Isomerization of glucose to fructose catalyzed by lithium bromide in water

Article abstract of DOI:10.1039/c7gc02145c

This study demonstrated that glucose could be isomerized to fructose in the concentrated aqueous solution of lithium bromide (LiBr) under mild conditions. The isomerization mechanisms were studied via isotopic labeling experiments. It was found that both the cation (Li⁺) and the anion (Br^-) in the system catalyzed the isomerization of glucose to fructose. Br^- catalyzed the isomerization through the proton transfer mechanism via an enediol intermediate, while Li⁺ did it through an intramolecular hydride shift mechanism from C2 to C1. The estimation using quantitative ¹³C-NMR analysis indicated that Br^- catalyzed approximately 85% of the isomerization, while Li⁺ was responsible for the remaining 15%. The results indicated that 31% of the fructose was produced from glucose in LiBr trihydrate at 120 °C for 15 min. The outcomes of this study provided not only a better understanding of and insights into the sugar transformations in the LiBr solution but also an alternative approach to produce fructose from glucose.

Full text of DOI:10.1039/c7gc02145c

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: C. Yoo, N. Li, M.
Swannell and X. Pan, Green Chem., 2017, DOI: 10.1039/C7GC02145C.
This is an Accepted Manuscript, which has been through the
Volume 18 Number
7 7 April 2016 Pages 1821–2242
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Accepted Manuscripts are published online shortly after
Cutting-edge research for a greener sustainable future
www.rsc.org/greenchem
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1463-9262
CRITICAL REVIEW
G. Chatel et al.
Heterogeneous catalytic oxidation for lignin valorization into valuable
chemicals: what results? What limitations? What trends?
rsc.li/green-chem

Page 1 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
Isomerization of Glucose to Fructose Catalyzed by Lithium Bromide in Water
Chang Geun Yoo†, Ning Li†, Melanie Swannell, and Xuejun Pan*
Department of Biological Systems Engineering, University of WisconsinꢀMadison, Madison, WI
53706, United States
† These authors contributed equally to this work.
* To whom the correspondence should be addressed. Email: xpan@wisc.edu
Abstract
This study demonstrated that glucose could be isomerized to fructose in the concentrated
aqueous solution of lithium bromide (LiBr) under mild conditions. The isomerization
mechanisms were studied via isotopic labeling experiments. It was found that both the cation
(Li+) and the anion (Br‾) in the system catalyzed the isomerization of glucose to fructose. The
Br‾ catalyzed the isomerization through the proton transfer mechanism via enediol intermediate,
while the Li+ did through the intramolecular hydride shift mechanism from C2 to C1. The
13
estimation using quantitative CꢀNMR analysis indicated that Br‾ catalyzed approximately 85%
of the isomerization, while Li+ was responsible for the rest 15%. The results indicated that 31%
of fructose was produced from glucose in LiBr trihydrate at 120 °C for 15 min. The outcomes of
this study provided not only better understanding and insights of the sugar transformations in the
LiBr solution but also an alternative approach to produce fructose from glucose.
1

Green Chemistry
Page 2 of 29
View Article Online
DOI: 10.1039/C7GC02145C
Introduction
Glucose is the most abundant carbohydrate in nature and the building unit of
polysaccharides such as cellulose and starch. It has been an essential feedstock in food, beverage,
pharmaceutics, and other industry sectors. It is also an important platform chemical for the
production of biofuels, biochemicals, and biomaterials. For example, glucose can be fermented
to ethanol, butanol, and lactic acid, dehydrated and decomposed to 5ꢀhydromethylfurfural (HMF)
and levulinic acid (LA), and isomerized to fructose.1ꢀ4
Fructose, an isomer of glucose, shares the same elemental composition, but has different
chemical structure, and thereby exhibits different characteristics from glucose. Fructose is a
widely used sweetener in food and beverage industries, because it is the sweetest naturally
occurring carbohydrate, twice as sweet as glucose.5 Fructose is also an excellent humectant,
helping to extend the shelfꢀlife of products by improving moisture retention6 and being used in
skin and hair moisturizing products.7 In addition, fructose is considered as a crucial intermediate
from glucose to furanꢀbased chemicals.8
Isomerization of glucose is the dominant approach to produce fructose. A variety of
catalysts and processes have been explored and developed to isomerize glucose to fructose,
which were nicely summarized and compared in a recent review.9 Biocatalytic conversion using
immobilized enzyme is the primary commercial method for glucose isomerization.10 This is a
highly selective method; however, high enzyme cost, restrictive reaction conditions (pH and
temperature), and periodic replacement of deactivated enzyme are major drawbacks. Microbial
contamination is also a potential problem of the enzymatic process. Direct isomerization of
glucose in an alkaline solution is another approach despite some operational and environmental
issues.11 Solid catalysts, including solid base and solid Lewis acid catalysts, were recently
2

Page 3 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
introduced to isomerize glucose to fructose.12ꢀ14 For example, RománꢀLeshkov et al. reported
that tin in solid Lewis acid effectively enhanced the isomerization of glucose.15 Yang et al.
designed heterogeneous base catalyst with good stability and recyclability.16 Although these
catalysts showed impressive performance in glucose isomerization, the synthesis and high cost of
these solid catalysts need to be improved. Besides these catalysts, ionic liquids and halide salts
could enhance the isomerization of glucose to fructose as well and thereby improve HMF yield.17
However, ionic liquids are not economically attractive yet, and the processes face technical
challenges in solvent recycling and product separation.
Glucoseꢀtoꢀfructose transformation/isomerization has drawn significant attention in the
area of biomass conversion, because fructose is a crucial intermediate in the glucose conversion
to furanꢀbased chemicals, such as HMF and subsequent derivatives, which can be further
converted to biofuels and biochemicals. Studies indicated that fructose could be more easily,
quickly, and completely dehydrated to furanꢀbased chemicals than glucose, suggesting that
effective isomerization of glucose to fructose could improve the yield and productivity of furanꢀ
based chemicals from biomass glucose.18, 19
In our previous studies, the conversion of lignocellulosic biomass in molten halide salt
hydrates (highꢀconcentration aqueous salt solutions, such as LiBr·3H2O) to furanꢀderived
precursors of hydrocarbon fuels20 and furanꢀbased chemicals19 was investigated. It was observed
that the isomerization of glucose to fructose was a crucial step in producing the furanꢀbased
precursors and chemicals from biomass sugars. Therefore, the objective of this study is to
investigate how glucose is isomerized to fructose in the molten salt hydrates, specifically in
lithium trihydrate (LiBr·3H2O). The outcomes of this study are expected to provide insights for
better understanding the glucose transformation to fructose in the molten salt hydrate and clues
3

