Accelerated decarbonylation of 5-hydroxymethylfurfural in compressed carbon dioxide: A facile approach

Article abstract of DOI:10.1039/c8gc00174j

Herein, decarbonylation of biomass-based 5-hydroxymethylfurfural (HMF) in compressed CO₂ with an unexpected acceleration of the reaction rate and excellent catalytic activity is reported. Without any additive, CO surrogates, or any organic solvents, via the developed method, an excellent conversion of 99.8% and highest selectivity of furfuryl alcohol (99.6%) in 4 h at 145 °C were achieved using an alumina-supported Pd catalyst (Pd/Al₂O₃). The superior activity is due to the unique characteristics (miscibility of reactant gases and high diffusivity) of compressed CO₂ and the synergy between CO₂ and Pd/Al₂O₃, where CO₂ plays an interesting role in accelerating the reaction by enhancing the diffusion of CO and furfuryl alcohol (both products have high solubilities in CO₂), consequently shifting the equilibrium to the forward direction. Characterisation of the catalyst suggested its direct interaction with the substrate and provided an indication of the possible reaction path. Thus, a mechanism was outlined. Compared to the results obtained using organic solvents, the results obtained using compressed CO₂ were superior in terms of activity, selectivity, and reaction rate. This strategy highlights easy product separation, improved catalyst life, and a simple sustainable process. The efficiency of this protocol is confirmed by its potential application to a series of aldehydes with various substituents to produce decarbonylated products in good to excellent yields.

Full text of DOI:10.1039/c8gc00174j

Green Chemistry
View Article Online
View Journal
PAPER
Accelerated decarbonylation of
-hydroxymethylfurfural in compressed
carbon dioxide: a facile approach†
5
Cite this: DOI: 10.1039/c8gc00174j
a
a
a,b
Maya Chatterjee, * Takayuki Ishizaka and Hajime Kawanami
*
2
Herein, decarbonylation of biomass-based 5-hydroxymethylfurfural (HMF) in compressed CO with an
unexpected acceleration of the reaction rate and excellent catalytic activity is reported. Without any addi-
tive, CO surrogates, or any organic solvents, via the developed method, an excellent conversion of 99.8%
and highest selectivity of furfuryl alcohol (99.6%) in 4 h at 145 °C were achieved using an alumina-sup-
ported Pd catalyst (Pd/Al
tant gases and high diffusivity) of compressed CO
CO plays an interesting role in accelerating the reaction by enhancing the diffusion of CO and furfuryl
alcohol (both products have high solubilities in CO ), consequently shifting the equilibrium to the forward
2
O
3
). The superior activity is due to the unique characteristics (miscibility of reac-
2
and the synergy between CO and Pd/Al , where
2
2 3
O
2
2
direction. Characterisation of the catalyst suggested its direct interaction with the substrate and provided an
indication of the possible reaction path. Thus, a mechanism was outlined. Compared to the results obtained
Received 17th January 2018,
Accepted 10th April 2018
using organic solvents, the results obtained using compressed CO were superior in terms of activity, selecti-
2
vity, and reaction rate. This strategy highlights easy product separation, improved catalyst life, and a simple
DOI: 10.1039/c8gc00174j
sustainable process. The efficiency of this protocol is confirmed by its potential application to a series of
rsc.li/greenchem
aldehydes with various substituents to produce decarbonylated products in good to excellent yields.
1
–3
improved activity, and acceptor-less conditions
have been
Introduction
investigated. The drawbacks associated with the developed
Biomass is a vast renewable resource suitable as a potential re- catalytic systems are mainly related to the recovery of the
placement of fossil fuels for the development of a sustainable catalysts after the reaction, harsh reaction conditions, and
society. Naturally produced lignocellulosic biomass (a dry the requirement of high-boiling solvents, which make the de-
plant-based material) with an intricate polymeric structure carbonylation process incompatible with a sustainable reaction
consists of various oxygenated functionalities. One of the procedure.
targeted strategies for the implementation of bio-feedstock is
In this regard, heterogeneous catalysts would be of signifi-
upgradation of the raw materials through the reduction of cant interest as compared to homogeneous catalysts because
oxygen content using defunctionalisation processes, which they provide straightforward catalyst/product separation
include hydrogenation, dehydration, and hydrodeoxygenation; through filtration, thus making catalyst recovery easier;
in addition, decarbonylation is an effective transformation however, heterogeneous catalysts exhibit low reaction rates;
process; however, it is currently less studied for biomass- hence, the selective decarbonylation of different aromatic
derived compounds. As a major step towards the development aldehydes using heterogeneous catalysts in the gas phase faces
of effective catalysts, metal-based homogeneous systems the difficulties of tedious reaction conditions, presence of
4
with different ligands, the use of CO scavengers for additives, catalytic deactivation, and low product yield. Thus,
the development of suitable decarbonylation process that
perform under milder reaction conditions in different organic
a
solvents using metal nanoparticles or supported metal nano-
particles as catalysts was continued; however, limitations
such as the requirement of metal additives, CO scavengers, use
of large excess of organic solvents, longer reaction times, and
product separation hindered these developments.
Microflow Chemistry Group, Research Institute for Chemical Process Technology,
5
AIST Tohoku, 4-2-1, Nigatake, Miyagino-ku, Sendai, 983-8551, Japan.
E-mail: c-maya@aist.go.jp, h-kawanami@aist.go.jp; Fax: +81 22 237 5388;
Tel: +81 22 237 5213
b
CREST, Japan Science and Technology (JST), 4-1-8, Honcho, Kawaguchi, Saitama,
3
32-0012, Japan
Furan-based compounds have the potential to generate a
variety of products and may thus be implemented in the bio-
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c8gc00174j
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
refinery concept; therefore, several strategies have been developed cesses for a sustainable extension to the organic synthesis.
for their transformation of bio-feedstock into speciality chemi- Geilen et al. first reported the decarbonylation of HMF using a
cals. As a multifunctional compound, 5-hydroxymethylfurfural homogeneous Ir catalyst ([IrCl (cod)]
2
) with different phos-
(
HMF) is beneficial for the production of a wide spectrum of fuel phine ligands in 1,4-dioxane with 5 MPa of compressed CO .
