Isobutanol and methanol synthesis on copper catalysts supported on modified magnesium oxide

Article abstract of DOI:10.1006/jcat.1997.1777

Alcohols are selectively produced from CO/H₂ on K-CuMgCeO_x catalysts, but synthesis rates are strongly inhibited by CO₂ formed during reaction. Reaction pathways involve methanol synthesis on Cu, chain growth to C₂₊ alcohols, and metal-base bifunctional coupling of alcohols to form isobutanol. Ethanol reactions on K-Cu_0.5Mg₅CeO_x show that Cu catalyzes both alcohol dehydrogenation and aldol condensation reactions. CeO₂ increases Cu dispersion and MgO surface area and K decreases Cu dispersion, but increases the density of basic sites. Reactions of acetaldehyde and ¹³C-labeled methanol lead to 1-¹³C-propionaldehyde, a precursor to isobutanol. The density and strength of basic sites were measured using a ¹²CO₂/¹³CO₂ isotopic jump method that probes the number and chemical properties of basic sites available at typical isobutanol synthesis temperatures. K or CeO₂ addition to CuMgO_x increases the density and strength of basic sites and the rates of base-catalyzed ethanol condensation reactions leading to acetone and n-butyraldehyde. The presence of CO in the He carrier during temperature-programmed surface reactions of ethanol preadsorbed on Cu_0.5Mg₅CeO_x decreases the rate of base-catalyzed condensation reactions of preadsorbed ethanol, possibly due to the poisoning of basic and Cu sites by the CO₂ formed from CO via water-gas shift reactions.

Full text of DOI:10.1006/jcat.1997.1777

JOURNAL OF CATALYSIS 171, 130–147 (1997)
ARTICLE NO. CA971777
Isobutanol and Methanol Synthesis on Copper Catalysts Supported
on Modified Magnesium Oxide
Mingting Xu, Marcelo J. L. Gines, Anne-Mette Hilmen, Brandy L. Stephens, and Enrique Iglesia1
Department of Chemical Engineering, University of California at Berkeley, Berkeley, California 94710
Received January 27, 1997; revised June 9, 1997; accepted June 10, 1997
for 2-methyl-1-propanol (isobutanol) because of its steric
–
hindrance and the lack of the two -hydrogens required
for aldol condensation reactions. Therefore, isobutanol be-
comes a preferred end product of alcohol chain growth re-
actions on basic oxides. Isobutanol synthesis appears to be
limited by the rate of ethanol formation from synthesis gas
(4, 6), which limits the concentration of the aldehydic inter-
mediates required for chain growth. High reactor pressures
favor the synthesis of higher alcohols and bimolecular con-
densation reactions during CO hydrogenation, but lead to
unfavorable thermodynamics for alcohol dehydrogenation
steps that may be required in order to form aldehyde inter-
mediates required in aldol coupling pathways.
Our study addresses the reaction pathways and catalytic
site requirements for the synthesis of branched alcohols
on (metal–base) bifunctional catalysts at low temperatures.
These bifunctional catalysts consist of Cu metal crystallites
supported on high-surface-area basic oxides based on mod-
ified MgO. Specifically, we describe the synthesis proce-
dures, catalyst structural characterization data, and surface
and catalytic properties for a class of materials based on
Alcohols are selectively produced from CO/H2 on K CuMgCeOx
catalysts, but synthesis rates are strongly inhibited by CO2 formed
during reaction. Reaction pathways involve methanol synthesis
on Cu, chain growth to C2+ alcohols, and metal–base bifunc-
tional coupling of alcohols to form isobutanol. Ethanol reactions
–
on K Cu0.5Mg5CeOx show that Cu catalyzes both alcohol dehydro-
genation and aldol condensation reactions. CeO2 increases Cu dis-
persion and MgO surface area and K decreases Cu dispersion, but
increases the density of basic sites. Reactions of acetaldehyde and
13C-labeled methanol lead to 1-13C-propionaldehyde, a precursor
to isobutanol. The density and strength of basic sites were mea-
sured using a 12CO2/13CO2 isotopic jump method that probes the
number and chemical properties of basic sites available at typical
isobutanol synthesis temperatures. K or CeO2 addition to CuMgOx
increases the density and strength of basic sites and the rates of
base-catalyzed ethanol condensation reactions leading to acetone
and n-butyraldehyde. The presence of CO in the He carrier during
temperature-programmed surface reactions of ethanol preadsorbed
on Cu0.5Mg5CeOx decreases the rate of base-catalyzed condensation
reactions of preadsorbed ethanol, possibly due to the poisoning of
basic and Cu sites by the CO2 formed from CO via water–gas shift
c
reactions.
1997 Academic Press
–
K-modified CuMgCeOx (6–8). These K CuMgCeOx cata-
lysts are among the best reported catalysts for alcohol
synthesis at low temperatures and pressures (6–8). MgO-
based solids also catalyze condensation reactions of al-
cohols and aldehydes with high activity and selectivity
(9, 10).
CuMgCeOx samples were prepared by controlled pH
coprecipitation methods designed to maximize the num-
ber of Cu surface atoms and basic sites (8). Site titration
and isotopic switch methods were used to determine the
density and chemical properties of exposed Cu and ba-
sic sites. Isobutanol and linear alcohol synthesis pathways
were studied using the adsorption and surface reactions
of alcohols and aldehydes in steady-state catalytic reac-
tors and in a temperature-programmed surface reaction
(TPSR) apparatus. The role of individual elements in multi-
component catalysts was explored by varying the composi-
INTRODUCTION
Isobutanol is a potential fuel additive and a precursor
to isobutene and methyl-tert-butyl ether. Currently avail-
able catalysts for the synthesis of isobutanol from CO/H2
mixtures require high temperatures (750–800 K) and pres-
sures (10–30 MPa) and produce isobutanol with relatively
low productivity and selectivity under milder reaction con-
ditions (1–5). Typical isobutanol yields reported in the lit-
erature are shown in Table 1.
Isobutanol synthesis requires the initial formation of
methanol and higher linear alcohols and their subsequent
chain growth processes leading to 2-methyl-branched alco-
hols with high selectivity (3). Chain growth rates are low
–
tion of K CuMgCeOx catalysts and determining the effect
of each component on surface properties and on reaction
steps.
1 To whom correspondence should be addressed. E-mail: iglesia@
cchem.berkeley.edu. Fax: (510) 642-4778.
130
0021-9517/97 $25.00
c
Copyright
1997 by Academic Press
All rights of reproduction in any form reserved.

ISOBUTANOL AND METHANOL SYNTHESIS
131
low-Cu catalysts in order to get the same amount of CuO
charge in the reactor. After the temperature was lowered to
300 K, a 5% H2/He mixture (100 cm3/min) was introduced
into the reactor and the temperature was then increased
to 623 K at a rate of 0.17 K/s. The evolution of H2O and
the consumption of H2 during reduction were followed by
on-line mass spectrometry (H200M Gas Analyzer, Leybold
Inficon Inc.).
TABLE 1
Representative Isobutanol Synthesis Catalysts (6, 7)
GHSV
Isobutanol
yield (g/kgcat/h)
1
Catalyst
T (K) P (MPa) CO/H2 (h
)
–
–
–
–
–
–
–
–
Cu Zn Cr Cs
583
593
9.1
5.1
5.1
10.0
32.5
25.0
1.0/0.45
1.0/1.0
1.0/1.0
0.5/1.0
1.0/2.2
1.0/1.0
5330
1832
1832
10000
15000
20000
51
7.2
–
–
Cu Mg Ce K
–
–
–
Pd Zr Zn Mn Li 593
1.1
–
–
Cu Zn Mn K
643
693
175a
125a
740a
–
Zn Cr K
–
–
–
Pd Zr Zn Mn Li 700
Site Density and Reactivity
a g/litercat/h.
Cu surface areas were measured by titrating Cu sur-
face atoms in prereduced samples with N2O (Matheson,
ultra high purity) at 363 K. N2O was introduced as pulses
(1.60 mol N2O/pulse) into a He stream. The number of
chemisorbed O atoms was obtained from the amount of
N2O consumed and the N2 produced. A saturation cover-
age of 0.5 : 1 O : Cus was used to estimate Cu surface area
and dispersion (13, 14). Cu dispersion is defined as the ratio
of surface Cu (Cus) to the total number of Cu atoms in the
catalyst.
In temperature-programmed CO2 desorption (TPD)
measurements, samples (50 mg) were first pretreated in
flowing He ( 100 cm3/min) at 723 K for 0.3 h and then
reduced in 5% H2/He at 623 K for 0.5 h. After exposure
to CO2 (Cambridge Isotope Laboratories, Inc.) for 0.15 h
at room temperature, the samples were flushed with He to
remove gas phase and weakly adsorbed CO2. The temper-
ature was then increased up to 723 K at a rate of 0.5 K/s,
and the desorption profile of CO2 was monitored by mass
spectrometry.
EXPERIMENTAL METHODS
Catalyst Synthesis
CuMgCeOx samples were prepared by coprecipitation
of mixed nitrate solutions with an aqueous solution of
KOH (2 M) and K2CO3 (1 M) at 338 K and a constant pH
of 9 in a well-stirred thermostated container. The solids
formed were filtered, washed thoroughly with 300–500 cm3
of deionized water at 333 K, and dried at 353 K overnight.
Dried samples were treated in flowing air at 723 K for 4 h
in order to form the corresponding mixed metal oxides.
Residual potassium in all dry samples was less than 0.1 wt.%
before alkali impregnation. The controlled pH coprecipita-
tion technique produces uniform precipitates, which lead,
after thermal treatment, to mixed metal oxides with high
surface area and intimate contact between components (8).
CuZnAlOx samples were prepared using the same proce-
dures as for CuMgCeOx, except that the pH was held at 7.0
and the dried powders were treated in flowing air at 623 K
for 4 h (11, 12). K and Cs were introduced by incipient
wetness of the oxidized samples using K2CO3 (0.25 M) and
CH3COOCs (0.25 M) aqueous solutions (K2CO3: Fisher
Scientific, A.C.S. certified; CH3COOCs: Strem Chemicals,
99.9% ).
The density of basic sites was determined using a 13CO2/
12CO2 isotopic exchange method. This method provides a
direct measure of the number of basic sites kinetically avail-
able for CO2 adsorption at typical reaction temperatures.
This technique also provides the reactivity distribution of
basic sites and a direct measure of surface nonuniformity. In
this method, a prereduced catalyst was exposed to a 0.1%
13CO2/0.1% Ar/He (Cambridge Isotope Laboratories, Inc.)
stream at 573 K. After 13CO2 reached a constant concen-
tration in the effluent stream (0.5 h), the flow was switched
Structural Characterization
Powder X-ray diffraction (XRD) patterns were collected to 0.1% 12CO2/He (Cambridge Isotope Laboratories, Inc.).
using a D5000 Siemens Diffractometer and monochromatic The decay in the concentration of 13CO2 as it was replaced
Cu-K radiation. Total surface areas were determined us- on the surface by 12CO2 was followed by mass spectromet-
ing the single point BET method by measuring N2 phy- ric analysis of 13CO2 and Ar concentrations as a function of
sisorption at 77 K using a continuous flow Quantasorb Sur- the time elapsed after the isotopic switch. The presence of
face Area Analyzer (Quantachrome Corp.). Bulk catalyst Ar permitted corrections for gas holdup and hydrodynamic
compositions were measured by atomic absorption spec- delays within the apparatus.
troscopy (AAS).
Temperature programmed surface reaction studies were
The reduction behavior of Cu catalysts was studied using carried out by exposing a prereduced catalyst (33.0 mg)
temperature-programmed reduction (TPR). This method to an ethanol/He stream at room temperature for 0.2 h.
involves treating a Cu sample (15–100 mg) in He at 723 K Ethanol (C2H5OH: Fisher Scientific, A.C.S. certified) was
for 0.3 h in order to remove adsorbed water and carbonates. introduced by passing He through a saturator containing
Different amounts of samples were used for the high- and ethanol at 273 K. The catalyst sample was then purged with

