Tunable and selective conversion of 5-HMF to 2,5-furandimethanol and 2,5-dimethylfuran over copper-doped porous metal oxides

Article abstract of DOI:10.1002/cssc.201402095

Tunable and selective hydrogenation of the platform chemical 5-hydroxymethylfurfural into valuable C₆ building blocks and liquid fuel additives is achieved with copper-doped porous metal oxides in ethanol. A new catalyst composition with improved hydrogenation/hydrogenolysis activity is obtained by introducing small amounts of ruthenium dopant into the previously reported Cu_0.59Mg_2.34Al_1.00 structure. At a mild reaction temperature (100 °C), 2,5-furandimethanol is obtained with excellent selectivity up to >99 %. Higher reaction temperatures (220 °C) favor selective deoxygenation to 2,5-dimethylfuran and minor product 2,5-dimethyltetrahydrofuran with a combined yield as high as 81 %. Notably, these high product yields are maintained at a substrate concentration up to 10 wt % and a low catalyst loading. The influence of different alcohol solvents on product selectivity is explored. Furthermore, reaction intermediates formed at different reaction temperatures are identified. The composition of these product mixtures provides mechanistic insight into the nature of the reduction pathways that influence product selectivity. The catalysts are characterized by elemental analysis, TEM, and BET techniques before and after the reaction. Catalyst recycling experiments are conducted in batch and in a continuous-flow setup.

Full text of DOI:10.1002/cssc.201402095

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201402095
Tunable and Selective Conversion of 5-HMF to 2,5-
Furandimethanol and 2,5-Dimethylfuran over Copper-
Doped Porous Metal Oxides
Angela J. Kumalaputri,[a, b] Giovanni Bottari,[c] Petra M. Erne,[c] Hero J. Heeres,*[a] and
Katalin Barta*[c]
Tunable and selective hydrogenation of the platform chemical
5-hydroxymethylfurfural into valuable C6 building blocks and
liquid fuel additives is achieved with copper-doped porous
metal oxides in ethanol. A new catalyst composition with im-
proved hydrogenation/hydrogenolysis activity is obtained by
introducing small amounts of ruthenium dopant into the previ-
ously reported Cu0.59Mg2.34Al1.00 structure. At a mild reaction
temperature (1008C), 2,5-furandimethanol is obtained with ex-
cellent selectivity up to >99%. Higher reaction temperatures
(2208C) favor selective deoxygenation to 2,5-dimethylfuran
and minor product 2,5-dimethyltetrahydrofuran with a com-
bined yield as high as 81%. Notably, these high product yields
are maintained at a substrate concentration up to 10 wt% and
a low catalyst loading. The influence of different alcohol sol-
vents on product selectivity is explored. Furthermore, reaction
intermediates formed at different reaction temperatures are
identified. The composition of these product mixtures provides
mechanistic insight into the nature of the reduction pathways
that influence product selectivity. The catalysts are character-
ized by elemental analysis, TEM, and BET techniques before
and after the reaction. Catalyst recycling experiments are con-
ducted in batch and in a continuous-flow setup.
Introduction
Inedible, carbon-neutral lignocellulosic biomass represents the
most promising renewable starting material for the synthesis
of high-value-added products. 5-Hydroxymethylfurfural (HMF)
was identified by the U.S. Department of Energy (DOE)[1] as
one of the top ten biobased platform chemicals with signifi-
cant market potential.[2] This molecule, which is derived from
the C6 sugar fraction of lignocellulosic biomass, mainly by a bi-
phasic solvent approach,[3] may undergo several chemical
transformations to give valuable chemicals and fuels.[4] For ex-
ample, it can be oxidized into 2,5-diformylfuran (DFF)[5] and
furan 2,5-dicarboxylic acid (FDCA)[6] to provide building blocks
for the polymer industry. It can also be reduced to 2,5-furandi-
methanol (FDM) or biofuel additive 2,5-dimethylfuran
(DMF).[7,16]
FDM has been identified as a useful building block in the
synthesis of macromolecules.[8] Stoichiometric reduction of
HMF with a twofold excess of sodium borohydride leads to
FDM in 97% yield.[9] Catalytic approaches are more attractive
and several examples have already been reported. Alamillo
and co-workers described 81% selectivity to FDM by using
a
Ru/CeOx catalyst, with 2,5-tetrahydrofurandimethanol
(THFDM) as the reduction coproduct.[10] Nakagawa et al. used
a nickel–palladium bimetallic catalyst supported on silica for
the hydrogenation of HMF with the prevailing formation of
mixtures of FDM/THFDM and THFDM in a yield of up to 96%
under optimized conditions.[11] Recently, the same research
group reported the use of a highly selective IrÀReOx/SiO2 cata-
lyst in the hydrogenation of HMF to form FDM (99% yield)
under mild reaction conditions.[12] Gold nanoclusters supported
on alumina showed good activity in the hydrogenation of HMF
to FDM (up to 96% yield).[13] Homogeneous catalysts, such as
[Ir(Cp*)(TsDPEN)] (Cp*=C5Me5; TsDPEN=H2NCHPhCHPhNTsÀ;
Ts =tosyl) proved to be very active and selective in the transfer
hydrogenation of HMF with formic acid, giving 99% yield of
FDM.[14] Furthermore, THFDM and the corresponding alcohol
derivatives [e.g., 1,2,6-hexanetriol (1,2,6-HT) or 1,2-hexanediol
(1,2-HD)] represent important precursors for the synthesis of
polymers.[15]
[a] A. J. Kumalaputri,+ Prof. Dr. H. J. Heeres
Department of Chemical Engineering
University of Groningen, Nijenborgh 4
9747 AG Groningen (The Netherlands)
E-mail: h.j.heeres@rug.nl
[b] A. J. Kumalaputri+
Department of Chemical Engineering
Parahyangan Catholic University
Ciumbuleuit 94, Bandung 40141 (West Java, Indonesia)
[c] Dr. G. Bottari,+ P. M. Erne, Dr. K. Barta
Stratingh Institute for Chemistry
University of Groningen, Nijenborgh 4
9747 AG Groningen (The Netherlands)
E-mail: k.barta@rug.nl
Reductive deoxygenation of HMF leads to DMF. Dumesic
and co-workers reported a high yield in the production of
DMF (up to 79%) directly from fructose in a biphasic process
by using a ruthenium–copper-based catalyst.[16] Bell and Chi-
dambaram used glucose as a feedstock in ionic liquids or ace-
tonitrile.[17] Hydrogenation of HMF to form DMF was achieved
[+] These authors contributed equally to this work.
