Electrocatalytic hydrogenation of 5-hydroxymethylfurfural in the absence and presence of glucose

Article abstract of DOI:10.1002/cssc.201300443

Electrocatalytic hydrogenation of 5-hydroxymethylfurfural (HMF) to 2,5-dihydroxymethylfuran (DHMF) or other species, such as 2,5-dimethylfuran, on solid metal electrodes in neutral media is addressed, both in the absence and in the presence of glucose. The reaction is studied by combining voltammetry with on-line product analysis by using HPLC, which provides both qualitative and quantitative information about the reaction products as a function of electrode potential. Three groups of catalysts show different selectivity towards: (1) DHMF (Fe, Ni, Ag, Zn, Cd, and In), (2) DHMF and other products (Pd, Al, Bi, and Pb), depending on the applied potential, and (3) other products (Co, Au, Cu, Sn, and Sb) through HMF hydrogenolysis. The rate of electrocatalytic HMF hydrogenation is not strongly catalyst-dependent because all catalysts show similar onset potentials (-0.5±0.2 V) in the presence of HMF. However, the intrinsic properties of the catalysts determine the reaction pathway towards DHMF or other products. Ag showed the highest activity towards DHMF formation (up to 13.1 mM cm^-2 with high selectivity> 85 %). HMF hydrogenation is faster than glucose hydrogenation on all metals. For transition metals, the presence of glucose enhances the formation of DHMF and suppresses the hydrogenolysis of HMF. On poor metals such as Zn, Cd, and In, glucose enhances DHMF formation; however, its contribution in the presence of Bi, Pb, Sn, and Sb is limited. Remarkably, in the presence of HMF, glucose hydrogenation itself is largely suppressed or even absent. The first electron-transfer step during HMF reduction is not metal-dependent, suggesting a non-catalytic reaction with proton transfer directly from water in the electrolyte. A clean sweep: The hydrogenation of HMF in neutral media has been studied on a wide range of solid metal electrodes both in the absence and in the presence of glucose. From HMF hydrogenation, three groups of catalysts show affinities towards (1) DHMF, (2) DHMF and other products, depending on applied potentials, and (3) other products. HMF hydrogenation is shown to be preferred to glucose hydrogenation on all metals. Copyright

Full text of DOI:10.1002/cssc.201300443

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201300443
Electrocatalytic Hydrogenation of 5-Hydroxymethylfurfural
in the Absence and Presence of Glucose
Youngkook Kwon,[a] Ed de Jong,[b] Saeed Raoufmoghaddam,[a] and Marc T. M. Koper*[a]
Electrocatalytic hydrogenation of 5-hydroxymethylfurfural
(HMF) to 2,5-dihydroxymethylfuran (DHMF) or other species,
such as 2,5-dimethylfuran, on solid metal electrodes in neutral
media is addressed, both in the absence and in the presence
of glucose. The reaction is studied by combining voltammetry
with on-line product analysis by using HPLC, which provides
both qualitative and quantitative information about the reac-
tion products as a function of electrode potential. Three
groups of catalysts show different selectivity towards:
(1) DHMF (Fe, Ni, Ag, Zn, Cd, and In), (2) DHMF and other prod-
ucts (Pd, Al, Bi, and Pb), depending on the applied potential,
and (3) other products (Co, Au, Cu, Sn, and Sb) through HMF
hydrogenolysis. The rate of electrocatalytic HMF hydrogenation
is not strongly catalyst-dependent because all catalysts show
similar onset potentials (ꢀ0.5ꢁ0.2 V) in the presence of HMF.
However, the intrinsic properties of the catalysts determine
the reaction pathway towards DHMF or other products.
Ag showed the highest activity towards DHMF formation (up
to 13.1 mmcmꢀ2 with high selectivity> 85%). HMF hydrogena-
tion is faster than glucose hydrogenation on all metals. For
transition metals, the presence of glucose enhances the forma-
tion of DHMF and suppresses the hydrogenolysis of HMF.
On poor metals such as Zn, Cd, and In, glucose enhances
DHMF formation; however, its contribution in the presence of
Bi, Pb, Sn, and Sb is limited. Remarkably, in the presence of
HMF, glucose hydrogenation itself is largely suppressed or
even absent. The first electron-transfer step during HMF reduc-
tion is not metal-dependent, suggesting a non-catalytic reac-
tion with proton transfer directly from water in the electrolyte.
Introduction
Catalytic conversion of biomass to fuels and chemicals has at-
tracted great attention as one of the future technologies for
mitigating global warming and for building a carbon-neutral
energy cycle.[1,2] 5-hydroxymethylfurfural (HMF), a new bio-
mass-derived platform chemical, has potential as an alternative
commodity chemical to fossil-fuel-based platform chemicals.[3,4]
The chemistry, production processes, and potential applica-
tions of HMF have recently been reviewed.[4] HMF has been
produced from fructose in water, in organic solvents (DMSO),[5]
in a biphasic system (water-MIBK),[6] and in ionic liquids[7–10] by
using acid catalysts (i.e., HCl, H2SO4),[11] solid acids (i.e., zeo-
lites,[6] ion exchange resins[12]), and salts (i.e., LaCl3, CrCl2).[13,14]
Recently, HMF was obtained in high yields directly from fruc-
tose, from glucose via isomerization to fructose, as well as di-
rectly from cellulose.[15–17]
oxidative products of HMF are 2,5-furandicarbaldehyde (FDC)
and 2,5-furandicarboxylic acid (FDCA), which are excellent can-
didates for monomers in the production of biomass-derived
polymeric materials.[18,19] 2-formyl-5-furancarboxylic acid (FFCA)
is an intermediate species formed during HMF oxidation that
can be further oxidized to FDCA, a potential replacement for
the fossil-fuel-based platform chemical terephthalic acid.[20] For
the oxidation of HMF, Vuyyuru and Strasser[21] reported 100%
conversion of HMF to DFCA with 80% selectivity by using Pt/C
at mild temperature (508C) and pressure (10 bar O2); in com-
parison, the electrochemical oxidation of HMF on a Pt elec-
trode is more selective to FDC with 18% yield.