Green Chemistry
Page 4 of 29
View Article Online
DOI: 10.1039/C7GC02145C
for improving the products yield in the aforementioned two processes. In addition, this study has
the potential to establish an alternative approach for producing fructose from glucose. Our two
major working hypotheses are that (1) glucose could be isomerized to fructose in the molten salt
hydrate (such as LiBr·3H2O) and (2) the isomerization of glucose to fructose is catalyzed by both
the cation (such as Li+) and the anion (such as Br‾) in the system. In this study, the isomerization
of glucose to fructose in different molten salt hydrates were first demonstrated and verified. Then,
the reaction conditions such as time, temperature, and type and concentration of the salt were
investigated and optimized to improve the fructose yield. Isotopic labeling experiments were
conducted to verify the proposed isomerization pathways from glucose to fructose in the molten
salt hydrate.
Experimental
Chemicals
Dꢀglucose, Dꢀmannose, and fructose were purchased from SigmaꢀAldrich (St. Louis,
MO). Lithium chloride (LiCl), lithium bromide (LiBr), lithium iodine (LiI), calcium bromide
(CaBr2), and zinc chloride (ZnCl2) were obtained from VWR (Radnor, PA). Furfural and HMF
were purchased from Fisher Scientific (Pittsburgh, PA) and VWR, respectively. Deuterium oxide
(D2O) and deuterated glucose (Dꢀglucoseꢀ2ꢀd1, deuterated at C2) were purchased from Sigmaꢀ
Aldrich (St. Louis, MO).
Isomerization of glucose in molten halide salt hydrates
Different molten salt hydrates (concentrated aqueous salt solutions), including LiCl, LiBr,
LiI, CaBr2, and ZnCl2, at varying saltꢀtoꢀwater molar ratios from 1:2 to 1:6 were investigated for
glucose isomerization. In brief, glucose (50 mg) was loaded with 5 mL of the molten halide salt
4

Page 5 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
hydrate (equivalent to 10 g/L) in a 20ꢀmL glass reactor with a magnetic stirring bar. The
isomerization was conducted at 110 – 140 °C for 5 – 120 min with stirring at 400 rpm. After the
isomerization reaction, the reactor was cooled down in a water bath for 10 min, and then the content
was diluted for quantitation of sugars using a high performance ion chromatography (HPIC), as
described below.
Analytical methods
Quantitation of glucose and its isomers including fructose and mannose was conducted
using a HPIC system (Dionex ICSꢀ3000, Sunnyvale, CA) equipped with an integrated
amperometric detector and CarboPac® PA1 guard and analytic columns (Dionex, Sunnyvale,
CA). Sugar dehydration products such as HMF, furfural, and organic acids (lactic, formic, and
levulinic acids) were quantitated using a high performance liquid chromatography (HPLC)
system (Dionex, Sunnyvale, CA) with a Supelcogel Cꢀ610H column and an ultraviolet (UV)
detector at 210 nm. The yield of each monosaccharide was calculated with the following
equation:
ꢇ
ꢎ
ꢀꢁꢂꢃꢄꢅꢆ ꢄꢈ ꢉꢈꢂꢃꢊꢄꢅꢆ ꢄꢈ ꢋꢌꢍꢍꢄꢅꢆ ꢏꢐꢆꢁꢑ ꢇ%ꢎ
ꢒꢈꢄꢑꢂꢃꢆꢑ/ꢈꢆꢓꢌꢐꢍꢆꢑ ꢓꢄꢍꢄꢅꢌꢃꢃℎꢌꢈꢐꢑꢆꢅ ꢌꢔꢊꢆꢈ ꢈꢆꢌꢃꢊꢐꢄꢍ ꢇꢕꢎ
=
× 100
ꢖꢍꢐꢊꢐꢌꢁ ꢕꢁꢂꢃꢄꢅꢆ ꢇꢄꢈ ꢔꢈꢂꢃꢊꢄꢅꢆ ꢄꢈ ꢓꢌꢍꢍꢄꢅꢆꢎ ꢁꢄꢌꢑꢐꢍꢕ ꢇꢕꢎ
Nuclear magnetic resonance (NMR) experiments
Two isotopic labeling experiments were conducted. In the first experiment, the unlabeled
Dꢀglucose was isomerized at 110 °C for 15 min with a 10% glucose loading in deuterated LiBr
trihydrate (prepared by dissolving LiBr in D2O). In the second experiment, deuterated glucose at
C2 position (Dꢀglucoseꢀ2ꢀd1) was isomerized in LiBr trihydrate under the same condition. After
5

Green Chemistry
Page 6 of 29
View Article Online
DOI: 10.1039/C7GC02145C
quenching the reactions above, the sugars in the reaction solution were separated from LiBr
using a column (400 mm × 15 mm) packed with strong cation exchange resin Dowex 50W×2
(Li+ form, 200ꢀ400 mesh, Dow Chemical Co., MI) on an Isolera Spektra One flash
chromatographic system (Biotage, Charlotte, NC). The collected sugar fractions were
concentrated using a rotary evaporator and then freezeꢀdried. The dried sugars were dissolved in
1
13
D2O for nuclear magnetic resonance (NMR) analysis. H and C spectra were obtained with a
Bruker AVANCE 500MHz NMR spectrometer equipped with a DCH cryoprobe. The
1
longitudinal relaxation (T1) of monosaccharides was 2.1 and 1.1 s in H and 13C dimensions,
1
respectively. H NMR parameters: pulse program (zg30), acquisition time = 3.276 s, number of
scan = 64, relaxation delay (D1) = 6.3 s; standard 13C NMR parameters: pulse program (zgpg30),
acquisition time = 1 s, number of scan = 2048, D1 = 1 s; quantitative 13C NMR parameters: pulse
program (zgig30), acquisition time = 1s, number of scan = 2048, D1 = 6 s.
Results and discussion
Isomerization of glucose in molten halide salt hydrates
A molten salt hydrate is an aqueous solution of an inorganic salt at high concentration.
The concentration is determined by the coordination number of the salt cation.21 In a perfect
molten salt hydrate, all water molecules are tightly bound to the inner coordination sphere of the
cation of the salt, leaving the anion naked (uncoordinated) in the system.22 Some molten salt
hydrates were found to have the ability to dissolve cellulose, such as ZnCl2ꢁ3ꢀ4H2O, LiClO4ꢁ3H2ꢀ
O, LiIꢁ2H2O, LiCl⋅3H2O, CaBr2⋅6H2O, and LiBrꢁ3H2O.22ꢀ26 In particular, LiBrꢁ3H2O has been
used as a medium for hydrolysis of cellulose,27 direct saccharification of lignocellulosic
biomass,28, 29 and depolymerization of lignin.30
6