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201402095.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2266

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
with a Pd/C catalyst. Catalytic transfer hydrogenation by
means of a Ru/C catalyst and isopropanol as the hydrogen
donor also led to a DMF selectivity of up to 81%.[18] Recently,
Wang and co-workers showed that Ru/Co3O4 catalysts led to
very selective formation of DMF (94.1% from HMF and 75.1%
from fructose, with full substrate conversion).[19] In a recent
paper, Chatterjee and co-workers reported the use of a Pd/C
catalyst in supercritical carbon dioxide/water with very high
DMF selectivities.[20] Also, Pd/Fe2O3 was used in a continuous-
flow setup and proved to be effective in the conversion of
HMF to DMF (up to 72% yield).[21] Elaborately designed bimet-
allic PtCo nanoparticles encapsulated in carbon nanospheres
afforded 98% yield of DMF from HMF; the best result related
to DMF production reported so far.[22]
CuÀMgÀAl HTC and PMO has not yet been synthesized or
characterized.
The synthesized HTCs were analyzed by powder XRD and
displayed distinct features consistent with a double-layered
structure (XRD traces are shown Figure S1 in the Supporting
Information).[27] The XRD patterns were recorded in the 2q
range of 10–708, with sharp diffraction peaks for lower 2q
values (10–378), and broad peaks at higher 2q values (37–708);
the doublet at 60–628 is a typical indication of a crystalline
double-layered structure.[27] No diffraction peaks related to
copper or ruthenium oxide crystallites could be detected; this
observation is in agreement with a homodisperse distribution
of both metals in the HTC structure. After calcination at 4608C
for 24 h, PMOs of different compositions were obtained. BET
surface area values are comparable for both PMOs. There was
very good agreement between the theoretical and empirical
composition, as determined by elemental analysis of the
PMOs, which indicated that all metals were incorporated into
the HTC structure. Catalyst codes, the corresponding composi-
tions, and BET surface area values are summarized in Table 1.
Additional pore volume and average pore size values for both
catalysts are listed in Table S8 in the Supporting Information.
In general, the clean and selective conversion of HMF is par-
ticularly challenging due to the inherent reactivity of this mole-
cule, especially at higher reaction temperatures. Riisager and
co-workers previously showed that HMF could be reduced in
supercritical methanol by using copper-doped porous metal
oxides (PMOs), in the temperature range of 240–3008C.[23] The
hydrogen equivalents needed for the diverse reductions origi-
nated from the solvent itself, upon reforming, and no higher
boiling side products were detected. Although a good com-
bined yield of DMF+DMTHF was obtained, product selectivity
suffered from side processes due to reactive intermediates
formed through the concomitant methanol reforming process.
Herein, we report on a tunable and highly selective system
for the conversion of HMF to either FDM or DMF by using
both previously reported noble-metal-free and novel copper-
doped PMOs, comprising a very low ruthenium loading, in
contrast to many commercial catalysts with a high noble-metal
content. Biocompatible ethanol is used as a solvent, and
milder reaction temperatures (100–2208C) and 50 bar (1 bar=
1ꢁ105 Pa) of H2 pressure are applied. Similarly to the already
reported protocol in supercritical methanol, there is no evi-
dence for humin formation, even at HMF concentrations as
high as 10%. Identification of the reaction intermediates
formed during reduction processes and their evolution in time
is discussed at various reaction temperatures.
Table 1. Compositions and BET areas of the PMO catalysts used in this
study.
Catalyst code[a]
Composition
empirical[b]
BETarea
[m2 gÀ1
]
theoretical
Cu20-PMO
Cu0.60Mg2.40Al1.00
Cu0.59Mg2.34Al1.00
196
208
Cu20-Ru2-PMO Cu0.60Mg2.40Al0.98Ru0.02 Cu0.61Mg2.33Al0.98Ru0.02
[a] The code indicates the Cu or Ru mole percentage, respectively, with
respect to the M2+ or M3+ total content. [b] The numbers refer to molar
ratios of individual metals, as determined by inductively coupled plasma
(ICP) analysis. All metals are normalized relative to Al.
Tunable and selective conversion of HMF to FDM or DMF
First, we were interested in finding the optimal conditions for
selective carbonyl reduction of HMF to give FDM (Scheme 1).
Catalytic runs were performed by using HMF (0.5 g) and Cu20-
PMO (0.1 g) in ethanol under 50 bar of H2 in the temperature
range of 80–1408C for 3 h. The selectivity values for the main
components are shown in Table 2 (see Table S1 in the Support-
ing Information for all components).
Results and Discussion
Catalyst preparation and characterization
Two different PMO catalysts were prepared by calcination of
hydrotalcite (HTC) precursors. These HTCs were synthesized by
coprecipitation of Na2CO3 and NaOH with aqueous solutions of
aluminum, magnesium, and additional copper salts. In one
case, the molar ratio of M3+/M2+ was kept at 1:3, while
20 mol% of the Mg2+ ions were replaced with Cu2+. This com-
position was previously described and the corresponding PMO
was successfully used to convert various biomass resources.[24]
In another case, a new composition was prepared by the same
method, additionally replacing 2% of Al3+ with Ru3+ ions. Al-
though RuÀMgÀAl[25] and RuÀCuÀAl[26] compositions are
known, to the best of our knowledge, the corresponding RuÀ
Scheme 1. Selective conversion of HMF into FDM at mild reaction tempera-
tures.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2267

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Table 2. Hydrogenation of HMF at mild temperatures for the optimized
production of FDM.
Entry[a]
Catalyst
T
[8C]
FDM[b]
[%]
MFM[b]
[%]
MF[b]
[%]
DMF[b]
[%]
1[c]
2
3
4
5[d]
6
Cu20
Cu20
Cu20
Cu20
Cu20
Cu20-Ru2
80
100
120
140
100
100
91
>99
89
78
91
0.2
0.1
1
5
1
4
0.1
5
4
6
–
0.1
0.7
4
0.1
0.2
98
0.2
0.8
[a] Reaction conditions: HMF (0.5 g), catalyst (0.1 g), toluene (0.250 mL;
internal standard), ethanol (20 mL), 3 h, 50 bar H2. Full HMF conversion.
[b] Selectivity determined by GC–flame ionization detection (FID). [c] Con-
version=95%. [d] Reaction time=6 h.
Figure 1. Plot of DMF selectivity as a function of reaction temperature with
the Cu20-PMO catalyst (6 h runs).