132
XU ET AL.
He (100 cm3/min) at room temperature in order to remove purity). The reactant mixture was circulated continuously
gas phase and weakly adsorbed species and the reactor tem- through the catalyst sample and reaction products were
perature was increased to 723 K at 0.5 K/s. Desorption prod- sampled by syringe extraction after various contact times.
ucts were continuously monitored by mass spectrometric Reactant and product concentrations were measured by
analysis of the effluent stream. The TPSR apparatus and the mass spectrometric and flame ionization detection, after
detailed experimental procedures will be described else- separation with a 5% phenyl-methyl-silicone capillary col-
where (15). H2 and He were purified by passing through umn. The effect of CO2 (Matheson: research grade) during
–
a molecular sieve trap (Matheson, Model 452 : 4A) in or- ethanol reactions on 1 wt.% K CuyMg5CeOx was explored
der to remove H2O and through an oxygen scavenging trap by adding different amounts of CO2 (0–20 kPa) to the
(Matheson, Model 64-1008A).
ethanol reactants. Cross-coupling reactions of 13CH3OH
(13C: 99% , 18O: 12% ; Isotec Inc.) with acetaldehyde (Fisher
Scientific, Reagent Grade) were carried out using the
following reactant mixture: 12C2H4O/13CH3OH/CH4/He =
4.0/8.0/2.7/86.6 kPa. 13CH3OH was purified by several
freeze–pump–thaw cycles before catalytic experiments.
Catalytic Evaluation and Mechanistic Studies
High-pressure isobutanol synthesis studies were per-
formed in a catalyticmicroreactor consistingofa single-pass
tubular reactor (1.27 0.071 cm, 304 stainless-steel tube,
length 33.0 cm) held within a three-zone heated furnace in
order to ensure uniform axial temperatures ( 1 K). The
catalyst (2 g) was reduced in situ in pure H2 (Matheson,
99.99% ) at 583 K and atmospheric pressure for 12 h be-
fore catalytic experiments were carried out at 4.5 MPa us-
ing synthesis gas (H2/CO/Ar = 0.45/0.45/0.10 at.; Matheson:
99.99% CO, 99.99% Ar). Ar was used as an internal stan-
dard in order to ensure accurate material balances. Syn-
thesis gas was purified using activated carbon (Sorb-Tech
RL-13) in order to remove metal carbonyls. H2 was pu-
rified by passing through a catalytic purifier (Matheson,
Model 64-1008). Traces of water were removed from H2 and
H2/CO mixtures using molecular sieves (Matheson, Model
452 : 4A).
Reactants and products were analyzed on-line using a
gas chromatograph (Hewlett–Packard, Model 5890 II Plus)
equipped with a 10-port sampling valve and two sample
loops. One sample loop was injected into a 5% phenyl-
methyl-silicone capillary column (HP-5, 50 m, 0.32 mm di-
ameter, 1.05 m film thickness) and the other into a packed
column (Porapak Q, 1.8m length, 0.32cm diameter). A ther-
mal conductivity detector (TCD) was used to measure the
concentration of Ar, N2, CO, CO2, and H2O in the efflu-
ent from the packed column. A flame ionization detector
(FID) was used to measure the concentrations of all or-
ganic compounds eluting from the capillary column. Con-
densable products were collected and used to confirm the
identity of species in reaction products by mass spectro-
scopic analysis after capillary chromatography (Hewlett–
Packard, Model 5890 II Plus GC; Hewlett–Packard, Model
5972 Mass Selective Detector).
RESULTS AND DISCUSSION
Structural and Textural Properties of
–
K CuMgCeOx Catalysts
The X-ray diffraction pattern of a dried precipitate sam-
ple shows the presence of Mg(OH)2 and Ce(OH)4/CeO2
phases (Fig. 1a). The poor sample crystallinity and the co-
incidence of diffraction lines for Ce oxide and hydroxide
phases make it difficult to determine conclusively the form
ofCe in dried samples. Precipitatesdid not showanydiffrac-
tion lines corresponding to layered hydrotalcite-type struc-
tures.
Separate MgO and CeO2 phases were detected by X-ray
diffraction after air treatment at 723 K (Fig. 1b). Diffrac-
tion lines occurred at the positions expected for the
pure monometallic oxides and provided no evidence for
MgCeOx mixed oxides. CuO crystallites were not detected,
suggesting that the Cu component is well dispersed, either
Alcohol and aldehyde coupling reactions were carried
out in a gradientless recirculating reactor unit (RRU). The
catalyst (18.0 mg) was first reduced in H2 (10% H2/He) at
623 K for 0.5 h. The temperature was lowered to 573 K
and ethanol was introduced along with a small amount of
methane, used as an unreactive internal standard (reaction
mixture: C2H5OH/CH4/He = 4.0/2.7/94.6 kPa; C2H5OH:
FIG. 1. Powder X-ray diffraction patterns of Cu/Mg/Ce = 0.5/5.0/1.0
Fisher Scientific, A.C.S. certified; CH4: Matheson, ultra high (atomic ratio) samples (a) uncalcined and (b) treated in air at 723 K.

ISOBUTANOL AND METHANOL SYNTHESIS
133
TABLE 2
Composition, Surface Area, and Basic Site Density of Mixed Metal Oxides
BET
Exchangeable CO2 desorbed during CO2 exchange
K contenta surface area
(wt.% )
Cu
CO2 at 573 K
dispersionb (10 6 mol/m2)
TPD below 573 K
(10 6 mol/m2)
rate constantc
1
Sample
MgO
Cu0.1MgOx
Mg5CeOx
Cu0.5Mg5CeOx
Cu0.5Mg5CeOx
Cu0.5Mg5CeOx
Cu7.5Mg5CeOx
Cu/ZnO/Al2O3
(m2/g)
(k, 103 s
)
0.01
0.2
0.01
0.1
1.0
3.5
1.2
0
191
118
188
167
147
62
—
0.06
—
0.23
0.14
0.062
0.047
0.05
0.38
—
0.50
—
18
—
0.95
1.2
2.3
5.2
3.3
1.1
0.84
0.62
0.64
0.65
0.91
0.72
5.0
4.8
2.4
2.6
3.0
13
92
62
a Bulk composition measured by atomic absorption.
b
Dispersion calculated from the ratio of surface Cu (determined by N2O decomposition at 363 K (13, 14)) to the
total number of Cu atoms in the catalyst.
c Obtained at the peak maximum of basic site distribution curve.
as small crystallites or as a solid solution within the CeO2 was rereduced at 473 K instead of 623 K after the titration
lattice. Mixed Cu/Ce oxide solid solution can form during measurements. Surface oxygen atoms on Cu surface can be
coprecipitation and remain after high temperature oxida- removed at 473 K in H2, but reduction of CeO2 is likely to
tive treatments (16, 17).
occur only at higher reduction temperatures. Thus, reduc-
The total surface area of Cu0.5Mg5CeOx samples de- tion at 473 K should remove oxygen from Cu surface but
creased with increasing K concentration (Table 2). not from CeOx. The Cu surface area of the sample reduced
Lunsford and co-workers (18, 19) have reported that the at 473 K was 19.3 m2/g, a value identical to that obtained on
addition of Li also decreases the surface area of MgO. the fresh sample after reduction at 623 K (19.0 m2/g). This
The copper dispersion measured by N2O decomposition shows that CeOx does not contribute to the measured cop-
on Cu0.5Mg5CeOx is 0.23 (Table 2); this value is among the per dispersion and that measured dispersion values indeed
highest reported in the literature for Cu-based methanol reflect the presence of very small Cu metal crystallites of di-
and higher alcohol synthesis catalysts. This copper disper- ameter about 5 nm. An additional exposure of this titrated
sion value corresponds to a crystallite size of about 5 nm as- sample to 5% H2 at 423 K resulted in a Cu surface area of
suming hemispherical crystals. A Cu0.5Mg5Ox sample with 17.9 m2/g, suggesting that oxygen atoms chemisorbed on Cu
similar Cu content as Cu0.5Mg5CeOx has a copper disper- can be almost completely removed by H2 at temperatures
sion of only 0.06. Thus, the presence of Ce oxide species as low as 423 K.
favors the formation of small Cu metal crystallites on MgO
supports.
Copper dispersion decreased with increasing K loading
(Table 2) because of a decrease in the surface area of the
The decomposition of N2O on reduced ZnO sites has Mg5CeOx support and also because K species may block
been found to corrupt Cu surface area measurements surface Cu atoms or inhibit their reduction to Cu metal,
on Cu/ZnO catalysts (20). It is possible that decompo- resulting in a decrease in the number of exposed surface Cu
sition of N2O on reduced CeOx species in prereduced atoms. CeOx addition to Cu0.1MgOx increased both surface
Cu0.5Mg5CeOx catalysts also corrupts N2O titration data area and copper dispersion (Table 2), suggesting that CeOx
and leads to observed high copper dispersion values. N2O acts as structural promoter.
did not decompose on Mg5CeOx samples prereduced at
The temperature-programmed reduction profiles of
623 K, suggesting that any reduced Ce species present do MgO-based Cu samples are shown in Fig. 2. The onset and
not decompose N2O at 363 K. CeO2 may only reduce, how- peak maximum temperatures for H2 consumption and H2O
ever, when Cu metal is available in order to dissociate H2. formation appeared at the same temperature for each sam-
In effect, the presence of Cu may promote the reduction of ple. The high-temperature tail in the H2O peak is caused by
CeO2.
a strong interaction between H2O and MgO. This tail was
A reduction–oxidation cycle was carried out in order not observed for the H2 peak, but the signal-to-noise ratio
to rule out any contribution from reduced Ce species of the H2 peak was lower than that of H2O because of the
to N2O titration measurements of Cu surface area. A high H2 background pressure in the mass spectrometer.
Cu0.5Mg5CeOx sample was reduced at 623 K and N2O titra-
The presence of CeOx in Cu0.5Mg5CeOx decreases the
tion data were obtained. The measured Cu surface area reduction temperature of CuO (from 508 to 436 K). CeOx
was found to be 19.0 m2/g (0.23 Cu dispersion). This sample addition also increases Cu dispersion and decreases Cu

