TiO2-supported molybdenum carbide: An active catalyst for the aqueous phase hydrogenation of succinic acid

Article abstract of DOI:10.1016/j.apcata.2018.11.009

TiO₂-supported Mo carbides “MoC/TiO₂” were prepared by impregnation of Mo salt followed by temperature programmed reduction-carburization using 20% v/v C₂H₆/H₂. Catalysts were characterized by XRD, XPS, TEM, STEM, ICP, Raman, BET, and carbon elemental analysis. The catalytic activity was evaluated for aqueous phase hydrogenation of succinic acid at 160–240 °C, and 90–150 bar of H₂ in batch reactor. MoC/TiO₂ is active for this reaction. The main products after 24 h are γ-butyrolactone, and more remarkably butyric acid. These intermediates are then converted to tetrahydrofuran, butanol, 1,4-butanediol and butane. The reaction conditions (temperature, pressure) do not impact the products distribution. A larger amount of butyric acid is formed when catalysts were synthesized with a higher gas hourly space velocity. The deactivation observed while recycling the catalyst was mainly attributed to a decrease in the amounts of carbidic molybdenum and carbidic carbon, as demonstrated by XPS analysis.

Full text of DOI:10.1016/j.apcata.2018.11.009

Accepted Manuscript
Title: TiO2-supported molybdenum carbide: an active catalyst
for the aqueous phase hydrogenation of succinic acid
´
Authors: Marwa Abou Hamdan, Stephane Loridant,
´
Mohamad Jahjah, Catherine Pinel, Noemie Perret
PII:
S0926-860X(18)30562-3
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.11.009
APCATA 16878
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
6 August 2018
2 October 2018
10 November 2018
Please cite this article as: Hamdan MA, Loridant S, Jahjah M, Pinel C, Perret
N, TiO2-supported molybdenum carbide: an active catalyst for the aqueous
phase hydrogenation of succinic acid, Applied Catalysis A, General (2018),
https://doi.org/10.1016/j.apcata.2018.11.009
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

TiO2-supported molybdenum carbide: an active catalyst for the aqueous
phase hydrogenation of succinic acid
a
b
a
Marwa Abou Hamdan , Stéphane Loridant a, Mohamad Jahjah , Catherine Pinel ,
Noémie Perret a,
aUniv Lyon, Université Claude Bernard Lyon 1, CNRS, IRCELYON-UMR5256, 2 Avenue
Albert Einstein, 69626 Villeurbanne, France
bLCIO, Laboratoire de Chimie de Coordination Inorganique et Organométallique, Université
Libanaise- Faculté des Sciences I, Beyrouth, Liban
*E-mail: noemie.perret@ircelyon.univ-lyon1.fr
HIGHLIGHTS

Small crystallites (< 5nm) of cubic MoC on TiO2 were obtained by temperature-
programmed carburization using 20%v/v C2H6/H2 at 700°C
Full succinic acid conversion is reached after 22h at 240°C and 150 bar H2 in
water




The preparation conditions affect the catalytic performance
Higher carbon content favors the formation of butyric acid over γ-butyrolactone
Recovering the catalyst under inert atmosphere limits the deactivation
ABSTRACT
TiO2-supported Mo carbides “MoC/TiO2” were prepared by impregnation of Mo salt followed
by temperature programmed reduction-carburization using 20% v/v C2H6/H2. Catalysts were
characterized by XRD, XPS, TEM, STEM, ICP, Raman, BET, and carbon elemental analysis.
The catalytic activity was evaluated for aqueous phase hydrogenation of succinic acid at 160-
240 °C, and 90-150 bar of H2 in batch reactor. MoC/TiO2 is active for this reaction. The main
products after 24 hours are γ-butyrolactone, and more remarkably butyric acid. These
intermediates are then converted to tetrahydrofuran, butanol, 1,4-butanediol and butane. The
reaction conditions (temperature, pressure) do not impact the products distribution. A larger
amount of butyric acid is formed when catalysts were synthesized with a higher gas hourly
space velocity. The deactivation observed while recycling the catalyst was mainly attributed
1

to a decrease in the amounts of carbidic molybdenum and carbidic carbon, as demonstrated by
XPS analysis.
Keywords: Succinic acid, butyric acid, hydrogenation, aqueous phase, Mo carbide, TiO2
supported catalyst.
1. Introduction
The worldwide environmental concerns about the sustainable production of fuels and
chemicals, combined with the diminishing fossil resources, urged the development of
environmentally friendly processes for the production of valuable chemicals based on
renewable raw materials. The fermentation of lignocellulosic biomass using bacteria (e.g.
Anaerobiospirillum succiniciproducens) produces succinic acid “SA” [1–3]. The catalytic
hydrogenation of the fossil-based maleic anhydride used to be the conventional route in
industries for SA production. However the cost of production of bio-based SA at the present
time is lower than that of the petrochemical one when the oil price is above $30 per barrel [4].
Currently, bio-SA is produced commercially by several companies including Myriant,
Reverdia, Succinity, and BioAmber. It has been estimated that the global SA market would
steadily grow at a compound annual growth rate around 27% and reach 768 million tons at
$2.3/kg in 2025 [3].
SA can be hydrogenated to valuable products including γ-butyrolactone (GBL), 1,4-
butanediol (BDO) and tetrahydrofuran (THF) [5–8]. These chemicals are mainly employed as
solvents (e.g. THF), food additives, intermediates for polymer synthesis (e.g. BDO and THF)
and pharmaceuticals (e.g. GBL) [2,9,10]. Water is preferably chosen as solvent for this
reaction since after fermentation, SA is obtained in aqueous phase. It is then necessary for the
catalysts to be stable under the severe hydrothermal conditions (T ≥ 160 °C, P ≥ 80 bar, water)
that are required for this reaction. Noble metal supported catalysts have principally been
studied for the hydrogenation of SA in aqueous phase [5–8]. The selectivity to the products
varies according to the catalyst used and the reaction conditions. The main limitations to the
use of noble-metal catalysts are their scarcity and high prices. For an economically viable
2

valorisation of biomass derivatives, it would be interesting to replace group VIII (8 to 10)
metals.
A review of the literature dealing with the formulation of catalysts for heterogeneous
processes revealed the development and use of transition metal carbides “TMCs” [11–13].
TMCs display catalytic activity approaching those of group VIII metals, but with unique
catalytic pathways and different products selectivity [13]. Experimental and theoretical
examinations of these materials in literature confirm that the introduction of carbon into the
lattice of the early transition metals results in an expansion of the lattice parameter. Density
functional theory (DFT) calculations have shown that the hybridization between metal d-
orbitals and the carbon s- and p-orbitals causes a broadening of the d-band structure,
providing characteristics approaching the d-band of Pt [14].
Among TMCs, molybdenum and tungsten carbides are known to be active in hydro-
desulfurization (HDS), hydro-denitrogenation (HDN), water–gas shift, isomerization, and
Fischer Tropsch reactions [15–18]. Recently, they have emerged as efficient
hydrodeoxygenation (HDO) catalyst for the selective deoxygenation of biomass-derived
oxygenates such as vegetable oils, phenolic compounds (e.g. guaiacol, m-cresol) and furanic
compounds (e.g. furfural, HMF) [12,19–21]. They are often employed for gas phase reactions
or in organic solvent, and their use for the aqueous conversion of biomass derivatives to
chemicals is limited. However, Teixeira da Silva et al., investigated molybdenum carbide
supported on activated carbon for the aqueous hydrogenation of levulinic acid. The catalyst
exhibited high conversion (~99%), with selectivity towards γ-valerolactone above 85% at
200 °C and 30 bar in aqueous phase [22].
The challenging aspect for the preparation of TMCs, is controlling their crystallographic
structure and morphology. Molybdenum carbide systems consist of different phases, most of
which have two compositions, MoC and Mo2C, which may co-exist [23]. The main
crystallographic phases observed are the hexagonal, cubic, and orthorhombic molybdenum
carbide phases [24]. The composition can directly affect the performance of the catalyst. For
example a study of CO2 hydrogenation reaction using molybdenum carbide showed that
hexagonal close packed structure is more active than the face centered cubic system.
Moreover, the former one showed higher selectivity to methane while the latter one showed
higher selectivity to methanol [25]. These materials are usually prepared by a temperature
programmed reduction-carburization method. The metal oxide powder is placed under a
reductive-carburizing gas flow, while raising the temperature slowly in a uniform heating
ramp until reaching the maximum temperature needed [26]. This temperature-programmed
3

