Reaction media dominated product selectivity in the isomerization of glucose by chromium trichloride: From aqueous to non-aqueous systems

Article abstract of DOI:10.1016/j.cattod.2014.02.038

While chromium trichloride hydrate (CrCl₃·6H ₂O) effectively catalyzed the isomerization of glucose into fructose in aqueous solution, a general product selectivity behavior was observed independent of reaction variables such as reaction time, temperature, catalyst loading, halide ion, and initial glucose concentration. By studying the mixed solution of dimethylsulfoxide (DMSO) and water at varied DMSO/H₂O ratios, it was found that deviation from the general water-phase fructose yield curve, with concomitant 5-hydroxymethylfurfral (HMF) formation, occurred when the fructose concentration in available water became sufficiently high in DMSO rich solvent mix. Therefore, tuning the solvent system was the most effective approach to change the product distribution from CrCl₃·6H ₂O catalyzed glucose conversion. The apparent activation energy of glucose conversion in the studied system was estimated to be 58.6 kJ mol ^-1. Special attention was also given to gain some mechanistic insights by control experiments with simple model compounds and additives, ¹³C NMR and UV-vis spectroscopic analyses.

Full text of DOI:10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Reaction media dominated product selectivity in the isomerization of
glucose by chromium trichloride: From aqueous to non-aqueous
systems
Songyan Jiaa, Kairui Liua,b, Zhanwei Xua, Peifang Yana, Wenjuan Xua,
Xiumei Liua, Z. Conrad Zhanga,c,∗
a Dalian National Lab for Clean Energy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
b University of Chinese Academy of Sciences, Beijing 100039, China
c State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
a r t i c l e i n f o
a b s t r a c t
Article history:
While chromium trichloride hydrate (CrCl3·6H2O) effectively catalyzed the isomerization of glucose into
fructose in aqueous solution, a general product selectivity behavior was observed independent of reaction
variables such as reaction time, temperature, catalyst loading, halide ion, and initial glucose concentra-
tion. By studying the mixed solution of dimethylsulfoxide (DMSO) and water at varied DMSO/H2O ratios,
it was found that deviation from the general water-phase fructose yield curve, with concomitant 5-
hydroxymethylfurfral (HMF) formation, occurred when the fructose concentration in available water
became sufficiently high in DMSO rich solvent mix. Therefore, tuning the solvent system was the most
effective approach to change the product distribution from CrCl3·6H2O catalyzed glucose conversion. The
Received 16 November 2013
Received in revised form 21 January 2014
Accepted 18 February 2014
Available online xxx
Keywords:
Glucose isomerization
Chromium trichloride
Fructose
apparent activation energy of glucose conversion in the studied system was estimated to be 58.6 kJ mol−1
.
Special attention was also given to gain some mechanistic insights by control experiments with simple
model compounds and additives, 13C NMR and UV–vis spectroscopic analyses.
© 2014 Elsevier B.V. All rights reserved.
General behavior
Aqueous media
Mixed solvents
1. Introduction
dimethylsulfoxide (DMSO) [8,11,13,19,20], dimethylacetamide
(DMA)-LiCl [21], water/organic solvent biphasic system [2,9,17,18],
With the growing consumption of fossil fuels and the conse-
quential emission of green house gases, the world faces increasing
pressure to develop technologies that enable the efficient utiliza-
tion of alternative sources of energies and renewable feedstocks.
Biomass has recently been proposed as a potential feedstock
toward sustainable development in both alternative energy and
chemicals [1].