134
XU ET AL.
Ea, was 84 kJ/mol for the reduction of pure CuO whereas Ea
–
decreased to 77 kJ/mol for the reduction of the CuO ZnO
catalysts (22), in agreement with the promoting effect of
ZnO on the reducibility of CuO. Similar processes are likely
to occur during CuO reduction on CuMgCeOx samples. A
better contact between CeOx and Cu is expected with in-
creasing CeOx/Cu ratio, leading to CuO reduction at lower
temperatures.
In contrast with CeOx , K addition to Cu-containing sam-
ples inhibits CuO reduction, as shown by the shift of the re-
duction peak to higher temperatures (Fig. 3). The effect of
K is stronger on low-Cu (Cu0.5Mg5CeOx; 1T = 79 K) than
on high-Cu (Cu7.5Mg5CeOx; 1T = 57 K) catalysts. Also, the
effect of K did not depend on the presence of CeOx; the re-
duction temperatures increased to a similar extent (1T =
–
–
70 K) on K Cu0.5Mg5Ox and K Cu0.5Mg5CeOx. K appears
to increase the stability of Cu2 + ions and to make them
more difficult to reduce by H2. A similar effect was re-
ported on Cs-promoted Cu/ZnO/Cr2O3 (25). In this study,
the presence of Cs retards CuO reduction by about 50 K.
Campos-Martin et al. (25) suggest that the inhibited reduc-
tion of CuO is associated to closer interaction between the
CuO and alkali phases retards H2 activation.
FIG. 2. Temperature-programmed reduction profiles obtained in 5%
H2/He of MgO-based Cu catalysts: (a) Cu0.5Mg5CeOx; (b) Cu7.5Mg5CeOx;
(c) Cu0.1Mg5Ox. [Heating rate: 0.17 K/s; 15–100 mg of sample, 100 cm3/min
5% H2/He mixture; pretreatment temperature, 723 K.]
particle size apparently because of strong interactions be-
tween Cu and CeOx. The large Cu particles in Cu7.5Mg5
CeOx, however, can also be reduced at temperature lower
than that on Cu0.1MgOx. The reduction profiles (Fig. 2)
suggest that the promoting effect of CeOx on copper ox-
ide reduction increases with increasing Ce/Cu ratio. CeOx
has been reported previously to promote metal oxide re-
duction for Pd/CeO2/Al2O3 catalysts (21). The presence
of CeOx shifts the reduction peak temperature of PdO
from 437 to 376 K. Moreover, the reduction behavior of a
Pd/CeO2/Al2O3 sample prepared by the coprecipitation of
Pd and cerium nitrates differs from that of a Pd/CeO2/Al2O3
prepared by successive impregnation of Al2O3 with CeO2
and Pd. Some reduction occurs at room temperature on the
sample prepared by the coprecipitation methods as a conse-
The observed inhibition of copper reduction by K can
be explained by inhibited activation of H2, as proposed
by Klier and co-workers (25), or by the strengthening of
–
Cu O bonds upon K addition. As reported in the literature
(26, 27), the bond energy of surface oxygen for CuO is about
42 kJ/mol and it increases to 63 kJ/mol upon the addition
–
of 10–25 at.% MgO. Addition of MgO weakens the Cu Cu
–
bonds and strengthens Cu O bonds. In a similar way, the in-
corporation of K2O into CuO during catalyst synthesis may
–
quence of a higher degree of PdO CeOx contact area (22).
Lamonier et al. (16) have found that Cu2+ insertion into
CeOx occurs during the synthesis of CuCeOx samples by
coprecipitation methods. Four different species, present as
CuO monomers, dimers, clusters, and small particles, were
detected in these CuO/CeO2 samples (16).
ZnO has a similar effect as CeOx on copper reduction.
Garcia-Fierro et al. (22, 23) reported that the fraction of
copper oxide strongly interacting with ZnO increases with
increasing Zn/Cu ratio and that such copper oxide species
reduced a low temperature. A kinetic model of reduction
kinetics of CuO/ZnO suggests that the promoting effect
of ZnO on copper reduction is caused by the dissociative
adsorption of H2 on ZnO surfaces or on Cu metal clus-
ters closely associated with ZnO (24). The spillover of the
hydrogen atoms formed enhances Cu2+ reduction. In fact,
FIG. 3. Temperature-programmed reduction profiles obtained in 5%
–
H2/He of MgO-based Cu catalysts: (a) 1.0 wt.% K Cu0.5Mg5CeOx; (b)
–
–
1.2 wt.% K Cu7.5Mg5CeOx; (c) 1.1 wt.% K Cu0.1Mg5Ox. [Heating rate:
0.17 K/s; 15–100 mg of sample, 100 cm3/min 5% H2/He mixture; pretreat-
kinetic analysis showed that the apparent activation energy, ment temperature, 723 K.]

ISOBUTANOL AND METHANOL SYNTHESIS
135
–
FIG. 4. Isotopic switch method for probing adsorption–desorption processes dynamics on nonuniform surface of K Cu0.5Mg5CeOx.
–
increase the bond strength of Cu O and therefore retard of Fig. 4a. The exchange capacity is determined from the
CuO reduction.
area of the 13CO2 desorption curve immediately after a
switch from 13CO2 to 12CO2. When corrected for gas holdup,
physical hydrodynamic delays, and CO2 mass spectromet-
ric response factors, this area corresponds to the number of
Density and Strength of Basic Sites
The density and strength of basic sites on modified MgO basic sites that are kinetically accessible for exchange reac-
samples were measured from the exchange capacity and tions at 573 K. Weakly interacting sites are mostly unoccu-
rates obtained using isotopic 13CO2/12CO2 switch experi- pied by CO2; strongly interacting sites do not exchange in
ments at typical reaction temperatures (573 K). Site den- the time scale of the isotopic relaxation experiment. These
sities obtained from the area under the 13CO2 relaxation strongly interacting and weakly interacting sites are also
curves are shown in Table 2 for several catalytic materi- unlikely to contribute to catalytic reactions at 573 K. Site
als. As a comparison, basic site densities determined by densities obtained from these measurements are shown in
TPD of preadsorbed CO2 are also listed in Table 2. CO2 Table 2.
TPD measurements involve the adsorption of CO2 on pre-
The local slope in the semi-logarithmic plots of Fig. 4a re-
reduced samples at room temperature, the removal of gas flects the dynamics of the first-order CO2 exchange reaction
phase and weakly adsorbed CO2 by flowing He, and a linear and thus the exchange rate constant on available basic sites.
temperature increase with on-line monitoring of the CO2 The curved semi-logarithmicplotsshowthat Cu0.5Mg5CeOx
evolution rate by mass spectrometry. CO2 TPD basic site surfaces contain sites with a distribution of exchange rate
densities were obtained by integration of the evolution rate constants, because uniform surfaces with only one type of
curves up to a desorption temperature of 573 K.
adsorption site would lead to straight lines in Fig. 4a. Ex-
Although temperature-programmed desorption of CO2 change rate constants depend on the thermodynamics of
is frequently used for measuring basic site densities and binding interactions between CO2 and basic sites through
strength, the area under TPD curves gives the total number linear free energy relations commonly used to estimate acti-
of basic sites and not the number of basic sites kinetically vation energies for chemical reactions (29). Large exchange
available for adsorption and catalytic reaction at reaction rate constants and the concomitant short relaxation times
temperature. Desorption peak temperatures provides only (e.g. on 0.1 wt.% K) reflect short CO2 surface lifetimes and
qualitative information about the basic strength. In con- weaker binding of CO2 molecules on available basic sites.
trast, 13CO2/12CO2 exchange methods can provide not only
The distribution of exchange rate constants was obtained
a direct measure of the number of these basic sites that re- for each catalyst sample from the relaxation dynamics us-
versible chemisorb CO2 at 573 K, but also a distribution of ing inverse Laplace transform deconvolution methods (30).
reactivity for such basic sites (28).
These kinetic distributions are shown for 0.1 and 1.0 wt.%
–
The relaxation responses obtained from 13CO2/12CO2 iso- K Cu0.5Mg5CeOx catalysts in Figs. 4b and 4c. The distri-
topic switch experiments on Cu0.5Mg5CeOx catalysts with bution curves were normalized to give an area of unity.
0.1 and 1.0 wt.% K are shown in the semi-logarithmic plots The y-axis represents the distribution function f(k), where