carburization process is sensitive to many parameters that make the choice of the preparation
conditions critical. The carbon sources, composition of the mixture (i.e. % v/v HC/H2),
heating rate, maximum carburization temperature, and holding time have direct influence on
the structure and physicochemical properties of the final material [26–31]. For example, the
nature of carbon source and its concentration along with the maximum temperature affect the
crystallographic structure, the carbon content of the material, and its crystallite size. On the
other hand, the heating rate and holding time affect the surface area and morphology of the
metal carbide formed.
The use of a support in the preparation of the catalyst can increase the exposed surfaces
that lead to a global increase of the activity. Highly dispersed molybdenum carbide supported
on varied supports such as Al2O3 [32], ZrO2 [33], and carbon nanotubes [22,34] may be
prepared by depositing a precursor on a high surface area support, followed by carburization.
TiO2 has previously been used as a support due to its stability under hydrothermal conditions
for the aqueous phase hydrogenation of SA [6].
In this paper, we report the preparation of TiO2-supported molybdenum carbide catalysts
via impregnation followed by temperature-programmed reaction using 20% v/v C2H6/H2.
These catalysts were studied for the aqueous phase hydrogenation of succinic acid. The effect
of the preparation conditions of the catalyst, i.e. Gas Hourly Space Velocity (GHSV) and
passivation, on its properties and its catalytic performance in terms of activity, and products
selectivity was investigated. The stability of the catalyst was also studied under harsh reaction
conditions (240 °C, 150 bar H2, in water).
2. Experimental section
2.1. Preparation of catalysts
Commercial TiO2 P25 (specific surface area 55 m2.g-1, supplied by Degussa-Evonik) was
used as the support. MoC/TiO2 catalysts were prepared by impregnation followed by
temperature programmed reduction carburization. To prepare 12% w/w MoC/TiO2, 1.2 g of
ammonium molybdate tetrahydrate (NH4)6Mo7O24.4H2O (Aldrich Chemical Co, 99.98% trace
metals basis) was mixed with 5 g of TiO2 and 60 mL of water and stirred for 2 h at room
temperature. Water was evaporated under vacuum and the solid was dried in an oven at 80 °C
overnight. For the second step, the powder was placed in a quartz cell under a reductive-
carburizing gas stream (20% v/v C2H6/H2, 60-140 mL.min-1). GHSV was calculated based on
gas flow rate (mL.h-1) divided by the volume of catalyst bed (1.1-5.5 mL). GHSV was varied
4

in the range of 1527-7636 h-1. The temperature was raised at 0.5 °C.min-1 until 700 °C, held
for 2 h, and then cooled down to room temperature under argon. The non-passivated (NP)
samples were transferred to a glovebox and kept under nitrogen. The other samples were
passivated for 4 h under 1% v/v O2/N2 mixture, and then kept in a vial, in N2 atmosphere. For
comparison, a MoO3/TiO2 solid was prepared by impregnation, followed by calcination under
air at 0.5 °C min-1 till 600 °C, held for 2 h, then cooled down to room temperature.
2.2. Characterization of catalysts
The Mo content of the catalysts was determined by inductively coupled plasma optical
emission spectroscopy (ICP-OES) by using an ACTIVA instrument (HORIBA Jobin Yvon).
Before analysis, the samples were mineralized by fusion with lithium tetraborate in Pt-Au
crucibles at 1100 °C, and then soaked with 20% HCl.
Carbon elemental analyses were conducted via a LECO micro-analyzer SC144. The total
combustion of the samples was done at 1050 °C under a stream of helium/oxygen and under
pressure. The carbon was converted to carbon dioxide and quantified by a thermal
conductivity detector.
Oxygen analysis was done with a EMGA 620 W analyzer (HORIBA Jobin Yvon). The
system uses fusion in a helium pulse furnace to extract oxygen. The sample falls into a
graphite crucible heated to a high temperature between 2500 °C and 2800 °C. Oxygen which
is converted into carbon monoxide by combining with the carbon of the crucible, is measured
by a non-dispersive infrared cell, then oxidized to CO2 and trapped.
BET surface areas (S) of the samples were determined from the N2 physisorption at -196
°C using an ASAP 2020 Micromeritics apparatus. Before the measurements, the catalysts
were desorbed at 150 °C for 3 h under ultra-high vacuum (10-4 mbar).
Powder XRD patterns of the catalysts were recorded in the range 2θ = 20-80° at 0.04°.s-1
using a Bruker D8A25 X-ray diffractometer and a CuKα radiation source (λ=1.54184 Å).
Phase identification, lattice parameters, and mean crystallite sizes (d = 4/3 * LVol-IB; with
LVol-IB the volume averaged column height) were obtained by performing Rietveld
refinement using the Topas 5 software.
TEM and STEM images were obtained using Environmental Transmission Electron
Microscope: FTI TITAN ETEM instrument operated at 80-300 kV, equipped with a X-MAX
SDD EDX detector from Oxford-Instrument and a Tridiem ERs GIF from Gatan. Samples
5

were prepared by dispersing the solid in ethanol and then depositing them onto carbon-coated
grids.
The surface chemical composition and oxidation states of Mo in catalyst samples were
analyzed by XPS with monochromatized AlKα source (һν = 1486.6 eV) using a commercial
instrument (AXIS Ultra DLD KRATOS). The binding energies (± 0.5 eV) were referred to
the C1s line set at BE = 284.6 eV. Peaks decompositions, fittings and quantitative
determinations were performed using the Casa XPS and Igor Pro software. The non-
passivated samples were transferred after synthesis to a glovebox for preparation and then to
the XPS chamber. There was no contact with oxygen as the samples stayed under inert gas
(Ar or N2).
Raman spectra were recorded at room temperature using a LabRam HR (Jobin Yvon–
Horiba) spectrometer equipped with a CCD detector cooled at -76 °C. Measurements were
carried out under microscope with a 50x objective that focuses the laser beam on the sample
surface and collects the scattered light. The exciting line at 514.53 nm of a Ar+ laser was used
with a power limited at 100 microW. It was previously checked that the samples laser heating
was negligible with such power. A 300 grooves/mm grating was used to disperse light leading
to band position accuracy within 4 cm-1.
2.3. Catalytic testing
SA with purity above 99% was purchased from Aldrich. The aqueous catalytic
hydrogenation experiments were performed using a 300 mL high-pressure batch (Parr 4560)
Hastelloy autoclave, equipped with a magnetically driven impeller and a liquid sampling
system. The reaction temperature is monitored by a thermocouple probe which sites inside a
thermowell, in the reactor. A 100 mL aqueous solution of SA (0.12-0.14 M; pH = 2.3) and 0.6
g of the catalyst were loaded into the reactor. After sealing, the autoclave was purged three
times with Ar, heated up to the desired reaction temperature (160-240 °C), pressurized with
H2 (90-150 bar) and stirred at 900 rpm. The reactions were conducted for 22-32 hours and
samples were collected periodically every 2 hours during the day in order to follow the
evolution of the reaction.
Analyses of the reaction products in the liquid samples were performed using a GC Agilent
Technologies 6890N with a flame ionization detector employing a VF-WAXms column (30
m × 0.25 mm × 0.25 μm). The temperature was increased up 190 °C, under helium as a carrier
gas. The concentration of SA was monitored using a Shimadzu LC 20A HPLC connected to
UV and RI detectors. The separation was achieved using ICSep Coregel 107H column (7.8 ×
6

