Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature CO and CO2 hydrogenation to methanol

Article abstract of DOI:10.1016/j.cattod.2015.08.021

Ester hydrogenolysis is a key step in the production of alcohols and new catalysts are being sought to improve the performance and sustainability of this process. Research described in this paper investigated the use of Cu, Mo₂C, and Mo₂N-based catalysts for the hydrogenolysis of ethyl formate at 105-150 °C and 30 bar H₂. The high surface area Mo₂C-based catalysts were more active and selective to methanol than the Mo₂N-based catalysts. The deposition of nanoscale Cu and/or Pd particles onto the Mo₂C or Mo₂N resulted in significant enhancements in the methanol selectivities. The methanol selectivities varied with the acidity or basicity of the support suggesting some degree of cooperation between the Cu or Pd particles and the support. Temporal variations in rates for the Cu/Mo₂C catalyst were consistent with first order kinetics with regard to ethyl formate and the activation energy was 44 kJ/mol. The best performance was achieved over a Pd-Cu/Mo₂C catalyst; this material outperformed, by a significant margin, the oxide supported Cu catalysts. Given their high activities and selectivities, the Mo₂C supported catalysts are attractive candidates for ester hydrogenolysis, including the intermediate step during the low temperature, cascade hydrogenation of CO and CO₂ to methanol.

Full text of DOI:10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Ethyl formate hydrogenolysis over Mo2C-based catalysts:
Towards low temperature CO and CO2 hydrogenation to methanol
Yuan Chen, Saemin Choi, Levi T. Thompson∗
Department of Chemical Engineering and Hydrogen Energy Technology Laboratory, University of Michigan, Ann Arbor, MI 48109-2136, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Ester hydrogenolysis is a key step in the production of alcohols and new catalysts are being sought to
improve the performance and sustainability of this process. Research described in this paper investigated
the use of Cu, Mo2C, and Mo2N-based catalysts for the hydrogenolysis of ethyl formate at 105–150 ◦C and
30 bar H2. The high surface area Mo2C-based catalysts were more active and selective to methanol than
the Mo2N-based catalysts. The deposition of nanoscale Cu and/or Pd particles onto the Mo2C or Mo2N
resulted in significant enhancements in the methanol selectivities. The methanol selectivities varied
with the acidity or basicity of the support suggesting some degree of cooperation between the Cu or
Pd particles and the support. Temporal variations in rates for the Cu/Mo2C catalyst were consistent
with first order kinetics with regard to ethyl formate and the activation energy was 44 kJ/mol. The best
performance was achieved over a Pd-Cu/Mo2C catalyst; this material outperformed, by a significant
margin, the oxide supported Cu catalysts. Given their high activities and selectivities, the Mo2C supported
catalysts are attractive candidates for ester hydrogenolysis, including the intermediate step during the
low temperature, cascade hydrogenation of CO and CO2 to methanol.
Received 9 April 2015
Received in revised form 29 July 2015
Accepted 11 August 2015
Available online xxx
Keywords:
Alkyl formate hydrogenolysis
Mo carbide/nitride-based catalysts
Methanol synthesis
CO/CO2 hydrogenation
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
(i) carbonylation of an alcohol to the corresponding ester over a
sodium methoxide catalyst and (ii) hydrogenolysis of the ester
The hydrogenolysis of esters is a key step in the production
of fatty alcohols (from natural fatty esters) and other lower alkyl
alcohols, such as methanol and ethanol [1–4]. Copper based cata-
lysts are commonly used because they possess high selectivities
for activation of the C–O bonds rather than the C–C bonds [4].
For example, Cu chromite catalysts are well-established for the
commercial production of fatty alcohols, although increasing envi-
ronmental concerns with regard to Cr have resulted in the search for
new formulations [5]. Other catalysts containing less toxic mate-
rials, such as Cu/SiO2 [6] and Cu/ZnO [7], have been reported to
possess ester hydrogenolysis activities and selectivities that are
similar to those for Cu chromite catalysts. Nevertheless, new cata-
lyst formulations are being sought to enhance the performance and
sustainability of this process.
to methanol and the parent alcohol over a reduced Cu oxide
catalyst (see Fig. 1a). Liu et al. explored the use of potassium
methoxide and Cu chromite catalysts for this process at 140–180 ◦C
[12]. They suggested the formate hydrogenolysis was rate-limiting
due in part to deactivation of the Cu chromite catalyst by CO.
Fan et al. [10] reported the liquid phase hydrogenation of CO2
to methanol over Cu-Zn or Cu-Cr based catalysts at 200 ◦C and
proposed that the reaction occurred via a multistep pathway.
The proposed reaction pathway involved (i) hydrogenation of
CO2 to formic acid, (ii) esterification of formic acid to alkyl for-
mate, and (iii) hydrogenolysis of the formate to methanol and
the corresponding alcohol (Fig. 1b). They also suggested that for-
mate hydrogenolysis was the rate-limiting step in the overall
reaction system, encouraging the development of more active cat-
alysts for this step. Recently Huff and Sanford [13] demonstrated
homogeneous catalysts for the cascade hydrogenation of CO2 to
methanol. This system included a (PMe3)4Ru(Cl)(OAc) catalyst for
the hydrogenation of CO2 to formic acid, Sc(OTf)3 catalyst for the
esterification of formic acid with ethanol, and (PNN)Ru(CO)(H)
catalyst for the hydrogenolysis of ethyl formate. A key limitation
for this homogeneous cascade system was incompatibility of the
ester hydrogenolysis catalyst with other components and reactant
species [13,14].
Of particular interest here is the low temperature (<200 ◦C),
liquid phase hydrogenolysis of esters, a key intermediate step
during the low temperature hydrogenation of CO or CO2 to
methanol [8–10]. Christiansen [11] first described a two-step pro-
cess for converting syngas (CO and H2) into methanol at 180 ◦C via
∗
Corresponding author. Tel.: +1 734 936 2015; fax: +1 734 763 0459.