The hydrogenation of the formyl group or the furan ring in
HMF to 2,5-dihydroxymethylfuran (DHMF) or dihydroxy-
methyl-tetrahydrofuran (DHMTHF) has been studied by using
heterogeneous catalysts based on Ni, Cu, Pt, Pd, and Ru,[22] or
by using bimetallic catalysts (i.e., Ni–Pd).[23] DHMF is used in
a range of materials, such as resins, polymers, and artificial
fibers, and as an intermediate in the synthesis of drugs.[3]
DHMTHF is a useful chemical with applications as a solvent,[24]
a monomer,[18] or as a precursor to the production of other
high-value chemicals.[25] The hydrogenation of the hydroxyl
group on HMF leads to the formation of 2,5-dimethylfuran
(DMF), which is of particular interest because of its high
energy content and potential use as a biofuel.[26] DMF can be
obtained on CuRu/C[26,27] catalysts with high yields (76–79%).
Although a large number of papers related to HMF hydrogena-
HMF contains two functionalities (ꢀOH, ꢀC=O) attached to
a furan ring, and it can be converted into several value-added
compounds through oxidation and reduction processes. The
[a] Y. Kwon, S. Raoufmoghaddam, Prof. M. T. M. Koper
Leiden Institute of Chemistry, Leiden University
P.O. Box 9502, 2300 RA Leiden (The Netherlands)
Fax: (+31)071-527-4451
E-mail: m.koper@chem.leidenuniv.nl
[b] Dr. E. de Jong
Avantium Chemicals, Zekeringstraat 29
1014 BV Amsterdam (The Netherlands)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201300443.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1659

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
tion have been published, the hydrogenation of HMF through
electrochemical methods has not yet been explored.
reduction in the absence of glucose: (1) metals mainly forming
DHMF (Fe, Ni, Ag, Zn, Cd, and In), (2) metals forming DHMF
and others, depending on the applied potential (Pd, Al, Bi, and
Pb), and (3) metals forming mainly other products (Co, Au, Cu,
Sn, and Sb). The same catalysts were compared for the reduc-
tion of a mixture of HMF and glucose.
Recently, we studied the electrocatalytic glucose hydrogena-
tion in neutral media on a large number of pure solid-metal
electrodes from across the Periodic Table and successfully
demonstrated the activity and selectivity of various catalysts
by correlating the voltammetry with on-line HPLC product
analysis.[28] Three groups of catalysts, regarding reaction prod-
ucts, could be distinguished: (1) hydrogen formation on early-
transition metals (Ti, V, Cr, Mn, Zr, Nb, Mo, Hf, Ta, We, and Re)
and platinum-group metals (Ru, Rh, Ir, and Pt), (2) sorbitol for-
mation on late-transition metals (Fe, Co, Ni, Cu, Pd, Au, and
Ag) as well as Al (sp metal), and (3) sorbitol and 2-deoxysorbi-
tol formation on post-transition metals (In, Sn, Sb, Pb, and Bi)
as well as Zn and Cd (d metals). As HMF is a product of subse-
quent glucose dehydration during acid-catalyzed cellulose hy-
drolysis,[29] there is a need to understand the effect of the pres-
ence of glucose on the hydrogenation of HMF, and vice-versa.
Herein, we demonstrate the electrocatalytic HMF hydrogena-
tion to DHMF and other products (i.e., 5-methylfurfural, 2,5-di-
methylfuran, 2,5-dimethyl-2,3-dihydrofuran), in the absence
and in the presence of glucose, on a large number of pure
solid-metal electrodes, which are all active for glucose hydro-
genation, aiming at understanding the activity and selectivity
of catalysts by correlating the voltammetry and on-line prod-
uct analysis. Reaction products are determined both quantita-
tively and qualitatively by using a combination of voltammetry
and on-line HPLC.[28–35] Finally, a general mechanism for electro-
catalytic hydrogenation of HMF and the effect of the presence
of glucose are discussed based on the observed activity and
selectivity profiles during electrolysis.
HMF reduction in the absence of glucose
Figure 1 shows the voltammograms of the HMF-free (i.e.,
blank) solution (dashed line) and a solution containing 50 mm
HMF (solid line) in 0.1m Na2SO4, alongside the concentration
profiles of the reaction products (DHMF and others), on (a) Fe,
Ni, and Ag, (b) Pd and Al, and (c) Co, Au, and Cu. The cathodic
current observed in the blank experiment was attributed to
the evolution of hydrogen, and the solid line represented the
formation of DHMF and other products from HMF reduction,
as well as hydrogen from the hydrogen evolution reaction
(HER). Note that all transition metals in Figure 1 showed differ-
ent onset potentials for hydrogen evolution in the blank scans,
within the potential ranges from ꢀ1 V to ꢀ0.4 V, which was
consistent with our previous study.[28] For all metals, the onset
potential in the presence of HMF was the same or more posi-
tive compared to the blank. In the presence of HMF, the catho-
dic currents at high overpotentials were lower than those from
blank scans, presumably caused by HMF adsorption onto the
electrode surface. Periodic spikes were visible in the voltammo-
grams with HMF on Cu and Ag, which were attributed to irreg-
ularities in the pumping flow rate during fraction collection.[30]
Although each metal showed a different catalytic activity, we
observed certain trends in product distribution. According to
the dominant reaction products, the transition metals could be
divided into three groups. The metals in Figure 1a showed
higher activity towards DHMF formation than to other prod-
ucts over the entire potential range, whereas the metals in Fig-
ure 1c preferred to generate other products through hydroge-
nolysis of HMF over the production of DHMF. The metals in
Figure 1b alternated their activity towards DHMF and others
depending on the applied potential.