2
or non-fuel materials through hydrogenation, oxidation, hydro- Although HMF was decarbonylated to furfuryl alcohol with
genolysis of the C–O bond, dehydrogenation, decarbonylation, 95% selectivity, the reaction required dioxane as a solvent, a
6
rearrangement, dissociation, and polymerisation.
very high temperature (220 °C), and a long reaction time
1
1
Decarbonylation of the –CHO group requires abstraction of (12 h).
the aldehydic hydrogen, which is an endothermic reaction;
In our previous study, we have successfully developed a
thus, elevated temperatures are essential for this process. In strategy to dehydrogenate alcohol over a heterogeneous Rh
this context, one of the main challenges is to avoid humin for- catalyst in compressed CO using HMF as a model com-
2
1
2
mation (soluble or insoluble polymeric species), which gener- pound. During dehydrogenation, decarbonylation of the
ally occurs at high temperatures and hampers the catalytic aldehyde group was detected as a competing reaction that pro-
7
activity. In addition, the presence of highly active multi-func- duced furfuryl alcohol depending on the reaction conditions.
tional groups results in different side reactions and conse- While screening different metal catalysts for dehydrogenation,
quently decreases the selectivity of the decarbonylated we observed that Pd had the ability to mediate both reactions
product. Moreover, the adsorption and desorption of CO (dehydrogenation of alcohols and decarbonylation of alde-
causes deactivation of the metal surface. Rauchfuss and co- hyde) in one transformation, but the tuning of reaction para-
workers employed an open system using the Pd/C catalyst at meters could direct the reaction in a desired direction.
comparatively lower temperatures (120 °C), but the presence of
Herein, we attempted the decarbonylation of HMF using a
catalyst, focusing on the use
to play the dual role of a solvent and a
oxygen was detrimental to the reduced catalysts as well as to commercially available Pd/Al
2 3
O
8
the highly active HMF. A heterogeneous Pd catalyst supported of only compressed CO
2
on mesoporous silica (SBA-15) provided promising results in surrogate to the generated gases, especially CO. Specifically,
9
cyclohexane, but necessitated the presence of a molecular the goal of this study was to ensure the suitability of com-
2
sieve (MS 4A) as a CO surrogate (a negative consequence was pressed CO in the development of a simple process through
the regeneration of MS 4A, which was a tedious process and the intelligent tuning of the advantageous physicochemical
changed the resin composition), required a prolonged reaction properties that could be helpful in building a compact system
time (12 h), and had restricted substrate scope and recyclabil- to achieve desired transformation under mild reaction con-
ity. To improve the yield, Meng et al. applied alkali metal- ditions without any additional components.
modified Pd catalysts; the decarbonylated product was
obtained at 180 °C with a yield of >90% in 16 h of reaction
1
0a
after purging N at an 8 h interval to prevent deactivation.
2
Results and discussion
The same group also employed 28% water along with 1,4-
dioxane, which improved the yield of furfuryl alcohol (95.3%)
under the same conditions as reported in their previous study,
but under a hydrogen atmosphere, i.e. fresh hydrogen was
flushed after 8 h; hydrogen was used as a carrier gas to dis-
tinguish between hydrogenation and decarbonylation as the
furfural moiety could be readily hydrogenated on a Pd catalyst,
which was not reported in the study. Despite these advances,
the decarbonylation reaction conducted using heterogeneous
catalysts still has limited success because of the requirement
of harsh reaction conditions, longer reaction times, as well as
the use of a CO acceptor.
As abovementioned, HMF can undergo different types of reac-
tions, which affect the selectivity of a particular product, and
eventually, a separation step is required to obtain the desired
compounds. However, the sensitivity of each process strongly
depends on the reaction conditions; hence, careful modulation
of the reaction parameters has an utmost importance in shift-
ing the reaction path towards the preferred direction.
Scheme 1 represents the general transformation routes of
1
0b
2
Compressed carbon dioxide (compressed CO ) is a promis-
ing alternative for rapid and selective organic synthesis owing
to its unique properties such as enhanced diffusion rates, easy
product separation, controlled mass transfer resistance, and
improvement of catalyst lifetime. However, reactions with
heterogeneous catalysts are still restricted mainly to hydrogen-
ation and oxidation because of the accelerated reaction rates,
different product distributions, as well as high selectivity
associated with the miscibility of the reactant gases (H and O ).
2
2
A combination of heterogeneous catalysts and compressed
CO can provide an immense opportunity to develop highly
efficient and environmentally and economically beneficial pro- Scheme 1 A general reaction pathway for the HMF transformation.
2
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
HMF, which can be converted to furfuryl alcohol and 2,5-difor- another route competes with decarbonylation. Thus, it is
mylfuran (DFF) via decarbonylation and dehydrogenation, reasonable to conduct the reaction in the absence of air to
respectively. In addition, DFF can be transformed to furfural avoid the formation of unwanted by-products.
again through decarbonylation followed by hydrogenation to
Fig. 1b shows the excellent pressure (density)-dependent
furfuryl alcohol. Furthermore, the dehydrogenation of furfuryl performance in the absence of air (an inert atmosphere). The
alcohol can produce furfural, which can be decarbonylated to conversion of HMF increased significantly even at a low
furan.
pressure of 4 MPa (79.9%). After reaching a maximum of >99%
at 6 MPa, the conversion remained almost constant until 8
MPa. Surprisingly, the conversion of HMF dropped substan-
tially to 54.6% when the pressure was increased to 16 MPa. As
Optimisation of different reaction parameters
CO pressure. In the first step of optimisation, selection of CO2 is a compressed gas, at a fixed temperature of 145 °C,
2
−
1
proper pressure is essential to tune the properties of com- there is a huge change in the density from 0.054 g ml
pressed CO . To elucidate the effect of CO pressure on the (vapour) to 0.259 g ml− (supercritical) (NIST chemistry web
1
2
2
catalytic performance of Pd/Al
2
O
3
, a series of reactions were book) with an increase in pressure (4 to 16 MPa) that sub-
performed with pressure varying from 4 to 16 MPa while main- sequently affects the solubility of the substrate. Thus, it is
taining a constant temperature (145 °C) and reaction time reasonable to understand the experimental environment
(4 h) in the presence (Fig. 1a) and in the absence of air (phase behaviour) inside the reactor that can provide an expla-
(
Fig. 1b). As shown in Fig. 1a, the reaction proceeded with a nation for the CO pressure-based activity of the present reac-
2
change in catalytic activity depending on the applied pressure; tion. Because of the technical constraints related to the reactor
a very low conversion of ∼9% was observed at 4 MPa, which used to perform the reaction, video monitoring of the phase
started to increase with the increasing CO2 pressure and behaviour was conducted in a separate setup. After replicating
reached 38% at 16 MPa. From the product distribution, it can the applied conditions, the video is expected to offer an actual
be seen that the selectivity of DFF is comparatively higher scenario of the number of phases present in the system during
(25% at 16 MPa); this suggests that dehydrogenation as the reaction. Fig. 2a–f show the images of the phase behaviour
Fig. 2 Images obtained during the phase observation of the HMF–CO
2
system in a view cell at 145 °C. (a) Empty, (b) HMF in cell at various CO
2
Fig. 1 Effect of CO
2 3 2
Pd/Al O catalyst; (a) presence of air and (b) N atmosphere. Reaction
conditions: Catalyst : substrate = 1 : 10; temperature = 145 °C, and reac-
tion time = 4 h.