In the past decade, saccharides have attracted tremendous
attention toward the development of a new biorefinery industry,
since they can be utilized to synthesize a variety of valuable chem-
icals [1]. Dehydration of fructose into 5-hydroxymethylfurfural
(HMF) was a mostly studied pathway for this purpose. The
fructose dehydration proceeds relatively easily, which can be
catalyzed by mineral acid [2–7], organic acid [8,9], metal salt
[10–13], solid acid [14–18] in varieties of reaction media, such as
ionic liquid [3,4,12,14–16], and even water [5,22]. Establishing a
protocol to produce HMF based chemicals from sugars offers the
potential to reduce society’s reliance on traditional fossil resources
to some extent. Unfortunately, fructose has a low abundance in
nature [1], with high cost [23], and is not a suitable feedstock for
industrial development. By contrast, glucose is the most abundant
monosaccharide and available from many non-edible cellulosic
materials. However, the effective conversion of glucose into HMF
based products always remained a great challenge until Zhang and
co-workers discovered that chromium chloride (CrCl2 or CrCl3) sig-
nificantly facilitated HMF production from glucose in ionic liquids
[24]. The key role of CrClx was to isomerize glucose into fructose
in situ [24]. Thereafter, a spate of researches focused on the con-
version of glucose based feedstocks into HMF by employing CrClx
in ionic liquids as well as DMSO or DMA-LiCl system [21,25–32].
Other metal salts, such as tin chloride (SnCl2 or SnCl4) and ger-
manium chloride (GeCl4), were also found to be effective for the
dehydration of glucose in ionic liquids via fructose intermediate
[33–35]. In general, the bottleneck of converting glucose into HMF
lies on effectively isomerizing glucose into fructose in a reaction
medium.
∗
Corresponding author at: Dalian National Lab for Clean Energy, Dalian Institute
of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian
116023, China. Tel.: +86 411 84379462; fax: +86 411 84379462.
E-mail address: zczhang@yahoo.com (Z.C. Zhang).
http://dx.doi.org/10.1016/j.cattod.2014.02.038
0920-5861/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
2
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
Up to now, Cr-catalyzed systems are still the most active for
HMF production from glucose based feedstocks, and almost all of
these systems employ ionic liquids or other non-aqueous solvents
[21,24–32]. However, the chemistry of Cr catalysts in isomerizing
glucose to fructose is shadowed by that of fructose dehydration
to HMF which is rapidly formed as the most abundant product in
the non-aqueous solvent systems. Some solvents used above are
expensive and some may display toxicity, so alternative cheaper
and greener solvents are more desirable. One may be concerned of
the potential toxicity of Cr catalyst. In fact, hexavalent chromium
is highly toxic while trivalent chromium, found in most foods and
nutrient supplements, is an essential nutrient with very low tox-
icity [36]. Thus, trivalent chromium catalysts should be avoided
in an oxidation reaction. Because glucose is a reducing sugar,
the oxidation of Cr(II, III) to Cr(VI) is unlikely to occur. Water is
the most environmentally benign and naturally abundant solvent.
Exploring the ability to utilize water for glucose conversion will be
highly attractive for reducing cost of solvent and potential pollu-
tion. Recently, a series of studies on the isomerization of glucose
into fructose have already been carried out in aqueous solution
with solid bases or zeolites as catalysts [37–40]. However, most
of the above systems have not been extended for the further use
of the in situ formed fructose yet. Nikolla and co-workers have
showed reasonable potential for one-pot synthesis of HMF from
glucose over Sn-Beta zeolite in water/tetrahydrofuran biphasic sys-
tem at 160–180 ◦C [41]. However, the Sn-Beta also has drawbacks,
for examples, the aging time for the preparation of this catalyst
is very long (>1 month) and it can be poisoned by a certain sol-
vent/additive, such as DMSO [42]. Choudhary et al. recently studied
the consecutive conversion of glucose into levulinic acid and xylose
into furfural, respectively, in aqueous solution with CrCl3·6H2O
and hydrochloric acid (HCl) as catalysts, and demonstrated that Cr
complexes isomerized glucose and xylose into fructose and xylu-
lose, respectively [43,44]. The results offer an opportunity to realize
efficient conversion of sugars in aqueous media.