300 mm) heated at 40 °C. A solution of H2SO4 (0.001 mol L-1) in ultra-pure ELGA water was
used as the mobile phase at a flow rate of 0.5 mL min-1. After some reactions, the gases were
collected in a gas-bag and analyzed using a micro-GC SRA with MS Agilent 5975 detector.
Three columns were set up: Alumina (10 m × 3 µm) at 90 °C for C3–C4 hydrocarbons,
Poraplot U (8 m × 30 µm) for C2–C3 compounds and CO2, and MolSieve 5A (10 m × 12 µm)
for H2, CO, and CH4.
SA concentration (%) and conversion (%) are based on initial concentration of SA ([SA]0)
and defined by,
and
(1)
where [SA]t refers to the concentration of SA at time t.
Selectivity and yield of a product i are given by,
and
(2)
where [Pi]t refers to the concentration of the product i at time t.
The carbon balance (CB) is given by,
(3)
where ni refers to the number of carbon of the product i.
The initial reaction rate was calculated based on the slope of the linear curve (up to 6-8h),
at low conversion (< 40%),
(4)
The total organic carbon (TOC) in solution was measured by using a Shimadzu TOC-
VCSH analyzer and use to check the carbon balance in the liquid phase and the possible
formation of gaseous products. The samples were diluted by a factor of 70 before TOC
analysis.
Repeated reactions delivered conversion and products yield reproducible within ± 3%.
Therefore the selectivity is given with an error of ± 3% and the reaction rate with an
-1
uncertainty of 1 .0 mmolSA gMo h-1. Carbon balance and TOC values are given with a
precision of 8%.
3. Results and discussion
7

3.1. Preparation and characterization of the catalysts
The supported molybdenum carbide catalysts were prepared by impregnation of
ammonium heptamolybdate tetrahydrate on TiO2 at room temperature. The impregnation
occurs at pH = 5.3 which is slightly lower than the PZC of TiO2 P25 (pHPZC = 6.3) [35]. The
predominant species in solution at that pH should be Mo7O246- [36,37]. The support must bear
a slight positive charge which should favor the interaction with the anionic species. Water was
removed by rotary evaporator in order to insure the total deposition of Mo, which was
confirmed by ICP-OES analysis of the final materials (~12 wt.% Mo, Table 1).
Table 1. GHSV employed during the catalysts synthesis. Mo and C content (% wt.), surface area (S), and
mean crystallite size of MoC, anatase TiO2, and rutile TiO2.
Catalysts
Synthesis
Characterization
Crystallite sizee (nm)
Sd(m2.g-1) Cubic
Anatase Rutile
MoC
GHSVa (h-1) % Mob % Cc
TiO2
TiO2
1527
2545
7636
11.8
12.0
11.9
2.2
3.7
5.3
45
47
46
4
50
45
MoC-I/TiO2
MoC-II/TiO2
MoC-III/TiO2
4
4
45
50
25
11
a GHSV of 20% v/v C2H6/H2 during the reduction-carburization process
b ICP analysis
c Carbon elemental analysis
d BET surface area
e Rietveld refinement
The powder obtained after impregnation and drying was then placed under a reductive-
carburizing gas stream (20% v/v C2H6/H2) while raising the temperature up to 700 °C. The
effect of GHSV has been widely studied for bulk molybdenum nitride [38]. It has been
demonstrated that a decrease in space velocity during the reaction causes a loss of surface
area. The use of high space velocities would facilitate the removal of H2O from the vicinity of
the solid surface, thus reducing the potential for hydrothermal sintering [38,39]. However we
could not find any report on the effect of GHSV on the synthesis of supported molybdenum
carbide. In the case of TMCs, the GHSV factor might also affect the carbon content of the
catalyst. A series of TiO2-supported molybdenum carbides, catalysts were synthesized by
changing the GHSV (1527 h-1, 2545 h-1, and 7636 h-1) and were denoted MoC-I/TiO2, MoC-
II/TiO2, and MoC-III/TiO2, respectively (Table 1). The three catalysts exhibit similar BET
surface area of 45-47 m2.g-1, slightly lower than the support TiO2 (Table 1).
8

Figure 1. XRD diffraction patterns of the support P25 TiO2 P25 and the catalysts: a) MoC-I/TiO2, b)
MoC-II/TiO2, c) MoC-III/TiO2. Note: peaks associated with ● Anatase, ○ Rutile, * Cubic MoC phase.
Rietveld refinement allowed us to conduct a quantitative identification of the different
crystalline phases present in the catalysts. Concerning the support, the main peaks
corresponding to TiO2 anatase and rutile phases were observed (Figure 1). For the three
catalysts, the composition anatase/rutile was the same as for the support P25 (75%/25%).
Anatase is known to convert to rutile at temperature between 600 and 700 °C [40]. However
in the presence of a dopant, such as metal oxide, this temperature is increased [41]. Therefore
the presence of the Mo species at the surface might stabilize the anatase phase during the gas
treatment at high temperature (700 °C). The lattice parameters of the anatase and rutile phases
are similar in the three samples (Table A.1 in appendices). It is however worth noting that the
mean crystallite size of the minor rutile phase decreased from ca. 45 nm (bare support and
MoC-I/TiO2) to 11 nm (MoC-III/TiO2) when increasing the GHSV during the preparation
(Table 1), while the anatase phase was not affected.
Regarding the Mo crystalline carbides, the main peaks attributed to cubic Mo carbide
phase at 2θ = 36.4°, 42.2° and 61.3° were detected for all the catalysts. They can be associated
with the planes (111), (002), and (022), respectively. The peaks were broad and overlapped
with the ones from the supports, which might alter the analysis. Rietveld refinement
confirmed the presence of the cubic Mo carbide phase (> 10 %) with crystallite size below 5
nm (Table 1). The presence of molybdenum oxides, such as MoO2 or MoO3, was not detected,
which suggests full carburization.
9

The reports in the literature on the effects of carburizing agent on the structure of bulk
TMCs show that the type and concentration of the carbon source affect the crystallographic
structure [27,28,31]. When increasing the chain length of the carburizing agent, lower
temperatures are required for the complete transformation from molybdenum oxide to carbide
[27]. The hexagonal hcp phase is usually obtained when conducting the carburization with
CH4, while carburizing agent with longer chain, such as C4H10 and heptane, or aromatics,
such as toluene, favor the formation of the cubic fcc phase [27,31,42,43]. Carburization with
C2H6/H2 generates a mixture of fcc and hcp phases at low concentration of carbon 10 % v/v
[28], while pure cubic phase is obtained at 20 % v/v [25].
The literature dealing with molybdenum carbides supported on TiO2 is really limited [44–
46]. One study reported the formation of large crystallite (~ 25 nm) of hexagonal Mo2C on
TiO2 using 20% v/v CH4/H2 at 650 °C [44]. The preparation of MoC1-x/TiO2 with 20% v/v
C3H8/H2 generated a mixture of particles of fcc and hcp phases with a mean crystallite size
around 15 nm [45]. It is worth noting that the carburization at 700 °C in CH4/H2 can generate
small crystallites (< 5 nm) of cubic MoC over activated carbon [34] and carbon nanotubes
[47]. Therefore, we report the first synthesis of small crystallites (< 5 nm) of cubic MoC over
TiO2, by employing 20 % v/v C2H6/H2 at 700 °C.
Figure 2. Representative TEM image (a, a.1), electron diffraction (a.2), and STEM image (b) of MoC-III/TiO2.
10