2
pressure on the decarbonylation of HMF using the
pressure (c) 6 MPa, (d) 8 MPa, (e) 12 MPa, and (e) 16 MPa. For 12 and 16
MPa, the image obtained taken during stirring is presented, which
clearly shows the status of HMF. Dotted red line represents the liquid
meniscus.
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
2
obtained at various CO pressures. Fig. 2a and b represent the
images of the cell before and after the introduction of HMF,
respectively. Naked eye observation indicates a sharp change
in the system environment with a change in CO
Fig. 2c–f); at 6 MPa, the view cell is mainly filled with vapour,
and a liquid phase is evident below the stirrer bar (Fig. 2c).
Interestingly, as the CO pressure was increased beyond 8 MPa,
2
pressure
(
2
the liquid phase started to diminish (as can be seen from the
change in the meniscus position) (Fig. 2d and e), and finally, a
single phase was obtained at 16 MPa. By correlating the phase
behaviour with the catalytic activity, it was logical to suggest
that the adsorption of the substrate on the catalyst surface con-
1
3
trolled the reaction because the substrate was more concen-
trated around the catalyst when the CO pressure was low, and
2
the reaction progressed with an accelerated rate. On the other
hand, at higher pressures, the increased solubility of HMF in
compressed CO2 caused a decrease in the substrate concen-
tration near the catalyst surface and generated a HMF-enriched
CO
2
phase (Scheme 2); this could explain the reduced catalytic
performance related to a change in the mass transfer
1
4
properties.
Effect of temperature. As abovementioned, decarbonylation
of aldehyde is an endothermic reaction; hence, most of the
studies have revealed an exceptionally high temperature
regime for catalytic decarbonylation. By conducting the reac-
tion in compressed CO , the temperature can tune the density
2
and consequently the solubility, which in turn affects the
Fig. 3 Effect of temperature on the (a) catalytic activity and selectivity
phase behaviour. The conversion of HMF was investigated in and the rate of (b) decarbonylation (TOFFA
the temperature range from 80 °C to 150 °C at a fixed pressure (TOFDFF) over the Pd/Al catalyst. Reaction conditions: catalyst :
) and dehydrogenation
2 3
O
substrate = 1 : 10, PCO2 = 6 MPa, and reaction time = 4 h. TOF =
Turnover frequency (TOF) = number of moles reacted/moles of metal x
time (calculated on the basis of the lowest conversion of around 6%).
of 6 MPa (Fig. 3). Notably, the vapour state of CO
2
was main-
tained in the temperature range studied herein, and there was
−
1
a nominal change in the density from 0.110 g ml (80 °C) to
1
0
.0835 g ml− (145 °C) at the fixed pressure of 6 MPa. Thus, no
significant change was observed in the phase behaviour. The
results presented in Fig. 3a revealed an unexpectedly poor con- from 80 °C (selectivity ∼80%) to 100 °C (selectivity 61%)
version of HMF (6.1%) at 80 °C that increased monotonically (Fig. 3a), but decarbonylation dominated above 130 °C.
with temperature and plateaued at or above 145 °C. Almost
quantitative conversion of HMF (>99%) was achieved at corresponding to dehydrogenation (TOFDFF) and decarbonyla-
45 °C. There was a significant difference in the product tion (TOFFA) in terms of turnover frequency (TOF) at the con-
We evaluated the temperature-dependent reaction rate
1
selectivity depending on the temperature; as seen in Fig. 3a, version level of ∼6%. Fig. 3b shows a comparison between the
the product selectivity transferred from dehydrogenation (DFF) rate of formation of furfuryl alcohol (TOFFA) and that of DFF
to decarbonylation (FA) at 130 °C; the formation of DFF was (TOFDFF) with the applied temperature. At 80 °C, TOFFA
favoured at low temperatures, and DFF was the major product (1.6 h− ) was lower than TOFDFF (4.8 h ), but started to
1
−1
−
1
increase with temperature and reached a maximum of 92.0 h
at 145 °C. When the temperature was further increased to
1
50 °C, the TOF was slightly improved to 94.1 h− . On the
1
other hand, except those at 80 °C, all TOFDFF values were lower
than the TOFFA values (Fig. 3b) and further reduced at higher
temperatures (12.4 h− ; 145 °C). This suggests that the acti-
vation energies of the two paths leading to the formation of
furfuryl alcohol and DFF are different. Therefore, the obtained
results underline the importance of the applied temperature to
dictate the dominance of the kinetically or thermodynamically
1
1
5
controlled path. Hence, kinetics is a straightforward explica-
tion for the increase in TOFFA with temperature rather than the
correlation of the reaction rate with the phase change model,
Scheme 2 Schematic of the possible scenario inside the reactor under
a) low and (b) high pressure conditions.
(
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
as described in Fig. 2. Based on the results presented in that at the lowest conversion, furfuryl alcohol and DFF were
Fig. 3a and b, a temperature of 145 °C was used as the obtained with a yield of 3.9% and 1.4%, respectively. The yield
optimum temperature for HMF decarbonylation.
of furfuryl alcohol increased steadily with conversion (major
Reaction time versus catalytic performance. Fig. 4a presents path), but DFF disappeared after reaching a maximum (4.4%)
the time profile of the reaction (temperature of 145 °C and at 20% conversion (minor path of the HMF transformation).
P
CO2 of 6 MPa). With the variation of reaction time from Notably, we also detected furfural with a very low yield (0.6%)
1
5 min to 8 h, the activity of the catalyst changed (Fig. 4a). at 50% conversion that disappeared in the final product.
There was no detectable transformation right after the intro- According to Scheme 1, furfural can be obtained from the dec-
duction of the substrate (0 min). At the shortest time (15 min), arbonylation of DFF or through the dehydrogenation of fur-
the conversion of HMF was low (6.1%), but started to increase furyl alcohol. Herein, we can ignore the possibility of the
with time, which reached a maximum of >99% in 4 h and then direct formation of furfuryl alcohol from HMF by hydrogenoly-
1
7
remained almost constant. The change in conversion was also sis, which requires the presence of water and CO
associated with the alteration of the product selectivity; after trolled experiment, DFF was completely converted mainly to
5 min, the product mixture contained 76.7% furfuryl alcohol furfural (86.3%) (Table 2; entry 7). On the other hand, furfuryl
2
.