Page 7 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
To investigate the glucose isomerization in molten salt hydrates, five halide salts, LiCl,
LiBr, LiI, CaBr2, and ZnCl2 were tested. Their molten salt hydrates were prepared as following:
LiCl⋅3H2O, LiBr⋅3H2O, LiI⋅3H2O, CaBr2⋅6H2O, and ZnCl2⋅3H2O, respectively. The results of the
isomerization of glucose in these halide salt hydrates under the same conditions are presented in
Table 1. In general, it seemed that the bromides (LiBr and CaBr2) were more efficient than the
iodide (LiI) and chlorides (LiCl and ZnCl2) in glucose isomerization, and Li+ performed better
than Ca2+ and Zn2+. In addition to the isomerization, the bromides and iodide (but not the
chlorides) resulted in the epimerization of glucose to mannose.
It was observed that glucose isomerization was affected by the halide anions in the
systems. For example, LiCl hydrate only isomerized 7.7% glucose to fructose and caused the
limited degradation (to HMF, levulinic acid, etc.) of glucose (>90% glucose retained) in 15 min.
Extending reaction time to 30 min increased the fructose yield to 17.4%, but further prolonging
the reaction to 60 min did not lead to higher fructose yield (Table S1 in ESI). On the other hand,
LiBr hydrate gave much higher fructose yield (>30%) and also caused more glucose degradation
than LiCl hydrate. Accordingly, it was reasonably expected that LiI might give higher fructose
yield than LiBr. However, it turned out that LiI led to lower fructose yield and much more
glucose degradation than LiBr. This could be attributed to the fact that LiI is unstable and easily
oxidized to iodine in air in particular at elevated temperature (2I‾ + H2O + 1/2O2
→ I2 + 2OH‾),
which was evidenced by the observation that the clear and transparent LiI solution became
brown at the end of reaction (Fig. S1 in ESI). In addition, the formed OH‾ from LiI
decomposition might affect the glucose isomerization. To verify this hypothesis, the
isomerization of glucose in LiCl⋅3H2O and LiBr⋅3H2O with addition of NaOH was conducted,
and results are summarized in Table S2 in ESI. No sugar or only negligible amount of sugars
7

Green Chemistry
Page 8 of 29
View Article Online
DOI: 10.1039/C7GC02145C
(glucose, fructose, and mannose) was detected in the products. Preliminary investigation
indicated that the decomposition products of glucose in basic LiCl and LiBr solutions were
mainly lactic, formic, and levulinic acids. It seems that the conversion of glucose catalyzed by
base is greatly dependent on reaction systems and conditions (such as temperature, reaction time,
and glucose concentration). For example, NaOH in water (0.1 M) could catalyze glucose
isomerization to fructose at 110 °C.15 Bases could also catalyze glucose conversion to organic
acids under harsh conditions (temperature 200ꢀ300 °C).31ꢀ34 A recent study found that glucose
could be quantitatively converted into lactic acid by dilute bases (such as NaOH, KOH, and
Ba(OH)2) at room temperature (25 °C),35 and the authors also reported that when initial glucose
concentration was below 0.05 M, glucose was almost quantitatively converted to lactic acid
(yield >92%) by dilute base (0.25 M Ba(OH)2), while when the glucose concentration was
increased to 1 M, isomerization became dominant, leading to 45% fructose with <1% lactic acid.
It seemed that the cations also had an effect on the glucose isomerization. For example, as
shown in Table 1, the aqueous CaBr2 solution gave a lower fructose yield and resulted in more
sugar degradation than the aqueous LiBr solution did. However, this was not sufficient to
conclude that Li+ is more efficient than Ca2+ in catalyzing glucose isomerization because the
concentration of the former was twice of the latter in the system (LiBr⋅3H2O and CaBr2⋅6H2O,
respectively) in this study. Differently, in the ZnCl2 hydrate, no glucose isomer (fructose) or
epimer (mannose) was detected, but significant amounts of decomposed products were observed.
These results suggested that the cation of the salts might play an important role in isomerization
of glucose, which needs further investigation.
Comparing the average glucose conversion rate and fructose formation rate during the
first 15 min in different molten salt hydrates (Table S3 in ESI), one can find that the fructose
8

Page 9 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
formation rate in LiBr hydrate was the highest (1.9×10ꢀ3 M/s). The glucose conversion was
faster in LiI and CaBr2 hydrates (5.6×10ꢀ3 and 5.9×10ꢀ3 M/s, respectively), but these did not
turn out as higher fructose rate (1.3×10ꢀ3 and 1.4×10ꢀ3 M/s, respectively).
The observation and discussion above suggested that both cation and anion in the molten
salt hydrates influenced the isomerization of glucose to fructose. Among the salts investigated,
LiBr had the best performance in catalyzing the glucose isomerization; therefore, the following
investigations focused on LiBr hydrate system.
Effect of reaction conditions on glucose isomerization in LiBr hydrate
For better understanding of the effect of LiBr concentration on glucose isomerization, the
LiBr solutions with varying molar ratios of LiBrꢀtoꢀwater (1:2 – 1:6) were tested under the same
reaction conditions (1wt.% glucose loading at 120 °C for 15 min). As shown in Fig. 1, increasing
LiBr concentration (reducing the LiBrꢀtoꢀwater ratio) resulted in higher glucose conversion.
Fructose yield increased from 19% to 31% when the LiBr concentration was elevated from 44.6%
to 61.7% (w/w) (correspondingly, LiBrꢀtoꢀwater molar ratio changed from 1:6 to 1:3); however,
further increasing LiBr concentration to 70.7% (w/w) (LiBrꢀtoꢀwater molar ratio1:2) did not
improve the fructose yield. A small amount (up to 1.5%) of mannose, a glucose epimer, was
detected in the products. In addition, high LiBr concentration accelerated the decomposition of
glucose to furans and organic acids (19ꢀ32%). These results and observations suggested that
lithium bromide trihydrate (LiBr⋅3H2O, 61.7%), at which LiBr forms a molten salt hydrate,
seemed to have a balanced performance (high fructose yield and relatively low decomposition
products).
The effects of reaction temperature and time on the glucose isomerization in LiBr⋅3H2O
9