Table 2. Lattice parameters obtained from TEM analysis.
Spot h k l
d (nm)
Exp.
°)
Exp.
Theo.*
Theo.*
2 0 0
2 2 0
0 2 0
-2 2 0
0.00
0.00
0.212
0.150
0.214
0.153
0.213
0.150
0.213
0.150
1
2
3
4
44.37
89.40
45.00
90.00
134.49 135.00
* Theoretical values corresponding to cubic MoC
Representative TEM and STEM images of MoC-III/TiO2 are shown in Figure 2 and Figure
A.1. Small particles (< 5 nm) are dispersed on the support. Local EDX analyses and
diffractogram patterns (Table 2, Table A.2) revealed that these are particles of MoC with
cubic phase, which is in agreement with XRD results. An excess of graphitic carbon was
clearly seen in the images (Figure 2-a). The TEM and STEM analyses of the three catalysts,
MoC-I/TiO2, MoC-II/TiO2, and MoC-III/TiO2, exhibited similar results, i.e. presence of cubic
MoC particles and excess of carbon on the surface (Figure A.2 and Table A.3).
Carbon elemental analysis was conducted to estimate the carbon content of the catalysts.
The results in Table 1 show that the samples contained between 2.2 and 5.3 wt. % of carbon
while the theoretical carbon content for MoC/TiO2 with 12 wt. % of Mo should be 1.5 %.
This reflects an excess of free-carbon in the catalysts, as observed on the TEM pictures with
the presence of graphitic carbon. Moreover the carbon content increased from 2.2% to 5.3%
while increasing the GHSV during the preparation from 1527 to 7636 h-1 (Table 1). The
presence of free carbon at the surface of TMCs is well known. It is deposited on the surface
during the synthesis, as ethane starts to decompose from 590 °C [28]. The temperature for
complete formation of bulk molybdenum carbide is around 630 °C when using 10% v/v
C2H6/H2, however, supported catalysts often require more severe conditions [26,28]. The
presence of Ti dopant can facilitate hydrocarbon decomposition and carburization [48]. Thus,
the excess of carbon is not surprising after synthesis at 700 °C with high concentration of
hydrocarbon (20% v/v C2H6/H2).
Due to their pyrophoricity, TMCs are typically passivated prior to use in order to facilitate
their handling. However, it has been shown that the passivation treatment can affect the
catalytic results. For example, Nagai et al., have shown that passivated molybdenum carbide
catalysts supported on alumina were less active for CO2 hydrogenation reaction than the non-
11

passivated catalysts [49]. On the contrary, Mehdad et al. showed that after the removal of the
passivation layer of molybdenum carbide, the catalyst exhibited identical performance as the
non-passivated one for toluene hydrogenation [50]. Therefore, two catalysts were also
synthesized without passivation treatment (NP is added to the same catalysts labelling) in
order to investigate the effect of this treatment on the properties and performance of the
catalyst in SA hydrogenation. Limited characterization could be performed on these two non-
passivated catalysts as they cannot be exposed to air, except for XPS analysis. For the latter
one the samples were transferred directly after synthesis to the XPS chamber.
Table 3. XPS analysis: atomic concentration of Mo and C, and the abundance of Mo and C species.
Atomic
concentration %
Mo species %
C species %
Ratio
C Carbide/Mo δ+
Mo
3.3
3.9
4.8
C
Moδ+
47
Mo4+ Mo6+ Carbide Graphite
Catalyst
MoC-III/TiO2
NP-MoC-III/TiO2
NP-MoC-I/TiO2
46.5
49.2
22.4
8
45
0
8
74
65
47
1.4
1.6
0.7
88
12
13
16
17
76
12
XPS analyses were conducted on 3 catalysts in order to assess the differences resulting
from the passivation (MoC-III/TiO2 vs. NP- MoC-III/TiO2) and from variation of the GHSV
(NP- MoC-I/TiO2 vs. NP- MoC-III/TiO2). The Mo 3d and C 1s spectra of MoC-III/TiO2, NP-
MoC-III/TiO2 and NP- MoC-I/TiO2 are presented in Figure 3. The chemical states of the
catalyst components and their relative abundance are compiled in Table 3.
In Figure 3-a-1, the Mo 3d spectrum of MoC-III/TiO2 exhibits two peaks at BE = 228.7 eV
+
and 231.9 eV which correspond to Mo 3d5/2 and Mo 3d3/2 peaks of Mo . The former value
falls within the range 227.6 - 228.9 eV which is attributed to carbidic Mo (Mo next to carbon)
in the literature [51,52]. In addition, contributions assigned to Mo4+ (BE = 230.2 eV and 233.5
eV) and Mo6+ (BE = 232.7 eV and 235.8 eV) were also observed [53,54]. The difference
between the passivated and non-passivated catalysts is clearly noticeable when comparing Mo
3d spectra of MoC-III/TiO2 and NP- MoC-III/TiO2 in Figure 3-a-1 and Figure 3-b-1,
respectively. The absence of Mo6+ for the fresh sample (non-passivated), confirms that the
presence of Mo6+ in the passivated sample is due to the passivation treatment. The absence of
12

molybdenum bulk oxides phases (by XRD) suggest that the passivation treatment results in a
thin layer of oxide on the surface. It is assumed that oxygen adsorbs on top of Mo atoms and
hollow sites that are not occupied by C [50]. From the literature, it is expected to see a small
amount of Mo4+, even for the non-passivated sample [51]. Oxygen elemental analyses were
also conducted and no difference was observed for the two catalysts. However the fact that
the support is an oxide limits the efficiency of the analysis. As expected, MoC-III/TiO2 and
NP- MoC-III/TiO2 present similar Mo and C atomic concentration knowing that the Mo
theoretical atomic concentration of 12% w/w MoC/TiO2 is 3.6 (Table 3). The C atomic
concentration is in excess due to the presence of graphite on the surface. The C 1s spectrum of
MoC-III/TiO2 displayed in Figure 3-a-2 exhibits a peak at BE = 283.4 eV which corresponds
to the carbon in the carbidic form [52]. The broad peak at BE = 284.6 eV along with the peak
at 289.5 eV refer to graphitic carbon. The additional peak at 286 eV can be attributed to
carbon in C-O or C=O groups [52,55]. The ratios of carbidic carbon to Mo carbide are shown
in Table 3 and are around 1.5 for the two samples. The Ti 2p spectrum of MoC-III/TiO2 is
shown in Figure A.3 as an example. The peak at BE = 459.1 eV refers to Ti4+ and no peak
corresponding to Ti3+ (BE = 457.5 eV) was observed.
The sample NP- MoC-I/TiO2 synthesized at lower GHSV (Figure 3-c, Table 3) presents
13
Figure 3. XPS spectra of catalysts a) MoC-III/TiO2, b) NP- MoC-III/TiO2, and c) NP- MoC-I/TiO2 over 1)
Mo 3d and 2) C 1s.

similar compositions as NP- MoC-III/TiO2. However, the presence of a small amount of Mo6+
suggests that the carburization was not complete. When looking at the atomic concentration
determined by XPS, it is noticeable that the carbon content on the surface of the catalyst is
lower (22% vs. 49%) when decreasing the GHSV. This result is in agreement with the carbon
elemental analysis. Moreover the fraction of carbon graphite is slightly lower (from 65% to
47%) in favor of carbon as carbonyl/carbonate groups. The lower Mo % surface concentration
on MoC-III/TiO2 might be related to the higher concentration of carbon in the sample. Indeed,
the excess of carbon on the surface might be masking the Mo particles because only few
nanometers are probed by XPS.
3.2. Catalytic results
3.2.1. Hydrogenation of succinic acid over MoC/TiO2
The catalytic performance of MoC-II/TiO2 catalyst was evaluated for the aqueous phase
hydrogenation of succinic acid (SA) at 240 °C and 150 bar. Without catalyst, no conversion
was observed (< 1%). A typical temporal evolution of the substrate and products in solution is
presented in Figure 4. In the presence of the catalyst, SA was progressively converted with
time and full conversion was achieved after 22 h. From the start of the reaction, parallel
formation of γ-butyrolactone (GBL) and unexpectedly butyric acid (BA) were observed. After
8 h of reaction, 60% conversion was achieved with 38% yield of GBL and 26% yield of BA.
Proceeding more in the reaction, the two intermediates started disappearing along with the
formation of butanol (BOL), tetrahydrofuran (THF), and 1,4-butanediol (BDO). After 32 h,
22% of GBL, 6% of BA, 21% of BOL, 14% of THF, and 4% of BDO are present in solution.
Traces of propionic acid were also observed, close to the detection limit (< 0.5%). The carbon
balance (CB) started decreasing after 18 h and only 66% of the products were still present in
14