In a con-
1
and 23.3% DFF. No other products were observed in the detect- alcohol was converted to 36.9% furfural although with a very
able range. As the reaction progressed, the selectivity of fur- poor conversion (5.1%) (Table 2; entry 8). Based on the results
furyl alcohol increased to 99.6%, whereas the selectivity of DFF of the controlled experiments, DFF has emerged as a possible
decreased to 0.4%. This observation suggests that the trans- source of furfural that can be further hydrogenated to furfuryl
formation of HMF to furfuryl alcohol can occur (i) via direct alcohol using surface hydrogen, which has also been con-
decarbonylation and (ii) through DFF. The yield vs. conversion firmed by another experiment with trace amounts of hydrogen
curve provides further possibility to distinguish between the (Table 2; entry 9). Hence, the transformation of HMF to fur-
primary and secondary reactions. The curve with a linear furyl alcohol can occur through direct decarbonylation (major)
approach from the origin represented the primary product; on as well as via DFF (minor) under the present conditions.
the other hand, the product recognised by an increased fol-
Catalyst/substrate ratio. Like the case of other parameters
lowed by a decreased yield corresponded to the secondary reac- (temperature, pressure, etc.), the optimisation of the catalyst to
1
6
tion. The conversion-dependent yield plot (Fig. 4b) revealed substrate ratio under the studied reaction conditions is also
important. Investigations on the variation of catalyst : substrate
(
wt) ratio was carried out from 1 : 1 to 1 : 30 at a fixed reaction
time of 1 h (Fig. 5a). At a low ratio of 1 : 1, complete conversion
was observed, but it dropped significantly (4.9%) as the sub-
strate amount increased 30 times that of the catalyst. At the
−
1
lowest conversion (∼5%), the TOF is enhanced from 18 h
(
1 : 1) to 116 h− (1 : 30); thus, to achieve a high performance
1
after maximising the substrate concentration, appropriate reac-
tion conditions are necessary. In this context, increasing the
reaction time can also enhance the conversion. Fig. 5b pre-
sents the catalytic activity and product selectivity after 4 h of
reaction. The result shows that except for the case of the 1 : 1
system, in which the conversion dropped (70.2%) because of
the generation of black polymeric substances, the conversion
increased after extending the reaction time to 4 h. A highest
conversion of >99% was achieved using the 1 : 10 ratio,
whereas a vast improvement (40.2%) was observed for
maximum substrate concentration (1 : 30). A comparison of
the product distributions (Fig. 5a and b) indicated that at a
low substrate concentration (1 : 1), the selectivity of furfuryl
alcohol dropped from 93.7% (1 h) to 80.2% (4 h); this might
be attributed to the longer reaction time that led to the sub-
strate scarcity near the catalyst surface and triggered other side
reactions. On the other hand, although a higher ratio (1 : 30)
showed improved selectivity of furfuryl alcohol from 52.8%
(
1 h) to 87% (4 h), unwanted products (dehydrogenated, con-
densation, etc.) were also detected. Thus, at high substrate con-
centrations, the overcrowded surface resulted in the increased
selectivity of DFF that might be related to a change in the
adsorption geometry that influenced the activity and selectivity
Fig. 4 Time profile of the (a) catalytic decarbonylation of HMF and (b)
yield vs. conversion on Pd/Al catalyst. Reaction conditions: catalyst :
substrate = 1 : 10, PCO2 = 6 MPa, and temperature = 145 °C.
2
O
3
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
for the decarbonylation of HMF under the present reaction
conditions. Notably, a nonpolar and a borderline (THF)
solvent were chosen to compare the activity with that of com-
pressed CO , and the results are shown in Table 1. Under the
2
same reaction conditions, the conversion of HMF varied
depending on the solvent used and followed the order: hexane
(
1
92.5%) > cyclohexane (85.2%) > THF (59.1%) (Table 1; entries
to 3). Furfuryl alcohol was detected as the major product
independent of the solvent used, and the selectivity order was
hexane (95.6%) > cyclohexane (76.9%) > THF (50.1%), showing
a clear trend depending on the solvent polarity. No DFF was
detected in hexane and cyclohexane (Table 1; entries 1 and 2),
but 43% of DFF-enriched product mixture was found in THF
(
Table 1; entry 3).
2
After applying a fixed pressure of CO (6 MPa) to the above-
mentioned solvents, a change was observed, especially in the
selectivity of furfuryl alcohol (Table 1; entries 4 to 6). The
selectivity increased to 90% and 92% in THF and cyclohexane,
respectively, (Table 1; entries 4 and 5) after suppressing the
other side reactions observed in the absence of CO ; however,
2
hexane maintained a constant selectivity (95%). A positive
influence was also observed on the conversion of HMF in THF,
which increased significantly to 70%, but slightly dropped in
the other two solvents. This might be attributed to the pres-
2
ence of a co-solvent that affected the solvation of CO mole-
1
9
cule, and the activity decreased. Hence, by comparing the
results with those obtained using non-polar solvents, it can be
Fig. 5 Effect of the catalyst : substrate ratio (wt.) on the activity of Pd/
2 3
Al O catalyst at (a) 1 h and (b) 4 h. Reaction conditions: PCO2 = 6 MPa
and temperature = 145 °C.
2
inferred that compressed CO has a decisive influence on the
catalytic activity, and it is possible to achieve a significantly
improved performance regarding the selectivity of furfuryl
of the reaction. As the reaction progressed, the transform- alcohol.
1
8
ation of DFF to furfuryl alcohol occurred, and the selectivity
Effect of the support materials. It has to be mentioned that
increased. Hence, in the applied protocol, the catalyst : the catalyst support has a wide range of effects that originate
substrate ratio is a critical parameter to control the activity and from direct involvement and influence the particle size (dis-
selectivity. To achieve maximum conversion and selectivity to persion), charge transfer between metal and support (metal
furfuryl alcohol, an optimum ratio of 1 : 10 was used through- support interaction), redox property etc., which can influence
20
out the experiment.
the activity of the catalyst. To elucidate the role of the
Different organic solvents and solvent-CO
2
. After optimizing support material in the decarbonylation of HMF in com-
2
the reaction parameters, different organic solvents, such as pressed CO , we examined Pd supported on different types of
tetrahydrofuran (THF), cyclohexane, and hexane, were tested materials (Table 2). In each case, the metal concentration was
2 2 3
Table 1 Decarbonylation of HMF in different organic solvents and in solvent + CO using the Pd/Al O catalyst
Product selectivity (%)
Furfuryl alc.
Entry
Solvent
Conv. (%)
DFF
Furfural
Others
Organic solvent
1
2
3
THF
Cyclohexane
Hexane
59.1
85.2
92.5
60.1
76.9
95.6
32.9
—
—
6.9
5.2
4.4
—
17.9
—
Organic solvent + CO
2
4
5
6
THF
Cyclohexane
Hexane
70.0
81.4
84.0
90.9
92.0
95.5
5.5
3.4
—
3.3
4.6
4.5
—
—
—
Reaction conditions: catalyst : substrate = 1 : 10; PN2 = 0.1 MPa; Temp. = 145 °C; reaction time = 4 h; metal content ∼5 wt%; entries 1–3: solvent =
7 ml (corresponds to the moles of CO used); entries 4–6: PCO2 = 6 MPa.