Results and Discussion
DHMF is the product after hydrogenation of the C=O group in
HMF, and the term ‘others’ refers to the mixture of other hy-
drogenolysis products (i.e., 5-methylfurfural, 2,5-dimethylfuran,
2,5-dimethyl-2,3-dihydrofuran) generated through the cleavage
of the ꢀOH functionality in HMF, as shown in Scheme 1. The
overall sum of the other products was obtained from HPLC
analysis, whereas GC–MS identified each individual species.
The details of hydrogenation pathways will be described later.
The term ‘selectivity’ is the percentage of the total reaction
products that comes from HMF, that is, DHMF, others, sorbitol,
and 2-deoxysorbitol, not taking into account hydrogen genera-
tion. The current density (mAcmꢀ2) and concentration yield of
reaction products (mmcmꢀ2) to compare the catalytic activities
were obtained based on the geometric area of the catalyst.
The metal catalysts studied could be divided into three
groups, based on the dominant reaction products from HMF
Although Fe, Ni, and Ag showed different onset potentials
in the blank scans, shown in Figure 1a, we noted that in the
presence of HMF, DHMF was generated from the same poten-
tial of approximately ꢀ0.4 V on each electrode. The maximum
production of DHMF on these catalysts was observed at a po-
tential of approximately ꢀ0.8 V. At more negative potentials,
hydrogen evolution was dominant. On Fe, the other hydroge-
nolysis products were observed from ꢀ0.63 V and their con-
centrations increased as the cathodic potential increased.
Therefore, the selectivity of DHMF on Fe decreased from 100%
at the onset potential to approximately 50% at ꢀ1.5 V. Ni
showed the most positive onset
potential in the blank solution,
and the highest current was ob-
served with HMF in solution.
However, the activity of Ni to-
wards HMF hydrogenation was
Scheme 1. HMF hydrogenation and hydrogenolysis pathways.
lower than that of Fe and Ag.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1660

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 1. Electrocatalytic HMF (50 mm) reduction on (a) Fe, Ni, and Ag, (b) Pd and Al, and (c) Co, Au, and Cu in 0.1m Na2SO4. Current-density profiles (upper
panels) with (solid line) and without (dashed line) HMF in the solution during linear sweep voltammetry with a scan rate of 1 mVsꢀ1, and concentration pro-
files (lower panels) of corresponding reaction products, that is, DHMF and others, as a function of potential.
The maximum concentration of DHMF on Ni was observed at
ꢀ0.75 V, whereas the other products on Ni were generated
from a potential of ꢀ0.69 V and their concentrations increased
gradually as the cathodic potential increased, leading to a com-
parable selectivity towards DHMF and others at ꢀ1.5 V. Inter-
estingly, the current density on Ag in the presence of HMF was
higher than that of blank from the onset potential until ꢀ1.3 V.
The enhanced current was mainly attributed to the generation
of DHMF because Ag shows the highest activity towards
DHMF formation of all metal electrodes tested. The maximum
concentration of DHMF on Ag was 13.1 mmcmꢀ2 at ꢀ0.81 V.
The selectivity of DHMF on Ag was higher than 85% at all
potentials.
least an order of magnitude more active than glucose hydro-
genation to sorbitol.
The activity and selectivity of the post-transition sp metals
(including Zn and Cd) are shown in Figure 2. An important dif-
ference between sp metals and d metals is the wider potential
window on the sp metals because of their poor HER activity.
Figure 2 shows the voltammograms of the blank solution
(dashed line) and 50 mm HMF (solid line) reduction in 0.1m
Na2SO4 alongside the concentration profiles of the reaction
products, DHMF and others, on (a) Zn, Cd, and In, (b) Bi and
Pb, and (c) Sn and Sb. In general, the onset potential for hydro-
gen evolution in the blank scans was more negative than
ꢀ1 V, which allowed for clear differentiation between the hy-
drogenation current of HMF and that of the HER. According to
Figure 2a, the cathodic current with HMF on Zn, Cd, and In
began to increase from approximately ꢀ0.7 V, which was ap-
proximately 300 mV less negative than the potential of the
blank. The current profile with HMF in solution was very similar
to that of the Ag electrode shown in Figure 1a. Zn, Cd, and In
reduced HMF primarily to DHMF starting at approximately
ꢀ0.65 V. Zn was the most active material until the production
of DHMF began to decrease after ꢀ1.07 V, that is, the potential
at which the HER starts on Zn. Therefore, we suspected that
the HER interfered with the hydrogenation of the C=O group
of HMF. On Cd and In, the production of DHMF remained con-
stant at negative potentials. On Cd, the selectivity towards
DHMF appeared to increases to 90% with the onset of the
HER.
In Figure 1b, it is shown how Pd and Al alternate their selec-
tivity towards DHMF and other products depending on the ap-
plied potential. Although the onset potential of each metal ap-
peared very different in the voltammetry, both electrodes gen-
erated DHMF from the same potential of ꢀ0.57 V. These elec-
trodes demonstrated selectivity towards DHMF formation until
approximately ꢀ0.9 V, below which potential the formation of
other products became dominant. As shown in Figure 1c, Co,
Au, and Cu preferentially reduced HMF through hydrogenoly-
sis. Cu was the most active electrode material for HMF hydro-
genolysis, even though it exhibited the lowest overall reduc-
tion current. Co and Au showed similar activity and selectivity
to both DHMF and other hydrogenolysis products.