136
XU ET AL.
f(k) dk gives the fraction of the total number of accessi-
ble basic sites with exchange rate constants between k and
k + dk.
TABLE 3
Activities and Selectivities for Isobutanol Synthesis
on K-Promoted Cu0.5Mg5CeOx
The density of sites that reversibly bind CO2 at typi-
cal catalytic turnover time obtained by isotopic exchange
Catalyst:
Composition:
Ref. (7)
0.9 wt.% K
This study
1.0 wt.% K
This study
2.9 wt.% Cs
6
methods on MgO at 573 K is 0.38 10 mol/m2; this is
6
lower than the value of 18.3 10 mol/m2 reported for
Cu0.5Mg5CeOx Cu0.5Mg5CeOx Cu/ZnO/Al2O3
Space velocity
the surface density of oxygen atoms in MgO (31). This
is not unexpected since the 13CO2/12CO2 isotopic switch
method only detects those basic sites that participate in ex-
change reactions near 573 K. The density of available basic
sites in Cu0.5Mg5CeOx increases with increasing K loading
(Table 2). Potassium also leads to slower exchange pro-
cesses, suggesting that stronger basic sites form when alkali
is present on modified MgO surfaces (Figs. 4b and 4c). An
increase in K loading from 1.0 to 3.5 wt.% increases basic
[cm3/g-cat.h]:
1832
15.5a
1500
14.7b
6000
15.4c
CO conversion [% ]:
CO reaction rate
3
3
3
[mol CO conv./g-cat.h]: 6.3 10
4.4 10
17.0 10
Carbon selectivitiesd [% ]
Methanol
Ethanol
1-Propanol
Isobutanol
Other higher alcohols
and oxygenates
Dimethylether
Hydrocarbons
CO2
57.7
1.9
2.6
10.4
2.2
62.7
1.7
3.0
6.2
9.2
67.6
4.4
6.9
6.9
9.5
6
site density from 2.3 to 5.2 10 mol/m2, but has no de-
tectable effect on the exchange rate constant and site dis-
tribution (Table 2). The presence of small amounts of CeOx
in MgO (Mg/Ce = 5 at.) increases both the density and the
strength of basic sites (Table 2).
1.2
24.3
31.0
6.8
9.9
23.4
0.9
2.5
22.6
a Reaction conditions: 593 K, 5.1 MPa, H2/CO = 1.
b Reaction conditions: 583 K, 4.5 MPa, H2/CO = 1.
c Reaction conditions: 580 K, 4.5 MPa, H2/CO = 1.
In contrast with the CO2 TPD method, in which both tem-
perature and surface coverage vary with time, 13CO2/12CO2
exchange methods probe the density and reactivity of basic
sites under conditions of chemical equilibrium and constant
temperature. Exchange methods also avoid contributions
from the decomposition of catalytically irrelevant and un-
reactive carbonates and from the desorption of strongly
adsorbed CO2. This method probes the number of sites
that can be made available for base-catalyzed reactions at
temperatures typical of isobutanol synthesis. As we discuss
below, competitive adsorption between CO2 and aldehy-
dic species on such sites renders basic sites unavailable for
alcohol synthesis reactions on Cu0.5Mg5CeOx catalysts.
d Hydrocarbon and alcohol selectivities normalized on a CO2-free basis
(except CO2).
drocarbons, and higher alcohols or by the readsorption of
the water formed along with these products.
The higher CO2 selectivity reported by Apesteguia et al.
(7) may reflect the slightly higher reaction temperature
in their studies and their higher selectivity to hydrocar-
bons. Also, higher alcohol-to-hydrocarbon ratios were ob-
tained on our samples than on those studied by Apesteguia
et al. (7). Our Cu0.5Mg5CeOx catalyst, however, shows
much higher selectivity to dimethylether. Both hydrocar-
bons and dimethylether are formed on acid sites but the
Isobutanol Synthesis from CO/H2 Mixtures
–
–
Cs Cu/ZnO/Al2O3 and K Cu0.5Mg5CeOx catalysts show formation of hydrocarbons requires subsequent reactions
the highest isobutanol synthesis rates and alcohol-to- of dimethylether. These results suggest that our samples
hydrocarbon product ratios among reported catalysts for contain fewer residualsitesthan catalystsofsimilar nominal
the synthesis of isobutanol at low temperatures (<623 K) composition reported by Apesteguia et al. (7), apparently
and pressures (6, 7, 32). Alcohol synthesis rates on K- as a result of more effective alkali dispersion. Potassium
modified Cu0.5Mg5CeOx catalystsdepend stronglyon potas- appears to titrate residual acid sites formed during copre-
sium concentration. Intermediate K concentrations (1.0– cipitation or CuO reduction processes. Potassium also in-
1.5 wt.% ) lead to the highest isobutanol synthesis rates and creases the density and strength of basic sites required for
selectivities (7).
Reaction products obtained on 1.0 wt.% K Cu0.5Mg5
chain growth reactions of alcohols (Table 2, Fig. 4).
Alkali dispersion appears to be sensitive to the details
–
–
CeOx (583 K and 4.5 MPa) and 2.9 wt.% Cs Cu/ZnO/Al2O3 of the impregnation procedure. For example, the addition
(580 K and 4.5 MPa) are shown in Table 3 at similar CO of potassium before oxidative treatment of dry precipitates
conversion values. The data reported by Apesteguia et al. (calcination) leads to catalysts with much higher selectivity
–
(7) on similar K Cu0.5Mg5CeOx samples are also shown to hydrocarbons than on samples impregnated with K after
in Table 3. Methanol, isobutanol, and CO2 were the most oxidation. Higher potassium contents (e.g., 3.5 wt.% ) are
abundant reaction products. CO2 is formed in CO oxidation undesirable because the more complete titration of residual
or water–gas shift reactions, using adsorbed oxygen atoms acid sites is accompanied by sintering of the metal oxide
formed directly during the synthesis of dimethylether, hy- support and in the number of accessible Cu atoms measured

ISOBUTANOL AND METHANOL SYNTHESIS
137
–
–
FIG. 5. CO conversion and product selectivities vs space velocity on (a) Cs Cu/ZnO/Al2O3 and (b) K Cu0.5Mg5CeOx. [583 K, 4.5 MPa, CO/H2 = 1.
Product selectivities are on CO2-free basis.]
by N2O decomposition methods. These surface Cu atoms
Methanol synthesis rates decrease with decreasing space
are required for methanol synthesis and alcohol conversion velocity on both catalysts, because the methanol synthesis
to aldehydes during CO hydrogenation to higher alcohols reaction
(Table 2).
CO + 2H2 = CH3OH
The effects of space velocity (bed residence time) on al-
cohol synthesis rates and selectivities are shown in Fig. 5
approaches equilibrium under our reaction conditions;
these conditions were chosen in order to maximize yields
of higher alcohols. Higher alcohol yields are obtained at
higher temperatures and long bed residence times. When
methanol yields become limited by equilibrium, methanol
synthesis rates become proportional to space velocity, be-
cause CO conversion to methanol remains constant with
increasing bed residence time.
On typical methanol synthesis catalysts (i.e., Cu/ZnO/
Al2O3), the presence of CO2 in modest concentrations
(1–2% mol) actually increases methanol synthesis rates
(33). High CO2 concentrations (>10% mol), however, lead
to higher steady-state oxygen surface coverages and inhibit
methanol synthesis on these catalysts (33). Cs-modified
methanol synthesis catalysts reach maximum rates at very
low CO2 concentrations (<2% ) (34). Steady-state oxygen
coverages during methanol synthesis on Cu surfaces de-
pend on the relative rates of oxidation and reduction of
surface Cu atoms. At high CO2 concentrations the rate of
oxidation increases relative to reduction, leading to higher
steady-state oxygen coverages and to a (reversible) de-
crease in the number of surface Cu atoms available for
methanol synthesis. These surface metal atoms on Cu crys-
tallites catalyze methanol synthesis reactions of CO/CO2/
H2 mixtures (33, 35).
–
–
for 2.9 wt.% Cs Cu/ZnO/Al2O3 and 1.0 wt.% K Cu0.5
Mg5CeOx. CO conversion should increase linearly with in-
creasing bed residence time at the low CO conversion levels
in this study. The observed decrease in the slope for the CO
conversion curve with increasing bed residence time (Fig. 5)
reflects the gradual approach to equilibrium by the primary
methanol synthesis step and possibly also the inhibition of
methanol synthesis reactions by CO2 and H2O formed as
side products of hydrocarbon and higher alcohol synthe-
sis. Isobutanol and higher alcohols form in secondary chain
growth reactions, such as methanol carbonylation or aldol
coupling reactions, which are favored by longer bed resi-
–
dence times. C2 C3 alcohols are intermediate products that
undergo chain growth; therefore, their selectivity reaches a
maximum value at intermediate values of residence time.
Isobutanol is a kinetic end product, because it lacks
the two -hydrogens required for facile chain growth via
aldol condensation pathways. Therefore, isobutanol se-
lectivity increases with increasing residence time. This
–
increase, however, is less marked on K Cu0.5Mg5CeOx
–
than on Cs Cu/ZnO/Al2O3, because base-catalyzed chain
growth reactions are inhibited by CO2 more strongly on
–
–
K Cu0.5Mg5CeOx than on Cs Cu/ZnO/Al2O3. These ef-
fects are discussed in detail below. CO2 appears to inhibit
both methanol synthesis on Cu sites and aldol coupling re-
actions on basic sites during isobutanol synthesis reactions
on K-promoted Cu0.5Mg5CeOx catalysts. The weaker basic
In order to examine the details of CO2 inhibition effects
on methanol and isobutanol synthesis, CO2 was systemati-
cally added to the synthesis gas feed. The results are shown
in Fig. 6 as the reciprocal of methanol and isobutanol syn-
thesis rates versus CO2 concentrations. Typical Langmuir–
Hinshelwood rate expressions would lead to straight lines
when data are presented in this form. Methanol synthesis
–
sites on Cs Cu/ZnO/Al2O3 (Table 2) appear to be less sen-
sitive to CO2 poisoning; as a result isobutanol selectivities
–
reach higher values than on K Cu0.5Mg5CeOx with increas-
ing bed residence time.