2.2. Typical reaction procedure
Glucose (50 mg), CrCl3·6H2O (3.7 mg, 5 mol% with respect to
glucose) and 1 mL of DI H2O were added into each reaction vial
(5 mL) with a magnetic stir bar. Then, the reaction vials were sealed,
inserted into a heating and stirring module (TS-18821, Thermo Sci-
entific, USA), and stirred at 500 rpm for a specified time at the
reaction temperature. After the reaction was quenched in ice water
bath, 1 mL of DI H2O and the internal standard were added into the
reaction mixture. Then, a small amount of sample was taken out
and further diluted by ∼15 times and analyzed by high performance
liquid chromatography (HPLC).
2.3. Analysis method
HPLC analysis was performed on an Agilent 1260 series
with a refractive index detector and a PL Hi-Plex H column
(300 mm × 7.7 mm, 8 ␮m). Diluted H2SO4 solution (0.005 M) at a
flow rate of 0.6 mL/min was used as the mobile phase. The col-
umn and detector temperatures were 65 and 50 ◦C, respectively.
Authentic chemical compounds were used to identify the retention
times. Glycerol was added as an internal standard for the quan-
titative calculations. For the control experiments with glycerol,
dl-glyceraldehyde or 1,3-dihydroxyacetone as simple model com-
pounds or additives, 1,2-propylene glycol was used as the internal
standard instead of glycerol. Glucose conversion and product yield
are defined as in Eqs. (1) and (2). The pH values were measured by
a Mettler-Toledo G20 Titrator with a DGi115-SC electrode or Core
Module 3 (CM3) System (Freeslate). 13C NMR spectra were recorded
on a Bruker DRX-400 spectrometer. Ultraviolet–visible (UV–vis)
spectra were collected on a Shimadzu UV-2600 spectrophotometer.
Glucose conversion (mol%)
ꢀ
ꢁ
Mole of remaining glucose detected by HPLC
=
1 −
× 100%
(1)
In this work, we investigated the effect of reaction variables
on the CrCl3·6H2O catalyzed isomerization of glucose into fruc-
tose in aqueous media in great detail. It is remarkable that these
results for the first time revealed a general behavior of glucose
isomerization in the studied system independent of the process
parameters. It is also noted that other published works involving
aqueous systems used for glucose conversion fit into this general
behavior. In this work, special attention was also given to gain some
mechanistic insights using simple model compounds and additives.
Ultraviolet–visible spectroscopic analyses provided information on
the interaction between chromium species and glucose during the
reaction process.
Mole of initially added glucose
Yield (mol%)
ꢀ
ꢁ
Mole of product detected by HPLC
Mole of ideal amount of product from added feedstock
=
× 100%
(2)
3. Results and discussion
3.1. Isomerization of glucose into fructose in aqueous solution
Chromium dichloride (CrCl2) has been mostly reported to be
the effective catalyst for HMF production in non-aqueous solvent
system [24,25,27,28,31]. However, to avoid CrCl2 oxidation by air,
especially in H2O which may cause the oxidation [45,46], chromium
trichloride (CrCl3) is used in this work. CrCl3 is also an effective
catalyst [24], of which the hydrated form, CrCl3·6H2O, has better
solubility in the solvents [30]. Since water is used as a reaction
medium in this work and anhydrous CrCl3 has poor solubility in
H2O [30], CrCl3·6H2O was selected as the catalyst for isomerization
of glucose into fructose.
Fig. S1 shows the results of glucose isomerization at 90, 110,
and 130 ◦C for 0.5–4 h in 1 mL of H2O each, to which 50 mg of
glucose and 3.7 mg of CrCl3·6H2O (5 mol%, with respect to glu-
cose) were added. The reaction proceeded very slowly at 90 ◦C,
at which glucose conversion and fructose yield reached 13.6% and
10.8%, respectively, after 4 h. Elevating the reaction temperature
boosted the isomerization of glucose. At 110 ◦C, the yield of fruc-
tose increased to 21.8% after 4 h, while it reached the maximum
2. Experimental
2.1. Materials
d-glucose (99%), d-fructose (99%), 1,6-anhydro-␤-d-glucose
(AGP, 99%), d-cellobiose (98%) and chromium bromide hexa-
hydrate (CrBr3·6H2O, 98%) were purchased from Alfa Aesar.
