Full Papers
À3
beyond 44.610 mmol decreased the yields of FFA and fur-
fural and increased the yields of 2-MF and furan. At the same
time, the yield of FA decreased continuously. These results sug-
gest that a higher amount of Pd(OAc) promoted the hydroge-
2
nation of FFA to 2-MF and the decarbonylation of furfural to
furan accompanied by the consumption of FA formed earlier.
Effect of water on the one-pot reaction
The presence of water is unfavorable for the decarbonylation
[
15a]
of aldehydes.
In our system, we investigated the possible in-
fluence of water on the tandem dehydration and decarbonyla-
tion of fructose to FFA by the removal or addition of water
from the catalytic system. As AlCl ·6H O contains water in the
3
2
Figure 3. Effect of the amount of 4 molecular sieves used on the yields of
crystal, we replaced it with anhydrous AlCl as the dehydration
3
FFA and HMF. Reaction conditions: 0.3 mmol of fructose, 2 mL of 1,4-diox-
À3
À3
catalyst. The results shown in Table 1 indicated that the total
ane, 4.110 mmol of AlCl
atmosphere.
3 2 2
·6H O, 44.610 mmol of Pd(OAc) , 1508C, 4 h,
N
2
yields of FFA and HMF were comparable over AlCl ·6H O
3
2
(
entry 8) and anhydrous AlCl catalysts (entry 9), but the yield
3
of FFA over anhydrous AlCl was slightly lower. Clearly, the de-
3
carbonylation efficiency of the anhydrous AlCl /Pd(OAc) cata-
amount of 4 molecular sieves was increased revealed the
promotion effect of the 4 molecular sieves on the tandem
decarbonylation of the HMF intermediate to FFA. The slight in-
crease of the HMF yield together with the decrease of the FFA
yield if the ratio of 4 molecular sieves to AlCl ·6H O was fur-
3
2
lytic system was lower than that of the AlCl ·6H O/Pd(OAc)
2
3
2
system, most probably because of the decrease of the dehy-
dration rate caused by the absence of water of crystallization
in the AlCl3 catalyst. For comparison, we added a certain
amount of water into the catalytic system before the one-pot
reaction, but no 4 molecular sieves were added to remove
water. Moreover, we performed another control experiment in
which no 4 molecular sieves were added if AlCl ·6H O was
3
2
ther increased to 250 (w/w) might imply the possible adsorp-
tion of FFA on the molecular sieves, but this was excluded im-
mediately as the molecular size of FFA (~7 ) is larger than the
mean pore size of the 4 molecular sieves (~4 ). Moreover,
the fact that no HMF or FFA was detected in the washing
liquor from the molecular sieves after the reaction also sup-
ported this deduction. Therefore, the decreased yield of FFA in
the presence of a larger amount of 4 molecular sieves was
most probably because of the relatively lower dehydration rate
caused by the reduction of the catalytically active Al species
after the removal of water. Furthermore, the color of the reac-
tion mixture became lighter after the addition of 4 molecular
sieves into the AlCl ·6H O/Pd(OAc) system. The yields of the
3
2
used as the dehydration catalyst (Table 1, entry 12). A compari-
son of the data shown in entries 10, 11, and 12 revealed a de-
crease of the yield of FFA with the increase of the water
dosage, whereas the total yields of FFA and HMF almost did
not vary. This result indicated that the presence of water was
unfavorable for the decarbonylation reaction, possibly because
the dehydration rate was decreased. We infer that some hydro-
lyzed Al species and HCl coexisted in the aqueous system
caused by the hydrolysis of AlCl ·6H O. If we used HCl as the
3
2
3
2
2
dehydration catalyst instead of AlCl ·6H O, we found that the
byproducts FA, furfural, and 5-MF also decreased if an in-
creased amount of 4 molecular sieves was used. These re-
sults demonstrate the alleviation of side-reactions such as the
direct decomposition of fructose and the polymerization of
any reaction intermediates by 4 molecular sieves by the re-
moval of water from the reaction mixture. A similar water-re-
moval role was observed for 5 molecular sieves, which have
3
2
total yields of FFA and HMF increased as the amount of HCl
used increased, but these yields were still lower than those
over AlCl ·6H O, even if we used three molar equivalents of
3
2
[17]
HCl relative to AlCl ·6H O (Table S2). This indicates that the
3
2
hydrolyzed Al species and Pd(OAc) were more active than HCl
2
and Pd(OAc) for the one-pot catalytic deoxygenation of fruc-
2
[
18]
[17]
tose. Additionally, the yields of furfural and FA increased if
the amount of HCl was increased, which demonstrates the
ease of fructose decomposition in the presence of a large
a similar pore size to 4 molecular sieves (Table S3).
Effect of PVP on the one-pot reaction
[13]
amount of Brønsted acid.
We tried to remove the water that existed in the solvent
and was formed during the dehydration step by adding 4
molecular sieves. If the amount of 4 molecular sieves in-
creased, we found that the yield of FFA increased from 7.6%
We investigated the role of PVP on the one-pot deoxygenation
of fructose. PVP itself had nearly no activity for the dehydration
of fructose (Table 1, entry 13). However, the addition of PVP
into the AlCl ·6H O/4 molecular sieves system resulted in an
3
2
(
without molecular sieves) to 19.8% (molecular sieves/
increase of the HMF yield from 23.4 to 48.1% (Table 1, entry 5
vs. entry 6), which illustrated the synergetic role of PVP on the
dehydration of fructose. Furthermore, the yields of FFA and
HMF were 19.8 and 1.2%, respectively, with a total yield of
21.0% if no PVP was used (Figure 4). After the addition of PVP,
AlCl ·6H O=150, w/w) and then decreased to 17.1% (molecu-
lar sieves/AlCl ·6H O=250, w/w; Figure 3). In contrast, the
yield of HMF exhibited the opposite trend. A decrease of the
HMF yield accompanied by the increase of the FFA yield if the
3
2
3
2
ChemCatChem 2016, 8, 1379 – 1385
1382
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim