Batch versus Continuous Flow Performance of Supported Mono- and Bimetallic Nickel Catalysts for Catalytic Transfer Hydrogenation of Furfural in Isopropanol

Article abstract of DOI:10.1002/cctc.201800530

Furfural takes an important position in hemicelluloses biorefinery platforms. It can be converted into a wide range of chemicals. One important valorization route is the catalytic hydrogenation. Whereas molecular hydrogen is mostly used in industrial hydrogenation processes, recent studies also showed that alcohols can be used as reductants from which hydrides can be transferred catalytically to furfural. This process is often assisted by the formation of significant amounts of side products, in despite of high yields to the hydrogenolysis product 2-methylfuran. The present work explores the catalytic behavior in batch and continuous flow of mono- and bimetallic nickel catalysts supported on activated carbon for the catalytic transfer hydrogenation of furfural in isopropanol.

Full text of DOI:10.1002/cctc.201800530

Accepted Article
Title: Batch vs. continuous flow performance of supported mono- and
bimetallic nickel catalysts for catalytic transfer hydrogenation of
furfural in isopropanol
Authors: Yantao Wang, Pepijn Prinsen, Konstantinos S. Triantafyllidis,
Stamatia A. Karakoulia, Alfonso Yepez, Christophe Len, and
Rafael Luque
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201800530
Link to VoR: http://dx.doi.org/10.1002/cctc.201800530
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
Batch vs. continuous flow performance of supported mono- and
bimetallic nickel catalysts for catalytic transfer hydrogenation of
furfural in isopropanol
[
a]
[b]
[c,d]
[d]
Yantao Wang, Pepijn Prinsen, Konstantinos S. Triantafyllidis, Stamatia A. Karakoulia, Alfonso
Yepez, Christophe Len and Rafael Luque*[b, f]
[
b]
[a,e]
Abstract: Furfural takes an important position in hemicelluloses
biorefinery platforms. It can be converted into a wide range of
chemicals. One important valorization route is the catalytic
hydrogenation. Whereas molecular hydrogen is mostly used in
industrial hydrogenation processes, recent studies also showed that
alcohols can be used as reductants from which hydrides can be
transferred catalytically to furfural. This process is often assisted by
the formation of significant amounts of side products, in despite of
high yields to the hydrogenolysis product 2-methylfuran. The present
work explores the catalytic behavior in batch and continuous flow of
mono- and bimetallic nickel catalysts supported on activated carbon
for the catalytic transfer hydrogenation of furfural in isopropanol.
be converted into a wide range of chemicals, either by catalytic
oxidation,[6] amination[7] or hydrogenation (Scheme 1a).[3] The
hydrogenation route is a highly reported research area, as it can
provide a wide range of bulk chemicals. Although promising
results have been reported on catalytic furfural hydrogenation in
vapor phase, most of the studies aim liquid phase hydrogenation,
mainly because catalyst deactivation through coke formation
occurs faster in the former.[3] High selectivity to furfuryl alcohol
(FA) has been demonstrated for numerous catalysts,[8-17] as the
product of the first step in the hydrogenation pathway (Scheme
1b). Only few studies reported high selectivity to 2-methylfuran
(MF).[12,18] The production of MF is attractive, as it has a very low
boiling point which enables efficient separation. MF has shown
superior performance as an additive in gasoline fuels compared
to other furfural hydrogenation products.[ The conversion to
MF occurs via hydrogenolysis of the C-O linkage in FA, which in
general requires higher temperature and longer reaction time.
19]
Introduction
Hemicelluloses are a heterogeneous carbohydrate fraction and
can potentially provide a wide range of compounds to replace
current chemicals derived from crude oil.[1] The main
disadvantage of their heterogeneity is the fact that C-5 sugars
are not converted biochemically as efficient as C-6 sugars (using
whole cells or enzymes). Therefore, the chemocatalytic
valorization of hemicelluloses and their co-generated waste
streams is an imminent research area. Recent advances in the
Beneficial effects on the selectivity to MF have been reported for
[12,18,20]
bimetallic catalysts.
The type of catalyst support also
plays an important role in the hydrogenation of furfural. Supports
,[14,21,22] TiO ,[12,16,23]
often include Lewis acids, such as γ-Al
2
[24]
O
3
2
,[8] CeO
,[14] SiO
[9,10]
ZrO
MgO and ZnO)
2
2
2
and Al-SBA-15, basic metal oxides
[11,14]
[10]
(
3
and base metal carbonates (CaCO ),
which assist in polarizing the furfural carbonyl bond and which
can actively take part in one of the reaction pathways. More
neutral supports such as carbons were also used in some
catalytic valorization of hemicelluloses envision the conversion
to levulinates, lactones and furans,[
2,3]
either in one-pot
[10,13,18,24,25]
studies.
[
4]
conversion step or in multiple steps via the hydrolysis to
oligomers and xylose (as the most abundant monomer).[5] The
catalytic route to furfural through dehydration and isomerization
today is still challenging route. Once furfural is produced, it can
2
Molecular hydrogen (H ) is mostly used as the reductant for
catalytic furfural hydrogenation.[8-15,21-25] The reduction by
catalytic hydrogen transfer in various alcohols (ROH), however,
is reported only in few cases.[17,18,20] The latter is attractive, not
only from the techno-economic point of view, but also because
the interphase transfer is improved (solid/liquid for ROH vs.
solid/liquid/gas for H ). Important mechanistic insights were
2
reported by Gilkey et al. (2015), who demonstrated that when
[
[
a]
b]
MSc, Y. Wang, Prof. C. Len
Sorbonne Universités, Université de Technologie de Compiègne,
Centre de Recherche Royallieu, CS 60 319, F-60203 Compiègne
cedex, France.
Dr. P. Prinsen, Dr. A. Yepez, Prof. R. Luque
Departamento de Química Orgánica, Universidad de Cordoba
Campus de Rabanales, Edificio Marie Curie (C-3), Ctra Nnal IV, km.
starting from FA the hydrogenolysis via ring activation (Scheme
2a) was predominant over direct metal-mediated hydrogenolysis
396, E-14014 Cordoba, Spain
(
Scheme 2b), eventually leading to the formation of water.
