A study of methyl formate production from carbon dioxide hydrogenation in methanol over a copper zinc oxide catalyst

Article abstract of DOI:10.1007/s10562-010-0500-3

A preliminary study of the production of methyl formate (MF) from CO ₂ hydrogenation in liquid methanol was carried out over a Cu/ZnO/Al₂O₃ based catalyst which was synthesized by a precipitation technique following a well established route. The effects of amine concentration, hydrogen pressure, temperature, CO and water addition on the activity and selectivity of MF were investigated. It is of interest to note that the addition of 1% trimethylamine can dramatically increase the initial turnover frequency with the MF being the major product. It is evident that the formation of CO₂-amine adduct promotes the catalytic hydrogenation of CO₂ on the surface of the catalyst.

Full text of DOI:10.1007/s10562-010-0500-3

Catal Lett (2011) 141:259–265
DOI 10.1007/s10562-010-0500-3
A Study of Methyl Formate Production from Carbon Dioxide
Hydrogenation in Methanol over a Copper Zinc Oxide Catalyst
K. M. Kerry Yu Shik Chi Tsang
•
Received: 30 April 2010 / Accepted: 9 November 2010 / Published online: 21 November 2010
Ó Springer Science+Business Media, LLC 2010
Abstract A preliminary study of the production of
methyl formate (MF) from CO hydrogenation in liquid
straightforward approach to this goal and has attracted
additional attention as a possible method of reversibly
storing hydrogen. There has been, however, limited work
in the literature reporting attempts to catalyze formic acid
formation from CO and H under industrially applicable
2
methanol was carried out over a Cu/ZnO/Al O based
2
3
catalyst which was synthesized by a precipitation technique
following a well established route. The effects of amine
concentration, hydrogen pressure, temperature, CO and
water addition on the activity and selectivity of MF were
investigated. It is of interest to note that the addition of 1%
trimethylamine can dramatically increase the initial turn-
over frequency with the MF being the major product. It is
2
2
conditions, although this process has been well recognized
as potentially resulting in a significant advance for the
chemical industry [1]. Most previous attempts were of
limited industrial interest as they either suffered from poor
product yield and low selectivity, or they were tested under
conditions far from those of relevance for industrial
application [2, 3].
evident that the formation of CO -amine adduct promotes
2
the catalytic hydrogenation of CO on the surface of the
2
catalyst.
It has been previously noted that the initial product
under typical methanol synthesis conditions from a H , CO
2
Keywords Carbon Dioxide ꢀ Catalyst ꢀ Fine Chemicals ꢀ
Formate ꢀ Hydrogenation ꢀ Trimethylamine
and CO gas mixture over a Cu colloid system was methyl
2
formate (MF) rather than formic acid [4]. Subsequently, we
have clearly demonstrated the catalytic hydrogenation of
CO2 to MF in the presence of liquid methanol by
employing a Pd/Cu/ZnO/Al O nanocatalyst system [5].
1
Introduction
2
3
The quantity of the MF produced was found to greatly
exceed the maximum amount of surface species covering
the catalyst by many folds ([45 times). This clearly sug-
gests that the catalyst was capable of activating gaseous
Chemical utilization of CO2 has attracted considerable
attention as a possible route for the manufacture of useful
chemicals. In the last few years carbon dioxide has become
the focus because of its position as the primary greenhouse
gas and the implication of its emissions on the problems of
climate change. Therefore, advances in the chemistry of
CO to produce high yields ([20%) of condensable MF
2
with excellent selectivity ([96%) over many catalytic
cycles. Mechanistic studies have indicated that the reverse
water gas shift reaction (CO ? H ) generates adsorbed
CO including its activation and its utilization will be a
2
2
2
necessary component of the development of a sustainable
formate intermediate species on the catalyst which further
couple with methanol (the solvent molecule) to produce
methyl formate. Thus, for the high yield synthesis of for-
mate/formic acid, the key strategy is to couple the surface
formate species with methanol before it shifts to CO and
H O. CO is known to form stable adducts with amine
economy in our society. The hydrogenation of CO to
2
formic acid or related species appears to be
a
K. M. Kerry Yu ꢀ S. C. Tsang (&)
2
2
Wolfson Catalysis Centre, Department of Chemistry, University
of Oxford, Oxford OX1 3QR, UK
e-mail: edman.tsang@chem.ox.ac.uk
molecules which hence may facilitate its capture. The use
of a CO -amine adduct, rather than CO , may therefore
2
2
123

2
60
K. M. Kerry Yu, S. C. Tsang
result in higher apparent catalytic activity. Here we report a
preliminary study of the effect of including an amine
additive for the above reaction. We have also investigated
the effects of the inclusion of various chemical species
under different conditions in order to acquire some
understanding of the reaction mechanism, which may lead
to rational design of optimised catalyst, reactor and
experimental conditions for this new reaction. Our initial
findings are detailed below.