Already at 808C, 95% of the starting material was converted
into FDM with 91% selectivity (Table 2, entry 1). Excellent re-
sults were achieved at 1008C, with full conversion of the start-
ing material and exceptionally high FDM selectivity (>99%;
Table 2, entry 2; see also Figure S2 in the Supporting Informa-
tion for the corresponding GC trace). At 120 and 1408C, FDM
selectivity decreased due to the formation of 5-methylfurfural
(MF) and 5-methyl-2-furanmethanol (MFM) intermediates
(Table 2, entries 3 and 4, respectively). These intermediates
were also seen in the run performed for a longer time (6 h) at
1008C, albeit to a much lesser extent (91% FDM; Table 2,
entry 5). Cu20-Ru2-PMO was also tested at 1008C for 3 h, and
gave rise to 98% FDM selectivity (Table 2, entry 6); a value very
similar to the one obtained with Cu20-PMO. In a separate ex-
periment pure FDM (0.489 g) was isolated in a yield of 97%
under the optimized reaction conditions (1008C, 50 bar H2,
3 h). For the corresponding NMR spectra, see Figures S3 and
S4 in the Supporting Information.
tion temperatures, a mixture of DMF and 2,5-dimethyltetrahy-
drofuran (DMTHF) was obtained, of which DMF was the pre-
vailing product. The yields of DMF and DMTHF were deter-
mined by using calibration curves and toluene as an internal
standard. The selectivity and yield values are in very good
agreement, which indicates that no higher boiling side prod-
ucts were formed and all components of the reaction mixtures
were detected by GC analysis.
After only 1 h, a 49% yield of DMF+DMTHF was obtained
with Cu20-PMO (Table 3, entry 1; see also Table S7 in the Sup-
porting Information for more details). The product yield gradu-
ally improved from 54% at 3 h to 66% at 6 h (Table 3, entries 2
Table 3. Dependence of DMF+DMTHF selectivity and yield on the reac-
tion time with the Cu20-PMO catalyst at 2208C.
Next, we turned our attention to the more challenging re-
ductive deoxygenation of HMF to demonstrate the utility of in-
expensive non-noble-metal-based Cu20-PMO catalyst, as well
as Cu20-Ru2-PMO catalyst, containing only small amount of
ruthenium, in the production of DMF, according to Scheme 2.
Catalytic runs were first conducted in the temperature range
140–2208C at 6 h with Cu20-PMO. The results are summarized
in Figure 1 (see Table S2 in the Supporting Information for
more details). In all cases, full HMF conversion was seen. In the
lower temperature range (100–1408C), FDM was the main reac-
tion product. At 1808C, more desired product was formed, but
intermediates FDM and MFM were still present in considerable
quantities. The best product selectivity was achieved at 2208C;
thus we chose this temperature for further studies. At all reac-
Entry[a]
t
[h]
DMF+DMTHF
DMF/DMTHF
ratio
selectivity[b] [%]
yield[c] [%]
1
2
3
4
5
1
3
6
52
54
69
67
69
49
54
66
65
65
3.3
5.8
3.6
3.2
3.1
12
18
[a] Reaction conditions: HMF (0.5 g), catalyst (0.1 g), toluene (0.250 mL;
internal standard), ethanol (20 mL), 2208C, 50 bar H2. Full HMF conversion
in all cases. [b] Determined by GC-FID. [c] Determined by calibration
curves and internal standard.
and 3). This already surpasses our previous results with the
same catalysts in supercritical methanol.[23] Prolonging the re-
action time to 12 and 18 h did not lead to any relevant
changes (Table 3, entries 4 and 5).
With these promising results in hand, we performed the
same series of reactions with the new Cu20-Ru2-PMO catalyst.
Ruthenium is highly active for CÀO bond hydrogenolysis;[16,18]
thus we expected a possible cooperativity effect between
copper and small amounts of ruthenium in the new catalyst
composition. We would like to point out that a low catalyst
Scheme 2. Production of DMF and DMTHF at 2208C.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2268

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
loading was maintained and the weight percentage of rutheni-
um in our new Cu20-Ru2-PMO was 0.6%. The ruthenium load-
ing with respect to the substrate is 0.12%. The results from
these runs are shown in Table 4 (see also Table S6 in the Sup-
porting Information for more details).
(Table 5, entry 5). Slight irregularities in yields were seen at 3 h
with Cu20-PMO and at 6 h with Cu20-Ru2-PMO, even after re-
peating these experiments several times. We tentatively attri-
bute this to adsorption phenomena, although further detailed
studies are required.
In all cases, the formation of DMF was accompanied by the
corresponding over-reduction product, DMTHF. The DMF/
DMTHF ratio remained relatively constant at all reaction times
when Cu20-PMO was used; however, a slight discrepancy
could be seen at 3 h. This is the same data point that corrobo-
rates the decrease in the overall DMF+DMTHF yield. The
DMF/DMTHF ratio gradually, but only slightly, decreased from
5.9 at 1 h to 3.3 at 18 h with Cu20-Ru2-PMO. This indicates
that DMF is not the main source of DMTHF in our system (see
also the section on Reaction pathways and intermediates). The
individual selectivity values of DMF and DMTHF for runs with
Cu20-Ru2-PMO are shown in Figure 3 (for the corresponding
values obtained with the Cu20-PMO catalyst, see Figure S5 in
the Supporting Information).
Table 4. Dependence of DMF+DMTHF selectivity and yield on the reac-
tion time with the Cu20-Ru2-PMO catalyst at 2208C.
Entry[a]
t
[h]
DMF+DMTHF
DMF/DMTHF
ratio
selectivity[b] [%]
yield[c] [%]
1
2
3
4
5
1
3
6
12
18
55
68
71
79
77
55
66
70
79
75
5.9
4.2
3.4
3.4
3.3
[a] Reaction conditions: HMF (0.5 g), catalyst (0.1 g), toluene (0.250 mL;
internal standard), ethanol (20 mL), 2208C, 50 bar H2. Full HMF conversion
in all cases. [b] Determined by GC-FID. [c] Determined by calibration
curves and internal standard.
Because the nature of the solvent may have a crucial influ-
ence on the studied reaction,[18] we evaluated different alco-
hols, including methanol, isopropanol, and MIBC, as reaction
solvents with Cu20-Ru2-PMO and Cu20-PMO (Table 5; for de-
tailed information, see Table S3 in the Supporting Information).
Interestingly, significant solvent effects were seen with Cu20-
PMO, for which the product yield improved from 66% in etha-
nol to 77% in isopropanol (Table 5, entries 1 and 2, respective-
ly). Because, during these runs, 0.250 g of PMO was used, this
result also shows that the catalyst loading did not affect the
overall product yield greatly, although the DMF/DMTHF ratio
slightly changed (Table 3, entry 3 versus Table 5, entry 1).