∼
2
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
Table 2 Activity of the supported Pd catalysts for the decarbonylation of HMF in compressed CO
2
Product selectivity (%)
a
Entry
Catalyst
Pd/Al
Pd/MCM-41
Pd/Al-MCM-41
Pd/C
Dispersion (%)
Conv. (%)
DFF
Furfuryl alc.
Other
1
2
3
4
5
6
2
O
3
21.4
5.6
4.8
12.1
20.2
—
99.8
22.8
45.6
86.0
65.2
—
>99.0
5.1
>99.0
0.4
26.1
9.1
10.4
58.3
—
—
—
—
99.6
73.9
90.9
86.6
37.9
—
—
—
97.8
—
—
—
2.7
3.8
—
>99.0
>99.0
Pd/hydrotalcite
2 3
Al O
b
c
7
Pd/Al
Pd/Al
Pd/Al
2
2
2
O
O
O
3
3
3
—
—
d
e
8
f
9
—
Reaction conditions: Catalyst : substrate = 1 : 10; PCO2 = 6 MPa; PN2 = 0.1 MPa; Temp. = 145 °C; reaction time = 4 h; 0.05 mol% of Pd.
a
d
b
c
Approximate expression of metal dispersion = 0.9/diameter (in nm); [see ref. 36]. DFF as substrate. 86.3% = furfural and 13.7% = other.
e
f
Furfuryl alcohol as substrate. 36.9% furfural and 63.1% condensation products. Controlled experiment with DFF in the presence of trace
amounts of hydrogen for 1 h.
fixed to ∼0.05 mol%. The dispersion of Pd was calculated from under the optimised reaction conditions (temperature of
the particle size distribution (ESI; Fig. S6†), which varied from 145 °C, reaction time of 4 h, and PCO2 of 6 MPa) (Table 3). In
5
% to ∼21% depending on the support used. An inert support each case, the decarbonylated product was detected as the sole
such as MCM-41 (only Si) provided <25% conversion (Table 2; product. Furfural and 5-methylfurfural were converted to their
entry 2); however, after modification with Al, the conversion corresponding decarbonylated compounds with an excellent
reached ∼46% (Table 2; entry 3). A very high conversion of yield of >99% and 92.5%, respectively (Table 3; entries 1 and 2).
8
6% was observed on Pd/C (Table 2; entry 4), whereas 65.2% Furthermore, in addition to furanic compounds, pyrrole alde-
HMF was converted over Pd/hydrotalcite, which was a basic hyde also underwent decarbonylation to produce pyrrole
support (Table 2; entry 5). The product distribution also high- (56.2% yield) (Table 3; entry 3). The method was then extended
lights the support effect. Under the applied conditions, fur- to a series of substituted benzaldehyde. Depending on the
furyl alcohol was the major product on MCM-41 (73.9%) as presence of electron-withdrawing and electron-donating
well as on Al-MCM-41 (90.9%) (Table 1; entries 2 and 3). In groups, the activity of the compound varied; nitro-substituted
addition, Pd/C also offers an excellent selectivity of furfuryl benzaldehydes exhibited good to excellent yields based on the
alcohol (86.6%) accompanied by DFF (10.7%) and 2.7% of position of the substituent, which followed the sequence
2 3
other by-products (Table 2; entry 4). Although Pd/Al O and p- (100%) > m- (90.6%) > o- (26.8%) (Table 3; entries 4, 5, and 6).
Pd/hydrotalcite exhibit comparable Pd dispersion, dehydro- The presence of –OH and –OMe groups in p-substituted benz-
genation predominates over decarbonylation on hydrotalcite; aldehyde also affords the corresponding decarbonylated pro-
DFF is obtained as a major product with a selectivity of 58.3% ducts without affecting the other functional groups; p-hydroxy-
(
Table 2; entry 5). The conversion and selectivity of furfuryl benzaldehyde shows tolerance to the –OH group and is suc-
alcohol among the catalysts were investigated; they followed cessfully converted to phenol with 64.4% yield (Table 3; entry 7).
the order Al > C > hydrotalcite > Al-MCM-41 > MCM-41 and On the other hand, 76% anisole was generated smoothly from
Al > C > Al-MCM-41 > MCM-41 > hydrotalcite, respectively. the decarbonylation of p-anisaldehyde (Table 3; entry 8).
2 3
O
2 3
O
From these results, it can be suggested that all the catalysts are Moreover, 55% toluene was obtained from p-tolylaldehyde
active under the studied conditions. Furthermore, the nature (Table 3; entry 9). Interestingly, styrene reached an excellent
of the support material plays an important role; acidic sup- yield of 99% from the decarbonylation of trans-cinnamalde-
ports are more efficient than basic supports; this can be hyde (Table 3; entry 10). Indole-3-carboxaldehyde, a bicyclic
explained by their role in the modification of active sites (Pd) compound, also decarbonylated to indole with a high yield of
through the enhancement of the electron deficiency of noble 68.3%. Thus, the efficiency of the developed method was con-
2
1
metals.
The excellent performance of Pd/Al O may be firmed by the powerful performance of a wide range of alde-
2
3
2
2
related to the possible interaction of Pd with Al
2
O
3
as Al
2
O
3
hydes to achieve quantitative selectivity of the targeted decar-
itself is completely inactive in the reaction (Table 2, entry 6). bonylation products.
However, the present results do not allow us to reach a precise
Catalyst recycling
conclusion, which requires a detailed study. Previously, a
support-dependent activity has also been observed in the gas- Recycling is one of the most advantageous factors of a hetero-
4
phase decarbonylation of furfural.
geneous catalyst for the environmental and economic sustain-
Scope of the developed method. After successful application ability of a process. In the decarbonylation process, it is con-
of the developed method in the decarbonylation of HMF, the sidered that the catalytic deactivation can originate from the
process was further extended to other heterocyclic aldehydes adsorption of CO on the metal surface. Thus, to assess the
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
Table 3 Substrate scope of the decarbonylation reaction on the Pd/
Al O catalyst
2 3
Transmission electron microscopy (TEM) images of the
fresh and used catalysts and the corresponding particle size
distributions are shown in Fig. 6a and b, respectively. In the
fresh catalyst, spherical Pd particles with an average size of
4.2 ± 0.4 nm were detected on the Al O surface (Fig. 6a).