E-mail address: ltt@umich.edu (L.T. Thompson).
http://dx.doi.org/10.1016/j.cattod.2015.08.021
0920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
2
Y. Chen et al. / Catalysis Today xxx (2015) xxx–xxx
CH3OH
+ 2H2
2. Material and methods
CH3OH
(ii)
CO2
+H2
CO
(i)
(i)
(iii)
- ROH
(ii)
- H2O
2.1. Catalyst preparation and characterization
O
O
O
H
OR
(a)
+ROH
The Mo2C and Mo2N materials were synthesized from an ammo-
nium molybdate (AM) precursor (Alfa Aesar) using a temperature
programmed reaction (TPR) technique. The Mo2C was prepared
by reducing AM in H2 at 350 ◦C for 12 h, followed by treatment
in 15% CH4/H2 at 590 ◦C for 2 h, and the resulting material was
then quenched to room temperature. The Mo2N was synthesized
by reacting AM with anhydrous NH3 as the temperature was lin-
early heated from room temperature to 350 ◦C (33 min), then to
450 ◦C (2.5 h) and finally to 700 ◦C (2.5 h) followed by a soak at this
temperature for 1 h before quenching to room temperature. Other
details regarding the synthesis procedures have been described in
previous reports [33,37]. Metals were deposited onto the Mo2C or
Mo2N supports using a wet impregnation technique. Typically, car-
bides and nitrides are passivated prior to exposure to air to avoid
bulk oxidation of the material [36]. To avoid passivation and deposit
metals directly onto the native Mo2C or Mo2N surfaces (as opposed
to a passivated surface), the freshly-synthesized materials were
transferred under an inert gas (CH4/H2 or He) into a deaerated
aqueous solution containing a target concentration of Cu(NO3)2
or Pd(NO3)2·4NH3 and then allowed to interact for at least 20 h.
Argon was continuously purging through the solutions during the
wet impregnation process to deaerate and agitate the solution. The
synthesized metal/Mo2C catalysts were pyrophoric, indicating the
absence of surface passivation on the final materials. Such−observa-
tion also suggests that neither exposure of H2O nor NO3 (during
the wet impregnation process) has passivated/oxidized the cata-
lyst surface. Schweitzer et al. reported that passivation/oxidation
with deaerated H2O did not occur and was not thermodynamically
favorable [38]. Our finding on the absence of surface passivation of
Mo2C catalysts, therefore, is consistent with the literature conclu-
sion. The resulting catalyst slurry was dried at 110 ◦C for 2 h and
reduced in flowing H2 (400 mL/min) at 300 ◦C for 4 h to decompose
the nitrate and produce the Cu or Pd domains. Preparation of the
Cu-Pd/Mo2C catalysts was achieved sequentially by first depositing
Cu onto Mo2C, followed by transferring the dried Cu/Mo2C under
an inert atmosphere to a deaerated aqueous solution containing the
target concentration of Pd(NO3)2·4NH3. The catalyst was allowed to
interact for 5 h and was dried and reduced in a quartz tube reactor
using the protocols described earlier.
Properties for the Mo2C- and Mo2N-based catalysts were
compared to those for bulk and oxide supported Cu catalysts. Nano-
Cu (QuantumSphere), Cu/ZnO/Al2O3 (Süd-Chemie/Clariant), and
Cu/Cr2CuO4 (Cu Chromite, Strem Chemicals Inc.) catalysts were
acquired from commercial vendors and used after pretreatment
(described later). The Cu/ZnO/Al2O3 and Cu/Cr2CuO4 catalysts are
referred to as “Cu-Zn-Al” and “Cu-Cr”, respectively. The Cu/␥-Al2O3,
Cu/SiO2, and Cu/C catalysts were synthesized via incipient wet-
ness impregnation of ␥-Al2O3 (Alfa Aesar, 0.4 cm3/g pore volume),
fumed SiO2 (AEROSIL®, 0.8 cm3/g pore volume), and carbon black
(Fuel Cell Store, Vulcan XC-72R, 0.8 cm3/g pore volume) with aque-
ous solutions of Cu(NO3)2 (5 wt% nominal loading). The supports
were pelletized, then ground and sieved to 125–250 ␮m prior to
impregnation with the Cu solution. After impregnation, the cata-
lysts were dried in vacuum at 110 ◦C for 16 h and then annealed
under N2 flow (100 mL/min) at 300 ◦C for 5 h to decompose the
nitrate from Cu(NO3)2. A Cu/ZrO2 (referred to as Cu-Zr) catalyst was
synthesized via a co-precipitation technique with a 1:1 molar ratio
of Cu(NO3)2 and Zr(NO3)2 precursors, based on a protocol described
elsewhere [42].
H
OH
H
OR
(b)
Fig. 1. Reaction pathways to achieve one-pot methanol synthesis in the presence
of alcohol through alkyl formate intermediate, (a) from CO hydrogenation [11], and
(b) from CO2 hydrogenation [10].
A variety of Cu-based catalysts, including Raney Cu [15], Cu/SiO2
[16], Cu chromite [9,17,18] and Cu/ZnO [19], have been reported
to be active for the liquid phase hydrogenolysis of alkyl formates
at temperatures in the range of 110–200 ◦C. Copper chromite cat-
alysts are among the most active; however, their poor tolerance
to CO reduces the feasibility of using this type of catalyst in pro-
cesses where CO or CO2 is present [15,20]. A variety of oxide
supported Pt and Pd catalysts (including ZnO, Ga2O3, In2O3, and
SiO2) have also been reported to be active for gas phase methyl
formate hydrogenolysis at 150 ◦C. The ZnO and In2O3 supported
catalysts provided the highest selectivities to methanol possibly
due to the formation of alloys (e.g. Pt–Zn and Pd–In), although a sub-
stantial Pd or Pt loading (10 wt%) was required to achieve the best
performance [21]. Braca et al. [22] reported that several Re-based
homogeneous and heterogeneous catalysts were highly active and
selective for the hydrogenolysis of formate at 200 ◦C, albeit at high
H2 pressures (∼100 bar). For the heterogeneous catalysts, the acid-
ity/basicity of the support (SiO2, TiO2, Nb2O5, and MgO) had a
significant effect on catalytic performance [22]. Rhenium, however,
is relatively expensive, which makes it less attractive for large-scale
applications.