Higher HMF concentrations were adopted to prove the val-
idity of our catalytic system in conditions usually affording
HMF self-condensation products. Increasing the HMF concen-
tration to 8 wt% (Table 5, entry 3) still led to a good product
yield of 60%. This reaction time was not enough to fully con-
vert all reaction intermediates (Table S3, entry 3, in the Sup-
porting Information). Next, we adopted a 10 wt% HMF con-
centration while prolonging the reaction time up to 18 h. To
our delight, a colorless solution, accounting for an excellent
75% yield of DMF+DMTHF and good DMF/DMTHF ratio
(Table 5, entry 4), was obtained. For both of these runs, a cata-
lyst loading of 0.250 g was maintained. To the best of our
knowledge, only Dumesic and co-workers successfully reported
DMF production from 10 wt% HMF, albeit in the vapor
phase.[16] It is noteworthy that no char formation was detected,
as evidenced by the clear and colorless solution obtained
upon dissolving the catalyst residue in HNO3 (Figure S14 in the
Supporting Information). This was additionally confirmed by
thermogravimetric analysis (TGA) of the solids (Figure S15 in
the Supporting Information).
A comparison of product yields obtained with Cu20-PMO
and Cu20-Ru2-PMO is shown in Figure 2. Indeed, the novel cat-
alyst composition is very efficient in HMF reduction/deoxyge-
nation. A product yield of DMF+DMTHF as high as 55% at
1 h, 66% at 3 h, and 70% at 6 h was reached with the Cu20-
Ru2-PMO catalyst (Table 4, entries 1, 2, and 3, respectively). The
DMF+DMTHF yield further increased to 79% at 12 h; a notable
improvement compared with the 66% yield obtained with
Cu20-PMO.
In general, the product mixtures were very clean, and almost
all intermediates that might have led to DMF were consumed
after 6 h (Table S6 in the Supporting Information). After 18 h,
the product yield only slightly decreased to 75%. The yield of
DMF+DMTHF could be further improved to 80% after only
6 h by increasing the Cu20-Ru2-PMO loading to 0.250 g
Next, we screened the Cu20-Ru2-PMO composition in differ-
ent solvents at higher catalyst loading. In contrast to Cu20-
PMO, no further improvement took place when using Cu20-
Ru2-PMO in isopropanol, for which an almost identical, high
yield of 79% was obtained compared with 80% in ethanol
(Table 5, entry 6 versus entry 5). Because no furan ring-contain-
Figure 2. Plot of DMF+DMTHF yield with Cu20-PMO and Cu20-Ru2-PMO at
different reaction times (dashed lines are added for visual interpretation of
data).
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2269

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
mined due to the overlap of the
internal standard with the sol-
vent peak), but the higher boil-
ing point difference between the
solvent (131.68C) and DMF (92–
948C) allowed an attempt to
separate the product by distilla-
tion. For this experiment, HMF
Table 5. Catalyst loading and solvent dependence in DMF+DMTHF production.
Entry[a]
Catalyst
Solvent
t
[h]
DMF+DMTHF
selectivity[b] [%]
DMF/DMTHF
ratio
yield[c] [%]
1
2
3[d]
4[e]
5
6
7
8
Cu20
Cu20
Cu20
EtOH
iPrOH
iPrOH
iPrOH
EtOH
iPrOH
iPrOH
iPrOH
MeOH
MIBC[h]
6
6
6
18
6
6
4
1
6
6
72
81
69
80
84
81
84
84
18
70
66
77
60
75
80
79
78
81
17
–
5.5
3.8
5.9
7.0
3.4
3.5
3.4
3.4
17.0
3.3
Cu20
Cu20-Ru2
Cu20-Ru2
Cu20Ru2
Cu20-Ru2
Cu20-Ru2
Cu20-Ru2
(1.0 g)
and
Cu20-Ru2-PMO
(0.4 g) were used, and indeed
DMF (0.195 g) could be isolated
in 85% GC purity (for 1H and
13C NMR spectra, see Figures S7
and S8 in the Supporting Infor-
mation). This procedure demon-
strates that MIBC could be, after
optimization, the solvent of
choice for easy separation of
pure DMF from the reaction mix-
ture.
9[f]
10[g]
[a] Experimental conditions: HMF (0.500 g), catalyst (0.250 g), solvent (20 mL), toluene (0.250 mL; internal stan-
dard), 50 bar H2, 2208C. [b] Determined by GC-FID. [c] Determined based on calibration curves and internal
standard. [d] 8 wt% HMF (1.258 g) was adopted. [e] 10 wt% HMF (1.572 g) was adopted and reaction time was
18 h. [f] Reaction was carried out with 0.100 g of catalyst. [g] The yield could not be determined due to overlap
of the solvent with the internal standard. [h] MIBC=methyl isobutyl carbinol.
Furthermore, the catalyst stability at 2208C was tested. As
a representative example, the Cu20-Ru2-PMO catalyst was re-
covered after the 4 h run in iPrOH (Table 5, entry 7). Elemental
analysis showed very good agreement between the fresh cata-
lyst and spent catalyst residue. In addition, trace-metal analysis
of the product solution confirmed no metal leaching (see
Table S4 in the Supporting Information for more details).
Reaction pathways and intermediates
The reaction of HMF with H2 may generate a mixture of several
partially reduced products, such as FDM, MF, and MFM, and
products of ring hydrogenation, such as THFDM and 5-methyl-
2-tetrahydrofuranmethanol (THMFM). Ring-opening products
2-hexanol, 1,2-HD, and 1,2,6-HT are also expected under the re-
action conditions used. We have identified the vast majority of
these reaction intermediates by GC-MS analysis and authentic
standards. The main products and pathways are summarized
in Scheme 3 (see also Scheme S2 in the Supporting Informa-
tion).
Figure 3. Individual DMF and DMTHF selectivities plotted separately versus
reaction time (2208C, 50 bar H2, EtOH).
ing reaction intermediates, by only ring-opening products,
were present as minor side products in these runs, we first
shortened the reaction time to 4 h and subsequently to 1 h to
enhance the product yields; however, this did not lead to sig-
nificant changes (Table 5, entries 7 and 8). Nonetheless, under
these conditions, the reaction afforded the desired product in
excellent yield (81%) and selectivity (84%) after only 1 h (the
corresponding GC trace is shown in Figure S6 in the Support-
ing Information).
We first studied the composition of reaction mixtures as
a function of time by using Cu20-PMO at 1408C to gain an in-
sight into the evolution of earlier, partially reduced reaction in-
termediates (Figure 4; see also Table S5 in the Supporting In-
formation). At 1408C, most of HMF is converted after 1 h and
the main process is carbonyl reduction to form FDM. FDM rep-
resents the main component of all reaction mixtures up to
12 h, whereupon its amount gradually decreases, giving rise to
the second main intermediate, MFM. This intermediate, upon
hydrogenolysis, is converted into the desired product, DMF.
The amount of MF stays consistently low at all reaction times,
which indicates that carbonyl hydrogenation is more rapid
than the loss of the alcohol functionality in HMF. These data
also indicate that the loss of one ÀOH from FDM is more rapid
than the loss of the second ÀOH from MFM, which is a more
persistent intermediate.
As expected, the reaction carried out in methanol gave rise
to significant amounts of ethers and only 17% yield of DMF+
DMTHF (Table 5, entry 9). Among these ethers, several, namely,
2-(methoxymethyl)-5-methylfuran; 2,5-bis(methoxymethyl)fur-
an; and [5-(methoxymethyl)furan-2-yl]methanol (A, B, and C,
respectively, shown in Scheme S1 in the Supporting Informa-
tion), were identified by mass and their characteristic fragmen-
tation pattern. Together they accounted for 56% GC selectivity.
The reaction performed in MIBC (Table 5, entry 10) showed
lower DMF+DMTHF selectivity (the yield could not be deter-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2270

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Scheme 3. Reaction pathways for HMF conversion with copper-doped PMO
catalysts.
Figure 5. Composition of the mixture at 2208C with the Cu20-Ru2-PMO cat-
alyst.
and S10 and Table S6 in the Supporting Information) at 2208C;
the temperature most ideal for DMF formation.
At 2208C with Cu20-Ru2-PMO, the amount of partially re-
duced reaction intermediates is much lower than that of the
runs at 1408C, and already after 1 h a significant amount of
product DMF+DMTHF is formed. One of the reaction inter-
mediates is MFM, which accounts for 14% selectivity, whereas
FDM is only present in negligible quantities due to its higher
reactivity under these conditions. Ring-opening product 1,2-
HD is also present, already after 1 h, in 6% selectivity. In addi-
tion, two components belonging to retention times of 7.57
and 7.85 min can be detected, which together account for 5%.
Their abundance shows a decreasing trend over time (2% after
6 h and <1% after 18 h; see Table S6 in the Supporting Infor-
mation). From all relatively abundant intermediates, these are
the only ones we have not yet convincingly identified. Howev-
er, they could be attributed to products formed from MFM be-
cause they are also present in significant quantities (ꢀ10%)
when MFM is used as substrate (see Table 6, entry 2). All other
reaction intermediates (included in ‘other compounds’) were
present in less than 1% selectivity.
Figure 4. Composition of the reaction mixture at 1408C with the Cu20-PMO
catalyst.
The relative amount of all other reaction intermediates
(products of ring opening or furan ring hydrogenation) stays
below 5%. However, their presence indicates that these com-
peting pathways cannot be completely prevented and repre-
sents the main bottleneck towards exclusive product forma-
tion.
Further studies were conducted by using both Cu20-PMO
(see Figures S11 and S12 and Table S7 in the Supporting Infor-
mation) and Cu20-Ru2-PMO catalysts (Figure 5 and Figures S9
At 3 and 6 h reaction time, remaining MFM is gradually con-
sumed, giving rise to 68 and 71% DMF+DMTHF selectivity, re-
spectively. We attribute the further increase in product selectiv-
Table 6. Conversion of key reaction intermediates with Cu20-PMO at 2208C.
Entry[a]
Substrate
Conversion[b]
[%]
Intermediates formed [%]
DMF+DMTHF[b]
THFDM[b]
MFM[b]
MF[b]
THMFM[b]
1,2-HD[b]
1,2,6-HT[b]
5
EMMF[b]
others[b,e]
1
2
3[c]
4
FDM
MFM
MFM
MF
DMF
DMF
THFDM
>99
61
96
>99
2
50 (40+10)
38 (33+5)
79 (60+19)
33 (31+2)
99 (98+1)
99 (97+2)
–
4
–
–
–
–
–
–
13
39
4
44
–
–
<1
–
5
1
2
1
–
–
–
14
2
<1
3
<1
1
3
1
–
10 (8)
18 (7)
12 (7)
17 (7)
1
–
–
–
–
–
<1
–
5
–
6[d]
7
3
–
–
–
–
–
–
–
1
–
–
–
[a] Reaction conditions: substrate (4 mmol), catalyst (0.100 g), toluene (0.250 mL; internal standard), ethanol (20 mL), 2208C, 3 h, 50 bar H2. [b] Determined
by GC-FID. [c] The experiment was run for 12 h. [d] Cu20-Ru2-PMO was used as catalyst. [e] Numbers in parentheses indicate selectivity for other com-
pounds, excluding the two main components at 7.58 and 7.86 min.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2271

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
ity (79%) seen at 12 h to desorption of excess material from
the catalyst surface (notably, the 6 h data point gave slightly
decreased product yields). Indeed, after 6 h, all intermediates
that might lead to the product are already consumed. The
slight decrease in product selectivity (from 79 to 77%) at 18 h
could be due to further, slow conversion of DMF (see also
Table 6, entry 6).
pound can be also seen in the HMF reactions at longer reac-
tion times with Cu20-PMO.
MF was highly reactive and was fully converted to its reduc-
tion product MFM. This reaction gave very similar results to
the reaction in which MFM was used as a substrate. This
shows that C=O reduction is much faster than the hydrogenol-
ysis of MFM to DMF and is in agreement with data obtained at
1408C.
Lastly, the reactivity of product DMF was investigated both
with Cu20-PMO and Cu20-Ru2-PMO; the former resulted in
only 2% and the latter in 3% conversion after 3 h. The corre-
sponding over-reduction product, DMTHF, was detected in
small quantities, which was in agreement with the slight
change in the ratio between these two components towards
longer reaction times. However, DMF is considerably more
stable under these reaction conditions than the partially re-
duced intermediates FDM or MFM. Furthermore, no conversion
of THFDM was seen after 3 h, which was in agreement with
previous data.
In conclusion, the main pathway identified in this study is
the conversion of MFM to the DMF+DMTHF product within
6 h at 2208C. Already in the first hour, a significant amount of
desired product is formed and its amount remains relatively
constant over time. Similarly, other fully hydrogenated inter-
mediates, such as THFDM, THMFM, and ring-opening product
alcohols, formed already after 1 h and are constant over time,
with variation in the relative amounts below 5% (for a full de-
scription of the reaction intermediates and graphs, see
Table S6, Figures S9 and S10, and Scheme S2 in the Supporting
Information).
Very similar product profiles were obtained with Cu20-PMO,
albeit with an overall lower DMF+DMTHF selectivity. The
product selectivity reaches its maximum (69%) at 6 h, with no
significant further changes with time (for a full description of
the reaction intermediates and graphs, see Table S7 and Figur-
es S11 and S12 in the Supporting Information).
Preliminary kinetic modeling[28] (of runs in ethanol) suggests
that the main role of Cu20-Ru2-PMO lies in increasing the rate
of reduction of the main reaction intermediate, MFM, to DMF.
Because reaction intermediate MFM is a source of ring-opening
products and engages in side reactions with ethanol, more
rapid reduction of MFM to DMF with Cu20-Ru2-PMO explains
the cleaner reaction mixtures and higher product yields.
To gain further insight into the formation pathways of the
ring-opening and furan-ring hydrogenation products, we con-
ducted reactions at 3 h with the partially reduced reaction in-
termediates, FDM, MFM, MF, and THFDM, and product DMF.