138
XU ET AL.
–
–
FIG. 6. Reciprocal methanol (a) and isobutanol (b) productivity vs average CO2 partial pressure for 1 wt.% K Cu0.5Mg5CeOx, 2.9 wt.% Cs Cu/
ZnO/Al2O3, and 0.9 wt.% K Cu7.5Mg5CeOx. [583 K, 4.5 MPa, H2/CO = 1, 3000 cm3/g/h.]
–
rates for catalysts with high Cu contents decreased slightly isobutanol from ethanol. Also, as we discuss later, aldol
as the CO2 concentration was increased. Methanol syn- coupling reactions of alcohols require both Cu and basic
–
thesis rates on 1.0 wt.% K Cu0.5Mg5CeOx were more sites; therefore any blocking of surface Cu atoms by oxy-
strongly inhibited by CO2. These results reflect the opera- gen can lead to a decrease in isobutanol synthesis rates.
tion of these catalysts near methanol synthesis equilibrium. CO2 inhibition of isobutanol synthesis reactions may also
Catalysts containing a larger number of Cu surface sites reflect the reversible titration ofbasicsiteson oxide surfaces
–
(e.g., 1.0 wt.% K Cu7.5Mg5CeOx) can maintain equilibrium by acidic molecules, such as CO2. The inhibiting effect of
methanol concentrations even after a significant fraction of CO2 is greater on 1 wt.% CuMg7.5CeOx than on 2.9 wt.%
–
such sites are covered with oxygen adatoms during steady- Cs Cu/ZnO/Al2O3, possibly because of the stronger basic-
state catalysis; therefore no effect of CO2 on the methanol ity of the former (detected by the dynamics of CO2 ex-
productivity is observed until larger amounts of CO2 are change). On low Cu catalysts, alcohol chain growth appear
added. Catalysts containing fewer Cu sites (e.g., 1.0 wt.% to be limited by the availability of minority Cu sites and
–
K Cu0.5Mg5CeOx) become unable to maintain methanol thus CO2 inhibition of Cu such sites decrease chain growth
synthesis equilibrium conversions at lower CO2 concentra- rates. Higher Cu contents lead to coupling rates controlled
tions than catalysts with higher Cu site densities.
by reactions on basic sites, which are blocked by CO2 most
The effect of CO2 on the rate of higher alcohol synthesis effectively on the more strongly basic MgO-based materi-
has not been as thoroughly studied as on methanol synthesis als. Basicsiteson 0.9wt.% CuMg7.5CeOx appearsto be more
–
rates. Tronconi et al. (36) observed lower yields of higher sensitive to CO2 than on 2.9 wt.% Cs Cu/ZnO/Al2O3. This
–
alcohols on K ZnCrOx at 673 K when CO2 (3–6% ) was is consistent with the faster dynamics of CO2 exchange on
added to the H2/CO feed. These authors proposed that wa- the latter catalysts.
–
–
ter, produced from CO2 in the reverse water gas shift re-
action, titrated active sites. Elliot (37) observed that higher
alcohol synthesis over Cu/ZnO was enhanced by the addi-
tion of small amounts of CO2 (6% ) at 557–561 K. Calver-
ley and Smith (38) observed that the higher alcohol yields
reach a maximum at intermediate CO2 concentrations
Alcohol Coupling Reaction Pathways
Alcohol and aldehyde coupling reactions were studied in
order to probe reaction pathways involved in branched and
linear higher alcohols synthesis from methanol, ethanol,
and propanol intermediates formed initially during CO hy-
drogenation reactions at high pressures. These studies also
address the effects of CO2 on alcohol dehydrogenation re-
actions and on chain growth reactions and the roles of Cu,
–
(0–9% CO2) on 0–0.5 wt.% K Cu/ZnO/Cr2O3 at 558 K.
They also observed that the effect of CO2 becomes stronger
as alkali concentration increases.
–
Figure 6b shows the reciprocal isobutanol synthesis rate
as a function of the average CO2 partial pressure within
the catalyst bed. The catalyst with low Cu content was in-
hibited more strongly by the presence of CO2 than cat-
alysts with higher Cu contents. To some extent, this may
reflect the reversible inhibition of Cu metal sites by CO2,
which leads to lower methanol yields on catalysts with low
Cu content. Although methanol is not involved in the ini-
tial formation of ethanol (39) on Cu/MgCeOx catalysts, it
acts as a C1 precursor in the synthesis of 1-propanol and
K, and Mg Ce oxides in the synthesis of branched alco-
hols. These experiments were carried out on Mg5CeOx,
–
–
Cu0.5Mg5CeOx, K Cu0.5Mg5CeOx, and K Cu7.5Mg5CeOx
under atmospheric pressures in a gradientless batch reac-
tor. Two types of reactants were used: pure ethanol and
13CH3OH–acetaldehyde mixtures.
Acetaldehyde was the initial product of reactions of pure
ethanol on Cu0.5Mg5CeOx (Fig. 7). Dehydrogenation reac-
tions occur much faster than chain growth reactions; the
latter lead to the formation of acetone, n-butyraldehyde,

ISOBUTANOL AND METHANOL SYNTHESIS
139
Acetaldehyde reachesa maximum concentration at inter-
mediate contact timesduringethanolreactionsand then de-
creases gradually, suggesting that acetaldehyde concentra-
tions approach equilibrium. Acetone and n-butyraldehyde
were the predominant condensation products. The nonzero
initial slopes on acetone and n-butyraldehyde curves are
inconsistent with products formed through consecutive re-
actions (Fig. 7). Therefore, condensation products are ei-
ther formed directly by ethanol self-condensation or via
acetaldehyde condensation reactionswith zero-order kinet-
ics. Methylethylketone (byacetaldehyde self-condensation
with oxygen retention reversal (40, 41)), 2-pentanone (by
–
acetaldehyde acetone condensation), and ethyl acetate (by
–
ethanol acetaldehyde reactions) were detected in much
smaller concentrations among reaction products.
Reaction pathways involved in the formation of all de-
K Cu0.5Mg5CeOx in ethanol reactions. [573 K, 101.3 kPa total pressure, tected products are shown in Fig. 8. Acetaldehyde is formed
FIG. 7. Site yields as a function of contact time on 1.0 wt.%
–
4.0 kPa ethanol, balance He.]
via ethanol dehydrogenation (step I). Aldol species (butan-
1-al-3-ol) formed from direct condensation reactions of
ethanol and from self-condensation of acetaldehyde pro-
and other oxygenates. Site yields (y-axis, Fig. 7) are defined
as the number of moles of ethanol converted to products
per g-atom Cu (total) in the catalyst charge. The total num-
ber of Cu atoms in the sample is a convenient normalizing
parameter. The local slope of the curve in Fig. 7 gives the
rate of formation of each product. When divided by the
fractional Cu dispersion, this becomes a site-time yield for
each product, at least for reactions with rate-determining
elementary steps requiring Cu surface atoms.
duce n-butyraldehyde (steps II–IV) and methyl ethyl ke-
tone (stepsII–V) (40). The aldolspeciesconvertsto the keto
form via H transfer (step VI); methyl ethyl ketone is also
formed by dehydration–dehydrogenation reactions of the
keto form (step VII). Acetone is formed via two pathways:
by reaction of aldol intermediates with surface oxygen fol-
lowed by decarboxylation (steps VIII–IX) and by reverse
aldol condensation of the keto form of reaction intermedi-
ates (step X). Formaldehyde formed in the latter reaction
–
FIG. 8. Reaction scheme for ethanol reactions on K CuMg5CeOx catalysts.