liquid phase at the end of the reaction. This was confirmed by qualitative GC-MS analysis of
the gas phase where butane as well as traces of propane were detected. All the catalytic results
were reproducible within 5%. The CB calculated based on GC and HPLC analysis were
always very close to the TOC measured, which means that all the products in liquid phase
were identified and analyzed (Figure A.4).
Figure 4. Hydrogenation of succinic acid over MoC-II/TiO2 at 240 °C and under 150 bar H2: temporal evolution
of the concentrations of SA, the products yield and the carbon balance (CB). Aqueous solution of SA (0.12 M,
100 mL), 0.6 g of catalyst.
The selectivity to the products was distinct from the results reported in literature for SA
hydrogenation over group VIII metals catalysts. For example, at 160 °C and 150 bar over
Pd/TiO2, SA is converted to GBL with 94% selectivity [6]; under the same reaction conditions,
introducing Re to this catalyst favours the conversion of GBL to BDO with 83% selectivity
[7]. On the other hand, using Re/C at 240 °C and under 80 bar of H2 generates THF with 86%
selectivity [8] (Scheme 1). In the literature dealing with the hydrogenation of SA, BA is
usually not observed. A couple of studies reports the formation of small amounts of BA (<
3%), e.g. over Au-Pt/TiO2 [9], Pd-ReOx/TiO2 [56], FeOx/C + Pd/C [57], and Pd/SiO2-NH2
[58]. A combined selectivity of 18% towards propionic acid and butyric acid (at 80%
conversion) was reported during the hydrogenation of SA over Pd/alumina xerogel, in
dioxane (P = 60 bar, T = 240 °C) [59]. In another study over Pd/alumina, they reported that
2
H
the selectivity towards the acids increased with increasing palladium dispersion on the surface
[60]. BOL was observed as by-product (< 5%) in a couple of studies, e.g. over Re/C [61] and,
with maleic acid as reactant, over Pd/TiO2 and Re/TiO2 [62]. The formation of butane has not
been reported in the literature before. These reflect major differences with the selectivity
obtained over carbide catalyst.
15

Scheme 1. Suggested reaction pathway for the aqueous phase hydrogenation of SA over MoC/TiO2.
GBL is formed as intermediate through hydrogenation and dehydration of SA [7].
Dehydration reactions are usually favored on acid sites while hydrogenation is favored on
metal sites. BDO is formed by further ring opening and hydrogenation of GBL. This might be
due to the simultaneous activation of both oxygens of GBL, as it has been proposed over Re-
Pd/TiO2 [7]. THF can be formed by dehydration of BDO in the presence of acid catalysts. The
direct hydrogenation of the carbonyl group of GBL is more likely in water. BOL, n-propanol,
BA and propionic acid can be formed as further by-products. It was proposed that BA and
propionic acid are formed from SA or GBL, while BOL and propanol are formed by
hydrogenation of BDO or the monoacids [63]. BOL is usually obtained from further
hydrogenolysis of BDO, favored by acid sites [64]. It can generate n-propanol by C-C bond
cleavage [8], e.g. over Re [63]. In a series of publications, Liang ad co-workers [8,61,63]
investigated the reaction pathway of the aqueous hydrogenation of SA over Pd/C, Re/C, Ru/C,
Re-Ru/C and Pd-Re/C, at 240 °C and under 80 bar H2. Starting from GBL, high selectivity
towards THF was observed over Re/C, while a mixture of BDO and THF was obtained over
Re-Ru/C. Starting from BA, 100% selectivity towards BOL was observed, over all the
catalysts. The conversion of BDO generated mainly THF over Re/C, while propanol was also
formed in the presence of Ru based catalyst. The conversion of THF was negligible.
16

Since the products selectivity observed during the hydrogenation of SA over MoC/TiO2 is
distinct from over noble metals catalysts, the reaction pathway was investigated in more detail.
The hydrogenation reactions of GBL, BA, and BDO were conducted under the same reaction
conditions as before. Figure 5 compares the product distributions at around 50-60%
conversion and the initial rates of reaction (V0) for the hydrogenation reactions of these
intermediates and SA. The temporal evolution of these reactions are included in Figure A.5.
Figure 5. Selectivity to the products at 50-60% conversion and initial rate of reaction (V0) for the hydrogenation
reactions of SA, GBL, BA, and BDO, at 240 °C and 150 bar. Aqueous solution of reactant (0.12 M, 100 mL),
0.6 g of MoC-II/TiO2 catalyst.
THF (31%) was the main product during the hydrogenation of GBL, in addition to BOL
(28%), BA (16%) and BDO (14%). BOL and BA were formed simultaneously which suggest
that BOL can be formed directly from GBL. The hydrogenation of BA generated solely BOL
in liquid phase, as reported elsewhere [63]. The hydrogenation of BDO gave mainly THF
(57%) in addition to BOL (10%). Lastly, THF was stable in these reaction conditions where
barely 7% was converted to BDO after 30 h of reaction. The hydrogenation of the
intermediates suggests that BA can be formed directly from SA as well as from GBL.
Moreover gaseous products (i.e. butane) are mainly formed from BA. The rate of reaction
follows the order: V0 (SA) > V0 (BA) > V0 (BDO) > V0 (GBL) >> V0 (THF) (< 0.1 mmolTHF
gMo-1 h-1). These reactions lead to the proposal of the reaction pathway shown in Scheme 1.
Molybdenum carbide catalysts are known to be active for C=O hydrodeoxygenation
reaction. It was proposed that the hydrogenation steps occur over metallic sites, while the
dehydration occurs over Brønsted acid sites [65]. The hydrodeoxygenation of acrylic acid to
propane has been studied in gas phase over Mo2C [32]. The experimental and theoretical
results suggested that the hydrogenation/dehydration of the carboxylic acid generates
17

adsorbed aldehyde, which undergoes further hydrogenation/dehydration and forms propane.
Therefore BA might be formed via successive hydrogenation/dehydration of SA, without
desorption of the intermediate aldehyde.
In the literature, the conditions for aqueous phase hydrogenation of SA in the presence of
noble metal catalysts are in the range of 80-150 bar of H2 and 160-240 °C [5,6,8,63]. The
effects of temperature and pressure were investigated over MoC-II/TiO2. The reaction was
conducted at different temperatures (160, 200, and 240 °C) and H2 pressures (90, 110, 150
bar) for 22 h (Table 4). The initial rate of reaction (V0) increased with temperature and
pressure. According to the Arrhenius plots (Figure A.6), the apparent activation energy is 55
kJ mol-1 for MoC-II/TiO2. It is worth to mention that the effects of temperature and pressure
were only limited on the activation of SA. Indeed, Figure 6 represents the selectivity towards
the products in function of conversion, for all the reactions conducted over MoC-II/TiO2. The
selectivity values were similar, independently of the reaction conditions, implying that the
products distribution is not affected by the change of temperature and pressure. When going
from 2 to 85% of conversion, the selectivity towards BA decreased slightly (from 52 to 43%).
Above 90% SA conversion, GBL and BA started to be converted and the products BOL, BDO
and THF appeared. It is well known that there is a competitive adsorption of SA and GBL.
Indeed, it is usually observed that GBL starts to be converted when SA has disappeared [56].
In the same manner, we observed that the conversion of BA to BOL occurred only when SA
was almost fully converted. It is also worth noting that fare less gaseous products were
observed when working at 90 or 110 bar of H2. This was confirmed by the fact that the TOC
measured and the carbon balance remained close to the initial value (Figure A.7). This is due
to the fact that at low conversion, hardly any butanol, hence butane, were formed.
Table 4. Effect of temperature and pressure on initial rate (V0), SA conversion and product yields
Products yields (%)a
T
P
V0
SA Conversiona
(°C) (bar) (mmolSA gMo-1 h-1)
(%)
4
GBL BA
BOL THF BDO
160
200
240
240
110
110
110
90
0.4
1.6
4.5
3.4
12.5
2
2
0
0
0
1
0
9
0
0
0
0
2
20
12
33
24
45
8
0
61
26
24
17
1
48
1
240
150
100
15
a
after 22 h reaction : aqueous solution of SA (0.12 M, 100 mL), 0.6 g of MoC-II/TiO2 catalyst.
18