Entry
Substrate
Product yield (%)
a
1
2
3
Similarly, from the calculation of the particle size distri-
butions, the TEM image of the spent catalyst (Fig. 6b) did not
reveal any sign of agglomeration of metal particles or depo-
sition of carbonaceous materials, and the particle size of 4.8 ±
2
3
0.3 nm was almost retained.
The FTIR spectra of the fresh and used catalysts were
obtained in the range of 4000–400 cm− and are depicted in
Fig. S2a and S2b,† respectively (details are provided in the ESI
section†). In the spectrum of the spent catalysts (Fig. S2b†), no
peak was detected in the 1800–2100 cm− range (dotted rec-
tangle in Fig. S2b†) corresponding to adsorbed CO on metallic
1
1
4
5
2
3
Pd, which is one of the possible factors of catalyst de-
2
4
activation during decarbonylation. The high miscibility of
CO in compressed CO might prevent the adsorption of CO on
2
the catalyst surface, as confirmed from the analysis of the
gaseous part after the reaction (ESI; Fig. S3†), eventually
improving the catalyst life. Thus, the decreased catalytic
activity can be related to the loss of catalyst because of hand-
ling during separation.
6
In the fingerprint region (1700–1100 cm− ), the FTIR spec-
1
trum of the used catalyst also shows a number of signals at
1
7
8
9
190, 1280, 1400, 1510, 1620, and 1676 cm− (an expanded
1
view is illustrated in ESI; Fig. S4a†). The described IR bands
2
5
belongs to HMF, and the assignments can be made from an
overlay spectrum of HMF (ESI: Fig. S4b; details are in the
−
1
ESI†). The band at 1676 cm is due to the carbonyl group of
HMF, which shifts to 1665 cm− in the used catalyst because
1
1
1
0
1
b
Reaction conditions: catalyst : substrate = 1 : 10; PCO2 = 6 MPa; PN2 = 0.1
MPa; Temp. = 145 °C; time = 4 h. a = 6.5 h, b = 12 h; in each case, the
decarbonylated product was detected as the sole product.
efficiency of the spent catalyst, it was recycled after separation
from the product mixture by filtration (ESI; Fig. S1†). The used
th
catalyst retained its activity until the 5 recycle, which then
dropped slightly without affecting the product selectivity. After
confirming the stability of the metal by a hot filtration test,
surface characterisation of the catalyst was conducted using
different spectroscopic techniques.
2 3
Fig. 6 TEM images and particle size distribution of the Pd/Al O cata-
lyst: (a) fresh and (b) used.
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
of the possible adsorption on Pd, and explains the interaction results; thus, only an outline is proposed (Scheme 3). Herein,
of HMF with the metal surface through the aldehyde function- the reaction was conducted in compressed CO , and there was
2
26
−1
28
ality. In addition, a very weak band appears at 1620 cm
,
every possibility of the chemical participation of CO
2
by
2
9
30
which may be attributed to the CvC stretching vibration of direct interaction with the substrate and with the catalyst.
the furan ring bound to the carbonyl group. A small shoulder Considering the experimental results, we observed that the
at 1720 cm− indicates the possible CvO stretching of alde- highest catalytic activity was obtained in the substrate-rich
1
hyde (marked with *). Regarding the peaks in the –OH region phase; hence, the possibility of substrate–CO
2
interactions was
3500–3200 cm− ), the signal of the spent catalyst is broader dismissed. The IR spectrum of the used catalyst shows no
1
(
−
1
than that of the fresh catalyst with a maxima at 3425 cm
bands related to the carbonate species that generally exist in
3
1
because of the possible change upon adsorption of HMF. In the presence of CO
2
.
2
Therefore, CO can be considered a CO
addition, the peaks related to the aromatic and aldehyde C–H acceptor in the present system, which causes a shift in the
−1
−1
stretching are observed at 2920 cm
respectively.
and 2860 cm
,
reaction equilibrium and enhances the reaction rate (ESI;
Fig. S3†).
To understand the change in the metal environment before
The best performance of the catalyst was achieved under
and after the reaction, XPS was performed on the fresh and the two-phase condition (CO
recycled catalysts (ESI; Fig. S5†). Fig. S5a and S5b† exhibit the controlled reaction; however, only the support surface (Al
d region spectra of Pd in the fresh and used catalyst, respect- is completely inactive in the reaction, which requires Pd
2
–substrate), suggesting a surface-
2 3
O )
3
ively. The Pd 3d core level spectrum of the fresh catalyst con- (Table 2; entry 6). Therefore, we propose that in the first step,
tains two well-separated components at 334.9 eV (3d5/2) and HMF is invariably adsorbed on the metal surface through the
0
3
40.3 eV (3d3/2) corresponding to Pd , which are also present aldehyde functionality, as confirmed from the product distri-
in the spectrum of the recycled catalyst (335.3 eV (3d5/2) and bution. From the spectral analysis of the used catalysts, we
40.6 eV (3d3/2), respectively). There is no peak related to the also detected a shift of the aldehyde –C–O band and a change
3
−
1
PdO species in the higher binding energy region. The atomic in the aromatic skeletal vibration region (1600 and 1500 cm ).
x
concentrations calculated in the fresh and recycled catalyst are The transformation of HMF into furfuryl alcohol can be postu-
4
(
.7% (Pd), 36.8% (Al), and 58.5% (O) and 4.4% (Pd), 37.2% lated as a consequence of the C–C bond scission between the
Al), and 58.7% (O), respectively. In addition, XPS revealed a aryl and aldehyde groups. Previously, Song et al. have reported
slight change in the amount of Pd atoms on the surface of Pd/ that the cleavage of a C–C bond is easier than that of a C–H
Al measured in terms of Pd/Al ratio, which increased bond in aryl aldehyde due to the strong interaction between
slightly from the fresh catalyst (0.072) to the recycled catalyst the aryl group and the catalyst surface that causes a weakening
2 3
O
3
2
(0.110) and was attributed to a nominal change in the dis- of the C–C bond. In addition, a strong interaction of the
2
7
persion of the metal on the catalyst surface. The XPS spectra aldehyde group and furan ring with the same metal surface
of the support materials confirmed the minor changes in the was also reported. In the next step, the cleavage of the C–C
3
3
binding energies of the Al 2p and O 1s spectra (ESI†).
bond causes decarbonylation followed by reductive elimin-
ation and results in the decarbonylated product. The minor
dehydrogenation reaction might be the consequence of the
Possible mechanism
Elucidation of the reaction mechanism is important to under- interaction of the –OH proton with the neighbouring surface
3
4
stand each reaction step and helpful for further implemen- oxygen. Furthermore, the role of the support material cannot
tation of the developed method. It is difficult to predict a clear be ignored, as described in the previous section, which needs
cut mechanism for the studied reaction based on the present further investigations.