140
XU ET AL.
TABLE 4
Effects of Cu- and K-Loading on Ethanol Consumption and Product Formation on Mg5CeOx
Rates of formation
Ethanol
dehydrogenation
Acetone
Butyraldehyde
Cu
(wt.% )
K
Areal rate
Turnover rate
Areal rate
Turnover rate
Areal rate
Turnover rate
a
b
c
d
c
d
(wt.% )
r1
r1
r2
r2
r3
r3
9
7
7
7
11
5
3
3
2
0
7
7
0
4.0 10
3.6 10
2.4 10
9.4 10
—
6.2 10
3.0 10
4.3 10
4.4 10
6.5 10
2.2 10
1.8 10
1.0 10
NDe
4.5 10
7.6 10
1.2 10
NDe
3.4 10
3.2 10
2.8 10
9
10
10
9
4
0.1
1.0
1.2
0.24
0.23
0.24
9
4
8
4
49
a r1 is the rate of ethanol consumption as mol/m2 total s.
1
b Turnover rates per Cu surface atom as s
.
c r2 and r3 are the rates of product formation as mol/m2 MgCeOx s.
d Turnover rates per accessible site from 13CO2/12CO2 as mol/mol CO2 exchangeable s.
e Not detected.
decomposes rapidly to CO and H2. 2-Pentanone is formed 0.24 s 1, respectively. Thus, ethanol dehydrogenation re-
by condensation of aldol-type intermediates formed in quires exposed Cu atoms, which become inactive for both
acetone–acetaldehyde self-condensation reactions (steps N2O decomposition and alcoholdehydrogenation byblock-
XI–XII). Ethylacetate forms via reaction of ethanol with age with K species. Dehydrogenation turnover rates are not
acetaldehyde (step XIII) (40).
Ethanol reaction rates and the selectivity to acetalde- surface atoms with KOx species.
hyde, n-butyraldehyde, and acetone were measured on Aldol coupling chain growth rates are lower on Mg5CeOx
affected by Cu crystallite size or by the titration of some Cu
–
K CuMgCeOx catalysts with a range of Cu and K concen- than on Cu0.5Mg5CeOx because of the lower acetalde-
trations in order to examine the role of individual cata- hyde steady-state concentrations in the former. At simi-
lyst components on dehydrogenation and condensation re- lar steady-state acetaldehyde concentrations on Cu0.5Mg5
–
action pathways. The results are shown in Table 4. Initial CeOx and 1.0 wt.% K Cu0.5Mg5CeOx catalysts, the pres-
reaction rates were obtained from the slopes of site-yield ence of K increases the rate of base-catalyzed aldol cou-
plots at short contact times using the total surface area, the pling reactions leading to acetone and butyraldehyde
MgCeOx surface area (estimated from the difference be- (Table 4). Thus, it appears that the higher basic site den-
tween the total BET surface area and the copper surface sity and strength detected by 13CO2/12CO2 isotopic switch
area), the number of basic sites from CO2 exchange, or the measurements lead to higher rates of base-catalyzed aldol
Cu metal surface area of each sample in order to normalize condensation reactions. The total rates of base-catalyzed
reaction rates.
aldol coupling reactions, when normalized by accessible ba-
Initial dehydrogenation rates were very low on Mg5CeOx sic sites, are similar on Cu0.5Mg5CeOx (2.5 10 3 mol/mol
–
catalysts and ethanol conversion reached a limiting value of CO2 exchangeable s) and 1 wt.% K Cu0.5Mg5CeOx
about 9% after 0.33 h; this conversion value is much lower (2.2 10 3 mol/mol CO2 exchangeable s). Aldol coupling
–
than the equilibrium conversion predicted from thermo- chain growth reaction rates on 1.2 wt.% K Cu7.5Mg5CeOx
dynamic data at these reaction conditions (90% ). Ethanol (49 wt.% Cu, 0.047 Cu dispersion) are, however, much
conversion rates decreased markedly with increasing con- higher than on low Cu-loading catalysts (7 wt.% Cu, 0.14
version because of (i) the formation of surface polymeric Cu dispersion), even though basic site densities are similar
species from acetaldehyde and (ii) the inhibition effect of on these two samples. The difference in chain growth rates
CO2 (and possibly H2O) on base-catalyzed dehydrogena- indicates that Cu metal sites participate in rate-determining
tion reactions on Mg5CeOx. Dehydrogenation rates were steps required for condensation reactions. This is confirmed
–
much higher on Cu0.5Mg5CeOx , 1.0wt.% K Cu0.5Mg5CeOx, by the lower aldol coupling chain growth reaction rates ob-
–
and 1.2 wt.% K Cu7.5Mg5CeOx catalysts than on Mg5CeOx, served on Cu-free catalysts, which is caused not only by the
suggesting the involvement of Cu metal sites in the dehy- low concentration of required acetaldehyde intermediates,
drogenation of ethanol to acetaldehyde. The presence of but also by the absence of Cu sites required to increase the
K decreased areal ethanol dehydrogenation rates because rate of condensation steps.
of a concomitant decrease in Cu dispersion. Ethanol de-
The role of copper suggests a bifunctional mechanism
hydrogenation turnover rates (normalized by exposed Cu for aldol condensation. Ethanol adsorbs dissociatively on
–
atoms) on Cu0.5Mg5CeOx, 1.0 wt.% K Cu0.5Mg5CeOx, and MgO surface to form ethoxide and hydrogen species (42).
–
1.2 wt.% K Cu7.5Mg5CeOx catalysts were 0.24, 0.23, and Hydrogen species are then removed by migration to Cu

ISOBUTANOL AND METHANOL SYNTHESIS
141
sites, reaction with another H species, and desorption as H2:
Ethoxide species then lose a -H via intermolecular hydrogen transfer to oxygen ions and form surface acetaldehydic
species:
H species can migrate from basic to Cu sites and recombine on Cu sites to form H2; as a result, basic sites become
–
available for another hydrogen abstraction event. H H recombination rates increase with increasing ratio of surface Cu
–
atoms to basic sites, which is reflected by the ratio of Cu surface area to oxide surface area on K Cu7.5Mg5CeOx (0.31) and
–
–
K Cu0.5Mg5CeOx (0.07) at similar basicsite density(Table 2). The high Cu-to-oxide surface area ratio on K Cu7.5Mg5CeOx
(0.31) leads to higher aldol-condensation rates. Adsorbed acetaldehyde species can desorb as acetaldehyde, or they can
undergo -H abstraction and reaction with neighboring species to form aldol condensation products:
Surface aldol species can undergo reactions via the path- products has zero-order kinetics. The nonzero initial slope
ways shown in Fig. 8. In these reactions, the presence of of acetone and n-butyraldehyde curves (Fig. 7) is consistent
copper sites enhances H transfer and provides H species with a direct pathway for the activation and conversion of
for the hydrogenation of the unsaturated species (steps IV, ethanol in aldol condensation pathways. Additional experi-
VII, and XII, Fig. 8). It should be pointed out that the rates ments have shown that condensation, hydrogenation, and
of the final products formed through consecutive reactions total conversion rates are approximately first order in ac-
should have zero initial slopes, unless the formation of the etaldehyde (43). Reactionsof12C2H5OH/13C2H4O mixtures

142
XU ET AL.
influenced by CO2 only at the higher CO2 pressures typical
TABLE 5
Effect of CO2 on Ethanol Reaction on K CuyMg5CeOx
of isobutanol synthesis reactions. As discussed previously,
competitive adsorption between CO2 and acetaldehyde on
basic sites may account for the observed inhibition of CO2
–
1.0 wt.%
1.2 wt.%
–
–
–
on condensation reactions on 1.0 wt.% K Cu0.5Mg5CeOx
K Cu0.5Mg5CeOx
K Cu7.5Mg5CeOx
catalysts under high-pressure isobutanol synthesis condi-
tions. The ratio of CO2 to aldehydic intermediates is much
higher during CO hydrogenation than during low-pressure
ethanol reactions. Because of the lower CO2/acetaldehyde
ratio in ethanol dehydrogenation reactions, CO2 competi-
tion for basic sites is not favorable and, therefore, the aldol
coupling reactions are not influenced strongly by CO2.The
rate of acetone formation, however, decreased when CO2
was present during ethanol reactions. CO2 inhibition ef-
fect on acetone production may reflect: (i) the reversal of
equilibrium-limited surface reaction steps that form ace-
tone and CO2 (steps IX and X, Fig. 8) or (ii) different par-
allel reaction for acetone synthesis using stronger basic sites
that are more strongly inhibited by CO2.
b
b
Product
Ethanola Ethanol/CO2 Ethanola Ethanol/CO2
7
7
6
8
9
7
Acetaldehyde (r1) 2.3 10
Acetone (r2)
n-Butyraldehyde 8.0 10
(r3)
2.1 10
1.6 10
7.2 10
1.2 10
7.2 10
1.7 10
9.1 10
1.1 10
3.8 10
9
9
8
3.9 10
10
10
10
Note. rj are the rate of ethanol consumption and products formation as
mol/m2 total s.
a C2H5OH/CO2/H2/CH4/He = 4.0/0/26.7/2.7/67.9 kPa, 573 K.
b C2H5OH/CO2/H2/CH4/He = 4.0/5.3/26.7/2.7/62.6 kPa, 573 K.
suggest that condensation reactions can proceed via direct
reactions of ethanol without the intermediate formation of
gas-phase acetaldehyde molecules (43).
CO2 Effects on Alcohol Coupling Pathways
Cross-Coupling Reactions of 13CH3OH 12C2H4O Mixtures
–
The effects of CO2 on ethanol dehydrogenation and cou-
Ethanol is a useful and simple probe molecule to test
the role of individual components in metal–base bifunc-
tional catalysts for isobutanol synthesis. Ethanol reactions,
however, lead to acetone and n-butyraldehyde, which form
only 2-propanol and 1-butanol by hydrogenation during
CO/H2 reactions. 2-Propanol and 1-butanol cannot form
isobutanol precursors (such as isobutyraldehyde and pro-
pionaldehyde) by condensation reactions during CO hy-
drogenation. 13C-tracer studies of methanol-acetaldehyde
cross-coupling reactions were carried out in order to exam-
ine reaction pathways leading to the formation of the C3 and
–
pling reactions were examined on 1.0 wt.% K Cu0.5Mg5
–
CeOx and 1.2 wt.% K Cu7.5Mg5CeOx catalysts by adding
CO2 (3.5 kPa, 20 kPa) to ethanol/H2 (4.0/29.3 kPa) re-
actant mixtures at 573 K. Ethanol dehydrogenation rates
were not affected by CO2 addition (Table 5, Fig. 9) on
both catalysts, suggesting that Cu surface atoms are not
blocked by CO2 at the conditions of our study. CO2 also
did not affect n-butyraldehyde formation rate on either
low- or high-Cu catalysts and acetone formation rates on
low-Cu catalysts. These data differ from the observed de-
crease in methanol and isobutanol synthesis rates as CO2 is
formed (or added) during high-pressure alcohol synthesis
reactions from CO/H2 (Fig. 6), suggesting that Cu sites are
C4 oxygenate precursors required for isobutanol formation.
13
Cross-coupling site yields for CH3OH 12C2H4O mix-
–
tures on Cu0.5Mg5CeOx are shown in Fig. 10. The initial
FIG. 9. CO2 effect on ethanol reactions. Rates of ethanol consump-
tion and of product formation as a function of CO2 initial pressure on
FIG. 10. Methanol–acetaldehyde reactions. Site yields as a function
of contact time on Cu0.5Mg5CeOx. [573 K, 101.3 kPa total pressure, 8.0 kPa
methanol, 4.0 kPa acetaldehyde, balance He.]
–
1.0 wt.% K Cu0.5Mg5CeOx. [573 K, 101.3 kPa total pressure, 4.0 kPa
ethanol, 26.7 kPa dihydrogen, balance He.]