Table 1 The average compositions (Al, Cu and Zn) of the copper
zinc oxide catalysis after a detailed EDX analysis with 5 randomly
selected areas after taking the ZAF correction into account
Al/atomic % Cu/atomic % Zn/atomic %
Site of Interest 1
Site of Interest 2
Site of Interest 3
Site of Interest 4
Site of Interest 5
Average atomic %
Standard Deviation
16.13
14.07
16.11
13.69
15.22
15.04
1.13
22.77
25.41
20.79
25.25
23.10
23.46
1.92
14.06
15.59
12.89
15.41
14.72
14.53
1.10
2
Experimental
Copper zinc oxide catalyst was prepared by a co-precipi-
tation method. Typically, 3.03 g of Cu(NO ) ꢀ2.5H O,
temperature (150 °C). The extent of the reaction was fol-
lowed once hydrogen was charged into the reactor. Product
gases in the autoclave were carefully released from a gas
port and directed to an online GC via a heated (*140 °C)
gas transfer line for gas phase analysis. Generally, the line
was adequately flushed out prior to sample injection. After
the injection was performed, nitrogen was allowed to flush
the heated transfer line in preparation for the next injection.
Also, 3 9 60 mL gas samples were collected before and
after testing for injection into an off-line GC equipped with
a TCD to provide an independent analysis. Following
reaction, the pressure in the autoclave reactor was slowly
released after it was cooled in a dry ice-acetone mixture (-
75 °C). Off-line GC–MS was also used to verify the
chemical composition of the condensable products, and
showed that only methyl formate was the main condens-
able product. Quantitative analysis of the methyl formate in
3
2
2
2
.40 g of Zn(NO ) and 3.15 g of Al(NO ) ꢀ9H O were
3
2
3 2
2
dissolved in 500 mL water. The solution was stirred,
heated to 80 °C in a round bottom flask fitted with a con-
denser and equilibrated with 20 mL/min air flowing
through the flask for 5 min. An aqueous solution of
Na CO (15 g in 200 mL water) was added to the contin-
2
3
uously stirred solution until pH 7.0 was attained. At this
stage, air was passed through the solution and the precip-
itate was allowed to age for two hours while the pH was
maintained constant at 7.0. After aging, the precipitate was
recovered by filtration and washed with 50 mL of hot
double DI water (* 80 °C) 10 times prior to being dried at
1
20 °C in air for 16 h followed by calcination at 350 °C in
air for 12 h. Selected area energy dispersive x-ray (EDX)
analysis indicated the absence of significant levels of
?
residual of Na in the product. Pre-reduction of the
1
material was undertaken in a temperature programmed
furnace which was ramped from room temperature up to
the condensate after the reaction was carried out by H-
NMR.