Conclusion
In conclusion, we have developed an efficient decarbonylation
methodology free from any organic solvent or CO scavenger for
the selective decarbonylation of biomass-based platform mole-
cules, such as HMF, using a commercially available Pd/Al O
2
3
2
catalyst in compressed CO . The unique properties of the reac-
tion medium can be beneficially applied in the conversion of
HMF to furfuryl alcohol with very high selectivity (99.6%).
Indeed, the incorporation of compressed CO
improved the catalytic activity. Neither CO
2
significantly
nor Pd alone
2
appears suitable for the reaction, but a combination of both
provides high catalytic performance and excellent product
yield through the proper control of the reaction variables. A
minor path through dehydrogenation was also detected, but
Scheme 3 Proposed reaction mechanism.
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
had nominal effect on the product selectivity. This process
allows an easy separation of the product simply by depressuri-
Conflicts of interest
sation. Based on the characterisation of the fresh and used There are no conflicts to declare.
catalyst, a reaction mechanism was proposed. The catalytic
system provides appreciable reusability and shows tolerance to
a series of different types of substrates with various substitu-
ents, which generate the corresponding decarbonylated com-
Acknowledgements
pounds with good to excellent yield without affecting the sub- H. K. is funded by the Japan-U.S. cooperation project for
stituents. Our future challenge is to investigate the role of the research and standardization of Clean Energy Technologies,
oxide support in the presence of compressed CO and to deter- The Ministry of Economy, Trade and Industry (METI), Japan
2
mine the exact reaction mechanism after the proper under- and CREST, Japan Science and Technology (JST).
standing of the status of the reaction medium.
References
Experimental
1
(a) J. Tsuji, J. Am. Chem. Soc., 1968, 90, 99;
b) D. H. Doughty and L. H. Pignolet, J. Am. Chem. Soc.,
Materials
-Hydroxymethylfurfural (HMF) (Aldrich) was used as received.
Carbon dioxide (>99.99%) was supplied by Nippon Sanso Co.
Ltd. 5% Pd/C and 5% and 1% Pd/Al O were obtained from
(
5
1
1
978, 100, 7083; (c) J. M. O’Connor and J. Ma, J. Org. Chem.,
992, 57, 5075; (d) P. Fristrup, M. Kreis, A. Palmelund,
2
3
P.-O. Norrby and R. Madsen, J. Am. Chem. Soc., 2008, 130,
206; (e) J. J. Adams, N. Arulsamy and D. M. Roddick,
Aldrich. Pd(II) sodium chloride was obtained from Aldrich, and
1% Pd/hydrotalcite was obtained from Wako Pure Chemicals.
Cetyltrimethylammonium bromide used as a template for the
MCM-41 synthesis was obtained from Aldrich.
5
∼
Organometallics, 2012, 31, 1439; (f) T. Iwai, T. Fujihara and
Y. Tsuji, Chem. Commun., 2008, 6215.
2
3
(a) A. Modak, A. Deb, T. Patra, S. Rana, S. Maity and
D. Maiti, Chem. Commun., 2012, 48, 4253; (b) K. Ding,
S. Xu, R. Alotaibi, K. Paudel, E. W. Reinheimer and
J. Weatherly, J. Org. Chem., 2017, 82, 4924.
(a) M. Kreis, A. Palmelund, L. Bunch and R. Madsen, Adv.
Synth. Catal., 2006, 348, 2148; (b) C. M. Beck,
S. E. Rathmill, Y. J. Park, J. Chen, R. H. Crabtree,
L. M. Liable-Sands and A. L. Rheingold, Organometallics,
Tetraethylorthosilicate, which was used as the silica source,
was obtained from Wako Pure Chemicals.
The metal catalyst supported on MCM-41 was synthesised
3
5
in our laboratory using a modified method. In each case, the
metal content was maintained as ∼5%. Unless otherwise
stated, all the chemicals were used without further purifi-
cation. Details of the catalyst characterisation techniques are
provided in the ESI.†
1999, 18, 5311.
4
(a) H. Singh, M. Prasad and R. D. Srivastava, J. Chem.
Technol. Biotechnol., 1980, 30, 293; (b) P. Lejemble, A. Gaset
and P. Kalck, Biomass, 1984, 4, 263.
Phase observation
Video monitoring of the phase behaviour of HMF in com-
pressed CO2 at 145 °C was performed separately in a 10 ml
high-pressure view cell. At first, HMF was introduced into the
cell followed by CO after stabilisation of the temperature. For
2
phase observation during the reaction, the content was stirred
continuously, and images were obtained at each step.
5 (a) P. K. Kundu, M. Dhiman, A. Modak, A. Chowdhury,
V. Polshettiwar and D. Maiti, ChemPlusChem, 2016, 81,
1142; (b) V. V. Pushkarev, N. Musselwhite, K. An,
S. Alayoglu and G. A. Somorjai, Nano Lett., 2012, 12, 5196;
(c) N. Pino, S. Sitthisa, Q. Tan, T. Souza, D. López and
D. E. Resasco, J. Catal., 2017, 350, 30.
Catalytic performance
6
(a) A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
2411; (b) O. Casanova, S. Iborra and A. Corma,
ChemSusChem, 2009, 2, 1138; (c) Y. Roman-Leshkov,
J. N. Chheda and J. A. Dumesic, Science, 2006, 312, 1933;
(d) H. B. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang,
Science, 2007, 316, 1597; (e) G. Yong, Y. G. Zhang and
J. Y. Ying, Angew. Chem., 2008, 120, 9485, (Angew. Chem. Int.
Ed., 2008, 47, 9345); (f) L. Vanoye, M. Fanselow,
J. D. Holbrey, M. P. Atkins and K. R. Seddon, Green Chem.,
2009, 11, 390; (g) S. Hu, Z. Zhang, J. Song, Y. Zhou and
B. Han, Green Chem., 2009, 11, 1746; (h) J. Lewkowski,
ARKIVOC, 2001, 2, 17.