Research described in this paper investigated the low tem-
perature, liquid phase hydrogenolysis of ethyl formate over
Mo carbide and nitride supported Cu catalysts. Mo carbides
and nitrides possess catalytic properties that can sometimes
resemble those for platinum-group-metals [23–26]; these sim-
ilarities in catalytic properties are highly dependent on the
reaction conditions and intermediate species [27]. Mo carbides
and nitrides have also been reported to catalyze
a variety
of reactions including Fischer–Tropsch Synthesis (FTS) [28,29],
water–gas shift (WGS) [30], alcohol steam reforming [31,32],
and hydrodenitrogenation [33]. Like other early transition metal
carbides and nitrides [23,24], these materials can be syn-
thesized with surface areas exceeding 100 m2/g [34–36]. The
carbides in particular have also been reported to be effec-
tive supports, facilitating the production of highly dispersed
Pt, Pd, Co and Cu domains, and in some cases, modifying the
electronic properties of these metals [31,37]. This modifica-
tion can result in superior catalytic performance for carbide
supported metals. For example, Schweitzer et al. reported a
Pt/Mo2C catalyst that significantly outperformed oxide sup-
ported Pt catalysts for WGS [38]. Lausche et al. described a
Pd/W2C catalyst which demonstrated higher rates and selec-
tivities than those for bulk Pd catalysts during the selective,
electrocatalytic hydrogenation of triglycerides [39]. Furthermore,
Mo carbides and nitrides, known CO [28,29] and CO2 hydro-
genation catalysts [40,41], are naturally tolerant to CO and CO2.
In addition to the supported Cu catalysts, we also synthesized
and evaluated a series of Pd/Mo2C and Cu-Pd/Mo2C catalysts, as
oxide supported Pd catalysts have been reported to be active
and selective for ester hydrogenolysis [21]. Performance char-
acteristics for the Mo carbide and nitride supported catalysts
were compared to those for conventional oxide supported Cu
catalysts.
Surface areas of the materials were determined from N2
physisorption isotherms collected using
a
Micromeritics
ASAP 2010 analyzer. The isotherms were analyzed using the
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
Y. Chen et al. / Catalysis Today xxx (2015) xxx–xxx
Table 1
3
Brunauer–Emmett–Teller (BET) method. Prior to the measure-
ments, the catalysts were degassed (<5 mm Hg) for at least 4 h
at elevated temperatures (350 ◦C for the Mo2C or Mo2N based
catalysts, 200 ◦C for the other Cu-based catalysts). The bulk crys-
talline structures of the catalysts were characterized using X-ray
diffraction performed using a Rigaku Miniflex diffractometer
Physical properties of Cu, Mo2C and Mo2N based catalysts.
Catalysts
Surface area
(m2/g)
Cu content
(wt%)
Cu crystallite
size (nm)
Nano-Cua
Mo2Ca
Mo2N
5.5
151
153
135
141
60
100
N/A
N/A
5.3
3.2
33
36
34
5.0
5.0
5.0
254
N/A
N/A
34
˚
with Cu K␣ radiation (ꢀ = 1.5418 A). The diffraction patterns were
5 Cu/Mo2Ca
3 Cu/Mo2N
Cu-Zn-Ala
Cu-Cra
obtained by scanning 2ꢁ from 10 to 90◦ at a scan rate of 5◦/min.
Inductively coupled plasma (ICP-OES, Varian 710-ES analyzer) was
used to determine the metal compositions for the Mo2C and Mo2N
based catalysts. As appropriate, the metal content is included
in the catalyst name; for example, 5 wt% Mo2C supported Cu is
referred to as “5 Cu/Mo2C”.
N/Ab
89
163
98
47
101
58
46
Cu-Zr
80
Cu/␥-Al2O3
Cu/SiO2
Cu/C
54
172
169
a
Values adapted from [39].
b
Cu crystallite size was not quantified for this catalyst, as the peak at ∼41◦ 2ꢁ
2.2. Activity and selectivity measurements
could not be isolated from that for the ␥-Mo2N support (Fig. 2).
Ethyl formate was selected as the reactant so that CH3OH pro-
duced from other pathways could be distinguished from the parent
alcohol of the formate. All of the experiments were performed in
a 50 mL autoclave Parr reactor, equipped with a variable-speed
impeller. A gas chromatograph (GC) (Varian 450, equipped with
flame ionization and thermal conductivity detectors) was con-
nected directly to the reactor gas outlet for online gas product
analysis. A filter was also added to separate the liquid products
from the catalyst particles when withdrawing liquid samples; the
liquids were analyzed offline using a gas chromatograph (Varian
450) with a flame ionization detector.
to their corresponding support materials are likely a consequence
of pore blockage by the Cu nanoparticles. The Cu/␥-Al2O3, Cu-Zn-
Al, Cu-Cr, and Cu-Zr catalysts exhibited moderate surface areas
(45–80 m2/g), while the Nano-Cu material had a relatively low sur-
face area.
X-ray diffraction patterns for the catalysts are displayed in Fig. 2.
The bulk and supported Mo2C-based catalysts contained a mixture
of ␣-MoC1 − x (face centered cubic) and ␤-Mo2C (hexagonal close
packed); each phase was determined by matching the XRD pat-
terns to the standard spectra reported in the literature [44,45]. The
two phases showed approximately a 1:1 ratio as calculated using
the whole pattern fitting Rietveld refinement technique (JADE 7.0).