From the above results, we concluded that the hydrogena-
tion of HMF was not strongly catalyst-dependent because all
of the catalysts showed almost identical onset potentials
(ꢀ0.5ꢁ0.1 V). However, the catalysts differed in their reaction
pathways towards DHMF or other products. Compared to the
glucose hydrogenation activity on transition metals, the maxi-
mum concentration of sorbitol was achieved on Au and Cu
(approximately 0.15 mmcmꢀ2 with 0.1m glucose),[28] showing
that HMF hydrogenation to DHMF and other products was at
The voltammograms for HMF reduction on Bi and Pb shown
in Figure 2b are similar to those in Figure 2a. DHMF and
others were generated simultaneously from approximately
ꢀ0.66 V. However, selectivity towards DHMF on both elec-
trodes was lower than those towards other hydrogenolysis
products before the HER started. In the potential region of the
HER, the concentration of DHMF significantly increased from
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1661

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 2. Electrocatalytic HMF (50 mm) reduction on (a) Zn, Cd, and In, (b) Bi and Pb, and (c) Sn and Sb in 0.1m Na2SO4. Upper panels: current-density profiles
(upper panels) with (solid line) and without (dashed line) HMF in the solution during linear sweep voltammetry with a scan rate of 1 mVsꢀ1. Lower panels:
concentration profiles of the corresponding reaction products, that is, DHMF and others, as a function of potential.
ꢀ1.29 V on Pb and from ꢀ1.41 V on Bi. On the other hand, the
formation of other products was suppressed during HER.
According to Figure 2c, Sn showed a higher overpotential
for the HER in the blank voltammogram compared to that of
Sb. However, HMF reduction started at the same potential.
Both electrodes showed low activity towards the generation of
DHMF. Other products on Sn were first observed from ꢀ0.63 V,
which was 120 mV earlier than that of DHMF. The concentra-
tion of DHMF was constant below a potential of ꢀ1 V, and its
selectivity was lower than 30%. The hydrogenolysis activity of
Sn during HMF hydrogenation was consistent with the result
from glucose reduction because Sn showed a 90% selectivity
towards 2-deoxysorbitol, which was also a hydrogenolysis
product.[28] DHMF and other products on Sb were observed
from a potential of ꢀ0.69 V. The initial selectivity of DHMF was
39%; however, it gradually decreased and no DHMF was ob-
served below ꢀ1.59 V.
genation described in the previous section, there is no clear
correlation between the electronic structure of the catalysts
and the product distribution. In addition, the onset potential
for HMF hydrogenation appeared to be almost constant and
not influenced by the HER. It was of interest to study the hy-
drogenation of a mixture of HMF and glucose to gain insight
into how they influence each other. Specifically, we selected
a relatively low concentration of HMF (10 mm) compared to
that of glucose (0.1m) to mimic their ratio during acid-cata-
lyzed cellulose hydrolysis. The concentration of HMF is lowered
from 50 to 10 mm because of the instability of the sampling
pump for a viscous solution.
Figure 3 shows the voltammograms of the blank solution
(dashed line) and of the 10 mm HMF+0.1m glucose (solid
line) reduction in 0.1m Na2SO4 alongside the concentration
profiles of the reaction products on transition metals (a) Fe, Ni,
and Ag, (b) Pd and Al, and (c) Co, Au, and Cu. We note that the
presence of glucose in the solution did not change the general
trend of current profiles compared to Figure 1, although a ten-
fold higher concentration of glucose (0.1m) with respect to
HMF (10 mm) was added. Surprisingly, the glucose hydrogena-
tion product, sorbitol, was not observed at any potential on
any of the transition metals in Figure 3. However, the presence
of glucose in solution changed the distribution of the HMF re-
duction products. According to Figure 1a, the onset potential
of DHMF on Fe, Ni, and Ag was ꢀ0.39 V. However, with glu-
cose in the solution it shifted to ꢀ0.45 V on Fe, ꢀ0.21 V on Ni,
and ꢀ0.51 V on Ag. The delay of the onset potential on Fe and
Ag could be explained by adsorbed glucose on the catalyst
surface, but glucose accelerated HMF hydrogenation on Ni.
The formation of other products, generated through hydroge-
nolysis of HMF, was strongly suppressed at negative potentials
and only traces were observed on Fe and Ag. Interestingly, the
HER influenced product selectivity so that the concentration of
DHMF decreased on Ni and Ag after hydrogen evolution
For the poor-metal electrodes, we could conclude that, in
agreement with the transition metals, HMF hydrogenation ap-
peared to be a non-catalytic reaction. However, there were still
differences between the catalysts in terms of their selectivity
towards DHMF and other products. The HER on poor metals
appeared to influence the generation of DHMF; DHMF forma-
tion on Bi and Pb was enhanced by the HER, whereas the HER
appeared to suppress DHMF formation on Zn and Sb.
HMF reduction in the presence of glucose
Glucose and HMF are generated from acid-catalyzed cellulose
hydrolysis;[29] however, their reactivity during hydrogenation is
quite different. From our previous work on glucose hydrogena-
tion,[28] we demonstrated different affinities of various catalysts
towards hydrogen formation (early-transition and Pt-group
metals), sorbitol formation (late-transition metals), as well as
sorbitol and 2-deoxysorbitol formation (post-transition metals
including Zn and Cd). However, from the trend of HMF hydro-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1662

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 3. Electrocatalytic reduction of HMF (10 mm) and glucose (0.1m) on (a) Fe, Ni, and Ag, (b) Pd and Al, and (c) Co, Au, and Cu, in 0.1m Na2SO4. Upper
panels: current-density profiles with (solid line) and without (dashed line) HMF+glucose in the solution during linear sweep voltammetry with a scan rate of
1 mVsꢀ1. Lower panels: concentration profiles (lower panels) of the corresponding reaction products, that is, DHMF and others, as a function of potential.
began. However, HMF reduction on Fe was less influenced by
the HER.
glucose was not hydrogenated. The hydrogenolysis of HMF on
Sb was accelerated in the presence of glucose, especially from
the potential at which the HER starts.
