G Model
CATTOD-8920; No. of Pages7
ARTICLE IN PRESS
H. Cai et al. / Catalysis Today xxx (2014) xxx–xxx
4
On the whole, one may clearly see that conversion attains a max-
imum with the increasing hydrogen pressure (Fig. 2a), while as it
is not affected by the catalysts concentration (Fig. 2b). Therefore,
mass transfer (or hydrogen solubility) is possibly controlling the
overall rate of the reaction.
To obtain high yield of furan-based diols, hydrogenation reac-
tion was also conducted at different temperature, the results are
shown in Fig. 2c. As a typical hydrogenation catalyst, Pd/C shows
HMF yield
DHMTF selectivity
DHMF selectivity
1
20
HMF conversion
100
80
6
0
0
strong activity in IL/water mixture for HMF hydrogenation reaction,
◦
under mild conditions of 25 C and 6 MPa H , HMF was consumed
2
4
nearly completely in 3 h to afford total diols selectivity 85.8% (39.2%
of DHMF and 46.6% of DHMTF). Raising the reaction temperature
◦
◦
◦
20
0
to 50 C, 80 C and 100 C could eventually simultaneously hydro-
genate all the unsaturated bonds in HMF to obtain exclusively
DHMTF, and the total selectivity of diols are 84%, 88.1% and 91.4%,
◦
1h 3h
Pd/C
1h 3h
1h 3h
1h 3h
Pd/TiO 2
1h 4h
Pd/HT
respectively. Much higher reaction temperature of 200 C could
Pd/Al O
Pd/SiO 2
2
3
lead to severe side reactions mentioned before that decrease the
yield of target products. It is worthy of notice that even at this high
temperature, only trace DMF was formed (less than 2%), implying
that dehydroxylation reaction is hard to occur in our system. There-
fore, it is easy to control the products to furan-based diols, which is
one of the unique features of this work over most existing methods.
Taking into account of the reaction efficiency and the convenience
of the operation, 50 C is selected for hydrogenation reaction since
this low temperature could avoid the formation of by-products and
without compromising the diols yield.
It should be noted that in most of the above hydrogenation reac-
tions, DHMTF was the main product; even under relatively mild
condition, it is hard to stop the reaction in the DHMF stage, sug-
gesting that Pd/C is an active catalyst for both the aldehyde group
and furan ring hydrogenation reactions to produce DHMTF.
Fig. 3. Production of furan-based diols from fructose via one-pot tandem reaction.
Reaction conditions: the dehydration reaction was first conducted in [BMIm]Cl at
◦
130 C with 0.18 g fructose for 20 min to afford HMF. Then 35 ml water and 0.05 g
catalyst were added into the reactor for hydrogenation reaction under 6 MPa H2 at
◦
50
C.
◦
oxygen vacancies) [30] at the palladium-titania interfacial region
which can coordinate the oxygen atom of the C O group via a lone
pair of electrons and thus, activate the C O group for the selec-
tive hydrogenation reaction [29,30]. Therefore, by dispersing on
the reductive support TiO , metal catalyst exhibits the potential
2
for selectively hydrogenating the carbonyl group while leaving the
C
C bond unchanged to afford DHMF. This viewpoint will be further
proved in Section 3.3.
As for Pd/HT catalyst, it is found that the basic HT support
not only restrains the dehydration reaction, but also significantly
inhibits the hydrogenation activity of Pd. Under the identified reac-
tion conditions, the conversion of HMF in 1 h is distinctly decreased
to 69.7% in comparison with the nearly quantitative conversion of
HMF in Pd/C catalyzed reaction. Even we prolonged the reacting
time to 4 h, HMF could not be completely consumed. On the other
hand, nitrogen physisorption showed that HT possessed the small-
est surface area (ESI, Table S1), implying the lowest metal active
sites were exposed. Accordingly, the activity of Pd/HT was relatively
low. In addition, it is reported that HT could induce ring-opening,
C–O dissociation reactions and other side reactions [23,31]. As a
result, the combined selectivity of DHMF and DHMTF was remark-
ably lower than 50%. Overall, due to its basic property and low
surface area, HT is not the preferred support in our reaction system
for fructose dehydration and subsequent hydrogenation reactions
to produce furan-based diols.
3
.2. One-pot tandem reactions of carbohydrates: support effects
with palladium catalyst
Previous studies indicate that the activity of the supported cata-
lyst strongly depends on the property of the support [29]. With the
aim to explore the support effect and further control the products
distribution, Pd dispersed on various supports, including ꢀ-Al O3
and SiO , TiO and Mg-Al hydrotalcite (HT), are synthesized and
applied in the direct transformation of fructose. It is known that
the fructose dehydration step is a typical acid-catalyzed reaction,
however, the weak acidic ꢀ-Al O3 and SiO2 supported palladium
catalysts did not enhance catalytic activity obviously compared to
the neutral supports TiO and active carbon. The yields of HMF were
between 74% and 80% over the four catalysts (Fig. 3), which were in
the same level with that of thermal reaction. This result is similar
to our previous study that the weak mineral acid could not pro-
mote fructose dehydration reaction in ILs [29]. On the other hand,
the basic HT inhibited the dehydration reaction with HMF yield
decreased to 60.7%.
2
2
2
2
2
3.3. Production of furan-based diols: effect of metal species
In the hydrogenation step, these supports showed different
Studies were carried out to control the product selectivity by
changing the nature of the metal component in the one-pot tandem
reactions. All the reactions are started from fructose via one-
pot two-step process: first, 20 min of the dehydration reaction in
effects on the catalytic activity. The acidic Pd/ꢀ-Al O3 and Pd/SiO2
2
embodied similar activities as that of Pd/C, HMF almost completely
transformed in 1 h. However, the selectivity to diols was a little bit
lower and the proportion of DHMTF was higher than those of Pd/C,
which means the product selectivity is sensitive to the acidity of
the support. If the reactions were prolonged to 3 h, the C C bond in
DHMF could be further saturated, and the final products are dom-
◦
[BMIm]Cl at 130 C, and then, 3 h of hydrogenation reaction under
◦
6 MPa H2 at 50 C. It is found in Table 1 that using either carbon
supported palladium or ruthenium as the catalyst, HMF could be
completely consumed to afford diols’ selectivity and yield of ca.
88% and 66% (entries 1–2), respectively. However, the distribu-
tions of the diols are different. Deep hydrogenated product DHMTF
(selectivity: 84.2%, yield based on fructose: 63.4%, see in entry 1)
dominated in the case of Pd/C, while carbonyl hydrogenated com-
based diols’ selectivity could reach 93% and 83.8% respectively,
inant DHMTF. In the case of Pd/TiO , it is important to point out
2
that although the conversion of HMF and the combined selectivity
to diols were similar to those of Pd/C, nearly half of the reaction
was stopped at DHMF stage under the same conditions in 1 h, a
longer reaction time (3 h) did not significantly change the products
distribution. It is possibly because that the interaction between the
3+
reductive support and metal created a special defect sites (Ti or
Please cite this article in press as: H. Cai, et al., Biomass into chemicals: One-pot production of furan-based diols from carbohydrates