␣-MoC1 − x and ␤-Mo2C have similar Mo:C ratios (∼2:1), as sug-
gested by the elemental analysis, therefore catalysts containing
these materials will be referred to as Mo2C for simplicity. Such
composition was also reported by several works in the litera-
ture on the same type of Mo2C material [32,46]. The Mo2N-based
All of the Cu-based catalysts were pretreated in 4% H2/N2
(100 mL/min) at 200 ◦C for 4 h in the reactor prior to measurement
of the rates and selectivities. The Mo2C or Mo2N-based catalysts
were used as synthesized without passivation. To avoid contact of
the materials with air, they were transferred under an inert atmo-
sphere and stored in an Ar filled glovebox. The reactant solution was
prepared by adding 0.6 mmol ethyl formate (Acro Organics, 99%) to
37.5 mL 1,4-dioxane (Acro Organics, 99+%, 100 ppm water); decane
(10 ␮L) was introduced as an internal standard for GC analysis. The
reactant solution and 200 mg catalyst were loaded into the reac-
tor. The Mo2C and Mo2N materials were loaded into the reactor
in the glovebox due to their pyrophoricity. After loading the cata-
lysts, 30 bar of H2 was charged into the reactor through a dip tube
(immersed in the solvent) to allow sufficient gas dissolution. The
solubility of H2 in 1,4-dioxane is 0.14 M at 135 ◦C and 30 bar H2
[43]. The reactor was then heated at a rate of 5 ◦C/min from room
temperature to 135 ◦C, and agitated at a constant rate of 300 rpm.
The start of the reaction was designated as the time at which the
reactor temperature reached 135 ◦C. The reactor was maintained
at 135 ◦C throughout the reaction. Each experiment was repeated
at least twice and the errors were calculated considering both the
standard deviations from replicates and GC analyses. Carbon bal-
ances closed to within 8% for all the experiments. The conversion
was defined as the percentage of the initial ethyl formate consumed
during the reaction, while the selectivity was defined as the molar
ratio of a specific product over the total products. The reaction rates
were evaluated from temporal variations in the amounts of ethyl
formate consumed or CH3OH formed.
3. Results and discussion
3.1. Physical properties
The surface areas, Cu contents, and Cu crystallite sizes for the
supported Cu, Mo2C and Mo2N catalysts are listed in Table 1. As
expected, the Cu/C, Cu/SiO2 and all of the Mo2C- and Mo2N-based
catalysts possessed high surface areas (>100 m2/g). The reduced
surface areas for the Cu/Mo2C and Cu/Mo2N catalysts compared
Fig. 2. X-ray diffraction patterns for the Cu, Mo2C and Mo2N catalysts. The open
circle indicates the Cu peaks. The standard patterns on the bottom include Cu
(JCPDF 01-085-1326), ␣-MoC1−x (JCPDF 00-015-0457), ␤-Mo2C (JCPDF 00035-
0787), ␥-Mo2N (JCPDF 00-025-1366), ZnO (JCPDF 01-080-3004), and Al2O3 (JCPDF
00-004-0877).
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
4
Y. Chen et al. / Catalysis Today xxx (2015) xxx–xxx
catalysts contained only ␥-Mo2N phase. Only a weak peak associ-
ated with Cu (41◦ 2ꢁ) was detected for the Cu/Mo2C catalyst and Cu
peaks could not be clearly defined for the Mo2N supported mate-
rial. These findings suggested that the Cu dispersion was relatively
higher on Mo2C and Mo2N compared to that on other supports,
although it was possible that the Cu and ␥-Mo2N peaks overlapped
at ∼41◦ 2ꢁ. Sharp Cu peaks are observed for the Nano Cu material,
indicating that it is phase pure with large crystallites. Cu peaks,
as well as peaks for the support oxides, were observed for most
of the oxide supported catalysts. Peaks attributable to the oxides
were weak or non-existent for the Cu-Cr, Cu-Zr and Cu/SiO2 cata-
lysts, suggesting that the oxide particles were very small (<5 nm)
or amorphous. In addition, a weak peak was detected for the Cu-
Cr catalyst at 54◦ 2ꢁ. This peak is consistent with the presence of
CuO [40]. The Cu crystallite sizes were smallest for the Cu/Mo2C
catalyst; the Cu crystallite size was not quantified for the Cu/Mo2N
catalyst, as the peak at ∼41◦ 2ꢁ could not be isolated from that for
the ␥-Mo2N support.
HCOOC2H5 + H2O → HCOOH + C2H5OH,
ꢃG◦ = 20 kJ/mol, ꢃH◦ = −6.8 kJ/mol
(2)
(3)
(4)
(5)
HCOOH → CO2 + H2,
ꢃG◦ = −48 kJ/mol, ꢃH◦ = 16 kJ/mol
HCOOC2H5 + CH3OH → HCOOCH3 + C2H5OH,
ꢃG◦ = −7.1 kJ/mol, ꢃH◦ = −26 kJ/mol
HCOOH + 3H2 → CH4 + 2H2O,
ꢃG◦ = −179 kJ/mol, ꢃH◦ = −237 kJ/mol
3.2. Catalytic activities and selectivities
On a gravimetric basis, catalysts containing Mo2C were the
most active, outperforming the oxide supported catalysts, includ-
ing the Cu-Cr catalyst, by a significant margin. In fact, the bulk
Mo2C catalyst exhibited the highest rate although its selectivity
to CH3OH (27%) was moderate. The deposition of Cu onto this cata-
lyst improved its selectivity to 41% (the maximum selectivity is 50%
based on the reaction stoichiometry). The Cu-Cr, Cu-Zn-Al, Cu-Zr,
and Mo2N catalysts possessed moderate activities, while the other
catalysts, including the Nano-Cu catalyst, exhibited low activities.