2
2
50 °C at 5 °C/min and held at 250 °C for 2 h under a
It was necessary to perform pre-calibrations for both gas
1
(by GC-FID) and liquid (by H-NMR) separately. For this,
0 mL/min flow of 10% H /N . The resulting catalyst was
2
2
retained in the sample tube under the 10% H /N atmo-
2
a series of known concentrations of methyl formate in
2
sphere. BET surface area analysis was performed using
isothermal adsorption of N2 at -196 °C. The reduced
copper zinc oxide catalyst was found to have a specific
reaction mixture were used (methanol, amine, H
and CO
2 2
using CH as the internal standard). For the gas analysis,
4
analysis of the gas-phase was performed five times by GC-
FID at the reaction temperature before the reactor was
cooled down by dry ice-acetone prior to discharge of the
gas. Subsequently, liquid sampling was conducted at 0 °C
where 1.20 mL of the liquid sample was mixed with
2
surface area of 67 m /g. Elemental analysis of the material
after reduction was performed by EDX analysis where a
atomic ratio of Al:Cu:Zn = 1.0:1.6:1.0 was found as
summarized in Table 1. This generally matches the ratio of
the starting precursor salts added.
0.80 mL 99.8% CDCl (with 0.2% CHCl as the external
3 3
1
Catalyst testing was carried out in a 100 mL stainless
steel Parr autoclave reactor. Typically, 0.20 g of catalyst
and 10 mL methanol (with and without amine additive)
were placed in the reactor, which was then sealed. The
reactor was then flushed with pure methane. This gas was
used as the internal standard (1 bar above atmosphere
pressure). Dried CO2 and then H2 were then pumped
accordingly into the autoclave reactor to reach their desired
pressure (total pressure = 160 bar) at the reaction
standard) as the solvent for the H-NMR analysis. As a
result, the data obtained from gas phase GC-FID analysis
can be combined with the data obtained from liquid phase
1
H-NMR analysis to provide a comprehensive chemical
analysis within the calibrated range. In our previous com-
munication [5], we reported a satisfactory carbon mass
balance for this reaction. For the triethylamine promotion
study, a nitrogen mass balance of over 95% was obtained
(from NMR). There was also no indication that the amine
1
23

A Study of Methyl Formate Production
261
was converted into different chemical species other than
1
carbamate according to our H-NMR analysis. Also, no
Thus, a mixture of a simple amine molecule, triethylamine,
and methanol was initially attempted in this work.
other gaseous product, for instance, ethane from triethyl-
amine, was observed in addition to the methanol, methyl
Figures 1a and 1b summarize the results obtained from
this study. From the Fig. 1a, it is interesting to observe that
1v/v% amine/methanol mixture gives the highest MF
production in 200 min of reaction. It would be desirable to
compare the initial turnover frequencies under various
levels of the amine loadings. However, we do not yet know
whether copper or copper-oxygen is the active phase under
our experimental conditions (some degree of oxygen cov-
erage on Cu which may assist the adsorption of amine).
Instead of using conventional methods for estimation of
surface areas of copper metal in our catalysts (i.e. nitrous
oxide decomposition to yield nitrogen), we adopted a more
simple estimation of ‘copper’ dispersion. The estimation of
the number of the copper sites was first based on the XRD
data to obtain the average copper particle size as small
spheres. By taking the total copper content from the cata-
lyst measured by EDX into account, the number of copper
sites can therefore be worked out. It should be noted that
this method may ignore the contributions from highly
dispersed copper or copper-oxygen species where the
severe peak broadening of the XRD prohibits their size
estimation. Ressler et al. [13] have recently stressed the
complex micro-structural interactions between copper and
zinc oxide, which is difficult to be standardized by any
simple method. Nevertheless, our XRD characterization
indicated that 15 nm was the average particle size. As a
result, 9.1% dispersion of Cu was deduced assuming mono-
dispersion and spherical morphology. This also means that
a maximum of 9.1% of the overall Cu atoms was available
formate and methane (where CH is the internal standard.)
4
3
Results and Discussion
It was found from the previous feasibility study [5] that
methanol can couple with formate species adsorbed on the
catalyst to produce MF. Esterification of formate species
with methanol possibly catalyzed by acid sites on the cat-
alyst was a probable pathway on the copper zinc oxide
surface under the reaction conditions applied. In this study,
one aim was to determine the effect of addition of a second
solvent, amine, to methanol upon catalytic performance.
The capture of carbon dioxide with amine molecules is a
mature technology, which is currently applied in power
plants based on combustion of non-renewables. By scrub-
bing the flue gases to form carbamic acid derivatives using
primary/secondary amines (Eq. 1) [6, 7] or to form trial-
kylammonium monoalkyl carbonate using tertiary amine
(
Eq. 2 ), CO can be captured from the gas-phase to form a
2
soluble species [8–10].