E-mail: q62alsor@uco.es
Based on isotopic kinetic effects, it was also found that in the
first step of the hydrogenation (formation of FA), the transfer of
hydrides from isopropanol to furfural mainly occurs through
formation of a complex between furfural, ROH and Lewis acid
[
[
[
[
c]
d]
e]
f]
Prof. K. S. Triantafyllidis
Department of Chemistry, Aristotle University of Thessaloniki,
University Campus, P.O. Box 116, GR-54124 Thessaloniki, Greece.
Dr. S. A. Karakoulia, Prof. K. S. Triantafyllidis
Chemical Process & Energy Resources Institute, CERTH, Thermi
P.O. Box 60361, GR-57001 Thessaloniki, Greece
PSL Research University, Chimie ParisTech, CNRS, Institut de
Recherche de Chimie Paris, 11 rue Pierre et Marie Curie, F-75231
Paris Cedex 05, France
2
sites (Scheme 2c), rather than production of H from ROH on
the metal surface (Scheme 2d). Gilkey et al. (2015) also
concluded that strong adsorption of furfural retards the metal
catalyzed pathway (competitive adsorption of furfural and ROH
on the metal site), but does not affect the Lewis acid-mediated
hydrogen transfer pathway because of its concerted bulky
Peoples Friendship University of Russia (RUDN University), 6
Miklukho-Maklaya str., 117198, Moscow, Russia
Supporting information for this article is given via a link at the end of the
document.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
nature in which ROH coordinates a hydrogen directly to the
carbonyl group in furfural (no competition on the metal site).
High selectivity to MF in catalytic hydrogen transfer can be
promoted by inducing a metal oxide phase (acting as Lewis
acid) co-existing along with the metal phase,[18,20] to catalyze the
hydride transfer and to bind OH groups released upon
hydrogenolysis of FA.
Based on these findings, the present study aims to explore the
extent of catalytic hydrogen transfer to furfural in isopropanol (i-
PrOH), using monometallic (Ni) and bimetallic (Ni-W) nickel
catalysts supported on neutral micro/mesoporous activated
carbon. The experiments were conducted both in batch and in
continuous flow to study the effect on the catalyst performance.
Fresh and used catalysts were analyzed with XRD, XPS and
ICP-MS to study the catalyst deactivation.
Scheme 1. (a) Catalytic valorization routes of lignocellulose derived furfural and (b) reaction pathways in furfural hydrogenation according to the literature.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
Scheme 2. (a) Furfuryl alcohol (FA) hydrogenolysis via ring activation, (b) direct FA hydrogenolysis, (c) Meerwein-Ponndorf-Verley reduction of furfural and (d)
metal-mediated intermolecular hydride transfer to furfural.
Results and Discussion
average crystal size 6.8 nm. As the Ni content increased to 10%,
the Ni(0) crystal size increased to 23.2 nm and a NiO phase
appears with crystal size 6.1 nm. In the case of bimetallic
catalysts, and more specifically in 5%Ni-15%W/AC, the
Catalyst characterization. A series of monometallic (5%Ni/AC
and 10%Ni/AC) and bimetallic (5%Ni-15%W/AC and 10%Ni-
1
5%W/AC) nickel-based catalysts were synthesized via classic
wet impregnation on commercial activated carbon (AC), followed
by thermal treatment in inert and H atmosphere. The AC
catalyst support) was a neutral carbon (pH 6-8) and contained
crystalline phases of tungsten oxide (WO
2
) and nickel tungsten
oxide (NiWO ) prevail with average crystal sizes 9.9 and 15.5
4
nm, respectively. However, for the 10%Ni-15%W/AC catalyst,
the dominant peak in the XRD pattern pertains to Ni(0) with
crystal size 10.7 nm, being markedly lower than the Ni(0) crystal
size in the monometallic 10%Ni/AC catalyst.
2
(
[26]
only small amounts of Lewis acid sites. The catalyst samples
were stored under ambient atmosphere and were used as such
for catalytic experiments and characterization. They were
characterized by nitrogen (N
2
) porosimetry; the corresponding
N
2
adsorption-desorption isotherms are shown in Figure 1a. The
parent AC is a microporous carbon with a high contribution of
meso/macroporosity (Table 1, entry 1). About 65 % of its total
surface area and 36 % of its total pore volume is attributed to
micropores and the remaining to meso/macropores and external
surface, as revealed from the isotherm shapes, where a
Table 1. Porosity of mono- and bimetallic Ni catalysts supported on
activated carbon (AC).