Interestingly, the Cu/C catalyst was inactive, possibly due to limited
formate adsorption on the carbon support, as was suggested by
Braca et al. for a Re/C catalyst [22]. Where possible, the rates were
also normalized by the surface areas and Cu loadings in order to
define more intrinsic activities. When normalized by their BET sur-
face areas, rates for the Nano-Cu as well as the Cu-Cr and Cu-Zn-Al
catalysts were the highest. The results make sense as the measured
total surface areas should be close to the active Cu surface areas for
these catalysts, given their relatively high Cu composition. For the
other catalysts with lower Cu loading (∼5 wt%), the active species
surface area may only account for a small portion of the total sur-
face area, therefore, the surface area normalized rates were smaller.
When normalized by the Cu content, all of the supported cata-
lysts yielded rates that are significantly higher than that for the
Nano-Cu catalyst, indicating that Cu is a key active species and its
dispersion is critical. The highest rates, when normalized by Cu con-
tents, were achieved for the Mo2C- and Mo2N-based catalysts. This
result was not unexpected given that Mo2C and Mo2N were active
without Cu.
When considering the Cu catalysts, the support appeared to
have a significant effect on selectivity, as illustrated by results listed
in Table 2. The Cu/Mo2C and Nano-Cu catalysts demonstrated high
methanol selectivities (>40%); selectivities for the Cu-Zn-Al, Cu-
Cr, and Cu/SiO2 catalysts were moderate (∼30%) and the Cu/Mo2N,
Cu/ZrO2 and Cu/␥-Al2O3 catalysts yielded relatively low selectivi-
ties (<10%). Differences in selectivities for these catalysts could be a
consequence of the activity or lack of activity for the support. Monti
et al. reported, based on results from deuterium labeling studies at
∼145 ◦C, that methyl formate hydrogenolysis over a Cu/SiO2 cat-
alyst involved the adsorption of methyl formate on SiO2 via the
carbonyl group and the dissociation of H2 on the Cu particles [47].
Hydrogen can then be added to the adsorbed formate sequentially
to produce methanol as shown in Fig. 3a. Braca et al. also reported
that the SiO2 supported Re catalyst was more selective to methanol
than catalysts produced from more acidic or basic supports [22].
In addition to facilitating the hydrogenolysis reaction, the support
The ethyl formate hydrogenolysis reaction rates and selectivi-
ties over the supported Cu, Mo2C and Mo2N catalysts are shown
in Table 2. Ethyl formate hydrogenolysis generates CH3OH and
C2H5OH in an equimolar ratio (Eq. (1)), i.e. a 50% selectivity should
be expected for each alcohol product. However, we observed that
the selectivities to methanol were less than 50% over all of the cat-
alysts, while the selectivities to ethanol were consistently around
50%. This finding indicated that either CH3OH was consumed dur-
ing a secondary reaction or additional ethanol was produced via
side reactions. To examine the possibility of CH3OH consumption,
control experiments were conducted with 0.6 mmol CH3OH as the
starting material at 135 ◦C and 30 bar H2 using 200 mg of each
catalyst; these conditions are similar to those used during for-
mate hydrogenolysis. No CH3OH was consumed during the control
trials, indicating that CH3OH was stable once produced in the reac-
tion system. Interestingly, CO2 was also observed during the ester
hydrogenolysis experiments and its selectivity was comparable to
the difference between the CH3OH and C2H5OH selectivities. It is
plausible that the excess ethanol was produced via hydrolysis of the
formate (Eq. (2)). Water was a contaminant in the 1,4-dioxane sol-
vent according to the product specification (100 ppm H2O), which
resulted in ∼0.04 mmol of H2O in the reactor system, while the
CO2 was produced in the amounts of 0.04–0.46 mmol for the ethyl
formate hydrogenolysis experiments (Table 2). This calculation
suggested that additional water may have come from moisture in
the air to balance the amount of CO2 production. This was not sur-
prising given the hygroscopic nature of 1,4-dioxane. Based on our
previous investigations, any formic acid produced during formate
hydrolysis would be rapidly converted to CO2 (Eq. (3)) under the
conditions used in the present work [40]. Meanwhile, the amount of
CO2 is always equivalent to the ethanol produced from the hydrol-
ysis (Eq. (2)). Therefore, formic acid is most likely the source of
CO2; a similar pathway was also suggested by Braca et al. [22]. For
one reactant (ethyl formate) with two pathways (hydrogenolysis
and hydrolysis) to a common product (ethanol), the selectivity for
that product should be 50% as was observed for most of the cata-
lysts. Small amounts of CH4 and methyl formate were also produced
over the Mo2C-based catalysts, possibly due to formic acid hydro-
genation (Eq. (4)) and/or transesterification (Eq. (5)), respectively.
HCOOC2H5 + 2H2 → CH3OH + C2H5OH,
ꢂG◦ = −38 kJ/mol, ꢃH◦ = −122 kJ/mol
(1)
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
Y. Chen et al. / Catalysis Today xxx (2015) xxx–xxx
5
Table 2
Reaction rates and selectivities for ethyl formate conversion over Cu, Mo2C and Mo2N based catalystsa.