Green Chemistry
Page 10 of 29
View Article Online
DOI: 10.1039/C7GC02145C
were investigated at four temperatures (110 – 140 °C) for 5 – 120 min. The product profiles at
each reaction temperature are presented in Fig. 2 (aꢀd). When the reaction proceeded, glucose
content in the system sharply decreased in the first 5 – 15 min, while the fructose content
increased accordingly and quickly reached a maximum (~30%). The average glucose conversion
rate and fructose formation rate at different reaction times and temperatures are summarized in
Table S4 in ESI. The rate results suggest that the isomerization of glucose to fructose was very
fast in the first 15 min, in particular in the first 5 min, beyond which the reaction slowed down
significantly. It was also observed that both glucose conversion and fructose formation were
enhanced by elevating reaction temperature. After the fast reaction phase, the consumption
(conversion) of glucose slowed down, and fructose content in the system did not increase further
but contrarily decreased, in particular at higher temperature. Meanwhile, mannose content in the
system kept increasing with reaction time. These observations suggested that (1) glucose
isomerization in LiBr⋅3H2O occurred readily under moderate conditions (Fig. 2a and 2b); (2)
fructose in the system was not stable, and overlong reaction, in particular at higher temperature,
hurt the yield of fructose. For example, no fructose could be detected at 140 °C and 120 min (Fig.
2d). Fructose might be either isomerized to mannose or further dehydrated to furans and
subsequently condensed to humins; (3) mannose was formed by either from the epimerization of
glucose or from the isomerization of fructose; and (4) in addition to fructose and mannose,
glucose was transformed to other nonꢀcarbohydrate products as well, such as furans, organic
acids, and unidentified compounds. The glucose degradation occurred dominantly under severe
conditions. For example, only less than 30% sugars (glucose, mannose, and fructose) were
detected in the system at 140 °C and 120 min (Fig. 2d). In summary, moderate temperature (such
as 120 °C) was favorable, because it could reach the maximum fructose yield within short time
10

Page 11 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
(<15 min) and meanwhile reduced the excessive degradation of the sugars. The highest fructose
yield (31%) was observed at 120 °C and 15 min, and this yield is comparable with the fructose
yields (20ꢀ45%) in the previous studies.12, 36ꢀ41 Excessive extension of reaction time or elevation
of temperature accelerated the degradation of the sugars.
The effect of higher starting glucose concentration on the isomerization was also
examined under the same reaction conditions of 120 °C and 15 min (Fig. S2 in ESI). Increasing
glucose concentration from 10 g/L to 20 g/L did not affect the fructose yield (31%), but further
raising the glucose concentration to 50 g/L and 100 g/L led to reduced fructose yield, 25% and
24%, respectively.
Sugar decomposition products were further investigated at 130 °C in the LiBr⋅3H2O. As
shown in Fig. 3, a significant amount of HMF was detected along with small amount of
bromomethylfurfural (BMF) and furfural. It is well known that HMF is formed from the
dehydration of glucose via fructose as in intermediate, and fructoseꢀtoꢀHMF transformation is
much readier and faster than glucoseꢀtoꢀHMF transformation.42 The HMF profile was well
correlated to the fructose profile in Fig. 2c. For example, fructose yield reached the maximum
(~30%) at 5 min and then dropped quickly with time (Fig. 2c), and while HMF was detected at
about 5 min and increased gradually with the reaction (Fig. 3). Less than 5% fructose retained in
the system at 120 min, while approximately 14% HMF was observed correspondingly. These
observations suggested that part of the generated fructose was dehydrated to HMF during the
reaction. Small amount of BMF (<1%) was observed in the system, which was formed via the
bromination of HMF, as reported in our previous study.19 Interestingly, trace amount of furfural was
detected as well. Generally, furfural is formed from the dehydration of pentoses, but formation of
furfural from glucose was also reported in previous studies.43, 44 In addition, under severe conditions
(such as high temperature and long reaction time), significant amount of precipitates were observed
11

Green Chemistry
Page 12 of 29
View Article Online
DOI: 10.1039/C7GC02145C
in the system, which were likely the humins from the condensation of sugar degradation products.45
Pathways of glucose isomerization in LiBr⋅3H2O
The results above explicitly demonstrated the formation of fructose via glucose
isomerization in the LiBr solution. However, the role of LiBr in the glucose isomerization was
still ambiguous. It has been known that glucose isomerization in water could be catalyzed either
by base catalysts via a proton transfer mechanism through an enediol intermediate or by Lewis
acid catalysts (such as metal ions) via an intramolecular C2 to C1 hydride shift mechanism.15, 46ꢀ
48
Examining the aqueous LiBr solution used in this study, it contains only Br‾ and Li+ in
addition to water. It is reasonable to hypothesize that both Br‾ and Li+ in the solution could
catalyze the isomerization of glucose to fructose. Metal ions of soluble Lewis acids (such as
CrCl3, SnCl4, and AlCl3) were reported to catalyze the isomerization of glucose to fructose in
water through the intramolecular hydride shift from C2 to C1.37, 38, 49ꢀ54 However, to our
knowledge, the role of the halogen anions of the Lewis acids (such as Br‾ and Cl‾) in glucose
isomerization in water was not reported yet. Herein, we proposed two plausible isomerization
pathways of the glucose isomerization in LiBr trihydrate (LiBr⋅3H2O), as summarized in Fig. 4.
In Path A, free halide ions (Br‾) abstract the H on C2 of glucose, resulting in a cisꢀ1,2ꢀenediol
intermediate.55, 56 The rearrangement of the enediol intermediate generates fructose, accompanied
with a proton transfer from medium (LiBr solution) to C1. Path B goes through an intramolecular
hydride transfer from C2 to C1 similar to that catalyzed by metal cations.56ꢀ58 Li+ is involved in a
simultaneous coordination with the hydroxyl group on C2 and the carbonyl group on C1,
forming a 5ꢀmember complex. This complex facilitates the subsequent intramolecular hydride
12