Micropore area
Meso/macro-
Catalyst
Total
Total
pore
2
SSA
(m /g) & volume
pore
&
external
2
[a]
(
m /g)
volume
(cc/g)[c]
2
progressive increase of adsorbed N
2
can be observed over the
area (m /g)
&
b]
(
cc/g)[
whole P/P range, indicating the presence of a relatively
o
volume (cc/g)[d]
disordered network of meso- and macropores. As a result of this
size inhomogeneity, no distinct peaks in the BJH pore size
distribution curves were observed, while only a weak peak was
identified in the DFT pore size distribution curve with an average
diameter of 3.4 nm. The total and micropore surface area, as
well as the pore volumes of the monometallic loaded AC
catalysts were essentially unchanged compared to those of the
parent AC (max. 10 % decrease). The porosity characteristics of
the bimetallic catalysts decreased by ca. 30 %, as determined
for the 10%Ni-15%W/AC catalyst. The catalysts were also
characterized by powder XRD analysis. The corresponding
diffraction patterns are shown in Figure 1b. The 5%Ni/AC
catalyst consisted essentially of metallic Ni(0) nanoparticles with
AC
1280
0.95
840 / 0.34
440 / 0.60
5%Ni/AC
1250
1250
0.88
0.90
830 / 0.34
810 / 0.33
420 / 0.54
440 / 0.57
10%Ni/AC
5%Ni-
1
020
20
0.72
0.64
680 / 0.28
610 / 0.25
350 / 0.44
310 / 0.39
15%W/AC
10%Ni-
9
15%W/AC
[
a] SSA: specific surface area from N
method). [b] Total pore volume at P/P
Meso/macropore external area
Meso/macropore volume = Total pore volume - Micropore volume (t-plot)
2
sorption at -196 °C (multi-point BET
= 0.99. [c] t-plot method. [d]
Total SSA Micropore area;
o
&
=
-
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
with the bimetallic Ni-W catalysts (entries 5-8). Furfuryl alcohol
(
FA) and 2-methylfuran (MF) were the only products produced in
significant amounts, whereas the corresponding ring
hydrogenated products tetrahydrofurfuryl alcohol (THFA) and 2-
methyltetrahydrofuran (MTHF) were produced only in small
amounts (0-3 %). The blank experiment (entry 1) gave only 2 %
conversion of furfural. With the monometallic 5%Ni/AC catalyst,
but in the absence of molecular hydrogen (H
was observed (entry 2). Much higher conversion (85 %) was
observed with the same catalyst when 30 bars of H was used
entry 3), with high selectivity (78 %) to MF. To the best of our
2
), 10 % conversion
2
(
knowledge, the corresponding MF yield is among the highest
reported in literature at temperatures ≤ 200 °C.[12,18,20] The
bimetallic Ni-W catalysts (entries 5 and 7) gave moderate
conversions (ca. 50 %), with 30-40 % selectivity to MF. At
2
260 °C without the addition of H (entries 6 and 8), the
conversions reached 95 and 83 % for the mono- and bimetallic
catalyst, respectively. The selectivity to MF, however, was much
higher for the monometallic (53 %) than for the bimetallic
catalyst (7 and 39 % for 5%-15%W/AC and 10%Ni-15%W/AC,
respectively). These results show that higher furfural conversion
can be obtained through catalyzed hydrogen transfer from i-
PrOH to furfural, by operating at 260 °C instead of 200 °C,
however with lower selectivity to MF. What did seem to occur is
2
that at 260 °C and without H , a substantial higher amounts of
side products were produced, as seen from the significant lacks
in the mass balance (entries 4, 6 and 8). Side products were
only detected using the GC-FID conditions used for product
analysis after batch experiments. In one experiment (entry 3),
after removing the solvent (i-PrOH) and MF by rotavaporisation,
Figure 1. (a)
2
N adsorption-desorption isotherms of the parent calcined
activated carbon (AC) and the supported mono- and bimetallic Ni catalysts
and (b) corresponding XRD patterns (symbols indicate the respective
crystalline phase: (*) for Ni(0) and (♦) for NiO, (▲) for WO , and (∆) for NiWO ).
2 4
2
-(isopropoxy)methylfuran (iPrOMF) was identified using 1H-
NMR and 13C-NMR (results not shown), in agreement with the
literature.[15-17,20,27] This is the alkyl ether side product from FA
and i-PrOH (Meerwein-Ponndorf-Verley reduction).
Batch furfural hydrogenation experiments. The experiments
in batch were carried out for 5 h using a 0.35 M furfural
feedstock in isopropanol (i-PrOH). Table 2 shows the results
obtained with the monometallic catalysts (entries 3-4), as well
Table 2. Conversion and product yields after 5 h batch hydrogenation of 0.35 M furfural in 60 mL i-PrOH (FA: furfure alcohol, MF: 2-methylfuran, THFA:
tetrahydrofurfuryl alcohol and MTHF: 2-methyltetrahydrofuran).
Yield (%)
T
H
(bars)
2
Conversion
(%)
Entry
Catalyst
Mass balance (%)
(°C)
FA
6
MF
1
THFA
MTHF
iPrOMF
1
2
3
4
-
200
200
200
260
30
0
2
0
0
1
1
0
0
2
1
0
0
105
101
93
5%Ni/AC
5%Ni/AC
5%Ni/AC
10
85
95
10
6
1
30
0
66
50
3
1
4
20
78
5
5%Ni-15%W/AC
200
30
50
32
15
1
2
104
6
7
8
c
5%Ni-15%W/AC
10%Ni-15%W/AC
10%Ni-15%W/AC
260
200
260
0
83
51
83
16
8
6
1
1
1
2
3
2
6
9
8
48c
90
30
0
20
6
25
59d
Unknown compounds U1 and U2 eluting at 10.4 and 12.0 min in GC-FID analysis not included (39 % of total peak area); d Unknown compounds U1
and U2 accounted for 21 % of total peak area
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
This side product probably corresponded with the unidentified
peak at 7.0 min in the corresponding chromatograms (see
Figure S1 in the Supporting Information). Only in the
experiments with bimetallic catalysts (entries 6 and 8), two other
peaks were detected at higher retention times (10.4 and 12.0
min). Other side products may include (hemi)-acetals (via
dehydration) and FA polymerization products, but these were
not identified in the present study. Based on the GC retention
times of the pure standards 1,4-pentanediol, 1,2- pentanediol,
Continuous flow hydrogen transfer experiments. All
continuous flow experiments were performed without the supply
of
continuously using
gradually moving from low (200 °C) to high (260 °C) temperature
H
2
.
First, the 5%Ni/AC catalyst (110 mg) was tested
a
0.2 furfural feedstock in i-PrOH,
M
-1
and for each temperature from high (0.4 mL min ) to low (0.1
-1
mL min ) flow rate (Figure 2a). The total time on stream was ca.
9 h. Conversions increased with lower flow rates and with higher
temperatures. During the first hour on stream the selectivity to
MF was fare (20-30 %), but then gradually dropped over time,
even at higher temperatures. The results suggested that
substantial catalyst deactivation occurred.
1,5- pentanediol, furan and tetrahydrofuran, it was confirmed
that no ring-opening neither decarbonylation occurred.