HMF reduction on Pd and Al is shown in Figure 3b. Hydro-
genolysis products were dominant on these two metals, al-
though glucose appeared to enhance DHMF production on Al.
Also, the onset potential for DHMF formation on Pd was 0.3 V
less negative than that shown in Figure 1b. As a similar trend
was observed on Ni in Figure 3a, we suspected that the pres-
ence of glucose enhanced the hydrogenation activity on Ni
and Pd, which are both known to absorb hydrogen.[36]
Based on our observations we could conclude that on transi-
tion and poor metals, the presence of HMF blocked the reduc-
tion of glucose. On the other hand, glucose enhanced the for-
mation of DHMF and suppressed the hydrogenolysis of HMF
on all transition metals and some (Zn, Cd, and In) poor metals.
However, the influence of glucose on HMF reduction on Bi, Pb,
Sn, and Sb was relatively limited.
Voltammograms of Co, Au, and Cu in Figure 3c are similar
to those in Figure 1c; however, the product distribution was
different in the presence of glucose. Firstly, DHMF formation
on these three metals was observed from the onset potential,
whereas the formation of other products was delayed. Interest-
ingly, the concentration of DHMF and others on Co in the pres-
ence of glucose was significantly higher than that shown in
Figure 1c, although the concentration of HMF was five times
lower. On Au and Cu, glucose also enhanced the formation of
DHMF with suppression of the other products. Suppression of
DHMF formation on Cu was observed below ꢀ0.69 V, when
the HER was initiated.
A mechanistic view of HMF hydrogenation
From the relationship between the voltammetry and the reac-
tion product distribution of HMF hydrogenation in the absence
and the presence of glucose, as shown in Figure 1-4, we con-
cluded that the hydrogenation of HMF was always dominant
and that the reduction of glucose was strongly suppressed by
the presence of HMF. Interestingly, the hydrogenation of HMF
did not appear to be dependent upon the nature of the elec-
trocatalyst, suggesting a non-catalytic reaction based on the
almost identical onset potentials (ꢀ0.5ꢁ0.1 V) for all metals.
This indicated that hydrogen adsorbed on the metal surface
(Hads) was unlikely to be involved in the initial HMF hydrogena-
tion process, which was different from general hydrogenation
mechanisms of organic molecules.[36] Therefore, we assumed
that the initial HMF hydrogenation occurred directly through
water molecules from the electrolyte:
Similar to the transition metals shown in Figure 3, poor
metals also exhibited different activities and selectivities in the
presence of glucose. The main difference was that poor metals
showed activity towards glucose hydrogenation for the gener-
ation of sorbitol and 2-deoxysorbitol, even in the presence of
HMF. Glucose enhanced the formation of DHMF on Zn, Cd,
and In, as shown in Figure 4a, but the hydrogenolysis of HMF
was blocked, which was also seen for the transition metals.
The effect of glucose on HMF reduction on Pb, Sn, and Sb
was also rather small, as shown in Figure 4. Note that the high-
est concentration of sorbitol (0.5 mmcmꢀ2) was observed on
Pb, which was consistent with the activity of Pb for the hydro-
genation of glucose to sorbitol.[28] On Sn, the selectivity of
DHMF was within 15–40% over the entire potential range, and
HMF þ 2 H2O þ 2 eꢀ $ DHMF þ 2 OHꢀ
ð1Þ
A similar hydrogenation mechanism, not involving Hads, has
been suggested for the NO adsorbate reduction on Pt.[37]
As the onset potential of HMF hydrogenation was not
strongly catalyst-dependent, we assumed that other conduc-
tive materials could also be used as potential catalysts for HMF
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1663

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 4. Electrocatalytic reduction of HMF (10 mm) and glucose (0.1m) on (a) Zn, Cd, and In, (b) Bi and Pb, and (c) Sn and Sb, in 0.1m Na2SO4. Upper panels:
current-density profiles with (solid line) and without (dashed line) HMF+glucose in the solution during linear sweep voltammetry with a scan rate of
1 mVsꢀ1. Middle and lower panels: concentration profiles of the corresponding reaction products, that is, DHMF and others (middle panel), as well as sorbitol
and 2-deoxysorbitol (lower panel), as a function of potential.
hydrogenation. To this end, we tested pyrolytic graphite (PG)
and polycrystalline Pt for HMF reduction in 0.1m Na2SO4, as
shown in Figure 5. PG is normally used as a catalyst for the
oxygen reduction reaction (ORR) in alkaline solution[38] or as
a stable catalyst support[39] in electrochemistry; however, it
does not show catalytic activity for glucose hydrogenation.[40]
Polycrystalline Pt shows very limited activity towards the hy-
drogenation of glucose to sorbitol[28] because of its high affini-
ty to generate hydrogen.[28] According to the blank voltammo-
grams, the HER started from approximately ꢀ0.9 V on PG and
ꢀ0.3 V on Pt. The onset potential of HMF reduction on PG was
around ꢀ0.65 V, which was a similar onset potential of HMF re-
duction on other metals. From approximately ꢀ0.7 V, we ob-
served other hydrogenolysis products on PG. In contrast,
DHMF was predominantly observed from approximately
ꢀ0.4 V on Pt. Therefore, these experiments confirmed our hy-
pothesis that the HMF reduction is a non-catalytic reaction.