ISOBUTANOL AND METHANOL SYNTHESIS
143
rates (site-time yields per total mole Cu) for methanol con-
TABLE 6
version were higher than those for the conversion of ace-
taldehyde to ethanol and to condensation products. The
13CH3OH component in the feed reacted predominantly to
form 13CO and H2; this reaction reverses methanol synthe-
sis because of the unfavorable methanol synthesis thermo-
dynamics at atmospheric pressures (but not during CO hy-
drogenation at high pressures). The ethanol products were
predominantly unlabeled (Table 6) and formed by hydro-
genation of acetaldehyde using Cu sites and hydrogen pro-
duced by methanol decomposition. Propionaldehyde con-
tains predominantly one 13C atom, suggesting that it formed
by condensation of acetaldehyde with reactive formyl or
methoxy-type surface species or with gas-phase formalde-
hyde, all formed from 13CH3OH. The analysis of the mass
spectra pattern showed that the labeled carbon was present
at both the CH3 and the CHO fragments of propionalde-
hyde, suggesting that aldol condensation reactions occur
with retention of either one of the two oxygen atoms in
aldol-keto species, as proposed by Nunan et al. (42, 44).
They used the term “oxygen retention reversal” to indi-
cate the oxygen atom in formaldehyde remains in the final
product, whereas the oxygen of the growing intermediate in
the normal aldol condensation reaction remains in the final
products. Both normal and oxygen retention reversal types
Carbon-13 Distribution in Products of 13CH3OH C2H4O
Reactions on Cu0.5Mg5CeOx
–
Number of carbon-13
0 13
C
1 13
C
2 13
C
3 13
C
4 13
C
Carbon dioxide
Acetaldehyde
Ethanol
Propionaldehyde
n-Butyraldehyde
Isobutyraldehyde
14
86
1
0
91
6
0
—
4
0
6
3
—
—
—
3
0
23
—
95
100
0
91
0
—
—
—
0
74
3
Note. 573 K, 101.3 kPa total pressure, 8.0 kPa methanol, 4.0 kPa acetal-
dehyde, balance He.
ethanol were detected among reaction products, suggest-
ing that methanol coupling reactions, previously proposed
as a route for C1 to C2 chain growth during higher alco-
hol synthesis (40), do not readily occur at the normal pres-
sure reaction conditions of our study. In contrast with high-
pressure CO hydrogenation conditions, our reaction con-
ditions favor the decomposition of methanol and minimize
its bimolecular chain-growth reactions.
These isotopic tracer studies are consistent with the se-
quence of steps shown in Fig. 11. Reverse condensation
reactions of aldol and keto species (formed from 13C1
aldehyde-type intermediate and acetaldehyde) form 1-13C
labeled acetaldehyde, which in turn lead to the observed
propionaldehyde with 2-13C atoms, isobutyraldehyde with
3-13C atoms, and traces of labeled n-butyraldehyde. Addi-
tional studies of propionaldehyde-13C-methanol coupling
reactions have confirmed these isobutyraldehyde synthesis
pathways. These studies have also detected the reversibility
–
of aldol condensation occur on K CuMg5CeOx catalysts.
These types of reactions are caused by interconversion of
aldol and keto species via intramolecular H-transfer:
Isobutyraldehyde contained predominantly two 13C
atoms, suggesting that it formed by condensation of pro-
pionaldehyde (with one 13C) with a C1 species formed
13
–
from methanol (Table 6). The CO2 product of CH3OH
12C2H4O mixtures on Cu0.5Mg5CeOx was predominantly
labeled and formed from 13CO via water–gas shift reac-
tions (Table 6). Acetone and n-butyraldehyde showed no
detectable 13C-enrichment and formed exclusively via self-
condensation reactions of acetaldehyde or ethanol, as ob-
served in reactions of pure ethanol (Fig. 7). Decarboxyla-
tion of aldol intermediates to acetone explains the presence
FIG. 11. Reaction scheme for methanol–acetaldehyde reactions on
–
of small amounts of unlabeled CO2. Only traces of labeled K CuMg5CeOx catalysts.

144
XU ET AL.
–
of aldol formation and dehydration decarbonylation path- (steps VI, X, Fig. 8) can also lead to the formation of ace-
ways. The details of these studies will be reported elsewhere tone. The other product (formaldehyde) decomposes to
(43).
CO and H2. CO reacts with surface oxygen species to form
CO2.
Temperature-Programmed Surface Reactions
of Preadsorbed Ethanol
CO2 and acetone desorb at similar temperatures (about
573 K), suggesting that their evolution is limited by the ki-
netics of decomposition of a common intermediate leading
to both products (Fig. 12). These intermediates are likely
to be surface acetate species formed by nucleophilic attack
of carbonyl groups within adsorbed acetaldehyde by elec-
tronegative O2 species in a reducible metal oxide (e.g.,
CeO2). Oxygen atoms are replenished (at least under cata-
lytic conditions) by the water and CO2 formed in conden-
sation reactions. Acetate can also lead to the formation of
CO, CO2, H2O, and H2, as described by
Steady-state catalytic rates reflect the average surface
properties of multifunctional and frequently nonuniform
surfaces. Transient chemical reaction methods, such as
TPSR, provide a useful complement to steady-state cata-
lytic studies and isotopic switch methods. These transient
methods often reveal the dynamics and “capacities” of var-
ious reactive pools available on nonuniform surfaces by
varying temporally the surface coverage of preadsorbed
molecules and the reactor temperature. Our initial TPSR
studies have addressed the reactions of preadsorbed
ethanol on Cu0.5Mg5CeOx catalysts.
Step I: Alkyl elimination to produce CO2 and surface
alkyl species (47, 52, 53).
The rate of product evolution during temperature-
programmed surface reactions of ethanol preadsorbed on
Cu0.5Mg5CeOx is shown in Fig. 12. The fraction of the ad-
sorbed ethanol that desorbs near 350 K (Fig. 12) is likely to
arise from recombination of ethoxide with H species on the
surface. Ethanol appears to adsorb dissociatively on ZnO,
MgO, and Cu to form surface ethoxide and hydrogen (41,
45–47). Undissociated ethanol desorbs at lower tempera-
tures. H2 desorbs between 373 and 523 K, leaving ethoxide
species stranded on the catalyst surface. The H2 desorption
temperature decreases and the amount evolved increases
Step II:Alkyldecomposition to form H2, H2O and carbon
on the surface. These surface carbon species can be oxidized
byoxygen to produce carbon oxidesasproposed byBarteau
et al. on ZnO (47 ).
== ==
Steps III, IV: Ketene (CH2
C O), observed on ZnO
(53) and on Cu(110) single crystals above 573 K (52), might
form in these two steps.
Step V: CH2COO(a) surface species, formed by -H ab-
== ==
C O, leading
straction, attacksthe carbonylgroup ofCH2
to the formation of acetone. A similar mechanism has been
proposed for the formation of acetone from acetic acid on
TiO2 (54).
–
when Cu is present (Fig. 12), as expected from the H H re-
combination function provided byCu and from the mobility
of H-atoms on oxides near room temperature (48, 49). The
CO is one of the reactants in the conversion of synthesis
broad nonsymmetric peak of H2 suggests the presence of gas to isobutanol. Therefore, we have also examined how
nonuniform desorption sitesor multiple sourcesofH-atoms CO influences the rate and selectivity in base-catalyzed
(Fig. 12b), because second-order recombinative desorp- aldol condensation reactions of preadsorbed alcohols.
tion processes would lead to symmetric H2 evolution peaks TPSR studies of preadsorbed ethanol were carried out
–
(50, 51).
on K Cu0.5Mg5CeOx using a carrier gas consisting of 5%
Acetaldehyde is formed by additional dehydrogenation CO in He. As mentioned earlier, the presence of K in
of adsorbed ethoxide species as hydrogen is depleted from Cu0.5Mg5CeOx decreases the number of surface exposed
the catalyst surface by Cu-assisted hydrogen desorption Cu atoms, consistent with the lower rate of H2 production
–
steps. Some acetaldehyde desorbs without further reaction, on K Cu0.5Mg5CeOx compared to that on Cu0.5Mg5CeOx
but some adsorbed aldehydic species undergo secondary (Figs. 12 and 13). Moreover, the maximum peak temper-
reactions, such as coupling, dehydration, decarbonylation, ature for acetone production shifted to higher values on
–
and decarboxylation reactions to form acetone (Fig. 12). K Cu0.5Mg5CeOx. It appears that the presence of K de-
The CeOx lattice is the likely source of oxygen atoms in- creases the surface concentration of acetone precursors,
volved in the decarboxylation of aldol intermediates to which in turn, causes an increase in maximum peak tem-
form acetone. The reverse aldol condensation reactions peratures for second-order reactions (50).

ISOBUTANOL AND METHANOL SYNTHESIS
145
FIG. 12. Temperature-programmed surface reactions of preadsorbed ethanol on (a) Mg5CeOx, (b) Cu0.5Mg5CeOx. [100 cm3 (STP)/min He; heating
rate, 0.5 K/s; 33.4 mg catalyst; ethanol was preadsorbed at room temperature for 10 min followed by He purging for 30 min.]
The presence of CO in the carrier gas enhanced the for- and the possible stabilization of unreactive Cu + species by
mation of H2 and CO2 via water–gas shift reactions above CeO2 (15, 16).
523 K, markedly decreased acetone formation, and inhib-
The presence ofCO duringTPSR ofpreadsorbed ethanol
ited the evolution of H2 in the low-temperature desorption decreases the amount of acetone produced, possibly be-
peak (Fig. 13). The inhibited evolution of H2 at low tem- cause basic sites responsible for the coupling of adsorbed
peratures suggests that fewer Cu surface atoms are avail- acetaldehyde species are also blocked by the CO2 formed
able when CO is present in He carrier gas. Cu sites may from CO in water–gas shift or Bouduard reactions. The low
be blocked by oxygen atoms formed by reactions of CO2 concentration of surface acetaldehyde (a precursor to ace-
–
(formed from CO) with Cu surfaces on K Cu0.5Mg5CeOx. tone) and the smaller number of available basic sites appear
Cu clusters on Mg5CeOx appear to be more sensitive to ox- to be responsible for the marked decrease in acetone pro-
idation (by CO2) both during catalytic CO hydrogenation duction when CO is present in the He carrier gas. More de-
and duringethanolTPSR because oftheir smallparticle size tailed TPSR results and reaction pathways of preadsorbed
–
FIG. 13. Temperature-programmed surface reactions of preadsorbed ethanol on K Cu0.5Mg5CeOx in (a) helium flow and (b) 5% CO in He flow.
[100 cm3 (STP)/min He; heating rate, 0.5 K/s; 33.4 mg catalyst; ethanol was preadsorbed at room temperature for 10 min followed by He purging for
30 min.]