Figure 2. Conversion (gray scale) and selectivity to FA (blue) and MF (red) in continuous flow transfer hydrogenation in i-PrOH, using
(
a) 5%Ni/AC, 0.2 M furfural and 0 bar, (b) 0.2 M furfural, 260 °C, 0 bar and (c) 0.1 M furfural, 230 °C and 30 bars.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
Among different fresh catalysts, each tested during 3 h on
(Figure 3c and 3d), no significant amounts of FA were produced
at any point in contrast with the batch results (Table 2, entries 6
and 8). As well in contrast to batch experiments, the bimetallic
catalysts showed very high selectivity to MF, ca. 85 and 84 % for
the 5%Ni-15%W/AC and 10%Ni-5%W/AC catalysts, respectively,
as the mean selectivity calculated over the 0.7-7.0 h period on
stream. The conversion in turn decreased from 85 to 27 % and
from 99 to 11 %, respectively, showing their lower operational
stability compared to the monometallic catalysts (conversions
dropped faster over time). Still, after 6 h the MF yields produced
with the bimetallic catalysts (13-32 %) were higher than the ones
produced with monometallic catalysts (6-20 %), because the
bimetallic ones retained their high selectivity to MF. Selectivities
slightly higher than 100 % were observed at some points with
both bimetallic Ni-W catalysts, which was probably the result of
stream at 260 °C (Figure 2b), 5%Ni/AC provided the highest
-1
conversion at low flow rate (0.1 mL min ) while 10%Ni/AC gave
-1
higher conversion at higher flow rate (0.2-0.4 mL min ). These
monometallic catalysts were more active compared to bimetallic
5%Ni-15%W/AC and to 5%Raney-Ni/AC, a catalyst with high
hydrogen transfer activity in combination with i-PrOH, as
demonstrated by Wang and Rinaldi for various bio-oil derived
compounds at temperatures already starting from 120 °C.28 The
selectivity to MF was however much lower, in strong contrast
with the batch results. Interestingly, the selectivity to MF with
-
bimetallic Ni-W in the continuous flow experiment at 0.1 mL min
1
reached 98 %. Improved conversions were obtained by
increasing the hydrodynamic pressure from 0 to 30 bars and
using a smaller furfural feed load (0.1 M) as demonstrated in
Figure 2c, where 5%Ni/AC gave the highest conversion (similar
to 5%Pd/C), while 5%Ni-15%W/AC gave the highest selectivity
to MF.
kinetic sorption effects.
A significant amount of products
remained unknown when using monometallic catalysts (mass
balance 35-75 % for the monometallic vs. 74-105 % for the
bimetallic catalysts). These results indicate that side product
formation is suppressed with the bimetallic catalysts, in favor of
MF formation. Acetone formation and gas bubbles production
was observed in most of the experiments, although with
decreasing trend over time (deactivation of metal-mediated
Further experiments were conducted under optimized conditions
to compare the operational stability of monometallic Ni and
bimetallic Ni-W catalysts. With the monometallic catalysts
(
Figures 3a and 3b) a peak in the FA production was observed
during the first hour on stream, which then dropped followed by
a slow increase over time. With the bimetallic catalysts in turn
hydrogen
transfer).
Figure 3. Conversion, product yield and mass balance over 6-10 h continuous flow hydrogen transfer at 0.2 mL min-1 using 0.1 M
furfural in i-PrOH at 230 ℃ and 30 bars using 110 mg of (a) 5%Ni/AC, (b) 10%Ni/AC, (c) 5%Ni-15%W/AC and (d) 10%Ni-15%W/AC.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
All the results observed in continuous flow are in line with the
findings reported by Gilkey et al., who demonstrated
experimentally the importance of the co-existence of a Lewis
acid site in the vicinity of a reduced metal phase in bimetallic
systems to accomplish the actual hydrogenolysis of the C-O
bonds in FA and who demonstrated the competitive formation of
MF (hydrogenolysis product) and 2-(isopropoxy)methylfuran
continuous flow using bimetallic catalysts, significant metal
leaching occurred which decreased the activity. Advanced
kinetic studies and further insights in the catalytic role of the
2 4
individual NiO, WO and NiWO phases may explain better the
highly different catalytic performance when moving from batch to
continuous flow conditions. Finally, batch processes deal with
diffusion mass transfer limitations. The poorer porosity
characteristics of the bimetallic Ni-W catalysts may also explain
their worse performance in batch conditions (as compared to
monometallic Ni).
[
20]
(
etherification product). When the same catalyst was used in
the presence of H , in similar continuous flow equipment, side
2
product formation still occurred (for details see Figure S2 in the
Supporting Information). The products shifted to lower retention
times when using ethanol or methanol instead of isopropanol.
Moreover, side product formation was fully suppressed (100 %
mass balance) by using aprotic polar solvents such as ethyl
acetate, methyl isobutyl ketone and methyl cyclopentyl methyl
ether, however, with considerable lower activity compared to
protic solvents. These results show that side product formation
was mostly the result of alcoholysis reactions, which can be
catalyzed by acid sites.20,27
The fresh mono- and bimetallic catalysts and the corresponding
used catalysts were also analyzed with XPS (Figure 5) and ICP-
MS to study the extent of metal leaching from the support. The
results (Table 3) show the relative elemental composition on the
surface (XPS) and in the whole catalyst material (ICP-MS).
Interestingly, the XPS analysis of the fresh catalysts showed that
the catalyst surface was enriched in metal oxide phase (NiO),
especially in the Ni-W catalysts (Table 3, entry 6). In the XRD
analysis only low intensity peaks for NiO were observed. This
means that a NiO phase had formed on the surface, which is not
surprising as the catalysts were stored under air atmosphere.