Figure 6. Selectivity to GBL, BA, BOL, and THF as function of SA conversion, for all reactions conducted over
MoC-II/TiO2 catalyst, at different pressure and temperature (see Table 4). Aqueous solution of SA (0.12 M, 100
mL), 0.6 g of MoC-II/TiO2 catalyst.
3.2.2. Effect of preparation conditions of the catalyst on its performance in succinic acid
conversion
The initial rate and products distribution of MoC-II/TiO2 were then compared to those of
MoC-I/TiO2 and MoC-III/TiO2, at 240 °C and 110 bar of H2 and the results are included in
Table 5. For these reactions the CB was always over 90%. By increasing GHSV of C2H6/H2
during the carburization of the catalyst from 1527 h-1 (MoC-I/TiO2) to 7636 h-1 (MoC-
-1
III/TiO2), V0 increased from 1.6 to 7.8 mmolSA gMo h-1 and the selectivity shifted towards
BA. Indeed, when comparing the selectivity at ca. 23% conversion, MoC-III/TiO2 gave
selectivity towards BA of 71% while 44 and 49% were obtained over the other two catalysts.
In the catalyst characterization section it was shown that as the GHSV increases, the carbidic
Mo and the free-carbon contents increase. Therefore the degree of carburization and/or the
presence of free carbon must have a strong impact on activity and selectivity. A supported Mo
oxide catalyst MoO3/TiO2 was also prepared and tested for comparison (Table 5). High
selectivity to GBL was observed (95%). This shows that the carbon content has direct effect
on shifting the selectivity towards BA.
Table 5. Effect of GHSV during catalyst carburization on the initial rate and selectivity at 110 bar and 240 °C.
Selectivitya (%)
V0
Timea
(h)
22
Catalyst
(mmolSA gMo-1 h-1)
GBL
56
BA
44
49
71
1
BOL
THF
1.6
4.5
7.8
4.1
0
0
0
0
0
0
0
0
MoC-I/TiO2
10
51
MoC-II/TiO2
MoC-III/TiO2
MoO3/TiO2
6
29
10
95
a at 22-25 % conversion
The effect of the passivation treatment on the catalytic results was investigated under 110
bar of H2. Figure 7-a presents the results obtained over MoC-III/TiO2, which was used
directly after passivation. The non-passivated catalyst “NP- MoC-III/TiO2” (Figure 7-b) was
transferred to the reactor in a glovebox, directly after synthesis, without passivation treatment.
It is obvious that the two catalysts exhibited really similar catalytic response in terms of
activity and selectivity. The conversion reached 87-92% after 24 h reaction for both samples
19

-1
and the initial rates was of 7.6 ± 0.2 mmolSA gMo h-1 (the difference is within experimental
error ± 5%). The yield of GBL and BA reached around 30 % and 50%, respectively.
Therefore the passivation layer does not affect the catalytic performance of the catalyst. As no
pre-activation was conducted on the passivated sample, the reduction of the passivated sample
must occur in situ, at the beginning of the reaction. The XPS results showed that no more
Mo6+ and a lower amount of Mo4+ species were present in NP- MoC-III/TiO2 (Figure 3-b;
Table 3). These results suggest that the oxides must be reduced at the beginning of the
reaction and the selectivity must depend on the Mo carbide content, which is similar in both
catalysts.
Figure 7. Hydrogenation of succinic acid over (a) MoC-III/TiO2 and (b) NP-MoC-III/TiO2 at 240 °C and under
110 bar of H2: temporal evolution of the concentrations of SA, the product yields and the carbon balance (CB).
Aqueous solution of SA (0.12 M, 100 mL), 0.6 g of catalyst.
3.2.3. Stability of the catalyst
The stability tests were conducted with MoC-II/TiO2 at 240 °C and 150 bar for 26 h, which
are the harshest conditions used in this work. The results obtained during the first run over
MoC-II/TiO2 are shown in Table 6 and Figure A.8. The catalyst was then filtered, washed
with distilled water (under air) and dried at 80 °C; this catalyst is denoted R1- MoC-II/TiO2.
The percentage of Mo in the solution recovered after the first run was below 0.1 ppm, which
means that no leaching occurred (< 0.1%).
Catalyst oxidation is a possible cause of deactivation for metal carbides catalysts [66]. The
hydrothermal stability of Mo2C in water at 250 °C for 48 h has been investigated in a previous
study [67]. The formation of MoO2 (by XRD) was observed and due to the oxidation by H2O.
In order to assess the presence of oxides in the solids (crystalline or amorphous), Raman
analysis were done for the catalyst before (MoC-II/TiO2) and after reaction (R1- MoC-
II/TiO2). In Figure 8, the bands at 145, 196, 397, 517, and 637 cm-1 correspond to anatase
20

phase of TiO2 and the rutile phase is evidenced by two extra broad bands at 440 and 605 cm-1
[68]. The two bands at 1345 and 1597 cm-1 correspond to the disorder D and tangential G
bands of graphitic carbon [31,69,70]. The presence of these two bands have previously been
reported for molybdenum carbides synthesized under 10% C2H6/H2 at 800 °C [31]. The weak
bands at 2654, 2932, and 3209 cm-1 were attributed to second order features [71–73]. As the
cubic MoC structure leads to no Raman active band, no bands were detected for MoC-II/TiO2.
Furthermore, the presence of the light absorbing carbon layer can hinder observation of
underlying phase [74]. This result implies that the catalyst is covered with an excess of
graphite during the synthesis, in agreement with the XPS, elemental analysis and TEM results.
The absence of bands attributed to (Mo=O) vibrations around 950-1000 cm-1 [31,75,76]
suggests that MoC-II/TiO2 is fully carburized. In the spectra of R1- MoC-II/TiO2, two
additional (Mo=O) and (Mo-O-Mo) stretching bands located at 968 cm-1 and 832 cm-1,
respectively were observed and attributed to the presence of molecular polymolybdates
[77,78]. The recovered catalyst (R1- MoC-II/TiO2) was tested under the same reaction
conditions (240 °C, 150 bar of H2). After 26 hours, only 24% of SA is converted, with 14%
yield of GBL and 7% yield of BA (Figure A.8).
21

Table 6. Initial rate (V0) and product selectivity after recycling
Fig
ure
8.
Ram
an
spec
tra
asso
ciate
d
Products selectivity (%)a
V0
Timea
Catalyst
(mmolSA gMo-1 h-1)
(h)
6
GBL BA BOL THF BDO
11.2
2.7
60
67
34
40
32
60
0
0
5
0
0
1
0
0
0
MoC-II/TiO2
26
6
R1- MoC-II/TiO2
R2- MoC-II/TiO2
a at 22-25 % conversion
8.5
with
MoC-II/TiO2 (before reaction), R1- MoC-II/TiO2 (recovered under air), R2solid- MoC-II/TiO2 (recovered under
N2) and R2solution- MoC-II/TiO2 (recovered in solution).
In a second stage, we aimed to assess when the oxidation of the catalyst was taking place.
Indeed, the oxidation could occur during the reaction or during the recovery due to the
exposure to air in the washing and filtration process. For that, another reaction was conducted
with MoC-II/TiO2 but the used catalyst denoted R2solution- MoC-II/TiO2 was recovered with the
reaction solution. Raman spectra of the solid in solution did not show any peaks associated
with molecular or crystalline molybdenum oxides (Figure 8), hence oxidation did not occur
during the reaction. The presence of amorphous carbon might inhibit the oxidation of the
particles, as observed for iron carbide [79]. In order to test the reusability of the catalyst, the
reactor was transferred to a glove bag where the filtration was set up. The catalyst was
recovered under inert atmosphere and without washing with water. Raman spectra associated
with this solid, R2solid- MoC-II/TiO2, showed complete absence of molybdenum oxides. It was
then concluded that the oxidation of the catalyst occurs during the recovery step, and not
during the reaction. However, the relative intensity of the bands of graphitic carbon was
higher after recovery suggesting that the quantity of such carbon was increased. R2solid- MoC-
II/TiO2 was then tested for the hydrogenation of SA (Figure A.8, Table 6). 64% conversion
was achieved after 26 h. It is clear that the catalyst R2solid- MoC-II/TiO2 was more active than
R1- MoC-II/TiO2, however it was still less active than fresh MoC-II/TiO2. The selectivity was
also affected as more BA and less GBL were formed. The catalyst exhibited some
deactivation, while the characterization of the bulk catalyst after the first run did not show any
drastic changes, in terms of XRD, ICP and elemental analysis (C and O contents). Therefore
XPS analysis of R2solid- MoC-II/TiO2 was conducted and the results are presented in Table 7
and Figure A.9. It can be seen that the amounts of carbidic molybdenum and carbidic carbon
have decreased in the recovered catalyst in comparison to the fresh one, i.e. from ca. 26 to
20% and from ca. 5 to 2% respectively. Moreover, despite the fact that the bulk carbon
content (by elemental analysis) was constant, the surface carbon atomic concentration has
increased. Therefore the partial oxidation of the catalysts on the surface, coupled with some
22

coke formation, could be responsible for the deactivation. A regeneration of the catalyst under
a flow of H2/C2H6 at 700 °C might provide a recovery of the catalyst performance.
Table 7. XPS analysis: atomic concentration of Mo and C, and the abundance of Mo and C
species.
Atomic
concentration %
Mo species %
C species %
Mo
3.6
4.7
C
Moδ+ Mo4+ Mo6+ Carbide
Graphite
89
Catalyst
MoC-II/TiO2
R2solid- MoC-II/TiO2
33.9
49.3
26
20
11
20
63
60
5
2
80
4. Conclusion
In this study, we investigated the hydrogenation of succinic acid in aqueous phase using
supported molybdenum carbide catalysts instead of the more conventional noble metals
catalysts. MoC/TiO2 were synthesized by impregnation with (NH4)6Mo7O24.4H2O followed
by hydrogenation / carburization under a flow of 20% v/v C2H6/H2. XRD and TEM showed
that particles < 5 nm of cubic phase were obtained. The catalysts were active for the reaction
and full conversion was achieved. The results also showed that MoC/TiO2 catalysts exhibit
high selectivity towards butyric acid and butanol, contrary to noble metal catalysts. The
effects of H2 pressure and temperature were investigated and the selectivity does not depend
on conversion. The passivation treatment does not have a noticeable effect on the
performance, while an increase in GHSV during the synthesis is associated with an increase
of in activity and a switch of selectivity towards BA. Molybdenum does not leach during the
reaction. After recovering the catalyst under air, deactivation was observed, while keeping it
under inert atmosphere limits the deactivation. The promising perspective of this project is
replacing noble metals by the non-precious and more abundant Mo carbides.
Acknowledgements: The authors acknowledge the Service Scientifiques of IRCELYON for
their help in the characterization of the catalyst. They also acknowledge Patrick Jame and Erik
Bonjour (ISA, Institute of Science Analytical) for the Carbon Elemental Analyses.
REFERENCES
[1] J.M. Pinazo, M.E. Domine, V. Parvulescu, F. Petru, Catal. Today 239 (2015) 17–24.
[2] A. Mazière, P. Prinsen, A. García, R. Luque, C. Len, Biofuels Bioprod. Biorefining 11
(2017) 908–931.
[3] N. Nghiem, S. Kleff, S. Schwegmann, Fermentation 3 (2017) 26–39.
23

[4] A. H. Tullo, Chem. Eng. News 95 (2017) 22–23.
[5] D.R. Vardon, A.E. Settle, V. Vorotnikov, M.J. Menart, T.R. Eaton, K.A. Unocic, K.X.
Steirer, K.N. Wood, N.S. Cleveland, K.E. Moyer, W.E. Michener, G.T. Beckham, ACS
Catal. 7 (2017) 6207–6219.
[6] B. Tapin, F. Epron, C. Especel, B.K. Ly, C. Pinel, M. Besson, ACS Catal. 3 (2013)
2327–2335.
[7] B.K. Ly, B. Tapin, M. Aouine, P. Delichere, F. Epron, C. Pinel, C. Especel, M. Besson,
ChemCatChem 7 (2015) 2161–2178.
[8] Z. Shao, C. Li, X. Di, Z. Xiao, C. Liang, Ind. Eng. Chem. Res. 53 (2014) 9638–9645.
[9] G. Budroni, A. Corma, J. Catal. 257 (2008) 403–408.
[10] M.J. Biddy, C. Scarlata, C. Kinchin, Chemicals from Biomass: A Market Assessment of
Bioproducts with near-Term Potential, National Renewable Energy Laboratory (NREL),
Golden, CO (United States), 2016.
[11] A.-M. Alexander, J.S.J. Hargreaves, Chem. Soc. Rev. 39 (2010) 4388–4401.
[12] C. Chan-Thaw, A. Villa, Appl. Sci. 8 (2018) 259–277.
[13] S.T. Oyama, ed., The Chemistry of Transition Metal Carbides and Nitrides, Springer
Netherlands, Dordrecht, 1996.
[14] W.-F. Chen, J.T. Muckerman, E. Fujita, Chem. Commun. 49 (2013) 8896–8909.
[15] D.-V.N. Vo, A.A. Adesina, Appl. Catal. Gen. 399 (2011) 221–232.
[16] G.M. Dolce, P.E. Savage, L.T. Thompson, Energy Fuels 11 (1997) 668–675.
[17] A. Goguet, S. Shekhtman, F. Cavallaro, C. Hardacre, F.C. Meunier, Appl. Catal. Gen.
344 (2008) 30–35.
[18] J. Patt, D.J. Moon, C. Phillips, L. Thompson, Catalysis Letters 65 (2000) 193–195.
[19] K. Xiong, W.-S. Lee, A. Bhan, J.G. Chen, ChemSusChem 7 (2014) 2146–2149.
[20] Y.-B. Huang, M.-Y. Chen, L. Yan, Q.-X. Guo, Y. Fu, ChemSusChem 7 (2014) 1068–
1072.
[21] S. Liu, H. Wang, K.J. Smith, C.S. Kim, Energy Fuels 31 (2017) 6378–6388.
[22] E.F. Mai, M.A. Machado, T.E. Davies, J.A. Lopez-Sanchez, V. Teixeira da Silva, Green
Chem 16 (2014) 4092–4097.
[23] C. Wan, N.A. Knight, B.M. Leonard, Chem. Commun. 49 (2013) 10409–10411.
[24] J. Pielaszek, B. Mierzwa, G. Medjahdi, J.F. Marêché, S. Puricelli, A. Celzard, G. Furdin,
Appl. Catal. Gen. 296 (2005) 232–237.
[25] W. Xu, P.J. Ramirez, D. Stacchiola, J.A. Rodriguez, Catal. Lett. 144 (2014) 1418–1424.
[26] S.T. Oyama, Catal. Today 15 (1992) 179–200.
[27] A. Hanif, T. Xiao, A.P.E. York, J. Sloan, M.L.H. Green, Chem. Mater. 14 (2002) 1009–
1015.
[28] T. Xiao, A.P. York, K.S. Coleman, J.B. Claridge, J. Sloan, J. Charnock, M.L. Green, J.
Mater. Chem. 11 (2001) 3094–3098.
[29] S. Li, W.B. Kim, J.S. Lee, Chem. Mater. 10 (1998) 1853–1862.
[30] T. Xiao, A.P.E. York, V.C. Williams, H. Al-Megren, A. Hanif, X. Zhou, M.L.H. Green,
Chem. Mater. 12 (2000) 3896–3905.
[31] T. Mo, J. Xu, Y. Yang, Y. Li, Catal. Today 261 (2016) 101–115.
[32] A.S. Rocha, L.A. Souza, R.R. Oliveira, A.B. Rocha, V. Teixeira da Silva, Appl. Catal.
Gen. 531 (2017) 69–78.
[33] P.M. Mortensen, H.W.P. de Carvalho, J.-D. Grunwaldt, P.A. Jensen, A.D. Jensen, J.
Catal. 328 (2015) 208–215.
[34] R. Ma, W. Hao, X. Ma, Y. Tian, Y. Li, Angew. Chem. Int. Ed. 53 (2014) 7310–7315.
[35] M. Kosmulski, J. Colloid Interface Sci. 337 (2009) 439–448.
[36] J. Moreau, O. Delpoux, E. Devers, M. Digne, S. Loridant, J. Phys. Chem. A 116 (2012)
263–270.
24