Page 13 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
shift from C2 to C1 to transform aldose (glucose) to ketose (fructose). Since Br‾ is generally
recognized as a weak Lewis base, we initially hypothesized that the Path A would be a minor
mechanism for glucose isomerization to fructose in the LiBr solution. However, this was proved
not true according to the following results.
To verify the proposed pathways in Fig. 4, two isotopic labeling experiments were designed
and conducted. The first experiment was designed to confirm Path A, in which unlabeled glucose was
treated in deuterated LiBr solution (prepared by dissolving LiBr in D2O). If Path A is valid,
hydrogenꢀ2 (deuterium) should be detected on C1 of resultant fructose because of the proton transfer
from the reaction medium (D2O) to C1 of the fructose during the isomerization, as shown in Fig. 4
(top). The 13CꢀNMR spectra of labeled glucoseꢀ2ꢀd1 (Fig. 5a), unlabeled glucose (Fig. 5b), unlabeled
fructose (Fig. 5c), and the sugars obtained after the isomerization of unlabeled glucose in deuterated
LiBr solution (Fig. 5d) are shown in Fig. 5. When glucose was labeled by deuterium at C2, the C2
carbon resonance changed from two singlets at δ = 74.1 ppm (βꢀpyranose) and 71.4 ppm (αꢀ
pyranose) in the unlabeled glucose (Fig. 5b) to two triplets with low intensity at δ = 73.7 ppm (βꢀ
pyranose) and 71.0 ppm (αꢀpyranose) (Fig. 5a), respectively. The multiplet changes were caused
by the broadband proton decoupling method universally applied to the standard 13C NMR
13
analysis. The decoupling method greatly simplified the C spectra by effectively removing the
1
coupling from H, but was ineffective to D (2H). As a result, the C2 carbon was coupled by a
deuterium with spin = 1 during FID acquisition, changing both the chemical shifts and
multiplities. The deuteriumꢀbonded carbon also had low resonance intensity (approximately 50%
reduction) due to the ineffective nuclear overhauser enhancement (NOE) effect from the
broadband decoupling (continuously saturating the proton spin).59 After the isomerization
reaction, no resonance change of the unreacted glucose was observed in the NMR spectra (Fig.
5d and Fig. S3 in ESI), indicating that there was no deuterium transfer/exchange from deuterated
13

Green Chemistry
Page 14 of 29
View Article Online
DOI: 10.1039/C7GC02145C
LiBr solution to the carbons of glucose. As expected, the corresponding 13C resonance at δ = 63.8
ppm (βꢀfuranose) and 62.6 ppm (βꢀpyranose) was greatly attenuated and a new triplet peak at δ =
63.5 ppm appeared, which proved the presence of the deuterium on C1 of the fructose yielded
from the isomerization of unlabeled glucose (Fig. 5d, top right offset).15, 50 The reason why the
resonance did not disappear completely, as observed previously,15 was due to the competition of
Path B, which will be further discussed below. Since the glucose was not deuterated via
deuterium transfer/exchange with the deuterated LiBr solution, it is apparent that the deuterium
on the C1 position of fructose was transferred from the deuterated LiBr solution during the
rearrangement of the enediol intermediate to fructose. These results and observations supported
the validity of Path A.
In the second experiment, isotopically labeled glucose (Dꢀglucoseꢀ2ꢀd1) was treated in
unlabeled LiBr solution. If the proposed Path B is valid, deuterium on the C2 position of the
glucose should be transferred onto C1 of resultant fructose because the intramolecular hydride shift
from C2 to C1 during the isomerization, as proposed in Fig. 4 (bottom). The 13CꢀNMR spectra of
labeled glucoseꢀ2ꢀd1 (Fig. 6a), unlabeled glucose (Fig. 6b), unlabeled fructose (Fig. 6c), and the
sugars obtained after the isomerization of labeled glucose in unlabeled LiBr solution (Fig. 6d) are
shown in Fig. 6. After the isomerization of glucoseꢀ2ꢀd1 in nonꢀdeuterated LiBr solution, the
carbon resonance of the unreacted glucoseꢀ2ꢀd1 in the system was identical to the starting
glucoseꢀ2ꢀd1 (top left offset in Fig. 6). This further confirmed that the CꢀD (or CꢀH) bond of
glucose was inert (no direct exchange of deuterium/hydrogen between glucose and reaction
medium) in the system under the reaction condition unless the isomerization occurred. If the
isomerization proceeds via the intermolecular 1,2 hydride shift mechanism (Path B in Fig. 4),
fructose produced from glucoseꢀ2ꢀd1 should have a deuterium labeled on the C1 position, which
results in the disappearance of signals at δ = 63.8 ppm and 62.6 ppm, as discussed above and in
14

Page 15 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
the previous study.15 In deed, decrease of the resonance signals at δ = 63.8 ppm and 62.6 ppm
was observed from the resultant fructose (top right offset in Fig. 6 d), although the decrease was
not as significant as in the first isotopic labeling experiment (Fig. 5 top right offset). The
observation implicated the presence of deuterium at C1 of fructose. The small decrease in
resonance at δ = 63.8 in the second experiment should be attributed to the low proportion of
fructoseꢀ1ꢀd1 in the total fructose produced due to the competition between the two paths of the
isomerization (Fig. 4). As discussed above, when glucoseꢀ2ꢀd1 is isomerized in unlabeled LiBr
solution, Path A leads to nonꢀdeuterated fructose at C1, and Path B leads to fructoseꢀ1ꢀd1. The
coꢀexistence of the two pathways was the reason why the C1 of the resultant fructose was not
fully deuterated in both of the isotopic labeling experiments, and thereby the NMR signals at δ =
63.8 ppm and 62.6 ppm did not completely disappear, as they should.15 The presence of nonꢀ
1
deuterated fructose was also supported by the H NMR analysis (Fig. S4 in ESI) of the products
from the isomerization of glucoseꢀ2ꢀd1 in unlabeled LiBr solution from the second experiment, in
which C1 proton of resultant fructose was detected at δ = 3.56 ppm and 3.54 ppm, respectively,
implying that C1 was not completely deuterated. The results and observations above suggested
that Path A seemed to be the dominant pathway for glucose isomerization to fructose in LiBr
solution, and Path B was the minor one.
To quantitatively estimate the relative contribution of Path A and Path B, the following
13
experiment was conducted. Since standard C NMR analysis is unable to reliably quantitate the
fructoseꢀ1ꢀd1, the inverse gated coupled quantitative 13C NMR was conducted to estimate the
selectivity (proportion) of fructoseꢀ1ꢀd1 in the total fructose formed from the glucoseꢀ2ꢀd1
isomerization. The integrals of C1 signals of nonꢀdeuterated fructose at δ = 63.8 ppm (pyranose)
and 62.6 ppm (furanose) were quantitated using the integrals of C6 signals of both fructoseꢀ1ꢀd1
15