The general trend in the results was that significant metal
leaching had occurred (ICP-MS). The leaching effect was
substantially higher in batch (40-44 %) than in continuous flow
(5-26 %) conditions. This was especially the case for the
bimetallic catalyst, from which most of the metal phases,
Effect of process conditions on catalyst performance. The
change in mean crystal size after utilization in batch and
continuous flow was compared, as seen from XRD analysis
(
Figure 4) using the Scherrer equation. A higher crystal size
indicates that the metal dispersion is affected negatively (when
no metal leaching occurs). In general no drastic effects were
observed, except for the crystal size of the Ni(0) phase after
utilization at 260 °C, either in batch or in continuous flow (11.2-
including Ni(0), NiO, WO
utilization in batch.
2 4
and NiWO , were leached after
11.8 nm vs. 6.8-7.8 nm). This effect can be the result of metal
nanoparticle aggregation, metal leaching (removal of smallest Ni
nanoparticles) or transition to another Ni phase. Note that for the
fresh Ni/AC and Ni-W/AC catalysts, much higher crystal sizes
were observed when increasing the Ni load from 5 to 10 % in
their synthesis, explaining their worse catalytic performance.
The NiO phase was enriched on the catalyst surface (XPS) as
well in the bulk catalyst (XRD) after utilization in continuous flow
whereas it was not after batch utilization. This effect was caused
by the higher leaching of metallic Ni(0) in batch, but additionally
phase transition to NiO occurred in continuous flow at 260 °C.
Panagiotopoulou and Vlachos obtained high MF yields by
x
inducing a metal oxide (RuO ) phase along with the metal (Ru)
phase by treating catalyst in a posterior oxidation treatment
under controlled conditions.[18] In line with these findings, the
results in the present study suggest that the Ni/NiO catalyst was
more effective in batch for promoting high MF selectivity
x
whereas the Ni/WO catalyst in continuous flow (with NiO and
WO as the Lewis acid sites, respectively). However, as the
x
metal oxide phase was not induced in a controlled way (as it
was produced on the carbon surface during storage and also in
the bulk carbon under continuous flow conditions), no
straightforward conclusions can be drawn on the catalytic
2 4
contribution of the NiO, WO and NiWO phases in the Lewis
acid-mediated hydride transfer and their capacity to bind
hydroxyl groups released in the hydrogenolysis of FA, in addition
to the nickel metal-mediated pathway. Although high and stable
selectivity to MF was demonstrated for 8 h on stream in
Figure 4. Crystal sizes of fresh and used mono- and bimetallic
Ni catalysts based on XRD analysis: (*) for Ni(0) and (♦) for NiO
2 4
(WO and NiWO signals are not indicated here).
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
Figure 5. (Left) Full and (Right) zoomed XPS spectra of (a) fresh 5%Ni/AC, (b) used 5%Ni/AC (batch, (c) fresh 5%Ni/15%W/AC and
d) used 5%Ni/15%W/AC catalysts, after 5 h batch in i-PrOH at 260 ℃ in the absence of H
(
2
.
Table 3. C, O, Ni and W content as determined by XPS and ICP-MS analysis in fresh and used mono- and bimetallic Ni catalysts.
XPS (wt%)[a]
NiO
ICP-MS (wt%)
Entry
Catalyst
C
O
Ni(0)
0.4
W
-
Ni
W
1
2
3
5% Ni/AC, fresh
86.8
12.3
0.5
0.3
0.6
4.1
-
5%Ni/AC, used (batch, 200 °C, 30 bars H
2
)
79.1
86.6
20.4
12.3
0.2
0.6
-
-
2.3
2.4
-
-
2
5%Ni/AC, used (batch, 260 °C, no H )
5
6
5%Ni/AC, used (continuous flow, 260 °C, 30 bars, no H
5%Ni-15%W/AC, fresh
2
)
82.5
67.4
16.7
15.4
0.7
0.5
0.2
0.1
-
3.9
4.3
-
17.4
0.1[b]
7
8
5%Ni-15%W/AC, used (batch, 260 °C, no H
2
)
75.4
69.6
18.8
14.3
0.2
0.7
0.0
0.1
5.6
2.6
3.2
0.1[b]
0.1[b]
5%Ni-15%W/AC, used (continuous flow, 260 °C, 30 bars, no H
2
)
15.4
[a] based on C (1s), O (1s), Ni(0) (2p3/2), NiO (2p3/2) and W(4f) signals. [b] Not quantitative (only small amounts of W were solubilized in
the acid digest due to complex formation)
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
rpm). Stirring was kept for 1 h before the water was removed in a rotary
evaporator. The resulting catalysts were dried overnight at 100 °C and
-
1
were then calcined at 500 °C for 3 h under N
followed by treatment at 450 °C for 3 h under H flow (50 mL min ). The
%Pd/C catalyst (4.9 wt% Pd) and the Ni Raney catalyst (92% Ni) were
provided by Alfa Aesar
2
flow (50 mL min ),
Conclusions
-
1
2
5
A wide range of chemicals can be produced through catalytic
hydrogenation of biomass derived furfural. The production of 2-
methylfuran (MF) in particular is attractive, but high yields were
reported only in few cases, as it requires higher temperature and
longer reaction time to break the C-O bond in furfuryl alcohol
Catalyst characterization. Nitrogen (N ) adsorption/desorption analysis
was performed at -196 °C for the determination of the specific surface
area (multi-point BET method), total pore volume (at P/P = 0.99),
o
mircopore area by t-plot analysis and pore size distribution (BJH method
using adsorption data or DFT analysis) of the samples, which were
previously outgassed at 150 °C for 16 h under 5x10-9 Torr vacuum, using
2
(
FA) being the primary product in the reaction pathway.
Whereas molecular hydrogen (H ) is mostly used as the
reductant, few studies demonstrated the catalytic hydrogen
transfer from alcohols to furfural in the absence of H . This
2
an
Automatic
Volumetric Sorption
Analyzer
(Autosorb-1MP,
Quantachrome). The powder X-ray Diffraction (XRD) experiments were
conducted on a Shimadzu X-ray 7000 diffractometer using a Cu Kα X-ray
radiation operating at 45 kV and 100 mA; counts were accumulated in
the range of 5-75º with 0.02º steps (2θ) with counting time 2 s per step.
The X-ray Photo-emission Spectrometry (XPS) measurements were
carried out with a Phi Quantera Scanning X-ray Microprobe instrument
2
catalytic system is not only different from a techno-economic
point of view, but also because this way Lewis acids can also
take part in the reaction mechanism in addition to the metal-
mediated hydride transfer, taking into account that strong
adsorptive competition exists on the metallic sites. The present
study reports the catalytic transfer hydrogenation of furfural in
isopropanol at 200-260 °C using monometallic nickel (5%Ni and
(Chevron Energy Technology Company, Richmond, CA, USA) using Al
Kα (h∙ν = 1486.7 eV) radiation. The instrument was equipped with a
hemispherical energy analyzer with multichannel detection and an
energy resolution of 1.1 eV. The catalysts were mounted on double-
sticky tape confining an area of ca. 0.8 cm x 0.8 cm. The tape was
completely covered by the catalyst powder and the sample surface was
carefully smoothed. Spectra were collected for C-1s, O-1s, Ni-2p3 and
W-4f photoelectron peaks. Total spectral accumulation times were 100
minutes per analysis area while irradiating with 100 W of X-ray irradiation.
The binding energies (BE) were referenced to the C-1s peak (284.8 eV)
to account for charging effects. The XPS spectra were deconvoluted
using Gaussian/Lorentzian shaped curves and an iterative least square
algorithm provided in Phi Multipak software. The peak areas were
computed according to the best possible fittings in order to quantify the
W and Ni species contents on the catalyst surface. The crystal size of the
metal nanoparticles was estimated using the Scherrer equation based on
the characteristic peaks (2θ) at 44.3° for Ni(0) and 43.2° for NiO, 26.1°
1
1
0%Ni) and bimetallic nickel-tungsten (5%Ni-15%W and 10%Ni-
5%W) catalysts supported on micro/mesoporous activated
carbon (AC), resulting in mainly FA, MF and side products,
originating from furfural and furfuryl alcohol alcoholysis and
polymerization reactions. High furfural conversion (85%) and
selectivity to MF (78 %) was obtained with 5%Ni/AC using 30
bars H
2
in batch at 200 °C. In comparison, higher conversion
(
95 %) was achieved through catalytic hydrogen transfer at
2
260 °C (without H ), however with lower selectivity to MF (53 %).
The bimetallic Ni-W catalysts resulted in lower conversion and
selectivity to MF. In continuous flow, in contrast, the bimetallic
Ni-W catalysts performed better than the monometallic Ni
catalysts, as they showed to be very selective to MF (> 90 %),
however with lower stability than the monometallic catalysts. The
selectivity to MF was inversely correlated with the yield to side
products. XPS and ICP-MS analysis of fresh and used catalysts
showed that significant metal leaching occurred, especially in
batch conditions. XPS analysis showed that a substantial
amount of a nickel oxide (NiO) phase co-existed with metallic
nickel (Ni(0)) on the surface of the fresh catalyst and that phase
transition from Ni(0) to NiO occurred under continuous flow,
which also may have promoted high selectivity to MF.
2 4
for WO and 31.1° for NiWO . The metal content of fresh and used
catalysts was determined via inductively coupled plasma optical emission
mass spectrometry (ICP-MS, Perkin Elmer, NexION 350X) using the
corresponding acid digests. Acid digestions were performed with nitric
acid in individual
2
N pressurized (40 bars) reaction chambers in
microwave equipment (Milestone, UltraWave), with the temperature
raising during 25 min. until reaching 220 °C, which was hold during 15
min. The total digestion time was 50 min. and the final HNO
concentration was 2 %.
3
Batch experiments. All the experiments were carried out in a 100 mL
stainless steel autoclave equipped with a thermocouple and magnetic
stirring (800 rpm), as illustrated in Figure S3 (see Supporting
Information). First 300 mg of catalyst was added, followed by 25 mmol
furfural and 12 mmol n-octane (external standard IS) dissolved in 60 mL
Experimental Section
2
isopropanol (i-PrOH). After 3 times purging, 30 bars of H was added.
Reagents. See Supporting Information.
The reaction mixture was stirred and heated to 200 ºC (60 bars total
pressure). The target temperature was reached after ca. 35-50 min. at
which time was set to zero. Additional experiments were carried out
Catalyst synthesis. The monometallic (5%Ni/AC and 10%Ni/AC) and
the bimetallic (5%Ni-15%W/AC and 10%Ni-15%W/AC) catalysts were
prepared by the classic wet impregnation method. First, Norit® SX-Plus
2
without the addition of H , both at 200 °C (30 bars total pressure) and
260 °C (92 bars total pressure). After the reaction, the stirring was
stopped and the autoclave was cooled in an ice-batch until reaching a
temperature below 35 °C before opening the lit. Around 2 mL was taken
from the reaction mixture under stirring, from which 1 mL aliquots were
analysed after filtration on a 0.22 µm syringe filter. The catalyst was
recovered by simple filtration on paper, washing with ethanol (0.5 h) and
carbon was thermally treated at 500 °C under N
carbon, AC). In a typical preparation procedure, the required amount of
the metal salt precursor ((NH O and/or Ni(NO .6H O),
2
flow for 3 h (activated
4
)
6 2
H W
12
O
40.xH
2
3
)
2
2
was dissolved in 20 mL of deionized water and the solution was added
dropwise to a suspension of 10 g AC in 100 mL water under stirring (300
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
drying at 100 °C. All results (conversion and selectivity) are expressed as
molar percentages.
The amount of moles was calculated from the corresponding
concentration (mg mL ), the molecular weight (g mol ) and the reaction
volume in batch experiments (64 mL).