Chromium chloride hexahydrate (CrCl3·6H2O, 96%) and dl-
glyceraldehyde (≥90%) were purchased from Sigma–Aldrich.
5-Hydroxymethylfurfural (HMF, 98%), levulinic acid (LA, 99%) and
1,3-dihydroxyacetone dimer (97%) were purchased from Aladdin.
Glycerol (99%), dimethyl sulfoxide (DMSO, 99%) and 1,2-propylene
glycol (99%) were purchased from Sinopharm (China). Hydrochlo-
ric acid (HCl, 36–38 wt%) was provided by a local supplier. All the
chemicals were used as received. Deionized water (DI H2O) was
produced by a Milli-Q Integral 5 system, having a resistivity of
18.2 Mꢀ cm.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
3
Fig. 1. Isomerization behavior of glucose at different reaction time intervals and
temperatures.
Fig. 2. Effect of CrCl3·6H2O loadings on the isomerization of glucose into fructose
at 110 ◦C for different time intervals (0.5, 1, 1.5, 2, 3, 4 h) in aqueous solution.
In reported work, CrBr3 and CrF3 were found more active than
CrCl2 or CrCl3 for the conversion of glucose into HMF in some non-
aqueous solvents [21,49], implying halide anions may influence
the isomerization of glucose. Similarly, CrBr3 and CrF3 could be
potential catalysts for this work. However, CrF3 may be hydrolyzed,
producing hydrofluoric acid (HF) which is poisonous and can also
corrode the used glass reaction vial. Therefore, only CrBr3·6H2O
was employed in the current system for comparison. As shown in
Fig. S2, compared with the results over CrCl3·6H2O, the ones with
CrBr3·6H2O not only followed the general curve for the fructose
production, but also give similar absolute values of fructose yield,
glucose conversion and HMF yield, indicating that CrBr3·6H2O is
also a good catalyst for isomerizing glucose into fructose. The halide
anions had little effect on the isomerization reaction.
As reported, the thermodynamic equilibrium between glucose
and fructose has a constant of ∼1 at 25 ◦C [47], implying that
the isomerization between glucose and fructose is reversible. We
then checked the effect of glucose concentration. The initial con-
centration of glucose was varied from 50 mg/mL to 120 mg/mL,
while maintaining the same concentration of CrCl3·6H2O. As illus-
trated in Fig. S3, the production of fructose slightly decreased with
increasing the initial glucose concentration, because the amount of
catalyst was reduced relatively to the initial concentration of glu-
cose. However, the results indeed behaved in the similar manner
as mentioned above. Moreover, the results imply the system may
have the potential for a process compatible with high concentration
on the other hand.
Based on the results discussed above, it is evident that, indepen-
dent of the process variables, such as reaction time, temperature,
CrCl3·6H2O loading, halide ion and initial glucose concentration,
the production of fructose always followed a general curve in the
water medium. Fructose yield increased in a linear manner at
the glucose conversion below about 30%, while the formation of
secondary and further products was accelerated once the concen-
tration of fructose became high enough in water at higher glucose
conversion.
The reported results in literatures on the catalytic conversion of
glucose in various solvents, e.g. ionic liquids [24,25,30,32], DMSO
[29,32], suggest that the property of solvent plays a dominant
role in the product distribution in glucose conversion, and the
above results are consistent with this statement. DMSO is a sol-
vent generally used for the preparation of HMF from fructose while
suppressing side reactions [2,20,50]. To correlate the deviation
point from the general curve with fructose concentration, we select
mixed solvents composed of DMSO and H2O at various DMSO/H2O
ratios (v/v). With increasing DMSO/H2O ratio, the available water
of ∼26% even after 1 h at 130 ◦C. The isomerization of glucose
is slightly endothermic (ꢁH = 3 kJ mol−1) [47], so it is expected
that higher temperature will accelerate the isomerization of glu-
cose. However, it is remarkable to observe that, when the fructose
yield is plotted against the glucose conversion as shown in Fig. 1,
as the primary product of glucose isomerization, fructose yield
essentially followed a general trend of linear increase with respect
to glucose conversion of about 3–30% at different temperatures.