Catalysts
Formate conversion rate
(␮mol/m2/h)
Selectivity (%)
CH3OH
kobs × 104 (g−ca1tal s−1
)
(␮mol/gcatal/h)
79
(mmol/gCu/h)
C2H5OH
CO2
CH4 +MFc
Nano-Cub
Mo2Cb
Mo2N
4
14.4 0.8
5.9 0.6
2.3 0.2
6.2 0.5
2.9 0.3
10.0 0.6
12.0 0.7
7.1 0.9
3.5 0.5
0.5 0.1
0
0.08 0.04
N/A
N/A
40
27
0
3
3
51
49
51
49
51
51
50
50
50
50
4
5
4
5
3
4
6
5
4
3
0
8.9 0.7
22
49
0
1.5
0
2.7
0
0
0
0
0
0
1.1 0.2
2.8 0.2
0
889 73
350 42
837 55
413 37
598 47
554 32
567 45
190 14
2
3
5 Cu/Mo2Cb
3 Cu/Mo2N
Cu-Zn-Alb
Cu-Crb
15.8
12.9
2
1
41
4
7.5 0.6
3.6 0.4
0.04 0.0
1.2 0.1
1.7 0.3
0.3 0.06
0.4 0.1
0.6 0.1
0
7.2 0.6
42
21
23
43
44
19
0
2
2
2
5
3
1
1.8 0.3
1.5 0.2
1.7 0.2
3.8 0.3
1.6 0.1
0
28
27
2
4
Cu-Zr
6.8 0.5
5.7 0.5
Cu/␥-Al2O3
Cu/SiO2
Cu/C
78
0
5
31
0
2
0
a
135 ◦C, 30 bar H2, 0.6 mmol ethyl formate and 37.5 mL 1,4-dioxane, rates calculated at 2 h and selectivities calculated at 8 h.
Rate data are adapted from [39].
MF = methyl formate.
b
c
Fig. 3. Proposed mechanisms for (a) ethyl formate hydrogenolysis over Cu-supported catalysts (adapted from [43] with permission, copyright 1986 Elsevier), and (b) ethyl
formate hydrolysis.
could, depending on its acidicity or basicity, catalyze hydrolysis of
the formate (see Fig. 3b) [48,49]. SiO2, and Cr2O3 tend to be ampho-
teric and hence their hydrolysis rates should be low. ␥-Al2O3 is
relatively acidic and is known to catalyze hydrolysis reactions [50].
ZrO2 and ZnO are relatively basic and would facilitate the hydroly-
sis reaction [51]. The Mo nitrides and carbides have been reported
to possess moderate acidity and/or basicity depending on the syn-
thesis and treatment conditions [52,53] and would also catalyze
the hydrolysis of ethyl formate.
5Cu/Mo2C > Mo2C > Cu-Cr > Cu-Zn-Al ≈ Nano-Cu > Cu/SiO2 ≈ Cu/␥-
Al2O3 ≈ Cu-Zr > 3Cu/Mo2N ≈ Mo2N ≈ Cu/C = 0 (Table 2).
As the Cu/Mo2C catalyst exhibited the highest reaction rates
and selectivities, additional experiments were carried out with this
material. Fig. 4 shows the reactant and product distributions as
a function of reaction time and Fig. 5 shows the Arrhenius plot.
The apparent activation energy (EA) for ethyl formate hydrogenol-
ysis to methanol was 44 3 kJ/mol. Evans et al. reported a value of
Temporal variations in the reaction rates were consistent with
first order kinetics with respect to ethyl formate (R2 ≥ 92%). The
first order dependence on formate is consistent with reports by
Monti et al. regarding the kinetics for liquid phase methyl formate
hydrogenolysis at 135–200 ◦C [9]. We used the following equation
Et Formate
Methanol
Ethanol
0.6
0.4
0.2
0.0
CO2
CO2
Me Formate
to determine the pseudo first order rate constants, kobs
:
rh = kobs · Cformate
(6)
where rh = rate of hydrogenolysis [mol/L/gcatal/s] based on the
formation of methanol, kobs = pseudo first order rate constant
[g−ca1tals−1] and Cformate = the average concentration of formate
[mol/L]. Because the concentration of H2 was in significant
excess of stoichiometric (∼5.3 mmol based on its solubility
in 1,4-dioxane at 135 ◦C, 30 bar [43]), the H2 concentra-
0
2
4
6
8
Time (hr)
Fig. 4. Reactant and products plot for ethyl formate hydrogenation over Cu/Mo2C
catalyst ( Ethyl formate, Ethanol, Methanol,
CO2, × Methyl formate). The
experiments were performed at 135 ◦C, 30 bar H2 and 37.5 mL 1,4-dioxane.
tion term was embedded in the rate constant. Based on kobs
,
hydrogenolysis activities for the catalysts decreased as follows:
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
6
Y. Chen et al. / Catalysis Today xxx (2015) xxx–xxx
Table 3
Reaction activities and selectivities for ethyl formate conversion over Mo2C supported Cu and/or Pd catalystsa.
Catalysts
Properties
Formate conversion rate
Selectivity (%)
kobs × 104
catal
−1
(g
s−1
)
Surface area
(m2/g)
Cu or Pd
content (wt%)
(␮mol/gcatal/h)
(␮mol/m2/h) CH3OH
C2H5OH
CO2
CH4 +MFb
1Pd/Mo2C
5Pd/Mo2C
4Cu-1Pd/Mo2C
143
138
132
Pd (1.4)
Pd (5.2)
Cu (3.8)
Pd (1.4)
Cu (5.2)
Pd (1.6)
775 61
783 58
795 76
5.4 0.3
5.7 0.4
6.0 0.7
31
43
45
4
3
2
50
50
50
5
5
4
17
5.8
3.4
1.8
1.4
1.5
3.2 0.1
3.5 0.3
4.2 0.4
5Cu-1Pd/Mo2C
127
784 64
6.2 0.6
49
4
49
4
1.8
0.6
4.6 0.4
a
135 ◦C, 30 bar H2, 0.6 mmol ethyl formate and 37.5 mL 1,4-dioxane, rates calculated at 2 h and selectivities calculated at 8 h.