-
1
-1
Continuous flow experiments. The experiments were conducted
without the addition of molecular hydrogen (H
high-pressure Phoenix Flow Reactor (ThalesNanoTM
Figure S4), connected to a HPLC pump to supply a continuous feed of a
.1-0.2 M furfural feedstock in i-PrOH. A 30 mm CatCart cartridge (0.88
2
) in a high-temperature
The furfural (F) conversion, the selectivity and the yield of each
hydrogenation product (P) in the continuous flow experiments were
calculated as the following:
,
Hungary, see
0
[
퐶퐹퐼ꢀꢁ푡ꢁ푎ꢂ − 퐶퐹ꢃꢁꢀ푎ꢂ
]
mL empty volume) was packed with 110 mg catalyst and placed in the
reactor module. The total flow through volume (including feed, reactor
and product sections) was 14.0 mL. First, pure i-PrOH was pumped
through the system and then the feed was changed to the furfural
feedstock. The flow was continued until the temperature (200-260 °C)
and pressurization (0-70 bars) of the reactor module were reached. Then,
퐶표푛푣푒푟푠푖표푛퐹 (%) =
× 100
퐶퐹퐼ꢀꢁ푡ꢁ푎ꢂ
퐶푃ꢀ
푌푖푒푙푑푃ꢀ (%) =
× 100
퐶퐹퐼ꢀꢁ푡ꢁ푎ꢂ
푌푖푒푙푑푃ꢀ
푆푒푙푒푐ꢄ푖푣푖ꢄ푦푃ꢀ
(
%
)
=
× 100
-
1
퐶표푛푣푒푟푠푖표푛퐹
in function of the flow rate (0.1-0.4 mL min ), the reaction proceeded
during a certain time (20-80 min.) before collecting the first sample
100 − 푐표푛푣푒푟푠푖표푛
ꢉꢀ
ꢆꢇ
ꢈ × 퐶퐹퐼ꢀꢁ푡ꢁ푎ꢂ + ∑ꢉꢊ(푌푖푒푙푑/100 × 퐶퐹퐼ꢀꢁ푡ꢁ푎ꢂ)ꢋ
퐶퐹퐼ꢀꢁ푡ꢁ푎ꢂ
100
(sample at time = 0 min). Further samples were collected after regular
푀ꢅ푠푠 푏ꢅ푙ꢅ푛푐푒 (%) =
× 100
time intervals (20-40 min). All results (conversion and selectivity) are
expressed as molar percentages.
CF and CP are the concentrations of furfural and hydrogenation product
-
1
(
mol mL ). The total volume in the system of the continuous flow
experiments was assumed to remain constant during the experiments
no evaporation loss).
Product analysis. The concentrations of furfural and hydrogenation
products were determined by GC-FID. First, calibration curves for furfural,
furfuryl alcohol (FA), tetrahydrofurfuryl alcohol THFA, 2-methylfuran (MF)
and 2-methyltetrahydrofuran (MTHF) were determined in the 0–25 mg
mL-1 range and using n-octane (8.1 mg mL ) as the internal standard for
the products obtained in batch experiments (Figure S5), and in the 0-40
(
-1
Acknowledgements
-
-1
mg mL 1 range and using n-octane (3.0 mg mL ) as the internal standard
for the products obtained in continuous flow experiments (Figure S6).
For the batch experiments, GC-FID analysis was performed on a
AutoSystem XL gas chromatograph (PerkinElmer, Norwalk, CT 06859,
United States) coupled with a FID detector equipped with a AT™-1HT
GC CAPILLARY COLUMNS (30m × 0.25mm i.d. and 0.1 μm film
The authors gratefully acknowledge the financial support from the
European COST Action FP1306 (´Valorisation of Lignocellulosic Biomass
Side Streams for Sustainable Production of Chemicals, Materials & Fuels
Using Low Environmental Impact Technologies´) via Short Term
Scientific Missions (STSMs). The authors wish to thank the Central
Service for Research Support of the University of Córdoba. Y. Wang
thanks the China Scholarship Council for financial support. R.L. gratefully
acknowledges MINECO as well as FEDER funds for funding under
project CTQ2016-78289-P and financial support from the University of
Cordoba (Spain). The publication has been prepared with support of
RUDN University Program 5-100.
thickness; Alltech Part No.16368). N
2
was used as carrier gas at a rate of
-
1
1
mL min . The samples were injected with an auto-injector directly onto
the column using septum-equipped programmable injector (SPI) system
in split mode (20.8:1 ratio). The temperature of the injector was set
3
50 °C and the oven started at 50 °C, held for 6 min, raised to 130°C at a
rate of 10°C min and then raised to 250°C at a rate of 30°C min-1 and
-1
held for 2 min at 250 °C. The ionisation mode was FID (70 eV, 300 μA,
300 °C). For the continuous flow experiments, the GC-FID analysis was
performed on a gas chromatograph (HP, 14009 Arcade, New York,
United States) coupled with a FID detector equipped with a Supelco 2-
8
047-U capillary column (60 m x 25 m x 25 µm, Alltech Part No.31163-
-
1
Keywords: Furfural • Isopropanol • Hydrogen transfer •
Methylfuran • Continuous flow
2
01). N was used as carrier gas at a rate of 1 mL min . The temperature
of the injector was set 250 °C and the oven started at 80 °C, held for 5
-
1
min, raised to 100°C at a rate of 10°C min , held for 5 min. and then
raised to 120°C at a rate of 10°C min-1 and held for 10 min at 120 °C.
The ionization mode was FID (70 eV, 300 μA, 250 °C). The identification
of the compounds was performed by comparison of the retention times
with pure standards. The furfural conversion, the yield and selectivity to
[1]
[2]
R. A. Sheldon, Green Chem. 2014, 16, 950-963.
L. Jiang, L. Zhou, J. Chao, H. Zhao, T. Lu, Y. Su, X. Yang, J. Xu, Appl.
Catal., B. 2018, 220, 589-596.
[3]
[4]
J.-P. Lange. E. van der Heide, J. van Buijtenen, R. Price,
ChemSusChem 2012, 5, 150-166.
each hydrogenation product (P
batch experiments were calculated as the following:
n
) and the corresponding mass balance in
R. Sahu, P. L. Dhepe, ChemSusChem 2012, 5, 751-761.
R. Carrasquillo-Flores, M. Kꢀldstrꢁm, F. Schꢂth, J. A. Dumesic, R.
Rinaldi, ACS Catal. 2013, 3, 993-997.