Fructose production gradually deviated from the linear increase
at higher glucose conversion (>∼30%), implying more secondary
and consecutive reaction products were produced. This is indeed
the case when the yields of other products are inspected. For
example, HMF is one such secondary reaction product, from the
dehydration of fructose. As illustrated in Fig. 1, the formation of
HMF remarkably accelerated at higher glucose conversion (>∼30%).
Choudhary et al. recently studied the consecutive conversion of
glucose into levulinic acid (LA) and the mechanism of glucose isom-
erization in aqueous media at 140 ◦C [43,48]. Their published data
also followed the relationship between fructose yield and glucose
conversion as established in this work (Fig. 1) to a large extent.
Except for fructose and HMF, other products such as cellobiose,
1,6-anhydro-␤-d-glucopyranose/glucofuranose (AGP/AGF) and LA,
were also detected in the system. Cellobiose, AGP and AGF are the
primary products by undesired intermolecular or intramolecular
dehydration of glucose, while LA is a downstream product from
the rehydration of HMF. The yields of cellobiose, AGP, AGF and LA
were very low, so they were not listed. Other products remained
unknown, possibly humins, accounting for some loss of fructose. It
should be noted that mass balance is an important indicator for a
reaction. To circumvent undesired side reactions, further study is
essential for reducing mass loss.
We next investigated the effect of CrCl3·6H2O loadings at
110 ◦C and the other conditions remained the same as above. As
shown in Fig. 2, fructose production still followed the general
curve as depicted above, with CrCl3·6H2O loadings from 2.5 mol%
to 12.0 mol% (with respect to glucose). At the glucose conver-
sion of about 8–30%, fructose formation showed a linear increase
trend, while it gradually deviated from the general curve again at
higher glucose conversion. As illustrated in Fig. 2, more HMF was
produced at higher glucose conversion, accounting for the devi-
ation of fructose yield from the general curve to some extent as
mentioned above. The deviation from the linear increase in fruc-
tose yield versus glucose conversion is indicative of consecutive
reactions consuming fructose when fructose concentration
becomes sufficiently high.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
4
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
Fig. 4. Formation of HMF from glucose with CrCl3·6H2O in different DMSO/H2O
mixtures at 110 ◦C.
Fig. 3. Isomerization of glucose into fructose with CrCl3·6H2O in different
DMSO/H2O mixtures at 110 ◦C.
accelerated when fructose concentration could become high
enough for the secondary reaction. Evidently, water is detrimen-
tal to the dehydration of fructose, so the more water, the lower
HMF yield. However, the formation of HMF behaved differently
in pure DMSO. Due to the facile dehydration of fructose in DMSO
(non-aqueous solvent), once fructose was produced, it undergoes
consecutive conversion into HMF immediately.
Moreover, since HMF is a main secondary product from fructose,
a plot of the total yield of fructose and HMF against glucose conver-
sion reveals that the conversion of glucose exhibited a similar trend
in DMSO/H2O solution, while the total fructose and HMF yield devi-
ated from the general curve in DMSO at the glucose conversion of
∼15%. As seen in Table S2, much cellobiose was produced in pure
DMSO system after longer time. In addition, DMSO also favored
in the formation of AGP and AGF as reported [53]. The undesired
products, such as cellobiose, AGP and AGF, explain why the total
production of fructose and HMF deviated from the general curve in
Fig. 5.