MF = methyl formate.
b
10.0
1.0
consistent with Pd and Cu facilitating hydrogenolysis over the
Mo2C catalyst perhaps via a pathway similar to that proposed
by Monti et al. for Cu/SiO2 catalysts (Fig. 3a) [47]. Mo2C still
differs from other oxide supports in that it is an active species
for ester hydrogenolysis itself. As all the Mo2C-based materi-
als were synthesized without surface passivation and used in
an oxygen-free environment, minimal amount of molybdenum
oxide species should be involved in the reaction. In comparing the
5Cu/Mo2C and 5Pd/Mo2C catalysts, one could conclude that Pd is
a slightly more effective promoter than Cu. This could be a con-
sequence of differences in the inherent properties of the metals
or differences in their dispersions; this will be explored in future
experiments.
0.1
2.3
2.4
2.5
2.6
2.7
1000/T (1/K)
4. Conclusions
Fig. 5. Arrhenius plot for ethyl formate hydrogenolysis rate over Cu/Mo2C. Experi-
ments were performed at 105–150 ◦C, 30 bar H2 and 37.5 mL 1,4-dioxane.
Research described in this paper demonstrated that Mo2C-based
catalysts are catalytically active for the low temperature (<150 ◦C)
hydrogenolysis of ethyl formate to methanol. This reaction is a key
step in the production of alcohols and in the cascade hydrogena-
tion of CO and CO2 to methanol. Other products, including CO2
and minor amounts of CH4 and methyl formate, are also formed,
implicating several side reactions (e.g., ethyl formate hydrolysis,
formic acid decomposition and transesterification). The reaction
rates were adequately fit to a first order model with respect to ethyl
formate; over the best-perform−1ing catalyst (1Pd-5Cu/Mo2C) the
s
−1. The selectivity to methanol
rate constant was 4.6 × 10−4 gcatal
appears to be a strong function of the acidity/basicity of the support.
The catalysts with amphoteric or weakly acidic/basic supports gen-
erally yielded higher methanol selectivities than those for catalysts
with strong acidic/basic supports. Materials with strong acid/base
sites likely catalyzed the hydrolysis reaction, producing substan-
tial amounts of CO2. The deposition of nanoscale Cu and/or Pd
particles onto the Mo2C significantly enhanced the catalytic perfor-
mance, yielding high conversion rates and methanol selectivities
approaching 50% (as opposed to 27% for the Mo2C catalyst). The
enhancement in CH3OH selectivity is likely a consequence of syn-
ergy between the metal and support. These catalysts are excellent
candidates for use in cascade catalytic systems for the one-pot syn-
thesis of methanol via CO or CO2 hydrogenation as well as other
types of ester hydrogenolysis reactions.
Fig. 6. Product distributions for ethyl formate hydrogenolysis over the Mo2C-based
catalysts. The experiments were performed at 135 ◦C, 30 bar H2, 8 h and 37.5 mL
1,4-dioxane.
48 kJ/mol for a commercial Cu-Cr catalyst in a comparable temper-
ature range (130–170 ◦C) [17].
Given results for the Cu/Mo2C catalysts, we synthesized Mo2C
supported Pd and Pd-Cu catalysts. Palladium has been reported
to be active for ester hydrogenolysis [21]. Table 3 lists the reac-
tion rates and selectivities for the Mo2C supported Pd and/or
Cu catalysts. Interestingly, the surface area normalized rates for
the Mo2C catalyst and for these metal/Mo2C materials were very
similar, ranging from 5.4 to 6.2 ␮mol/m2/h at 135 ◦C. Given the
significant variations in the metal loading, this observation sug-
gested that Pd, Cu and Mo2C possessed similar intrinsic activities
for formate conversion. While the rates were similar, products
formed during formate conversion changed significantly when Pd
and/or Cu were deposited onto the Mo2C support. Fig. 6 compares
selectivities for the catalysts after 8 h on stream. The methanol
selectivity was lowest for the Mo2C catalyst and increased
with increases in the Pd and/or Cu loading. These results are
Acknowledgments
The authors acknowledge financial support from the NSF under
the CCI Center for Enabling New Technologies through Catalysis
(CENTC) Phase II Renewal, CHE-1205189. The authors also thank Dr.
Jason Gaudet, Dr. Tanya Breault, Brian Wyvratt, and Amin Bazyari
for assistance with data analysis and Evan Roberts and Tianshu Ji for
helping with the experiments. Finally, we thank Dr. Chelsea Huff,
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021

G Model
CATTOD-9751; No. of Pages7
ARTICLE IN PRESS
Y. Chen et al. / Catalysis Today xxx (2015) xxx–xxx
7
Dr. Tim Brewster, Prof. Alex Miller, Prof. Karen Goldberg, and Prof.
Melanie Sanford for helpful discussions.
[24] R.B. Levy, M. Boudart, Science 181 (1973) 547–549.
[25] G. Ranhotra, A. Bell, J. Reimer, J. Catal. 108 (1987) 40–49.
[26] J. Lee, S. Locatelli, S. Oyama, M. Boudart, J. Catal. 125 (1990) 157–170.
[27] A.J. Medford, A. Vojvodic, F. Studt, F. Abild-Pedersen, J.K. Nørskov, J. Catal. 290
(2012) 108–117.
[28] P.M. Patterson, T.K. Das, B.H. Davis, Appl. Catal., A 251 (2003) 449–455.
[29] A. Griboval-Constant, J.-M. Giraudon, G. Leclercq, L. Leclercq, Appl. Catal., A
260 (2004) 35–45.
[30] J. Patt, D. Moon, C. Phillips, L. Thompson, Catal. Lett. 65 (2000) 193–195.
[31] W. Setthapun, S. Bej, L. Thompson, Top. Catal. 49 (2008) 73–80.
[32] J.A. Schaidle, A.C. Lausche, L.T. Thompson, J. Catal. 272 (2010) 235–245.
[33] J.-G. Choi, J.R. Brenner, C.W. Colling, B.G. Demczyk, J.L. Dunning, L.T.
Thompson, Catal. Today 15 (1992) 201–222.
[34] J.B. Claridge, A.P. York, A.J. Brungs, M.L. Green, Chem. Mater. 12 (2000)
132–142.