146
XU ET AL.
alcohols and aldehydes on these samples will be reported
elsewhere (15).
7. Apesteguia, C. R., Soled, S. L., and Miseo, S., U.S. Patent 5,387,570
(1995). [Assigned to Exxon Research and Engineering Co.]
8. Apesteguia, C. R., DeRites, B., Miseo, S., and Soled, S. L. Catal. Lett.
44, 1 (1997).
9. Tsuji, H., Yagi, F., Hattori, H., and Kita, H., J. Catal. 148, 759
(1994).
CONCLUSIONS
10. Radlowski, C. A., and Hagen, G. P., U.S. Patent 5,159,125. [Assigned
to Amoco Corporation, Chicago]
11. Brideger, G. W., and Spencer, M. S., in “Catalyst Handbook,” 2nd ed.
(M. V. Twigg, Ed.). Wolfe Press, London, 1989.
12. Li, J.-L., and Inui, T., Appl. Catal. A 137, 105 (1996).
13. Iglesia, E., and Boudart, M., J. Catal. 81, 204 (1983).
14. Narita, K., Takeyawa, N., and Toyoshima, I., React. Kinet. Catal. Lett.
19 (1982).
K-promoted Cu0.5Mg5CeOx catalysts are active and se-
lective for isobutanol synthesis from CO/H2 mixtures with
high alcohol-to-hydrocarbon ratios at relatively low tem-
peratures (583 K) and pressures (4.5 MPa). CO2 decreases
the rates of methanol and isobutanol synthesis from CO/H2
–
on catalysts with low Cu concentration (K Cu0.5Mg5CeOx).
Methanol production was not affected by CO2 on catalysts
15. Xu, M., and Iglesia, E. [Unpublished manuscript]
16. Lamonier, C., Bennani, A., Duysser, A., and Wrobel, G., J. Chem. Soc.
Faraday Trans. 92(1), 131 (1996).
17. Soria, J., Solid State Ionics 63–65, 755 (1993).
18. Driscoll, D. J., Martin, W., Wang, J-X., and Lunsford, J. H., J. Am.
Chem. Soc. 107, 58 (1985).
19. Lunsford, J. H., Cisneros, M. D., Hinson, P. G., Tong, Y., and Zhang,
H., Faraday Discuss. Chem. Soc. 87, 1 (1989).
20. King, D. S., and Nix, R. M., J. Catal. 160, 76 (1996).
21. Souza-Monteiro, R., Noronha, F. B., Dieguez, L. C., and Schmal, M.,
Appl. Catal. 131, 89 (1995).
22. Garcia-Fierro, J. L., Lo Jacono, M., Inversi, M., Porta, P., Cioci, F., and
Lavecchia, R., Appl. Catal. 137, 327 (1996).
23. Garcia-Fierro, J. L., Lo Jacono, M., Inversi, M., Porta, P., Lavecchia,
R., and Cioci, F., J. Catal. 148, 709 (1994).
24. Annesini, M. C., Lavecchia, R., Marrelli, L., Lo Jacono, M., Campa,
M. C., Garcia-Fierro, J. L., Moretti, G., and Porta, P., Solid State Ionics
63–65, 281 (1993).
25. Campos-Martin, J. M., Garcia-Fierro, J. L., Guerrero-Ruiz, A.,
Herman, R. G., and Klier, K., J. Catal. 163, 418 (1996).
26. Davydova, L. P., Boreskov, G. K., Popovskii, V. V., Yurieva, T. M., and
Anufrienko, V. F., React. Kinet. Catal. Lett. 18, 203 (1981).
27. Aleksandrov, V. Yu., Popovskii, V. V., and Bulgakov, N. N., React.
Kinet. Catal. Lett. 8, 65 (1978).
–
–
with high Cu (K Cu7.5Mg5CeOx and Cs Cu/ZnO/Al2O3)
because they operate near methanol synthesis equilibrium.
CeO2 is a structural promoter increasing both surface area
and Cu dispersion. The addition of K to CuMgCeOx sam-
plesdecreasesCu surface area and methanolsynthesisrates,
but titrates residual acid sites leading to the undesired for-
mation of dimethylether and hydrocarbons during alcohol
synthesis.
A
13CO2/12CO2 isotopic switch method was used to probe
basic site density and strength and the kinetic availability
of basic sites at temperatures typical of isobutanol synthe-
sis reactions. The addition of K increases basic site den-
sity and strength and the areal rates of aldol condensation
chain growth reactions. Ethanol dehydrogenation reaction
is a structure-insensitive reaction that requires Cu surface
atoms. Cu sites play a critical role in condensation reac-
tion. Alcohol chain growth reactions occur via bifunctional
(metal–base) pathways requiring both Cu and basic sites.
Isotopic tracer and TPSR studies of alcohol coupling reac-
tions suggest that chain growth leading to propionaldehyde
and isobutyraldehyde occurs by condensation reactions in-
volving the addition of a C1 species to adsorbed oxygenates.
28. Xu, M., Hu, Z., Gines, M. J. L., and Iglesia, E. [Unpublished
manuscript]
29. Boudart, M., and Dje´ga-Mariadassou, G., “Kinetics of Heterogeneous
Catalytic Reactions.” Princeton Univ. Press, Princeton, NJ, 1984.
30. DePontes, M., Yokomizo, G. H., and Bell, A. T., J. Catal. 104, 147
(1987).
ACKNOWLEDGMENTS
31. McKenzie, A. L., Fishel, C. T., and Davis, R. J., J. Catal. 138, 547
(1992).
32. Nunan, J. G., Himmelfarb, P. B., Herman, R. G., Klier, K., Bodgan,
C. E., and Simmons, G. W., Inorg. Chem. 28, 3868 (1989).
33. Klier, K., Chatikavanij, V., Herman, R. G., and Simmons, G. W.,
J. Catal. 74, 343 (1982).
34. Vedage, G. A., Himmelfarb, P. B., Simmons, G. W., and Klier, K., ACS
Symp. Ser. 279, 295 (1985).
This work was supported by Division of Fossil Energy, the United States
Department of Energy, under Contract DE-AC22-94PC94066. The au-
thors thanks Mr. Zhengjie Hu for his technical assitance with isotopic
switch experimentsand Dr. Bernard A. Toseland (Air Productsand Chem-
icals, Inc.) for helpful technical discussions and suggestions.
35. Chinchen, G. C., Waugh, K. C., and Whan, D. A., Appl. Catal. 25, 101
(1986).
36. Tronconi, E., Ferlazzo, N., Forzatti, P., and Pasquon, I., Ind. Eng. Chem.
Res. 26, 2122 (1987).
REFERENCES
37. Elliot, D. J., J. Catal. 111, 445 (1988).
1. Keim, W., and Falter, W., Catal. Lett. 3, 59 (1989).
2. Keim, W., D.E. Patent 3810421 (1988).
3. Lietti, L., Forzatti, P., Tronconi, E., and Pasquon, I., J. Catal. 126, 401
(1990).
4. Forzatti, P., Tronconi, E., and Pasquon, I., Catal. Rev. Sci. Eng. 33, 109
(1991).
38. Calverley, E. M., and Smith, K. J., J. Catal. 130, 616 (1991).
39. Xu, M., and Iglesia, E., J. Phys. Chem. [Submitted for publication]
40. Elliot, D. J., and Pennella, F., J. Catal. 119, 359 (1989).
41. Takezawa, N., Hanamaki, C., and Kobayashi, H., J. Catal. 38, 101
(1975).
42. Nunan, J. G., Bogdan, C. E., Klier, K., Smith, K., Young, C., and
Herman, R., J. Catal. 116, 195 (1989).
43. Gines, M. J. L., and Iglesia, E. [Unpublished manuscript]
5. Sofianos, A., Catal. Today 15, 119 (1990).
6. Slaa, J. C., van Ommen, J. G., and Ross, J. R. H., Catal. Today 15, 129
(1992).

ISOBUTANOL AND METHANOL SYNTHESIS
147
44. Nunan, J. G., Bogdan, C. E., Herman, R. G., and Klier, K., Catal. Lett. 50. Rieck, J. S., and Bell, A. T., J. Catal. 85, 143 (1983).
2, 49 (1989).
51. Somorjai, G. A., “Chemistry in Two Dimensions: Surfaces,” Cornell
Univ. Press, Ithaca, NY, 1981.
45. Bowker, M., and Madix, R. J., Surf. Sci. 116, 549 (1982).
46. Bowker, M., Houghton, H., and Waugh, K. C., J. Chem. Soc. Faraday 52. Bowker, M., and Madix, R. J., Appl. Surf. Sci. 8, 299 (1981).
Trans. 1 78, 2573 (1982).
53. Bowker, M., Houghton, H., and Waugh, K. C., J. Catal. 79, 431
47. Vohs, J. M., and Barteau, M. A., Surf. Sci. 21, 590 (1989).
(1983).
48. Vannice, M. A., Boudart, M., and Fripiat, J. J., J. Catal. 17, 359 (1970). 54. Gonzalez, F., Munuera, G., and Prieto, J. A., J. Chem. Soc. Faraday
49. Levy, R. B., and Boudart, M., J. Catal. 32, 304 (1974). Trans. 1 74, 1517 (1978).