is decreased. As shown in Fig. 3, the production of fructose showed
different behaviors. In the system with DMSO/H2O (v/v) of 2/8,
the production of fructose essentially followed the general curve
as that in pure H2O. When the volume proportion of DMSO was
increased to 50% and 80%, respectively, the production of fruc-
tose deviated from the general curve at lower and lower glucose
conversion. In pure DMSO system, fructose production did not
follow the general curve as those in aqueous solution at all; the
maximum fructose yield was only ∼5% in DMSO solvent. Table
S1 gives the approximate fructose concentration at each devia-
tion point in different DMSO/H2O solutions, which decreased from
∼0.056 mmol/mL to ∼0.035 mmol/mL with respective DMSO/H2O
(v/v) ratios of 0/10–8/2. As shown in Fig. 3, only a little fructose
could stably exist in pure DMSO, because the dehydration potential
of DMSO is very high that made fructose undergo consecutive con-
version. If DMSO and water components in the mixed solvents had
displayed their individual intrinsic solvent properties, the volume
of the available water would be the real volume of water fraction.
On this assumption, the fructose concentrations in available water
fraction of the DMSO/H2O solutions at different ratios were esti-
mated to vary from ∼0.056 mmol/mL to ∼0.175 mmol/mL, with
increasing DMSO proportion from 0 to 80 vol%. The published stud-
ies on the DMSO/H2O binary mixture, on one hand, showed that the
bonding between water and DMSO molecules is non-linear with
varying DMSO/H2O ratio and water–water interaction remains sig-
nificant, even at water concentration as low as 5% [51], and on the
other hand, the mixing of the two solvents is strongly exother-
mic [52], indicating H-bonding formation that would passivate the
dehydration potential of DMSO by the presence of water. The weak-
ened dehydration potential of DMSO, due to the solvation effect of
water by H-bonding formation, appears to be responsible for the
increased fructose concentration at the deviation point over that
solely determined by the volume fraction of available water. In
general, it could be expected that a certain fructose concentration
will account for the deviation of fructose yield versus glucose con-
version. Fructose yield increased in a linear manner at the fructose
concentration below ∼0.056 mmol/mL (estimated by pure H2O sys-
tem), while the formation of secondary and further products was
accelerated once the concentration of fructose became high enough
in available solvent at higher glucose conversion. In addition, the
deviation from the general curve reflects fructose conversion into
HMF. As illustrated in Fig. 4, the production of HMF in DMSO/H2O
(v/v) of 0/10–8/2 showed a similar trend. HMF formation was
The suppression of dehydration reaction by water can be ben-
eficially utilized to suppress undesired side reactions. When an
adequate amount of water was present in the DMSO mixed sol-
vent, water was found to significantly suppress the intermolecular
and intramolecular dehydration of glucose, leading to low yields
of cellobiose, AGP and AGF. The results therefore indicate that if
Fig. 5. Conversion behaviors of glucose toward the total formation of fructose and
HMF in different DMSO/H2O mixtures at 110 ◦C.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
5
an appropriate aqueous system is employed, more desired prod-
uct could be produced. For example, as seen in Table S2, more HMF
could be obtained in the aqueous solution with 80% of DMSO, which
circumvented other by-products, such as cellobiose, AGP and AGF.
In general, tuning the composition of DMSO and H2O can change the
behavior of fructose production and the distribution of products.
Aside from DMSO, other solvents, such as N,N-dimethylformamide,
N,N-dimethylacetamide, N-methyl pyrrolidone and so on, are also
potential co-solvents to control the distribution of products.
end aldehyde groups in aqueous solution, so it is more competitive
than glucose to react with chromium species. When glyceralde-
hyde was treated in a HCl solution with a pH of 2.1 (Entry 6), low
conversion was observed, and 1,3-dihydroxyacetone yield was neg-
ligible. The results thus further demonstrate that chromium species
is the active component for the isomerization of glucose and glyc-
eraldehyde. Comparing the data in Entries 5 and 7, it is evident that
chromium species did predominantly interact with glyceraldehyde
in the competition reaction with respect to their interaction with
glucose.