The results are summarized in Table 6 and Scheme S3 in the
Supporting Information.
Overall, the above-described results indicate that ring-open-
ing products and DMTHF are likely to originate from partially
reduced, furanic intermediates and are not from the product
DMF or ring hydrogenation products THFDM and THMFM.
Ring opening of HMF and other highly functionalized furan de-
rivatives was also observed by Dumesic and co-workers.[10]
It appears that substrates with relatively high amounts of
polar ÀOH groups, such as FDM, have more affinity to the cata-
lyst surface.
Recyclability tests
Catalyst recycling experiments were performed at 1008C with
Cu20-PMO (0.1 g) and HMF (0.5 g) for 3 h of reaction time. The
catalyst was recovered at the end of each run, washed with
ethanol and THF, and reused after drying overnight at 1008C.
The catalyst activity was maintained for 3 cycles, then the FDM
selectivity gradually decreased to 61, 53, 38, and 19% in the
4th to 7th cycles, respectively (Figure 6). After the 7th cycle,
the HMF conversion was only 20%. After this run, 72 mg of
catalyst was recovered and subsequently calcined at 4608C for
24 h. After calcination, the catalyst residue (63 mg) was used in
a further experiment at 1008C for 3 h, resulting in 93% HMF
conversion and 92% FDM selectivity. Thus, upon recalcination,
the catalyst regained its activity.
Surprisingly, some of these intermediates resulted in the for-
mation of considerably more ring-opening products than HMF
itself. For example, after 3 h, all FDM was consumed and yield-
ed 50% DMF+DMTHF and 13% MFM, whereas the amount of
various ring-opening alcohol products combined was as high
as 23% (Table 6, entry 1). This result indicates that FDM is the
main source of these side products in this system.
Reactions with MFM revealed that this substrate was con-
verted at a slower rate (61% conversion), resulting in a lower
quantity of DMF+DMTHF, but also less ring-opening products
after 3 h compared with FDM. The selectivity to the unidenti-
fied products is 18%. Among these, 11% can be attributed to
the same two major components at 7.58 and 7.86 min as
those previously seen during HMF conversion with Cu20-Ru2-
PMO.
We next set up a gram-scale continuous-flow experiment for
the production of FDM by using similar experimental condi-
tions to those used in the corresponding batch reactions. The
catalyst remained active for 10 h without remarkable loss of
performance and FDM formed with very high selectivity (see
Figure S13 in the Supporting Information).
TEM measurements
Interestingly, at a prolonged reaction time (12 h, Table 6,
entry 3), most MFM was consumed and delivered higher prod-
uct selectivity at 12 h than HMF itself (Table 6, entry 6 versus
Table 3, entry 4). The ethyl ether of MFM, 3% of 2-(ethoxy-
methyl)-5-methylfuran (EMMF), was also detected. This com-
We carried out TEM measurements for both PMO catalysts
before and after reaction. Images before the reaction for Cu20-
PMO (Figure 7, right) and Cu20-Ru2-PMO (Figure 7, left)
showed irregular flakes, which is a morphology observed in re-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2272

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 8. TEM images of Cu20-Ru2-PMO (left) and Cu20-PMO catalysts (right)
after 6 h of reaction time (Table 4, entry 3, and Table 3, entry 3, respectively).
Scale bar, 50 nm.
Figure 6. Catalyst recycling experiments (HMF (0.5 g), Cu20-PMO (0.1 g),
50 bar H2, 1008C, 3 h).
the Cu20-PMO catalyst in supercritical methanol.[32] Further lit-
erature precedent attributes the formation of particles to sin-
tering of copper under reducing conditions.[33,4a] The influence
of this phenomenon on catalyst activity and stability should be
further investigated with different catalysts supports. No ap-
parent difference in morphology between the catalyst contain-
ing or lacking ruthenium (Figure 8) could be detected; this was
likely to be due to an insufficient amount of ruthenium
dopant.
Conclusions
We developed a tunable and highly selective method for the
conversion of HMF to either FDM or DMF (Scheme 4) by using
known and new HTC-based catalysts. These catalysts consisted
of inexpensive, earth-abundant starting materials and no or
a small amount of ruthenium as the hydrogenation metal, and
were an economically competitive alternative to commercial
noble-metal-based catalysts frequently used in these reactions.
Reactions carried out at 1008C with Cu20-PMO only led to car-
bonyl reduction with high selectivity (>99%) and the isolation
of FDM in 97% yield. The catalyst could be recycled and was
stable under continuous-flow conditions. Higher temperatures
(2208C) favored deoxygenation to DMF and DMTHF; the
former was the prevalent product. In isopropanol, a DMF+
DMTHF yield of 81% was obtained only after 1 h. In addition,
we were able to isolate DMF from a reaction mixture by using
MIBC under optimized reaction conditions. No catalyst leach-
ing was observed under these conditions. Finally, our method-
ology proved to be efficient in the clean conversion of HMF to
DMF, even at very high substrate concentrations (up to
10 wt%) and low catalyst loadings; this represents a distinct
advantage of the copper-doped PMOs in the clean conversion
of HMF.
Figure 7. TEM images of fresh Cu20-Ru2-PMO (left) and Cu20-PMO catalysts
(right). Scale bar, 50 nm.
lated systems, such as Cu/Cr HTC-like materials[29] and Cu/
MgAl2O4.[4a] No nanoparticles that could be CuO or RuO2 were
detected by TEM analysis, even at higher magnification. We
believe that highly dispersed copper and ruthenium species
are present in the PMO structures, as evidenced by the lack of
peaks related to these metals in the XRD pattern. Additionally,
parent Cu/Mg catalysts with a copper content comparable to
that of our systems (16 wt%) showed that very small aggre-
gates (<3 nm) could be formed;[30] the method of preparation
in our study (coprecipitation to form of regular HTC structure
additionally involving aluminum) was very likely to improve
copper dispersion to prevent the formation of visible crystalli-
tes of CuO. Ternary CuMgAl mixed oxides with a lower copper
content did not show any crystalline aggregation either.[31]
Upon imaging the Cu20-Ru2-PMO (Figure 8, left) and Cu20-
PMO (Figure 8, right) catalysts after 6 h of reaction time in
EtOH, both samples showed, in addition to the irregular
sheets, dark spherical particles with diameters of 10(Æ1) and
12(Æ2) nm, respectively. These
could be attributed to clusters of
Cu metal, formed by the reduc-
tion of CuO originally present in
the fresh catalyst. Similar behav-
ior was previously observed by
Ford and co-workers, who used Scheme 4. Tunable conversion of HMF to FDM and DMF+DMTHF.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2273

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Continuous-flow experiment for FDM production: The experiment
was performed first for 6 h in a 30 cm reactor tube charged with
Cu20-PMO catalyst (0.3 g) and fed with a solution of HMF (1.5 g in
Current efforts are focused on the development of new cata-
lyst compositions, with the aim of minimizing the ring-opening
processes, and one-pot procedures to allow the production of
DMF directly from fructose or glucose.