O
C
H
R1
R1
N
+
CO2
N
O
H
R2
R2
ð1Þ
for the CO activation reaction. EDX spectra showed that
2
the Cu loading was 23.46 atomic % which corresponded to
a Cu loading of 41.38 weight %. Therefore, there was
R'OCO - R NH+
2 3
R N
3
+
CO2
+
R'OH
0
.1185 mmol of surface Cu per 0.20 g of catalyst used in
ð2Þ
The basis of this chemistry is also used for separating
each reaction and this was subjected to different concen-
trations of amine. Thus, the initial TOF of the reaction was
evaluated based on the amount of methyl formate pro-
duced/amount of activated copper sites for the first 20 min
on-stream where time zero was defined as being the time at
CO from industrial waste streams [11]. The above reac-
2
tions are reversible (CO2 is evolved by heating to
8
0–100 °C for simple amines) and the carbamic acid
derivatives and intermediates have been characterized in
some simple systems by a number of methods including X-
ray and electron diffraction [12]. The formation of car-
bamic acid derivatives/trialkylammonium monoalkyl car-
bonate as liquid species may facilitate further catalytic
hydrogenation of the gaseous carbon dioxide. The transi-
tory nature of these derivatives and their interaction(s) with
solid catalyst material appear to be rather complex. Also,
there is very limited literature, if any, in this area. Never-
theless, a number of different simple amine molecules can
which the CO was charged.
2
Figure 1a indeed clearly shows that a highest TOF
-
1
towards MF (43.6 h ) when 1% amine is used than those
without the amine inclusion. But, higher amine concen-
tration gives lower TOFs.
It was originally proposed [5] that the formation of
methyl formate proceeds via the reaction between one of
the intermediates, which is likely to be the surface formate,
during the reverse water gas shift reaction and methanol
(see Scheme 1).
react with gaseous CO to form reactive intermediates, and
2
In order to reduce the extent of the reverse water gas
shift reaction, a low reaction temperature at 150 °C was
these have not been explored to any significant degree.
123

2
62
K. M. Kerry Yu, S. C. Tsang
Fig. 1 a A plot of catalytic
activity for Cu/ZnO/Al
a 12
-
1
Triethylamine Effect
2
O
3
100% amine ( 0 h
)
catalyst over different % of
amine quoted with initial TOF
values in brackets. (Reaction
-1
1
0
8
6
4
2
0
1
4% amine ( 30.8 h
)
)
-
-
-
1
1
1
7% amine ( 33.1 h
% amine ( 43.6 h
0% amine ( 28.2 h
conditions: 150 °C, 20 bar H
2
,
1
)
)
1
40 bar CO , MF production
2
was monitored continuously by
online GC-FID for initial
3
00 min; TOF towards MF
production was determined at
0 min of reaction time). b A
plot of catalytic activity for Cu/
ZnO/Al catalyst in terms of
2
0
20
40
60
80
100
120
140
160
180
200
2
O
3
Time / minutes
total methyl formate and CO
productions with increasing
amine concentrations at the end
of 1500 min reaction time.
b
14
12
1.00
MF Production
CO Production
0
.80
(
2
Reaction conditions: 150 °C,
0 bar H , 140 bar CO
1
0
8
6
4
2
0
2
2
,
reaction products monitored by
GC-FID and NMR after
1
0.60
500 min reaction)
0
0
.40
.20
0.00
0
1
7
14
100
%
amine
Scheme 1 Possible surface
reactions taking place under the
catalyzed reaction conditions
maintained in this study since the reverse water gas shift
reaction is favoured by a high reaction temperature.
Figure 1b shows a plot of total MF and CO formation
at various amine concentrations after 1500 min reaction
time. It is clearly demonstrated that when there was no
methanol in the reaction mixture (i.e. for pure triethyl-
amine), no methyl formate was produced but CO forma-
tion was evident. The formation of CO was the product of
the reverse water gas shift reaction between CO and H .
though the reverse water gas shift reaction pathway to
produce CO and H O.