[37] P. Célestin Bakala, E. Briot, J.-Y. Piquemal, J.-M. Brégeault, P. Beaunier, Catal.
Commun. 8 (2007) 1447–1451.
[38] J.-G. Choi, R.L. Curl, L.T. Thompson, J. Catal. 146 (1994) 218–227.
[39] E.J. Markel, J.W. Van Zee, J. Catal. 126 (1990) 643–657.
[40] D.A.H. Hanaor, C.C. Sorrell, J. Mater. Sci. 46 (2011) 855–874.
[41] S.C. Pillai, P. Periyat, R. George, D.E. McCormack, M.K. Seery, H. Hayden, J. Colreavy,
D. Corr, S.J. Hinder, J. Phys. Chem. C 111 (2007) 1605–1611.
[42] H.J. Guzmán, W. Xu, D. Stacchiola, G. Vitale, C.E. Scott, J.A. Rodríguez, P. Pereira-
Almao, Can. J. Chem. 91 (2013) 573–582.
[43] Q. Bkour, C.M. Cuba-Torres, O.G. Marin-Flores, S. Tripathi, N. Ravishankar, M.G.
Norton, S. Ha, J. Mater. Sci. 53 (2018) 12816–12827.
[44] H. Li, W. Hong, Y. Cui, S. Fan, L. Zhu, J. Alloys Compd. 569 (2013) 45–51.
[45] D.-V.N. Vo, A.A. Adesina, Fuel Process. Technol. 92 (2011) 1249–1260.
[46] S. Boullosa-Eiras, R. Lødeng, H. Bergem, M. Stöcker, L. Hannevold, E.A. Blekkan,
Catal. Today 223 (2014) 44–53.
[47] B. Frank, K. Friedel, F. Girgsdies, X. Huang, R. Schlögl, A. Trunschke, ChemCatChem
5 (2013) 2296–2305.
[48] Q. Bkour, K. Im, O.G. Marin-Flores, M.G. Norton, S. Ha, J. Kim, Appl. Catal. Gen. 553
(2018) 74–81.
[49] M. Nagai, K. Oshikawa, T. Kurakami, T. Miyao, S. Omi, J. Catal. 180 (1998) 14–23.
[50] A. Mehdad, R.E. Jentoft, F.C. Jentoft, J. Catal. 347 (2017) 89–101.
[51] K. Oshikawa, M. Nagai, S. Omi, J. Phys. Chem. B 105 (2001) 9124–9131.
[52] A.A. Smirnov, Z. Geng, S.A. Khromova, S.G. Zavarukhin, O.A. Bulavchenko, A.A.
Saraev, V.V. Kaichev, D.Y. Ermakov, V.A. Yakovlev, J. Catal. 354 (2017) 61–77.
[53] N. Perret, X. Wang, L. Delannoy, C. Potvin, C. Louis, M.A. Keane, J. Catal. 286 (2012)
172–183.
[54] R.A. Mir, P. Sharma, O.P. Pandey, Sci. Rep. 7 (2017) 3518–3529.
[55] A.C. Lausche, J.A. Schaidle, L.T. Thompson, Appl. Catal. Gen. 401 (2011) 29–36.
[56] B.K. Ly, D.P. Minh, C. Pinel, M. Besson, B. Tapin, F. Epron, C. Especel, Top. Catal. 55
(2012) 466–473.
[57] X. Liu, X. Wang, G. Xu, Q. Liu, X. Mu, H. Liu, J. Mater. Chem. A 3 (2015) 23560–
23569.
[58] C. You, C. Zhang, L. Chen, Z. Qi, Appl. Organomet. Chem. 29 (2015) 653–660.
[59] U.G. Hong, S. Hwang, J.G. Seo, J. Lee, I.K. Song, J. Ind. Eng. Chem. 17 (2011) 316–
320.
[60] U.G. Hong, J. Lee, S. Hwang, I.K. Song, Catal. Lett. 141 (2011) 332–338.
[61] X. Di, Z. Shao, C. Li, W. Li, C. Liang, Catal. Sci. Technol. 5 (2015) 2441–2448.
[62] A. Bhattacharyya, M. D. Manila, Catalysts for Maleic Acid Hydrogenation to 1,4-
Butanediol, US 2006/0004212 A1, 2006.
[63] X. Di, C. Li, B. Zhang, J. Qi, W. Li, D. Su, C. Liang, Ind. Eng. Chem. Res. 56 (2017)
4672–4683.
[64] Z. Huang, K.J. Barnett, J.P. Chada, Z.J. Brentzel, Z. Xu, J.A. Dumesic, G.W. Huber,
ACS Catal. 7 (2017) 8429–8440.
[65] M.M. Sullivan, A. Bhan, ACS Catal. 6 (2016) 1145–1152.
[66] L. Souza Macedo, D.R. Stellwagen, V. Teixeira da Silva, J.H. Bitter, ChemCatChem 7
(2015) 2816–2823.
[67] J.-S. Choi, V. Schwartz, E. Santillan-Jimenez, M. Crocker, S. Lewis, M. Lance, H.
Meyer, K. More, Catalysts 5 (2015) 406–423.
[68] P. Kitsiou, N. Lagopati, E.-P. Tsilibary, P. Falaras, P. Papazafiri, E.P. Pavlatou, E.
Kotsopoulou, Int. J. Nanomedicine 9 (2014) 3219–3230.
25

[69] A.C. Ferrari, J. Robertson, Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 362 (2004)
2477–2512.
[70] M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus, L.G. Cançado, A. Jorio, R. Saito, Phys
Chem Chem Phys 9 (2007) 1276–1290.
[71] J. Jehlička, O. Urban, J. Pokorný, Spectrochim. Acta. A. Mol. Biomol. Spectrosc. 59
(2003) 2341–2352.
[72] S. Vollebregt, R. Ishihara, F.D. Tichelaar, Y. Hou, C.I.M. Beenakker, Carbon 50 (2012)
3542–3554.
[73] Y.-J. Lee, J. Nucl. Mater. 325 (2004) 174–179.
[74] S. Loridant, I.C. Marcu, G. Bergeret, J.M.M. Millet, Phys. Chem. Chem. Phys. 5 (2003)
4384–4389.
[75] R. Radhakrishnan, C. Reed, S.T. Oyama, M. Seman, J.N. Kondo, K. Domen, Y.
Ohminami, K. Asakura, J. Phys. Chem. B 105 (2001) 8519–8530.
[76] E. de B. Santos, F.A. Sigoli, I.O. Mazali, Mater. Res. Bull. 60 (2014) 242–246.
[77] M. Dieterle, G. Mestl, Phys. Chem. Chem. Phys. 4 (2002) 822–826.
[78] L. Kumari, Y.-R. Ma, C.-C. Tsai, Y.-W. Lin, S.Y. Wu, K.-W. Cheng, Y. Liou,
Nanotechnology 18 (2007) 115717–115724.
[79] X.Q. Zhao, Y. Liang, Z.Q. Hu, B.X. Liu, J. Appl. Phys. 80 (1996) 5857–5860.
26

Figure A.1. Representative TEM image (a, a.1), and electron diffraction (a.2) of MoC-III/TiO2.
Figure A.2. Representative TEM image (a, a.1), and electron diffraction (a.2) of MoC-I/TiO2.
27

Figure A.3. XPS spectra of MoC-III/TiO2 over Ti 2p.
28

Figure A.4. Typical hydrogenation of succinic acid at 240 °C and under 150 bar H2: temporal evolution of (a) the
concentrations of SA and the products yield, (b) the carbon balance (CB) and TOC measured. Aqueous solution
of SA (0.14 M, 100 mL), 0.6 g of MoC-II/TiO2.
29