C. Li et al. / Carbohydrate Research 345 (2010) 1846–1850
1847
O
OH
OH
_
H
3
H O
+ 2H O HOOC-CH2-CH2-C-CH3
2
2
O
O
OH
O
HO
+
+
+
[H ]
[H ]
HO
OH
O
HC-OH
(Fructose)
(HMF)
Scheme 1. General transformation pathway of fructose.
2
2
. Experimental
.1. Materials and instruments
Fructose was purchased from Sigma (St. Louis, USA), and was
the standard curve method. The mass of HMF (M
as follows,
H
) was calculated
ðmgÞ ¼ HMF concentrationðꢁ10ꢀ3 mg=mLÞ ꢁ 4:025 ðmLÞ
M
H
ꢁ ðV
The yield of HMF was calculated as follows,
HMF yield ¼ M =ðinitial fructose mass ꢁ 0:7Þ ꢁ 100%
In which, M is the mass of HMF, V is the volume of the sample,
1
=0:025Þ ꢁ ðM
0
=M
1
Þ
dried under vacuum at 80 °C for 24 h before use. N-Methylimidaz-
ole (99%) was obtained from J&K Chemical Ltd (Beijing, China). 1-
Chlorobutane (98%), 1-chlorohexane (99%), and 1-bromobutane
H
(
99%) were purchased from ABCR Gmbh & Co. (Karlsruhe, Ger-
H
1
many) and were freshly distilled before use. Hydrochloric acid
M
0
is the total mass of the reaction solution, and M is the mass of
1
(
36 wt %), sulfuric acid (98 wt %), nitric acid (65ꢀ68 wt %), phos-
sample.
phoric acid (85 wt %), maleic acid (99%), acetic acid (99%), citric
acid (99%), allyl chloride (95%), pyridine (99%), and other chemicals
were all supplied by local suppliers. Unless otherwise specified,
acids were directly used as received. NMR spectra were measured
The results obtained by this quantification assay were in good
agreement with the isolated yields achieved by column chroma-
tography on silica gel.
1
with a Bruker DRX-400 spectrometer (400 MHz for H, 100 MHz
for C).
2
.5. Fructose quantification procedure
1
3
Fructose concentration was determined through DNS method
described elsewhere.
2
0
2
.2. Preparation of ionic liquids
1
-Butyl-3-methylimidazolium chloride ([C
methylimidazolium bromide ([C mim]Br), 1-ethyl-3-methyl-imi-
dazolium bromide ([C mim]Br), N-butyl pyridinium chloride
[C Py]Cl), 1-allyl-3-methylimidazolium chloride ([Amim]Cl), 1-
butyl-3-methylimidazolium bisulfate ([C mim]HSO ), 1-(4-sul-
fobutyl)-3-methylimidazolium bisulfate ([Sbmim]HSO ), 1-butyl-
-methylimidazolium tetrafluoroborate ([C mim]BF ), 1-butyl-3-
methyl imidazolium hexafluorophosphate ([C mim]PF ), and 1-
hexyl-3-methylimidazolium chloride ([C mim]Cl), were prepared
4
mim]Cl), 1-butyl-3-
3
3
. Results and discussion
.1. Preliminary dehydration of fructose in ILs
The dehydration of 10 wt % fructose in [C mim]Cl in the pres-
4
2
(
4
4
4
4
4
ence of 9 mol % hydrochloric acid was studied initially. The reac-
tion was completed in 8 min at 80 °C and HMF was obtained in
3
4
4
4
6
9
7% yield (Table 1, entry 1). By comparison, performing the same
reaction without acid resulted in a HMF yield of only 65% after
00 min (entry 2). The dehydration reaction catalyzed with nitric
6
1
9
according to the methods described in our previous work. All IL
samples were dried under vacuum at 90 °C for 3 d before use.
9
+
acid (5 min) and sulfuric acid (3 min) (both with an H concentra-
tion close to 9 mol % of fructose) at 80 °C afforded HMF in 93% and
2
.3. Typical procedure for fructose dehydration
9
1% yields, respectively (entries 3 and 4), indicating that strong
mineral acids were near equally effective. This suggests that a cat-
alytic amount of strong mineral acids in [C mim]Cl is an excellent
After an appropriate amount of fructose was dissolved in
4
4
[C mim]Cl (4.0 g) at 80 °C, hydrochloric acid (9 mol %, based on
fructose) was added. The reaction mixture was stirred at 80 °C;
samples were withdrawn at various time intervals, weighed (re-
1
corded as M , usually 50 mg), quenched with cold water, and sub-
jected to analysis. Alternatively, the entire reaction mixture was
loaded on silica gel and purified by column chromatography (ethyl
acetate/petroleum ether = 1: 10 to 1: 1) to afford HMF as a yellow
oil, which was transformed to a yellow solid upon storage below
0
1
system for the selective dehydration of fructose to HMF. Interest-
ingly, the reaction was quite slow in the presence of a weaker acid,
phosphoric acid (entry 5), indicating that the nature of the acid
played an important role. We found that organic acids also cata-
lyzed the reaction effectively. Under otherwise identical condi-
tions, the HMF yield was 88% after 50 min in the presence of
7
.6 mol % maleic acid (entry 6) and 78% after 720 min in the pres-
ence of acetic acid (entry 7).
Excellent results in [C mim]Cl encouraged us to test other ILs.
1
°C. H NMR (400 Hz, CDCl
3
): d 9.70 (s, 1H), 7.35 (d, J = 2.8 Hz,
4
13
3
H), 6.65 (d, J = 2.8 Hz, 1H), 4.84 (s, 2H). C NMR (100 Hz, CDCl ):
As shown in Table 1, regardless of the type of cation in the IL,
the reactions catalyzed with hydrochloric acid in halide anion-con-
d 178.2, 161.2, 152.8, 123.4, 110.5, 58.0.
4 4 2
taining ILs, [C mim]Br, [C Py]Cl, [C mim]Cl, [Amim]Cl, and
[C mim]Cl, afforded reasonably high HMF yields, although the
6
2
.4. HMF quantification procedure
optimal reaction time varied (entries 8–12). However, reactions
The water-quenched samples were neutralized with 0.5 mol/L
NaOH and centrifuged at 10,000 rpm for 5 min to give the superna-
tant (volume recorded as V ), from which a sample (25 L) was
pipetted and diluted with deionized water to a final volume of
in acidic ILs [Sbmim]HSO and [C mim]HSO tended to form insol-
4
4
4
uble products. On the other hand, if [Sbmim]HSO or [C mim]HSO
4
4
4
1
l
was applied as a catalyst at 5 mol % loading, the HMF yields
reached over 80% within 30 min (entries 13 and 14). These results
indicate that [Sbmim]HSO4 and [C mim]HSO behaved as strong
4
5
.025 mL. The HMF concentration was measured on a JASCO V-
30 Model spectrophotometer (JASCO Inc., Japan) at 282 nm using
4
4
acids, and likely promoted extensive side reactions when used as