2
In addition, as seen in Fig. 1b, the amount of MF col-
lected after 1500 min reaction progressively decreases
when higher concentrations of amine are applied. It is
logical to assume that when the reaction is taking place at
high amine content, there is a concomitant decrease in the
extent of esterification on the catalyst’s surface. For an
optimum catalytic rate, there should be an appropriate
2
2
Thus, the coupling reaction for the formation of methyl
formate depends on the availability of methanol in the
reaction mixture. Figure 1b also shows that there is a
clear increase in MF production at the 1% amine level but
a further increase in amine concentration from 1 to 14%
balance between the adsorption of CO precursor (car-
2
bamic acid derivative) to form surface formate, and
methoxy, generated from methanol adsorption, on the
catalyst surface. The strong adsorption of amine due to the
nitrogen electron lone pair availability on the metal surface
may account for its negative effect on the surface esterifi-
cation reaction, which could particularly attenuate the
adsorption of methanol leading to the higher degree of CO
formation from reverse water gas shift over the surface
esterification reaction. It is clear that when pure amine was
used there was no production of MF and CO was the only
detected carbon containing product (Fig. 1b). Also, no
methyl formate, HCHO or HCOOH was identified from our
(
a decrease in methanol availability) clearly promotes the
formation of CO. We would anticipate that a higher
amine concentration would provide a higher concentration
of activated CO as the amine-adduct to supply a larger
2
quantity of formate intermediate species on the catalyst
surface. To account for the rapid promotion of CO for-
mation, one possibility is that the slow surface coupling
reaction may not be rapid enough to cope with the surplus
in formate species, and that this excess formate proceeds
1
GC-FID, GC–MS and H-NMR analyses in this case.
1
23

A Study of Methyl Formate Production
263
In this type of hydrogenation, hydrogen adsorption
could be limited under the reaction conditions particularly
when catalyst poisons such as water and CO during the
production of MF are produced. Therefore, different
hydrogen pressures were investigated to determine the
optimum value in the absence of amine and the results were
shown in Fig. 2.
16
14
Water Effect
1
1
2
0
8
6
4
2
0
0
1
2
% water
% water
% water
It was found that 30 bar hydrogen was optimal under the
current reaction conditions for the reaction rate and MF
yield. Thus, the catalyst might indeed have suffered from
insufficient availability of hydrogen at 20 bar, but inter-
estingly a higher pressure of hydrogen at 40 bar led to
poorer activity.
0
200
400
600
800
1000 1200 1400 1600
Time / minutes
Fig. 3 (Left). A plot of activity for Cu/ZnO/Al
with different initial water contents against time. (Reaction condi-
, topped up with CO to 160 bar total
2
O
3
catalyst at 150 °C
It was also thought that the catalyst surface could be
poisoned by ‘carbonate’ species from the reaction between
CO and H O since carbonic acid can be formed if any
tions: 150 °C, 30 bar H
2
2
pressure, reaction products monitored by online GC-FID for
1500 min reaction)
2
2
water by-product was produced under our elevated pressure
conditions. Thus, the Cu/ZnO/Al O catalyst was tested
bidentate formate and bicarbonate species were formed on
the catalyst surface (Fig. 4). It is noted that these surface
species can either be on single metal, M site, as shown in
the figure or bridged between two M sites. In order to
reduce the negative effect due to the formation of water
2
3
under water enriched reaction conditions (1 and 2% v/v
water in 10 mL methanol) and the results are presented in
Fig. 3.
Figure 3 shows clearly that the activity was strongly
affected by the presence of water in terms of the TOF and
during the reaction, a 1% w/w Pd modified Cu/ZnO/Al O
2 3
final MF yield. It was observed that water and CO indeed
2
catalyst was prepared by a traditional wet impregnation
technique where palladium (II) nitrate was used as the
precursor salt. It is generally known in the literature that Pd
can provide an effective hydrogen spillover surface which
may be able to reduce the poisoning effect arising from
surface carbonaceous species. The 1% w/w Pd modified
produced surface carbonate which blocked the catalyst
sites. As demonstrated in Fig. 4, FT-IR spectrum A was
obtained from the catalyst after the reaction and spectrum
B was obtained from the catalyst before the reaction.
Spectrum C was the difference spectrum determined by
subtraction of spectrum B from spectrum A. As seen in
Cu/ZnO/Al O catalyst was tested under water enriched
2 3
-
1
spectrum C, 1588 and 1372 cm adsorption peaks are
reaction conditions (2% v/v water in methanol) and the
catalytic performance was shown in Fig. 5.
apparent which correspond well to the tas (CO ) and d
2
(
CH) vibrations of the surface bidentate form of adsorbed
Figure 5 shows that the Pd modified Cu catalyst was not
as active as the non-modified Cu catalyst. Presumably the
Pd atoms reduced the effective surface of area of the
copper zinc oxide catalyst. However, the negative effect of
water was clearly attenuated using the Pd promoter. It
should be particularly noted that the Pd modified Cu cat-
alyst does not show a rapid level-off in catalytic activity in
the absence of added water even after the 1500 min reac-
tion. However, the catalyst still suffered a degree of
deactivation as the catalytic performance was lowered after
-
formate’. The observed 1432 and 1291 cm peaks cor-
1
‘
respond to the t (CO ) and d (OH) vibrations of surface
bicarbonate. Thus, the results suggest that adsorbed
s
3
Hydrogen Effect
1
1
1
1
6
4
2
0
8
6
4
2
0
4
20 min. Higher reaction temperature (190 °C) was
2
3
0 bar
0 bar
investigated (Fig. 6) in order to see whether a higher rate
40 bar
could be obtained.
As shown in Fig. 6, the initial reaction rate was greatly
improved at 190 °C but the final MF yield was still leveled
at *10 mmol.
0
200
400
600
800
1000 1200 1400 1600
Time / minutes
The unselective CO species produced from the reverse
water shift reaction may also act as a poison leading to
catalyst deactivation. In order to determine its effect, 1
atmosphere CO was introduced into the reaction with Cu/
ZnO/Al O as the catalyst.
Fig. 2 A plot of activity for Cu/ZnO/Al
hydrogen pressure against time. (Reaction conditions: 150 °C,
variations in H pressure as stated, topped up with CO to 160 bar
total pressure, reaction products monitored by online GC-FID for
500 min reaction)
2 3
O catalyst with different
2
2
1
2
3
123

2
64
K. M. Kerry Yu, S. C. Tsang
H
C
OH
C
14
Temperature Effect
1
1
2
0
190 °
C
O
O
O
O
M
M
bidentate formate
bicarbonate
8
°
6
4
2
0
150
C
A
1
432
0
50
100
150
200
250
300
350
400
1
291
1
588
1372
Time / minutes
Fig. 6 A plot of activity for 1% w/w Pd modified Cu/ZnO/Al
catalyst at different temperature against time. (Reaction conditions:
0 bar H , topped up with CO to 160 bar total pressure, reaction
products monitored by online GC-FID for 1500 min reaction)
2 3
O
C
B
3
2
2
1
1
1
1
6
4
2
0
CO Effect
1
600
1500
1400
1300
1200
Wavenumber / cm-1
Fig. 4 (Right). Typical FT-IR spectra of adsorbed carbonaceous
species found as intermediates of the CO activation. B is the
background spectrum (catalyst); A is the spectrum of adsorbed species
on the catalyst; and C is their differential spectrum
8
6
4
2
0
2
no CO
1 atm CO
0
200
400
600
800
1000 1200 1400 1600
1
1
1
1
6
4
2
0
Pd Effect under Water
Time / minutes
Fig. 7 A plot of activity for Cu/ZnO/Al
2
O
3
catalyst with and without
, topped up
with CO to 160 bar total pressure, reaction products monitored by
CO against time. (Reaction conditions: 150 °C, 30 bar H
2
8
2
6
4
2
0
online GC-FID for 1500 min reaction)
no initial water
no initial water with Pd modification
2% v/v initial water
catalytic reaction. In order to obtain an optimum catalytic
activity and selectivity an initial study of their influence
has been performed. Further optimization is necessary but
this initial study provides an indication of their relative
effects. Typically, the addition of triethylamine to metha-
nol is capable of enhancing the initial TOF for the MF
production. This is attributed to the facilitated delivery of
gaseous carbon dioxide to the catalyst surface in form of
2% v/v initial water with Pd modification
0
200
400
600
800
1000
1200
1400
1600
Time / minutes
Fig. 5 A plot of catalytic activity to compare the effect of with/
without initial water content for the 1% w/w Pd modified Cu/ZnO/
Al
2
O
3
catalyst against time. (150 °C, 30 bar H
2 2
, topped up with CO
to 160 bar total pressure, reaction products monitored by online GC-
FID for 1500 min reaction)
soluble CO -amine adduct. However, cautious control of
2
Figure 7 clearly demonstrates that a significantly lower
initial reaction rate and a poorer final MF yield were
observed when CO was included in the reaction feed. This
result clearly suggests that the non-selective production of
CO from reverse water gas shift reaction negatively affects
catalytic activity.