[35] M. Ledoux, C. Pham-Huu, Catal. Today 15 (1992) 263–284.
[36] S. Oyama, Catal. Today 15 (1992) 179–200.
[37] J.A. Schaidle, N.M. Schweitzer, O.T. Ajenifujah, L.T. Thompson, J. Catal. 289
(2012) 210–217.
[38] N.M. Schweitzer, J.A. Schaidle, O.K. Ezekoye, X. Pan, S. Linic, L.T. Thompson, J.
Am. Chem. Soc. 133 (2011) 2378–2381.
[39] A.C. Lausche, K. Okada, L.T. Thompson, Electrochem. Commun. 15 (2012)
46–49.
[40] Y. Chen, S. Choi, L.T. Thompson, ACS Catal. 5 (2015) 1717–1725.
[41] W. Xu, P. Ramirez, D. Stacchiola, J. Rodriguez, Catal. Lett. 144 (2014) 1–7.
[42] C. Fröhlich, R. Köppel, A. Baiker, M. Kilo, A. Wokaun, Appl. Catal., A 106 (1993)
275–293.
[43] E. Brunner, J. Chem. Eng. Data 30 (1985) 269–273.
[44] S. Nagakura, M. Kikuchi, S. Oketani, Acta Crystallogr. 21 (1966) 1009–1010.
[45] J. Schuster, H. Nowotny, Monatshefte für Chem. 110 (1979) 321–333.
[46] T. Hyeon, M. Fang, K.S. Suslick, J. Am. Chem. Soc. 118 (1996) 5492–5493.
[47] D. Monti, N. Cant, D. Trimm, M. Wainwright, J. Catal. 100 (1986) 28–38.
[48] K. Yates, R.A. McClelland, J. Am. Chem. Soc. 89 (1967) 2686–2692.
[49] R. Koster, B. van der Linden, E. Poels, A. Bliek, J. Catal. 204 (2001)
333–338.
[50] A. Fukuoka, P.L. Dhepe, Angew. Chem., Int. Ed. 45 (2006) 5161–5163.
[51] R. Aelion, A. Loebel, F. Eirich, J. Am. Chem. Soc. 72 (1950) 5705–5712.
[52] S.K. Bej, C.A. Bennett, L.T. Thompson, Appl. Catal., A 250 (2003) 197–208.
[53] R.C.V. McGee, S.K. Bej, L.T. Thompson, Appl. Catal., A 284 (2005) 139–146.
References
[1] L.G. Hess, H.W. Schulz, US Patent 2,782,243 (1957).
[2] W.A. Lazier, US Patent 2,079,414 (1937).
[3] T. Turek, D. Trimm, N. Cant, Catal. Rev. Sci. Eng. 36 (1994) 645–683.
[4] D.S. Brands, E.K. Poels, A. Bliek, Appl. Catal., A 184 (1999) 279–289.
[5] H.B.M. Hoyng, Oleochemicals: green and clean, in: T.H. Applewhite (Ed.),
Proceedings: World Conference on Oleochemicals, The American Oil Chemists
Society, Kuala Lumpur, Malaysia, 1991, pp. 211–217.
[6] F.T. van de Scheur, B. van der Linden, M.C. Mittelmeijer-Hazeleger, J.G.
Nazloomian, L.H. Staat, Appl. Catal., A 111 (1994) 63–77.
[7] F.T. van de Scheur, L.H. Staal, Appl. Catal., A 108 (1994) 63–83.
[8] J. Evans, P. Casey, M. Wainwright, D. Trimm, N. Cant, Appl. Catal. 7 (1983)
31–41.
[9] D. Monti, M. Kohler, M. Wainwright, D. Trimm, N. Cant, Appl. Catal. 22 (1986)
123–136.
[10] L. Fan, Y. Sakaiya, K. Fujimoto, Appl. Catal., A 180 (1999) L11–L13.
[11] J.A. Christiansen, US Patent 1,302,011 (1919).
[12] Z. Liu, J.W. Tierney, Y.T. Shah, I. Wender, Fuel Process. Technol. 23 (1989)
149–167.
[13] C.A. Huff, M.S. Sanford, J. Am. Chem. Soc. 133 (2011) 18122–18125.
[14] C.A. Huff, J.W. Kampf, M.S. Sanford, Chem. Commun. 49 (2013) 7147–7149.
[15] R. Gormley, V. Rao, Y. Soong, E. Micheli, Appl. Catal., A 87 (1992) 81–101.
[16] A. Agarwal, N. Cant, M. Wainwright, D. Trimm, J. Mol. Catal. 43 (1987) 79–92.
[17] J.W. Evans, P.S. Casey, M.S. Wainwright, D.L. Trimm, N.W. Cant, Appl. Catal. 7
(1983) 31–41.
[18] S. Ohyama, H. Kishida, Appl. Catal., A 172 (1998) 241–247.
[19] G. Braca, A.R. Galletti, N. Laniyonu, G. Sbrana, E. Micheli, M. Di Girolamo, M.
Marchionna, Ind. Eng. Chem. Res. 34 (1995) 2358–2363.
[20] Z. Liu, J. Tierney, Y. Shah, I. Wender, Fuel Process. Technol. 18 (1988) 185–199.
[21] N. Iwasa, M. Terashita, M. Arai, N. Takezawa, React. Kinet. Catal. Lett. 74
(2001) 93–98.
[22] G. Braca, A.R. Galletti, G. Sbrana, M. Lami, M. Marchionna, J. Mol. Catal. A:
Chem. 95 (1995) 19–26.
[23] S.T. Oyama, The Chemistry of Transition Metal Carbides and Nitrides, 1st ed.,
Chapman & Hall USA, New York, 1996.
Please cite this article in press as: Y. Chen, et al., Ethyl formate hydrogenolysis over Mo2C-based catalysts: Towards low temperature
CO and CO2 hydrogenation to methanol, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.08.021