EtOH (0.4 L)) at
a
rate of 1.5 mLminÀ1 at 50 bar pressure,
30 mLminÀ1 H2 flow, and 1008C, as in the corresponding batch
conditions. The next day, additional solution of HMF (1.0 g in EtOH
(0.2 L)) was added to the reactor and the experiment was per-
formed for an additional 4 h under the same experimental condi-
tions as those previously described with no relevant changes to
the catalytic performance.
Experimental Section
Catalyst preparation: The HTC catalyst precursors were prepared
by
dures.[23,32] The new Cu20-Ru2-HTC composition was prepared ac-
cording to the following procedure: solution containing
a coprecipitation method, according to reported proce-
A
AlCl3·6H2O (11.83 g, 0.049 mol), Cu(NO3)2·2.5H2O (6.98 g, 0.03 mol),
MgCl2·6H2O (24.40 g, 0.12 mol), and RuCl3·H2O (0.22 g, 0.001 mol)
in deionized water (0.2 L) was added to a solution containing
Na2CO3 (5.30 g, 0.05 mol) in water (0.3 L) at 608C under vigorous
stirring. The pH was maintained between 9 and 10 by addition of
small portions of a 1m solution of NaOH. The mixture was vigo-
rously stirred at 608C for 72 h. After cooling to room temperature,
the dark-green solid was filtered and resuspended in a 2m solution
of Na2CO3 (0.3 L) and stirred overnight at 408C. The catalyst precur-
sor was filtered and washed with deionized water until it was chlo-
ride-free. After drying the solid under vacuum for 6 h at 1008C,
Cu20-Ru2-HTC (15.8 g) was obtained. Cu20-Ru2-HTC (8.75 g) of this
material was calcined at 4608C for 24 h in air to give Cu20-Ru2-
PMO (4.98 g).
Acknowledgements
We thank the Indonesian Directorate General of Higher Education
(DIKTI) for financial sponsorship. We gratefully acknowledge
Chris Bernt of the Center for Sustainable Use of Renewable Feed-
stocks (NSF CHE-1240194) and the University of California, Santa
Barbara (UCSB), for the precious support in the kinetic modeling.
We also thank Arjan Kloekhorst, Erwin Wilbers, Marcel de Vries,
Anne Appeldoorn, Jan Henk Marsman, Leon Rohrbach, and
Maarten Vervoort for their technical support.
Keywords: biofuels
· biomass · copper · heterocycles ·
hydrogenation
Catalytic tests: A stainless-steel autoclave (100 mL) equipped with
a mechanical stirrer was charged with catalyst (0.100 g), HMF
(0.500 g), and toluene (0.250 mL; internal standard) in the appro-
priate solvent (20 mL). The reactor was sealed and pressurized with
H2 (50 bar), heated to the desired temperature, and stirred at
800 rpm. After the reaction, the autoclave was cooled to room
temperature, the content was transferred to a centrifuge tube, cen-
trifuged, and the reaction mixture was separated from the solids
by decantation. Samples of the filtered solutions were injected into
a Hewlett Packard 5890 GC-MS-FID with a Restek RTX-1701 capilla-
ry column. Selectivity was defined as the ratio of the peak area of
a given compound to the total area of all the compounds pro-
duced in the reaction. Yield values were calculated based on the
use of internal standard and calibration curves.
[1] a) Top Value-Added Chemicals from Biomass, Vol. I—Results of Screening
for Potential Candidates from Sugars and Synthesis Gas (Eds.: T. Werpy,
G. Petersen), U.S. Department of Energy (DOE) by the National Renewa-
ble Energy Laboratory a DOE national Laboratory, Oak Ridge, TN 2004
http://www.eere.energy.gov/biomass/pdfs/35523.pdf,
Bozell, G. R. Petersen, Green Chem. 2010, 12, 539–554.
2004;
b) J. J.
[2] a) 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; b) A. A. Rosatel-
la, S. P. Simeonov, R. F. M. Fradea, C. A. M. Afonso, Green Chem. 2011, 13,
754–793.
[3] a) B. Saha, M. M. Abu-Omar, Green Chem. 2014, 16, 24–38; b) V. V. Or-
domsky, J. van der Schaaf, J. C. Schouten, T. A. Nijhuis, ChemSusChem
2013, 6, 1697–1707; c) V. V. Ordomsky, J. van der Schaaf, J. C. Schouten,
T. A. Nijhuis, ChemSusChem 2012, 5, 1812–1819; d) T. Okano, K. Qiao, Q.
Bao, D. Tomida, H. Hagiwara, C. Yokoyama, Appl. Catal. A 2013, 451, 1–
5.
For the experiment aimed to DMF distillation, HMF (1.0 g,
0.008 mol), Cu20-Ru2-PMO (0.4 g), and MIBC (20 mL) were placed
in a Parr reactor, pressurized with 50 bar H2, and heated at 2208C
for 4 h. After cooling, the crude mixture was quantitatively trans-
ferred to a centrifuge tube. The residual solid was washed with ad-
ditional MIBC (5 mL) and the solutions were combined. Distillation
of DMF was performed by using a 50 mL round-bottomed flask
equipped with a cooler and Vigreaux condenser. Two colorless
fractions were collected at 1008C (0.072 g and 0.123 mg, respec-
tively). After confirming that the composition of both fractions was
identical, these were combined. GC measurements were in accord-
ance with the retention time determined for DMF with known
[4] a) K. Pupovac, R. Palkovits, ChemSusChem 2013, 6, 2103–2110; b) D. M.
Alonso, S. G. Wettstein, J. A. Dumesic, Chem. Soc. Rev. 2012, 41, 8075–
8098; c) BASF Se (Ludwigshafen, Germany), US20130345450, 2013;
Dalian Institute of Chemical Physics, Chinese Academy of Sciences
(Dalian, Liaoning Province, P.R. China), US20130281719, 2013;
Du Pont De Nemours and Co. (Wilmington, DE, USA), US20130172580,
2013.
E I
[5] a) N.-T. Le, P. Lakshmanan, K. Cho, Y. Han, H. Kim, Appl. Catal. A 2013,
464–465, 305–312; b) T. S. Hansen, I. Sadaba, E. J. Garcia-Suarez, A. Riis-
ager, Appl. Catal. A 2013, 456, 44–50; c) A. Takagaki, M. Takahashi, S.
Nishimura, K. Ebitani, ACS Catal. 2011, 1, 1562–1565.