amine content is required to achieve maximum beneficial
effect. Thus, further studies utilizing different amines
should be carried out. It was also found in this study that an
appropriate H
hydrogenation of CO
pressure is important for the catalytic
2
in methanol. Higher hydrogen
2
pressure may reduce the adsorption of other species leading
to lower activity. The formation of water, which is pro-
duced from esterification of surface formate species with
methanol as well as from the unselective reverse water gas
shift reaction, can produce a strong negative effect on
catalytic activity. Its presence results in ‘carbonate’ species
on the surface of the catalyst, as evidenced in the FTIR
4
Conclusion
In summary, this preliminary study shows clearly that some
chemical species play an important role in this new
1
23

A Study of Methyl Formate Production
265
analysis. Doping 1wt% Pd onto the Cu/ZnO/Al O catalyst
2
4. Vukojevic S, Trapp O, Grunwaldt JD, Kiener C, Sch u¨ th F (2005)
Angew Chem Int Ed 44:7978
3
is clearly capable of attenuating this negative effect pre-
sumably by providing beneficial hydrogen spillover on
catalyst surface. The non-selective production of CO from
the reverse water gas shift reaction can also contribute to
the deactivation of the metal catalyst due to its strong
adsorption, and its formation should be avoided during
catalysis.
5
. Yu KMK, Yeung CMY, Tsang SC (2007) J Am Chem Soc
29:6360
1
6. Bonenfant D, Mimeault M, Hausler R (2003) Ind Eng Chem Res
42:3179
7
. Dell’Amico DB, Calderazzo F, Labella L, Marchetti F, Pampa-
loni G (2003) Chem Rev 103:3857
. Sasaki Y (1996) Chem Lett 825
8
9. Gray ML, Champagne KJ, Fauth D, Baltrus JP, Beckman E
2006) Preprints of Symposia - American Chemical Society. Div
Fuel Chem 51:107
(
Acknowledgments We are grateful to the UK EPSRC for sup-
porting this work.
10. Crooks JE, Donnellan JP (1990) J Org Chem 55:1372
1
1
1
1. Aresta M, Dibenedetto A (2003) Greenhouse Gas Control
Technologies. In: Proceedings of the 6th international conference
on greenhouse gas control technologies, Kyoto, Japan, p 1599
2. Aresta M, Ballivet-Tkatchenko D, Dell’Amico DB, Bonnet MC,
Boschi D, Calderazzo F, Faure R, Labella L, Marchetti F (2000) J
Chem Soc Chem Commun 2000:1099
References
1
2
3
. Behr A, Ebbinghaus P, Naendrup F (2003) Chem Ing Tech
5:877
. Gassner F, Leitner W (1993) J Chem Soc Chem Commun
993:1465
. Graf E, Leitner W (1992) J Chem Soc Chem Commun 1992:623
7
3. Kniep BL, Ressler T, Rabis A, Girgsdies F, Baenitz M, Steglich
F, Schlogl R (2004) Angew Chem Int Ed 43:112
1
123