Green Chemistry
Page 16 of 29
View Article Online
DOI: 10.1039/C7GC02145C
and nonꢀdeuterated fructose at δ = 63.3 ppm (pyranose) and 62.3 ppm (furanose), respectively, as
the references, as shown in Figure S5 in ESI. If no fructoseꢀ1ꢀd1 was formed via Path B, the C1
integral should be the same as the C6 integral. It turned out that the C1 integrals were 15.1 and
13.7% less than the corresponding C6 integrals in pyranose and furanose, respectively. These
results indicated that approximately 15% of the fructose had a deuterium on the C1 position,
which was from the glucoseꢀ2ꢀd1 isomerization via Path B. To verify the validity and reliability
of the estimation method, quantitative 13C NMR of unlabeled fructose in D2O was also
conducted, and the results are given in Fig. S6 in ESI. As expected, the C1 integral and C6
integral in unlabeled fructose were almost the same, confirming the validity of the method.
In summary, the isotopic labeling NMR experiments above verified that the glucose
isomerization to fructose in lithium bromide trihydrate was catalyzed by both Br‾ via Path A and
Li+ via Path B. The former pathway was responsible for approximately 85% of fructose, and the
latter one was for 15%. As mentioned above, Li+ in the molten salt hydrate system of LiBr
(LiBr⋅3H2O, equivalent to 61.7% LiBr in water) was highly hydrated, while Br‾ was left
coordinationꢀfree at high concentration in the system, which are notably different from those
cations and anions in dilute solutions of Lewis acids. The free Br‾ at high concentration in the
system was probable the reason why Br‾ dominated the catalysis of glucose to fructose.
Equilibrium of glucose, fructose and mannose in LiBr⋅3H2O
Mannose is an isomer of glucose and fructose. As shown above in Fig. 2, mannose was
detected in the system when glucose was isomerized to fructose in LiBr trihydrate. For a better
understanding of transformation among the three isomers (glucose, fructose, and mannose), both
fructose and mannose were tested as starting material in LiBr⋅3H2O under the same conditions
16

Page 17 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
(120 °C for 5 min). It was observed that the isomerization of these three sugars was reversible,
and one sugar could be transformed to other two in the system. They could reach equilibrium in
the system. However, as shown in Fig. 7, the isomerization yield varied from sugar to sugar.
Compared to the isomerization of glucose and that of mannose (Figs. 2b and 7b, respectively),
the conversion of fructose to glucose and mannose (Fig. 7a) under the same conditions seemed to
be slow and gave lower yield of the products. For example, glucose and mannose could yield
~30% fructose with varying amount of mannose and glucose, respectively, while fructose only
generated ~15% glucose and ~5% mannose when treated in LiBr⋅3H2O under the same
conditions. These were supported by the average conversion rates of the sugars in the first 5 min
(Table S5 in ESI). For example, the conversion rate of fructose (4.2×10ꢀ3 M/s) was slower than
those of glucose and mannose (6.4×10ꢀ3 and 7.2×10ꢀ3 M/s, respectively). The fructose
formation rates from glucose (0.45 ×10ꢀ4 M/s) and mannose (0.49 ×10ꢀ4 M/s) were significantly
higher than those of glucose formation (0.22 ×10ꢀ4 M/s) and mannose formation (0.08 ×10v M/s)
from fructose under the same reaction conditions. These observations suggested that more
fructose was degraded to other products rather than isomerized to glucose and mannose. These
results also explain why fructose yield rapidly reached a maximum, and then declined over time
during the glucose isomerization (Fig. 2). Both glucose and mannose gave fructose as a primary
product, while fructose yielded more glucose but less mannose. Interestingly, the behavior of
mannose in LiBr⋅3H2O was apparently different from that of glucose and fructose. The yield of
glucose from mannose isomerization continued increasing with time, and as a result, the system
had more glucose than mannose and fructose at 120 min (Fig. 7b).
Conclusion
17

Green Chemistry
Page 18 of 29
View Article Online
DOI: 10.1039/C7GC02145C
This study successfully demonstrated that glucose could be isomerized to fructose in
concentrated aqueous solution of lithium bromide (LiBr) alone as solvent and catalyst under mild
conditions. The results indicated that up to 31% of glucose was isomerized to fructose under the
investigated reaction conditions (120 °C for 15 min in lithium bromide trihydrate). The
isomerization mechanisms of glucose to fructose were studied via isotopic labeling experiments.
It was verified that not only the cation (Li+) but also the anion (Br‾) catalyzed the isomerization
of glucose to fructose, and the former catalyzed the isomerization through the intramolecular
hydride shift mechanism from C2 to C1, while the latter through the proton transfer mechanism
13
via enediol intermediate. The estimation using quantitative CꢀNMR analysis indicated that Br‾
catalyzed approximately 85% of the isomerization reaction, while Li+ was responsible for the
rest 15%. The outcomes of this study provided better understanding and insights of the sugar
transformations to fructose and further to furanꢀbased platform chemicals in LiBr molten salt
hydrate.19, 20 In addition, this study demonstrated an alternative isomerization approach to
produce fructose directly from glucose.
Acknowledgements
This work was supported by National Science Foundation grant CBET1159561 and the
USDA National Institute of Food and Agriculture, McIntire Stennis project 1006576.
References
1.
P. C. Badger, Trends in new crops and new uses, J. Janick and A. Whipkey (Eds.), ASHS
Press, Alexandria, VA, 2002, 17ꢀ21.
2.
3.
S. S. Tükel and D. Alagöz, Food Chem., 2008, 111, 658ꢀ662.
P. Mercier, L. Yerushalmi, D. Rouleau and D. Dochain, J. Chem. Tech. Biotechnol., 1992,
55, 111ꢀ121.
4.
R. Weingarten, J. Cho, R. Xing, W. C. Conner, Jr. and G. W. Huber, ChemSusChem, 2012,
18