[5]
[
푀표푙퐹퐼ꢀꢁ푡ꢁ푎ꢂ − 푀표푙퐹ꢃꢁꢀ푎ꢂ
푀표푙퐹퐼ꢀꢁ푡ꢁ푎ꢂ
]
퐶표푛푣푒푟푠푖표푛퐹 (%) =
× 100
[
6]
G. Yi, S. P. Teong, Y. Zhang, Green Chem. 2016, 18, 979-983.
M. Chatterjee, T. Ishizakaa, H. Kawanami, Green Chem. 2016, 18, 487-
496.
[
푀표푙푃ꢀ]
[7]
푌푖푒푙푑푃
ꢀ
(%) =
× 100
푀표푙퐹퐼ꢀꢁ푡ꢁ푎ꢂ
[
8]
9]
R. V. Sharma, U. Das, R. Sammynaiken, A. K. Dalai, Appl. Catal., A
2013, 454, 127-136.
푌푖푒푙푑푃
ꢀ
푆푒푙푒푐ꢄ푖푣푖ꢄ푦푃
ꢀ
(%) =
× 100
[
퐶표푛푣푒푟푠푖표푛퐹]
[
M. M. Villaverde, N. M. Bertero, T. F. Garetto, A. J. Marchi, Catal.
Today 2013, 213, 87-92.
1
00 − 푐표푛푣푒푟푠푖표푛
ꢉꢀ
ꢆꢇ
ꢈ × 푀표푙퐹퐼ꢀꢁ푡ꢁ푎ꢂ + ∑ꢉꢊ(푌푖푒푙푑/100 × 푀표푙퐹퐼ꢀꢁ푡ꢁ푎ꢂ)ꢋ
푀표푙퐹퐼ꢀꢁ푡ꢁ푎ꢂ
100
푀ꢅ푠푠 푏ꢅ푙ꢅ푛푐푒 (%) =
× 100
[10] N. S. Biradar, A. A. Hengne, S. N. Biradar, R. Swami, C. V. Rode, Org.
Process Res. Dev. 2014, 18, 1434-1442.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
[
11] K. Fulajtárova, T. Sotáka, M. Hronec, I. Vávra, E. Dobročka, M.
[21] S. Srivastava, G. C. Jadeja, J. Parikh, J. Mol. Catal. A: Chem. 2017,
426, 244-256.
Omastová, Appl. Catal., A 2015, 502, 78-85.
[
12] B. Chen, F. Li, Z. Huang, G. Yuan, Appl. Catal., A 2015, 500, 23-29.
13] W. Gong, C. Chen, H. Zhang, Y. Zhang, Y. Zhang, G. Wang, H. Zhao,
Mol. Catal. 2017, 429, 51-59.
[22] Z. Xinghua, W. Tiejun, M. Longlong, W. Chuangzhi, Fuel 2010, 89,
2697-2702.
[
[23] O. F. Aldosari, S. Iqbal, P. J. Miedziak, G. L. Brett, D. R. Jones, X. Liu,
J. K. Edwards, D. J. Morgan, D. K. Knight, G. J. Hutchings, Catal. Sci.
Technol. 2016, 6, 234-242.
[
[
[
14] M. J. Taylor, L. J. Durndell, M. A. Isaacs, C. M. A. Parlett, K. Wilson, A.
F. Lee, G. Kyriakou, Appl. Catal., B 2016, 180, 580-585.
15] X. Chen, L. Zhang, B. Zhang, X. Guo, X. Mu, Sci. Rep. 2016, 6, 28558-
[24] A. J. Garcia-Olmo, A. Yepez, A. M. Balu, P. Prinsen, A. Garcia, A.
Maziere, C. Len, R. Luque, Tetrahedr. 2017, 73, 5599-5604.
[25] A. J. Garcia-Olmo, A. Yepez, A. M. Balu, A. A. Romero, Y. Li, R. Luque,
Catal. Sci. Technol. 2016, 6, 4705-4711.
28571.
16] P. Reyes, D. Salinas, C. Campos, M. Oportus, J. Murcia, H. Rojas, G.
Borda, J. L. García Fierro, Quím. Nova 2010, 33, 777-780.
[
[
[
17] M. Koehle, R. F. Lobo, Catal. Sci. Technol. 2016, 6, 3018-3026.
18] P. Panagiotopoulou, D. G. Vlachos, Appl. Catal., A 2014, 480, 17-24.
19] H. Wei, D. Feng, G. Shu, M. Pan, Y. Guo, D. Gao, W. Li, Appl. Energ.
[26] S. S. Barton, M. J. B. Evans, E. Halliop, J. A. F. MacDonald, Carbon
1997, 35, 1361-1366.
[27] M. Y. Chernyak, V. E. Tarabanko, A. А. Morosov and A. А.
Kondrasenko. Journal of Siberian Federal University, Chemistry 2 2016,
9, 146-151.
2014, 132, 317-324.
[20] M. J. Gilkey, P. Panagiotopoulou, A. V. Mironenko, G. R. Jenness, D. G.
Vlachos, B. Xu, ACS Catal. 2015, 5, 3988-3994.
[
28] X. Wang and R. Rinaldi, Energy Environ. Sci. 2012, 5, 8244-8260.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201800530
FULL PAPER
FULL PAPER
Mono- and bimetallic nickel-based
catalysts are compared in batch- and
continuous flow mode for the
catalytic transfer hydrogenation of
furfural in isopropanol. The catalytic
performance depends mostly on the
reactions conditions.
Yantao Wang, Pepijn Prinsen,
Konstantinos S. Triantafyllidis,
Stamatia A. Karakoulia, Alfonso Yepez,
Christophe Len, Rafael Luque*
Page No. – Page No.
Batch vs. continuous flow
performance of supported mono-
and bimetallic nickel catalysts for
catalytic transfer hydrogenation of
furfural in isopropanol
This article is protected by copyright. All rights reserved.