A 50 ml batch reactor placed in a hot air circulating oven was
used to test the reaction on the laboratory scale. In a typical
experiment, a specified amount of the catalyst and substrate
were introduced into the reactor. The reactor was flushed
5
times with nitrogen to remove air. The reactor was heated for
a specified amount of time to stabilise to the required temp-
erature (145 °C). After stabilisation, CO was charged into the
2
reactor using a high-pressure liquid pump and then com-
pressed to the desired pressure. The content of the reactor was
stirred with a magnetic stirrer bar during the reaction. After
the reaction, the reactor was quenched using an ice bath, fol-
lowed by careful depressurisation and the separation of the
solid catalyst from the liquid product simply by filtration. The
detailed analytical method is provided in the ESI.†
7 (a) G. Tsilomelekis, M. J. Orella, Z. Lin, Z. Cheng,
W. Zheng, V. Nikolakis and D. G. Vlachos, Green Chem.,
2016, 18, 1983; (b) A. Herbst and C. Janiak, New J. Chem.,
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
2
016, 40, 7958; (c) B. Girisuta, L. P. B. M. Janssen and 22 (a) J. H. Kwak, J. Z. Hu, D. H. Mei, C. W. Yi, D. H. Kim,
H. J. Heeres, Green Chem., 2006, 8, 701.
C. H. F. Peden, L. F. Allard and J. Szanyi, Science, 2009, 325,
1670; (b) D. H. Mei, J. H. Kwak, J. Z. Hu, S. J. Cho, J. Szanyi,
L. F. Allard and C. H. F. Peden, J. Phys. Chem. Lett., 2010, 1,
2688.
8
9
J. Mitra, X. Zhou and T. Rauchfuss, Green Chem., 2015, 17, 307.
Y.-B. Huang, Z. Yang, M.-Y. Chen, J.-J. Dai, Q.-X. Guo and
Y. Fu, ChemSusChem, 2013, 6, 1348.
1
0 (a) Q. Meng, C. Qiu, G. Ding, J. Cui, Y. Zhu and Y. Li, Catal. 23 (a) F. M. Hoffmann, Surf. Sci. Rep., 1983, 3, 107;
Sci. Technol., 2016, 6, 4377; (b) Q. Meng, D. Cao, G. Zhao,
C. Qiu, X. Liu, X. Wen, Y. Zhu and Y. Li, Appl. Catal., B,
(b) C. Keresszegi, T. Burgi, T. Mallat and A. Baiker, J. Catal.,
2002, 211, 244.
2
017, 212, 15.
24 (a) R. D. Srivastava and A. K. Guha, J. Catal., 1985, 91, 254;
(b) M. D. Argyle and C. H. Bartholomew, Catalysts, 2015, 5,
145.
25 N. B. Colthup, L. H. Daly and S. E. Wiberley, Introduction to
Infrared and Raman Spectroscopy, Academic Press,
New York, 1964.
1
1 F. M. A. Geilen, T. vom Stein, B. Engendahl, S. Winterle,
M. A. Liauw, J. Klankermayer and W. Leitner, Angew. Chem.,
Int. Ed., 2011, 50, 6831.
2 M. Chatterjee, T. Ishizaka, A. Chatterjee and H. Kawanami,
Green Chem., 2017, 19, 1315.
1
1
3 (a) A. Milewska, A. M. Banet Osuna, I. M. Fonseca and 26 H.-W. Chen, C.-S. Chen and S.-J. Ham, J. Phys. Chem., 1995,
M. Nunes da Ponte, Green Chem., 2005, 7, 726; 99, 10557.
b) D. Chouchi, D. Gourgouillon, M. Courel, J. Vital and 27 B. Kucharczyk, W. Tylus and L. Kepinski, Appl. Catal., B,
(
M. Nunes da Ponte, Ind. Eng. Chem. Res., 2001, 40, 2551.
2004, 49, 27.
1
1
4 N. J. Meehan, A. J. Sandee, J. N. H. Reek, P. C. J. Kamer, 28 M. A. Blatchford, P. Raveendran and S. L. Wallen, J. Am.
Piet W. N. M. van Leeuwen and M. Poliakoff, Chem.
Commun., 2000, 1497.
Chem. Soc., 2002, 124, 14818.
29 Y. Jing, Y. Hou, W. Wu, W. Liu and B. Zhang, J. Chem. Eng.
Data, 2011, 56, 298.
5 P. Sykes, A guidebook to mechanism in organic chemistry,
Longman Scientific & Technical, Essex, UK, 6th edn, 1986, 30 (a) J. Szanyi and J. H. Kwak, Phys. Chem. Chem. Phys., 2014,
pp. 42–43.
16, 15126; (b) K. Fottinger, R. Schlogl and G. Rupprechter,
Chem. Commun., 2008, 320.
1
1
1
6 A.-N. Ko and B. Wojciechowski, Prog. React. Kinet., 1983, 12,
2
0.
7 M. Chatterjee, T. Ishizaka and H. Kawanami, Green Chem.,
014, 16, 1543.
8 (a) F. Delbecq and P. Sautet, J. Catal., 1995, 152, 217;
b) D. Loffreda, F. Delbecq, F. Vigné and P. Sautet, J. Am.
31 X. Wang, H. Shi, J. H. Kwak and J. Szanyi, ACS Catal., 2015,
5, 6337.
32 R. Song, D. Ostgard and C. V. Smith, in Catalysis of organic
reactions, ed. J. L. Davis and M. A. Barteau, Surf. Sci., 1987,
vol. 187, pp. 387.
2
(
Chem. Soc., 2006, 128, 1316.
9 (a) T. Aizawa, S. Janttarakeeree, Y. Ikushima, N. Saitoh,
K. Arai and R. Smith, J. Supercrit. Fluids, 2003, 27, 247;
33 (a) S. Sitthisa, T. Pham, T. Prasomsri, T. Sooknoi,
R. G. Mallinson and D. E. Resasco, J. Catal., 2011, 280, 17;
(b) S. H. Pang and J. Will Medlin, ACS Catal., 2011, 1,
1272.
34 (a) L. Nondek and J. Sedlacek, J. Catal., 1975, 40, 34;
(b) B. Xu, X. Liu, J. Haubrich and C. M. Friend, Nat. Chem.,
2010, 2, 61.
1
2
2
(
b) A. I. Frolov and M. G. Kiselev, J. Phys. Chem. B, 2014,
18, 11769.
0 (a) T. Ishihara, K. Harada, K. Eguchi and H. Arai, J. Catal., 1992,
36, 161; (b) M. Shekhar, J. Wang, W.-S. Lee, W. D. Williams,
1
1
S. M. Kim, E. A. Stach, J. T. Miller, W. N. Delgass and 35 M. Chatterjee, T. Iwasaki, Y. Onodera and T. Nagase, Catal.
F. H. Ribeiro, J. Am. Chem. Soc., 2012, 134, 4700. Lett., 1999, 61, 199.
1 (a) A. Yu. Stakheev and L. M. Kustov, Appl. Catal., A, 1999, 36 M. Boudart and G. Djega-Mariadassou, Kinetics of
1
88, 3; (b) R. A. Della Betta, M. Boudart, P. Gallezot and
Heterogeneous Catalytic Reaction, Princeton University
R. S. Weber, J. Catal., 1981, 69, 514.
Press, Princeton, NJ, 1984.
This journal is © The Royal Society of Chemistry 2018
Green Chem.