Page 19 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
5, 1280ꢀ1290.
5.
6.
7.
8.
9.
R. M. Pangborn, J. Food Sci., 1963, 28, 726ꢀ733.
L. M. Hanover and J. S. White, Am. Soc. Clin. Nutr., 1993, 58, 724ꢀ732.
H. Hadi, S. Omar and A. Awadh, Br. J. Pharm., 2016, 1.
M. J. Antal, W. S. Mok and G. N. Richards, Carbohydr. Res., 1990, 199, 91ꢀ109.
I. Delidovich and R. Palkovits, ChemSusChem, 2016, 9, 547ꢀ561.
10. R. DiCosimo, J. McAuliffe, A. J. Poulose and G. Bohlmann, Chem. Soc. Rev., 2013, 42,
6437ꢀ6474.
11. M. H. Gaily, B. M. Elhassan, A. E. Abasaeed and M. AlꢀShrhan, Int. J. Eng. Tech., 2010, 10,
1ꢀ5.
12. M. Moliner, Y. RomanꢀLeshkov and M. E. Davis, Proc. Natl. Acad. Sci. USA, 2010, 107,
6164ꢀ6168.
13. S. Lima, A. S. Dias, Z. Lin, P. Brandão, P. Ferreira, M. Pillinger, J. Rocha, V. Calvinoꢀ
Casilda and A. A. Valente, Appl. Catal., A, 2008, 339, 21ꢀ27.
14. R. BermejoꢀDeval, R. S. Assary, E. Nikolla, M. Moliner, Y. RománꢀLeshkov, S.ꢀJ. Hwang,
A. Palsdottir, D. Silverman, R. F. Lobo and L. A. Curtiss, Pro. Natl. Acad. Sci, 2012, 109,
9727ꢀ9732.
15. Y. RomanꢀLeshkov, M. Moliner, J. A. Labinger and M. E. Davis, Angew. Chem. Int. Ed. ,
2010, 49, 8954ꢀ8957.
16. Q. Yang, S. Zhou and T. Runge, J. Catal., 2015, 330, 474ꢀ484.
17. J. B. Binder and R. T. Raines, J. Am. Chem. Soc., 2009, 1979ꢀ1985.
18. H. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang, Science, 2007, 316, 1597ꢀ1600.
19. C. G. Yoo, S. Zhang and X. Pan, RSC Adv., 2017, 7, 300ꢀ308.
20. L. Shuai and X. Pan, US Patent 9,487,712, 2016.
21. H. Leipner, S. Fischer, E. Brendler and W. Voigt, Macromol. Chem. Phys., 2000, 201, 2041ꢀ
2049.
22. S. Sen, J. D. Martin and D. S. Argyropoulos, ACS Sustain. Chem. Eng., 2013, 1, 858ꢀ870.
23. R. J. Wilcox, B. P. Losey, J. C. Folmer, J. D. Martin, M. Zeller and R. Sommer, Inorg.
Chem., 2015, 54, 1109ꢀ1119.
24. Y.ꢀJ. Yang, J.ꢀM. Shin, T. H. Kang, S. Kimura, M. Wada and U.ꢀJ. Kim, Cellulose, 2014, 21,
1175ꢀ1181.
25. N. Cao, Q. Xu and L. Chen, Appl. Biochem. Biotechnol., 1995, 51, 21ꢀ28.
26. S. Fischer, W. Voigt and K. Fischer, Cellulose, 1999, 6, 213ꢀ219.
27. W. Deng, J. R. Kennedy, G. Tsilomelekis, W. Zheng and V. Nikolakis, Ind. Eng. Chem. Res.,
2015, 54, 5226ꢀ5236.
28. X. Pan and L. Shuai, US Patent 9,187,790, 2015.
29. N. Li, X. Pan and J. Alexander, Green Chem., 2016, 18, 5367ꢀ5376.
30. X. Yang, N. Li, X. Lin, X. Pan and Y. Zhou, J. Agric. Food Chem., 2016, 64, 8379ꢀ8387.
31. X. Yan, F. Jin, K. Tohji, T. Moriya and H. Enomoto, J. Mater. Sci., 2007, 42, 9995ꢀ9999.
32. X. Yan, F. Jin, K. Tohji, A. Kishita and H. Enomoto, AIChE J., 2010, 56, 2727ꢀ2733.
33. D. Esposito and M. Antonietti, ChemSusChem, 2013, 6, 989ꢀ992.
34. Z. Huo, Y. Fang, D. Ren, S. Zhang, G. Yao, X. Zeng and F. Jin, ACS Sustain. Chem. Eng.,
2014, 2, 2765ꢀ2771.
35. L. Li, F. Shen, R. L. Smith and X. Qi, Green Chem., 2017, 19, 76ꢀ81.
36. V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N. S. Marinkovic, A. I.
Frenkel, S. I. Sandler and D. G. Vlachos, J. Am. Chem. Soc., 2013, 135, 3997ꢀ4006.
19

Green Chemistry
Page 20 of 29
View Article Online
DOI: 10.1039/C7GC02145C
37. S. Jia, K. Liu, Z. Xu, P. Yan, W. Xu, X. Liu and Z. C. Zhang, Catal. Today, 2014, 234, 83ꢀ
90.
38. J. Tang, X. Guo, L. Zhu and C. Hu, ACS Catal., 2015, 5, 5097ꢀ5103.
39. Y. C. Lee, C. T. Chen, Y. T. Chiu and K. C. W. Wu, ChemCatChem, 2013, 5, 2153ꢀ2157.
40. R. BermejoꢀDeval, R. S. Assary, E. Nikolla, M. Moliner, Y. RomanꢀLeshkov, S. J. Hwang,
A. Palsdottir, D. Silverman, R. F. Lobo, L. A. Curtiss and M. E. Davis, Proc. Natl. Acad.
Sci. USA, 2012, 109, 9727ꢀ9232.
41. Q. Yang and T. Runge, ACS Sustain Chem. Eng., 2016.
42. S. Eminov, A. Brandt, J. D. WiltonꢀEly and J. P. Hallett, PloS one, 2016, 11, e0163835.
43. T. M. Aida, Y. Sato, M. Watanabe, K. Tajima, T. Nonaka, H. Hattori and K. Arai, J.
Supercrit. Fluids, 2007, 40, 381ꢀ388.
44. M. Watanabe, Y. Aizawa, T. Iida, C. Levy, T. M. Aida and H. Inomata, Carbohydr. Res.,
2005, 340, 1931ꢀ1939.
45. S. J. Dee and A. T. Bell, ChemSusChem, 2011, 4, 1166ꢀ1173.
46. R. Nagorski and J. P. Richard, J. Am. Chem. Soc., 2001, 123, 794ꢀ802.
47. Y. RománꢀLeshkov and M. E. Davis, Acs Catal., 2011, 1, 1566ꢀ1580.
48. I. Delidovich and R. Palkovits, ChemSusChem, 2016, 547ꢀ561.
49. V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N. S. Marinkovic, A. I.
Frenkel, S. I. Sandler and D. G. Vlachos, J. Am. Chem. Soc., 2013, 135, 3997ꢀ4006.
50. H. Nguyen, V. Nikolakis and D. G. Vlachos, ACS Catal., 2016, 6, 1497ꢀ1504.
51. K. R. Enslow and A. T. Bell, Catal. Sci. Technol., 2015, 5, 2839ꢀ2847.
52. S. H. Mushrif, J. J. Varghese and D. G. Vlachos, Phys. Chem. Chem. Phys., 2014, 16,
19564ꢀ19572.
53. S. De, S. Dutta and B. Saha, Green Chem., 2011, 13, 2859ꢀ2868.
54. X. Zhang, P. Murria, Y. Jiang, W. Xiao, H. I. Kenttämaa, M. M. AbuꢀOmar and N. S.
Mosier, Green Chem., 2016, 18, 5219ꢀ5229.
55. C. A. Collyer and D. M. Blow, Proc. Natl. Acad. Sci. USA, 1990, 87, 1362ꢀ1366.
56. R. W. Nagorski and J. P. Richard, J. Am. Chem. Soc., 2001, 123, 794ꢀ802.
57. C. A. Collyer, K. Henrick and D. M. Blow, J. Mol. Biol., 1990, 212, 211ꢀ235.
58. M. Moliner, Y. RománꢀLeshkov and M. E. Davis, Proc. Natl. Acad. Sci. USA, 2010, 107,
6164ꢀ6168.
59. T. D. Claridge, Highꢀresolution NMR techniques in organic chemistry, Elsevier, Cambridge,
MA, 2016.
20