1
standards; H and 13C NMR spectroscopic analysis in CDCl3 (see Fig-
[6] a) H. Ait Rass, N. Essayem, M. Besson, Green Chem. 2013, 15, 2240–
2251; b) A. Villa, M. Schiavoni, S. Campisi, G. M. Veith, L. Prati, ChemSu-
sChem 2013, 6, 609–612; c) S. E. Davis, B. N. Zope, R. J. Davis, Green
Chem. 2012, 14, 143–147; d) O. Casanova, S. Iborra, A. Corma, ChemSu-
sChem 2009, 2, 1138–1144; e) T. B. Rauchfuss, T. Todsapon (University of
Illinois, USA), US8324409, 2011; f) C. R. Adams (Shell Oil Co.),
US3228966, 1963.
[7] Y. Nakagawa, M. Tamura, K. Tomishige, ACS Catal. 2013, 3, 2655–2668.
[8] a) N. R. Janga, H.-R. Kimb, C. T. Houc, B. Soo Kim, Polym. Adv. Technol.
2013, 24, 814–818; b) S. Goswami, S. Dey, S. Jana, Tetrahedron 2008, 64,
6358–6363.
ures S7 and S8 in the Supporting Information) was consistent with
known spectral data for DMF.
Catalyst recycling tests: For reusability tests, catalytic runs were
performed as described above. The catalyst was separated from
the reaction solution by centrifugation and subsequent decanta-
tion, additionally washed with ethanol (1ꢁ10 mL), then with THF
(1ꢁ10 mL), and dried overnight at 1008C prior to the next run.
After the 7th cycle, catalyst residue (72 mg) was recovered. This
was additionally calcined at 4608C for 24 h, resulting in a solid
(62 mg). This solid was reused in the eighth run.
[9] L. Cottier, G. Descotes, Y. Soro, Synth. Commun. 2003, 33, 4285–4295.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2274

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[10] R. Alamillo, M. Tucker, M. Chia, Y. Pagꢂn-Torres, J. Dumesic, Green Chem.
2012, 14, 1413–1419.
[11] Y. Nakagawa, K. Tomishige, Catal. Commun. 2010, 12, 154–156.
[12] M. Tamura, K. Tokonami, Y. Nakagawa, K. Tomishige, Chem. Commun.
2013, 49, 7034–7036.
[24] a) T. D. Matson, K. Barta, A. V. Iretskii, P. C. Ford, J. Am. Chem. Soc. 2011,
133, 14090–14097; b) K. Barta, T. D. Matson, M. L. Fettig, S. L. Scott, A. V.
Iretskii, P. C. Ford, Green Chem. 2010, 12, 1640–1647; c) K. Barta, G.
Warner, E. S. Beach, P. T. Anastas, Green Chem. 2014, 16, 191–196.
[25] a) K. Motokura, D. Nishimura, K. Mori, T. Mizugaki, K. Ebitani, K. Kaneda,
J. Am. Chem. Soc. 2004, 126, 5662–5663; b) S. K. Sharma, P. A. Parikh,
R. V. Jasra, J. Mol. Catal. A 2010, 317, 27–33; c) S. K. Sharma, K. B. Sidh-
puria, R. V. Jasra, J. Mol. Catal. A 2011, 335, 65–70.
[13] J. Ohyama, A. Esaki, Y. Yamamoto, S. Arai, A. Satsuma, RSC Adv. 2013, 3,
1033–1036.
[14] T. Thananatthanachon, T. Rauchfuss, ChemSusChem 2010, 3, 1139–1141.
[15] a) T. Buntara, I. Melian-Cabrera, Q. Tan, J. L. G. Fierro, M. Neurock, J. G.
de Vries, H. J. Heeres, Catal. Today 2013, 210, 106–116; b) T. Buntara, S.
Noel, P. Huat Phua, I. Meliꢂn-Cabrera, J. G. de Vries, H. J. Heeres, Top.
Catal. 2012, 55, 612–619; c) T. Buntara, S. Noel, P. Huat Phua, I. Meliꢂn-
Cabrera, J. G. de Vries, H. J. Heeres, Angew. Chem. Int. Ed. 2011, 50,
7083–7087; Angew. Chem. 2011, 123, 7221–7225.
[26] H. B. Friedrich, F. Khan, N. Singh, M. van Staden, Synlett 2001, 0869–
0871.
[27] F. Cavani, F. Trifirꢄ, A. Vaccari, Catal. Today 1991, 11, 173–301.
ˇ
[28] P. Kuzmic, Anal. Biochem. 1996, 237, 260–273.
[29] Q. Jiao, H. Liu, Y. Zhao, Z. Zhang, J. Mater. Sci. 2009, 44, 4422–4428.
[30] B. M. Nagaraja, V. Siva Kumar, V. Shashikala, A. H. Padmasri, S. Sreevard-
han Reddy, B. D. Raju, K. S. Rama Rao, J. Mol. Catal. A 2004, 223, 339–
345.
[16] Y. Romꢂn-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature 2007, 447,
982–986.
[17] M. Chidambaram, A. T. Bell, Green Chem. 2010, 12, 1253–1262.
[18] J. Jae, W. Zheng, R. F. Lobo, D. G. Vlachos, ChemSusChem 2013, 6, 1158–
1162.
[19] Y. Zu, P. Yang, J. Wang, X. Liu, J. Ren, G. Lu, Y. Wang, Appl. Catal. B 2014,
146, 244–248.
[31] M. Dixit, M. Mishra, P. A. Joshi, D. O. Shah, J. Ind. Eng. Chem. 2013, 19,
458–468.
[32] G. S. Macala, T. D. Matson, C. L. Johnson, R. S. Lewis, A. V. Iretskii, P. C.
Ford, ChemSusChem 2009, 2, 215–217.
[33] a) P. L Gai, E. D Boyes, Electron Microscopy in Heterogeneous Catalysis,
CRC Press, Boca Raton, 2003, p. 180; b) A. Martꢅnez-Arias, A. B. Hungrꢅa,
G. Munuera, D. Gamarra, Appl. Catal. B 2006, 65, 207–216.
[20] M. Chatterjee, T. Ishizaka, H. Kawanami, Green Chem. 2014, 16, 1543–
1551.
[21] D. Scholz, C. Aellig, I. Hermans, ChemSusChem 2014, 7, 268–275.
[22] G.-H. Wang, J. Hilgert, F. H. Richter, F. Wang, H.-J. Bongard, B. Spliethoff,
C. Weidenthaler, F. Schꢃth, Nat. Mater. 2014, 13, 293–300.
[23] T. S. Hansen, K. Barta, P. T. Anastas, P. C. Ford, A. Riisager, Green Chem.
2012, 14, 2457–2461.
Received: February 22, 2014
Revised: March 25, 2014
Published online on June 12, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2266 – 2275 2275