The potential degradation of HMF at high pH values was
also investigated. 10 mm HMF was added to 0.1m NaOH in the
absence and presence of oxygen. After one hour, we did not
observe any isomerization, aldol-condensation, or expected
HMF oxidation products (i.e., FDC, FFCA, FDCA) under both
conditions, although it is known that HMF is not stable at ele-
vated temperatures in alkaline conditions.[4] As we have shown
previously, a high pH value was essential for glucose hydroge-
nation because the high pH environemnt near the electrode
surface could cause mutarotation of the cyclic (inactive) form
of glucose to the linear (electroactive) form, which is consid-
Figure 5. Electrocatalytic HMF (50 mm) reduction on pyrolytic graphite (PG)
and a polycrystalline Pt coil in 0.1m Na2SO4. (a) Current-density profiles with
(solid line) and without (dashed line) HMF in the solution during linear
sweep voltammetry with a scan rate of 1 mVsꢀ1, and (b) concentration pro-
files of the corresponding reaction products, that is, DHMF and others, as
a function of potential.
[41–43]
ered as the rate-determining step.
Based on this observa-
tion, we concluded that HMF was more stable than glucose at
high pH values at room temperature, and that a high pH value
was not an essential factor for HMF hydrogenation.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1664

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
In addition to the effect of pH, there was also a thermody-
namic reason why the hydrogenation of HMF was more favor-
able compared to that of glucose. The enthalpy change dur-
ing glucose hydrogenation to sorbitol was approximately
ꢀ10 kcalmolꢀ1, as hydrogenation reactions are typically exo-
thermic; however, the hydrogenation of HMF to DHMF had an
enthalpy change of approximately ꢀ20 kcalmolꢀ1,[44] which
could explain the preferential hydrogenation of HMF instead
of glucose. Under electrochemical conditions, the identical
onset potentials (ꢀ0.5ꢁ0.2 V) for HMF hydrogenation should
be related to the equilibrium potential of the reaction. Howev-
er, the equilibrium potential of HMF hydrogenation to DHMF
was around 0.1 V, based on the Gibbs free energy changes in
the gas phase at 298 K,[45] which implied that the onset poten-
tial for the reaction was mainly attributed to the activation
energy (ꢂ0.5 V), which was hardly influenced by the catalyst.
Although the hydrogenolysis of HMF was energetically neutral,
it was clear that proton- and electron-transfer reactions were
involved in the hydrogenolysis process. Additionally, the hydro-
genation of the C=C bond in the furan ring of HMF was ther-
modynamically more favored compared to DHMF formation
(i.e., the enthalpy change, DH=ꢀ35 kcalmolꢀ1 during DHMF
hydrogenation to DHMTHF[44]). Therefore, we could observe
the formation of 2,5-dimethyl-2,3-dihydrofuran as one of the
other products during HMF electrocatalysis, as shown in
Scheme 2.
an was extremely fast. Interestingly, DHMF was stable and not
reduced to any significant extent on Ag, Cu, Pb, or Sn, which
are highly active catalysts for HMF hydrogenation (see the Sup-
porting Information, Figure S1). Therefore, we concluded that
2,5-dimethylfuran was derived from the hydrogenation of 5-
methylfurfural through 5-methylfurfuryl alcohol, and not from
DHMF. In a final step, one of the C=C bonds on 2,5-dimethyl-
furan was hydrogenated to 2,5-dimethyl-2,3-dihydrofuran, the
most reduced reaction product observed in the GC–MS spectra
(see Figure S2 in the Supporting Information). In addition to
hydrogenated reaction products, we could also observe a cer-
tain amount of the oxidized species (i.e., 2,5-furandicarbalde-
hyde) from HMF. We believe that the oxidation products were
base-catalyzed reaction products, generated during electrolysis
near the electrode surface. Although all the detailed concentra-
tions of 5-methylfurfural, 2,5-dimethylfuran, and 2,5-dimethyl-
2,3-dihydrofuran were not presented herein in the Figures, the
general reaction pathway shown in Scheme 2 is consistent
with the experimental data.
Conclusions
This work has studied the electrochemical HMF hydrogenation
in the absence and in the presence of glucose on solid mono-
metallic electrodes in neutral solution. We studied the activity
and selectivity of the catalysts for HMF hydrogenation by cor-
relating the voltammetric scan data with on-line HPLC product
analysis, which provided qualitative and quantitative informa-
tion regarding DHMF, other products from HMF hydrogena-
tion, as well as sorbitol and 2-deoxysorbitol from the HMF and
glucose mixture, as a function of potential. Three groups of
catalysts were distinguished with respect to the reaction prod-
ucts from HMF hydrogenation. DHMF was mainly formed on
Fe, Ni, Ag, Zn, Cd, and In. DHMF and other hydrogenolysis
products were generated on Pd, Al, Bi, and Pb, depending on
the applied potential. Hydrogenolysis products of HMF were
mainly formed on Co, Au, Cu, Sn, and Sb. The hydrogenation
of HMF was not strongly catalyst-dependent because all cata-
lysts exhibited an almost identical onset potential (ꢀ0.5ꢁ
0.2 V), mainly attributed to the activation energy, which was
not influenced by the catalyst. However, the intrinsic property
of the catalysts determined the reaction final product.
Ag showed the highest activity among all metals towards
DHMF formation (with a high selectivity of >85%). Some poor
metals (Zn, Bi, Pb, and Sb) exhibited a change in selectivity for
HMF reduction when hydrogen evolution occurred simultane-
ously. In the presence of glucose, the hydrogenation of HMF
was always easier than glucose hydrogenation, on all metals.
On transition metals, glucose enhanced the formation of
DHMF and suppressed the hydrogenolysis of HMF, whereas
the glucose reduction itself was suppressed. On poor metals,
glucose enhanced DHMF formation on Zn, Cd, and In. Howev-
er, its influence on HMF reduction on Bi, Pb, Sn, and Sb was
relatively limited. Because HMF could also be reduced on
a graphite electrode, and the onset potential for HMF reduc-
tion was not very dependent on the nature of the electrode
material, we concluded that the initial (reduction) step of HMF
Scheme 2. HMF hydrogenation pathways.