Page 21 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
Table 1. Products of glucose isomerization in different halide salt hydratesa
Halide salt Glucose [%]
Fructose [%]
Mannose [%] Decomposition productsb [%]
LiCl
LiBr
LiI
CaBr2
ZnCl2
90.5
48.2
9.8
5.6
67.1
7.7
30.3
21.4
22.4
0.0
0.0
1.4
3.8
2.2
0.0
1.9
20.2
65.0
69.8
32.9
a
Concentration of halide salt hydrates: LiCl⋅3H2O, LiBr⋅3H2O, LiI⋅3H2O, CaBr2⋅6H2O, and
ZnCl2⋅3H2O; reaction conditions: starting glucose concentration 10 g/L, 120 °C, and 15 min.
b Include unidentified products and decomposition products (HMF, BMF, and furfural).
21

Green Chemistry
Page 22 of 29
View Article Online
DOI: 10.1039/C7GC02145C
Decomposed products
Fructose
Mannose
Glucose
100
90
80
70
60
50
40
30
20
10
0
2
3
4
5
6
H2O-LiBr Molar Ratio
Figure 1. Isomerization of glucose in aqueous LiBr solution with different H2OꢀLiBr molar
ratios. Reaction conditions: starting glucose concentration 10 g/L, 120 °C, and 15 min.
22

Page 23 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
100
90
80
70
60
50
40
30
20
10
0
100
90
(a)
(b)
Glucose
Glucose
Mannose
Fructose
Mannose
Fructose
80
70
60
50
40
30
20
10
0
0
15 30 45 60 75 90 105120
0
15 30 45 60 75 90 105120
Reaction Time [min]
Reaction Time [min]
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
Glucose
Glucose
(c)
(d)
Mannose
Fructose
Mannose
Fructose
0
15 30 45 60 75 90 105120
0
15 30 45 60 75 90 105120
Reaction Time [min]
Reaction Time [min]
Figure 2. Isomerization of glucose under different reaction temperatures in lithium bromide
trihydrate: starting glucose concentration 10 g/L; a) 110 °C; b) 120 °C; c) 130 °C; and d) 140 °C.
23

Green Chemistry
Page 24 of 29
View Article Online
DOI: 10.1039/C7GC02145C
16
14
12
10
8
HMF
BMF
Furfural
6
4
2
0
0
15
30
45
60
75
90
105
120
Reaction Time (min)
Figure 3. Formation of furan compounds during glucose isomerization in lithium bromide
trihydrate (starting glucose concentration 10 g/L and 130 °C).
24

Page 25 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
H
O
H
O
H
H
H
H
H
H
O
H
Li
O
O
H
O
H
H
Br
H
H
H
R
Br
H
O
OH
O
H
OH
O
O
Br
H
OH
O
H
H
Path (A)
H
R
R
R
H
Enediol
O
OH
H
H
OH
O
H
HO
H
HO
H
H
OH
= R
HO
H
H
H
OH
CH2OH
OH
OH
OH
H
H
OH
CH2OH
D-Glucose
H
H
H
H
H
H
H
H
H
Br
CH2OH
H
H
R
H
H
H
H
O
O
H
H
H
O
Path (B)
O
O
O
O
O
D-Fructose
O
O
H
O
H
Li
O
Li
O
Li
O
H
O
O
O
O
H
H
H
H
R
H
H
R
H
H
H
H
H
Br
Figure 4. Proposed pathways for glucose isomerization in lithium bromide trihydrate.
25

Green Chemistry
Page 26 of 29
View Article Online
DOI: 10.1039/C7GC02145C
Figure 5. 13C spectra of (a) glucoseꢀ2ꢀd1, (b) unlabeled glucose, (c) unlabeled fructose, and (d)
mixed sugars (glucose and fructose) from isomerization of unlabeled glucose in deuterated LiBr
trihydrate.
26

Page 27 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
13
Figure 6. C spectra of (a) glucoseꢀ2ꢀd1, (b) unlabeled glucose, (c) unlabeled fructose, and (d)
mixed sugars (glucose and fructose) from the isomerization of glucoseꢀ2ꢀd1 in LiBr trihydrate.
27

Green Chemistry
Page 28 of 29
View Article Online
DOI: 10.1039/C7GC02145C
100
90
80
70
60
50
40
30
20
10
0
100
(b)
(a)
Glucose
Glucose
90
80
70
60
50
40
30
20
10
0
Mannose
Fructose
Mannose
Fructose
0
15 30 45 60 75 90 105120
0
15 30 45 60 75 90 105120
Reaction Time [min]
Reaction Time [min]
Figure 7. Isomerization of (a) fructose (starting concentration 10 g/L) and (b) mannose (starting
concentration 10 g/L) in LiBr⋅3H2O at 120 °C.
28

Page 29 of 29
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02145C
Graphical Abstract
Both Br‾ and Li+ catalyzed isomerization of glucose in aqueous lithium bromide solution.
29