxide"; Springer-Veriag: New York, 1984. Herman, R. G.; Kiier, K.; Simmons, G. W.; Finn, B. P.;Buiko. J. 8. J . Catal. 1979, 56, 407. Kagan, Y. B.; Rozovskii, A. Y.; Liberov, L. W.; Siivinskii, E. V.; Lin, G. I.; Loktev, S. M.; Bashkirov, A. N. Dokl. Akad. Nauk SSSR 1975, 224, 1081. Kiier, K.; Chatikavanij, V.; Herman, R. G.; Simmons, G. W. J . Catal. 1982. 7 4 , 343. Kortuem. G.; Knehr, H. Z . phvs. Chem. (Frankfurtam Main) 1974, 89, 194. Kung, H. H. Catal. Rev. Scl. Eng. 1980, 22, 235. Leonov, V. E.; Karabaev, M. M.; Tsybina, E. N.; Petrishcheva, G. S. Klnet. Katal. 1973, 14, 970. Nagarjunan, T. S.; Sastri, M. V. C.; Kuriacose, J. C. J . Catal. 1963, 2, 223. Rozovskii, A. Y.; Kagan, Y. 6.; Lin, G. I.; Siivlnskii, E. V.; Loktev, S. M.; Liberov, L. G.; Bashkirov, A. N. Kinet. Katal. 1975, 76, 810. Saida, T.; Ozaki, A. Bull. Chem. SOC.Jpn. 1964, 37, 1817. Saussey, J.; Lavalley, J.; Lamotte, J.; Rais, T. J . Chem. SOC. Chem. Commun. 1982, 278. Shimomura, K.; Ogawa, K.; Oba, M.; Kotera, Y. J . Catal. 1978, 5 2 , 191. Tawarah, K. M.; Hansen, R. S. J . Catal. 1964, 8 7 , 305. Truchiya, S.; Shiba, T. Bull. Chem. SOC.Jpn. 1985, 38, 1726. Ueno, A.; Onishi, T.; Tamaru, K. Trans. Faraday SOC. 1971, 6 7 , 3585. Van Herwijnen, T.; Dejong, W. A. J . Catal. 1980, 63, 83. Vannice, M. A. J . Cats/. 1975, 37, 462.
Received for review November 19, 1984 Accepted December 8, 1984