According to Scheme 2, the C=O group in HMF was hydro-
genated to DHMF with 2H and the other products were gener-
ated through the hydrogenolysis of the ꢀOH group in HMF
during the hydrogenation process, leading first to 5-methylfur-
fural. To generate 2,5-dimethylfuran, the C=O moiety on 5-
methylfurfural presumably first needed to be hydrogenated to
5-methylfurfuryl alcohol, although 5-methylfurfuryl alcohol was
not observed from the GC–MS spectra. This indicated that the
conversion rate of 5-methylfurfuryl alcohol to 2,5-dimethylfur-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1665

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
collected samples was placed in an autosampler (SIL-20A) holder,
and 30 mL of sample was injected into the column. The column
(Shodex SP0810) for the analysis used pure water as the eluent.
The temperature of the column was maintained in an oven at
308C for the mixture of HMF and glucose, and at 508C for HMF
(CTO-20A). The separated compounds were detected with a refrac-
tive index detector (RID-10A). The concentrations of DHMF and
other products were obtained by using a Shodex SP0810 column
and the detailed compositions of other products were confirmed
by using an Agilent Technologies 7820A+5975 MSD GC–MS
instrument.
reduction was a non-catalytic reaction. In comparison to glu-
cose, HMF hydrogenation did not require a high pH value,
such as that generated close to the electrode surface during
hydrogen evolution, which is a known requirement for glucose
hydrogenation. As a result, the hydrogenation of glucose re-
quired a higher overpotential than that of HMF.
We note that the main aim of this paper was to scan the
ability of various metals towards electrochemical HMF hydro-
genation through a combination of voltammetry and on-line
HPLC. More quantitative information on the efficiency and the
stability of the catalysts would require additional long-term
electrolysis experiments. The current study indicated the most
interesting materials to choose for such a study.
Acknowledgements
This research has been performed within the framework of the
CatchBio program. The authors gratefully acknowledge the sup-
port of the Smart Mix Program of the Netherlands Ministry of
Economic Affairs and the Netherlands Ministry of Education, Cul-
ture and Science.
Experimental Section
Electrochemical procedures
All measurements were performed in a conventional single-com-
partment three-electrode glass cell, which was cleaned by follow-
ing a standard procedure for removing traces of organic and inor-
ganic contaminants.[46] Oxygen was removed by bubbling argon
through the solution prior to the voltammetric experiments. Transi-
tion and post-transition metals (ꢃ99.5%) that were used as work-
ing electrodes in the experiments were polycrystalline wires/rods/
plates, which were mechanically polished with alumina (up to
0.05 mm) and cleaned ultrasonically in pure water (MilliQ gradient
A10 system, 18.2 MW) followed by electropolishing by cycling
three times between ꢀ2 and 0 V (vs. reversible hydrogen elec-
trode, RHE) in 0.1m Na2SO4 solution to remove surface impurities
and oxides before use.[28] The surface area (cm2) of the catalysts
(see the Supporting Information, Figure S3), used to calculate the
current density and concentration yield of reaction products
(mMcmꢀ2), was obtained based on the geometric area, except for
the Pt electrode for which we considered the electrochemically
active surface area by the hydrogen desorption charge in 0.5m
H2SO4. In all experiments, a large gold coil was used as a counter
electrode and a RHE was used as a reference electrode. HMF
(50 mm, Sigma–Aldrich), HMF (10 mm) and Glucose (0.1m, Merck)
were dissolved into a solution of 0.1m Na2SO4 (Merck, pH 7). Elec-
trochemical cell potentials were controlled with a potentiostat/gal-
vanostat (m-Autolab Type III). All experiments were performed at
room temperature.
Keywords: biomass
·
carbohydrates
·
electrocatalysis
·
hydroxymethylfurfural · hydrogenation
[1] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098.
[2] A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502.
[3] E.-S. Kang, D. W. Chae, B. Kim, Y. G. Kim, J. Ind. Eng. Chem. 2012, 18,
174–177.
[4] R.-J. van Putten, J. C. van der Waal, E. de Jong, C. B. Rasrendra, H. J.
Heeres, J. G. de Vries, Chem. Rev. 2013, 113, 1499–1597.
[5] R. M. Musau, R. M. Munavu, Biomass 1987, 13, 67–74.
[6] C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier, P.
Ros, G. Avignon, Appl. Catal. A 1996, 145, 211–224.
[7] M. E. Zakrzewska, E. Bogel-Lukasik, R. Bogel-Lukasik, Chem. Rev. 2011,
111, 397–417.
[8] Z. Yuan, C. Xu, S. Cheng, M. Leitch, Carbohydr. Res. 2011, 346, 2019–
2023.
[9] Q. Cao, X. Guo, J. Guan, X. Mu, D. Zhang, Appl. Catal. A 2011, 403, 98–
103.
[10] S. P. Simeonov, J. A. S. Coelho, C. A. M. Afonso, ChemSusChem 2012, 5,
1388–1391.
[11] F. Salak Asghari, H. Yoshida, Ind. Eng. Chem. Res. 2006, 45, 2163–2173.
[12] D. Mercadier, L. Rigal, A. Gaset, J. P. Gorrichon, J. Chem. Technol. Biotech-
nol. 1981, 31, 489–496.
[13] K. Seri, Y. Inoue, H. Ishida, Bull. Chem. Soc. Jpn. 2001, 74, 1145–1150.
[14] H. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science 2007, 316, 1597–
1600.
[15] A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Afonso, Green
Chem. 2011, 13, 754–793.
Fraction collection and product analysis
For the detection of intermediates and products through HPLC,
samples of the electrolyte solution were collected with a small
Teflon tip (inner diameter, 0.38 mm) positioned close (10 mm) to
the center of the electrode surface, the tip was connected to
a PEEK capillary with inner/outer diameters of 0.13:1.59 mm.[30] The
tip was cleaned in a solution of 0.2m K2Cr2O7 and thoroughly
rinsed with ultrapure water before use. The sample volume collect-
ed in each well was 60 mL on a 96-well microtiter plate (270 mL per
well, Screening Devices) by using an automatic fraction collector
(FRC-10 A, Shimadzu). The flow rate for sample collection was ad-
justed to 60 mLminꢀ1 with a Shimadzu pump (LC-20AT). Collected
samples were immediately neutralized by adding 60 mL of 10 mm
phosphate buffer. After collecting all of the samples, the microtiter
plate was covered by a silicon mat to prevent the evaporation.
Samples collected during voltammetry were analyzed by using
a Shimadzu Prominence HPLC. The microtiter plate containing the
[16] P. M. Grande, C. Bergs, P. D. de Maria, ChemSusChem 2012, 5, 1203–
1206.
[17] S. P. Simeonov, J. A. S. Coelho, C. A. M. Afonso, ChemSusChem 2013, 6,
997–1000.
[18] C. Moreau, M. N. Belgacem, A. Gandini, Top. Catal. 2004, 27, 11–30.
[19] E. de Jong, M. A. Dam, L. Sipos, G.-J. M. Gruter in Biobased Monomers,
Polymers, and Materials (Eds.: P. B. Smith, R. A. Gross), ACS Symposium
Series, 2012, pp. 1–13.
[20] A. Gandini, Green Chem. 2011, 13, 1061–1083.
[21] K. R. Vuyyuru, P. Strasser, Catal. Today 2012, 195, 144–154.
[22] V. Schiavo, G. Descotes, J. Mentech, Bull. Soc. Chim. Fr. 1991, 704–711.
[23] Y. Nakagawa, K. Tomishige, Catal. Commun. 2010, 128, 154–156.
[24] R. Alamillo, M. Tucker, M. Chia, Y. Pagan-Torres, J. Dumesic, Green Chem.
2012, 14, 1413–1419.
[25] T. Buntara, S. Noel, P. H. Phua, I. Melian-Cabrera, J. G. de Vries, H. J.
Heeres, Angew. Chem. 2011, 123, 7221–7225; Angew. Chem. Int. Ed.
2011, 50, 7083–7087.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1666

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[26] Y. Romꢁn-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature 2007, 447,
982–985.
[39] O. Brylev, M. Sarrazin, D. Belanger, L. Roue, Appl. Catal. B 2006, 64, 243–
253.
[27] J. B. Binder, R. T. Raines, J. Am. Chem. Soc. 2009, 131, 1979–1985.
[28] Y. Kwon, M. T. M. Koper, ChemSusChem 2013, 6, 455–462.
[29] Y. Kwon, S. E. F. Kleijn, K. J. P. Schouten, M. T. M. Koper, ChemSusChem
2012, 5, 1935–1943.
[40] S. Fei, J. Chen, S. Yao, G. Deng, L. Nie, Y. Kuang, J. Solid State Electro-
chem. 2005, 9, 498–503.
[41] K. Park, P. N. Pintauro, M. M. Baizer, K. Nobe, J. Electrochem. Soc. 1985,
132, 1850–1855.
[30] Y. Kwon, M. T. M. Koper, Anal. Chem. 2010, 82, 5420–5424.
[31] Y. Kwon, S. C. S. Lai, P. Rodriguez, M. T. M. Koper, J. Am. Chem. Soc. 2011,
133, 6914–6917.
[32] Y. Kwon, K. J. P. Schouten, M. T. M. Koper, ChemCatChem 2011, 3, 1176–
1185.
[42] A. Bin Kassim, C. L. Rice, A. T. Kuhn, J. Appl. Electrochem. 1981, 11, 261–
267.
[43] P. N. Pintauro, D. K. Johnson, K. Park, M. M. Baizer, K. Nobe, J. Appl. Elec-
trochem. 1984, 14, 209–220.
[44] Breaking the Chemical and Engineering Barriers to Lignocellulosic Biofuels:
Next Generation Hydrocarbon Biorefineries (Ed.: G. H. Huber) NSF, 2008,
pp. 66–71.
[33] A. Santasalo-Aarnio, Y. Kwon, E. Ahlberg, K. Kontturi, T. Kallio, M. T. M.
Koper, Electrochem. Commun. 2011, 13, 466–469.
[34] Y. Kwon, Y. Birdja, I. Spanos, P. Rodriguez, M. T. M. Koper, ACS Catal.
2012, 2, 759–764.
[45] S. P. Verevkin, V. N. Emel’yanenko, E. N. Stepurko, R. V. Ralys, D. H. Zait-
sau, A. Stark, Ind. Eng. Chem. Res. 2009, 48, 10087–10093.
[46] S. C. S. Lai, M. T. M. Koper, Faraday Discuss. 2008, 140, 399–416.
[35] P. Rodriguez, Y. Kwon, M. T. M. Koper, Nat. Chem. 2012, 4, 177–182.
[36] J. M. Chapuzet, A. Lasia, J. Lessard, Electrocatalysis (Eds.: J. Lipowski,
P. N. Ross), Wiley-VCH, Weinheim, 1998, pp. 155–196.
[37] A. C. A. de Vooys, G. L. Beltramo, B. van Riet, J. A. R. van Veen, M. T. M.
Koper, Electrochim. Acta 2004, 49, 1307–1314.
Received: May 8, 2013
[38] D. Qu, Carbon 2007, 45, 1296–1301.
Published online on July 15, 2013
Part of a Special Issue on ”CatchBio“. To view the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cssc.v6.9/issuetoc
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1659 – 1667 1667