3.2. Kinetic analysis
13C NMR spectra of glucose in D2O were collected under differ-
ent conditions to probe the structural details of glucose in various
stages of catalytic transformation in the solvent systems of this
study. As shown in Fig. S5a, when glucose was dissolved in D2O at
room temperature for 20 min, it existed in ␣ and ␤-glucopyranose
forms simultaneously. The ␣-glucopyranose was a little more dom-
inant with ␤/␣ ratio of 0.7/1. After the solution was heated at
110 ◦C for 2 min, it was found that the higher temperature signifi-
cantly promoted the mutarotation of glucose, leading to ␤/␣ ratio
of 1.48/1 (Fig. S5b). In previous literatures, it was reported that
almost all the glucose showed the ␣-glucopyranose form in ionic
liquid or dimethylacetamide before reaction [24,60], and catalysts
therein facilitated the mutarotation to produce ␤-glucopyranose.
In contrast, there is a mutarotation equilibrium between ␣ and
␤-glucopyranose in H2O or DMSO, and elevating temperature can
accelerate the equilibrium process [61], which explains why both
␣ and ␤-glucopyranose were detected in Fig. S5. Glucose favors
the ␤-glucopyranose form at equilibrium [59,61]. After glucose
and CrCl3·6H2O dissolved in D2O at room temperature for 20 min,
the proportion of ␣-glucopyranose slightly increased, implying
chromium species may have an interaction with glucose under
those conditions. However, as shown in Fig. S5d, heating promoted
the mutarotation again. By extending reaction time at 110 ◦C,
␤-glucopyranose appeared more and more dominant, indicating
thermodynamic equilibrium was the major factor that affected
the distribution of ␣ and ␤-glucopyranose. In consideration that
acid may be produced by CrCl3·6H2O hydrolysis in D2O, we then
measured the 13C NMR of glucose in D2O with a pH of ∼2.7. As
shown in Fig. S6a, when glucose was dissolved in D2O with a pH of
∼2.7 at room temperature for 20 min, it showed a higher propor-
tion of ␣-glucopyranose like that in Fig. S5c, indicating that acidic
medium accounts for the reduced mutarotation of glucose to ␤-
glucopyranose. As illustrated in Fig. S6b and c, heating promoted
the mutarotation, and extending reaction time led to increased
mutarotation to ␤-glucopyranose. In general, 13C NMR spectra
mainly confirmed that the mutarotation of glucose did occur in
the studied system. Glucose must undergo a ring opening process
to realize mutarotation, accompanied with the formation of open
chain glucose [59]. The open chain glucose is likely the key compo-
nent to react with chromium species, offering the opportunity for
the isomerization as discussed in the last paragraph.
The kinetic parameters for CrCl3·6H2O catalyzed isomeriza-
tion of glucose in water were calculated by assuming a rate
expression that was first order in the concentration of glu-
cose. The data on glucose conversion at 90, 110 and 130 ◦C
were used for the calculation. As shown in Fig. S4a, the reac-
tion rate was fitted well with the assumption of first order
reaction. Fig. S4b gives the Arrhenius plot, and the apparent acti-
vation energy (Ea) is estimated to be 58.6 kJ mol−1, close to the
reported values. For example, Choudhary et al. reported that the Ea
was 15.3 kcal mol−1 (∼64.0 kJ mol−1) for glucose conversion with
CrCl3·6H2O, while the Ea was 15.5 kcal mol−1 (∼64.9 kJ mol−1) for
xylose (aldopentose) conversion [43,44]. The estimated Ea is much
lower than those for base catalyzed isomerization of glucose, about
121–129 kJ mol−1 [54–56], indicating that the unique mechanism
involved in chromium mediated isomerization of glucose is differ-
ent from that of base catalyzed one.
3.3. Mechanistic insights
To understand the mechanism involved in the chromium
mediated isomerization in water medium, a series of control
experiments were carried out with different feedstocks, catalysts
or additives, and the results are listed in Table 1. It is known
that CrCl3·6H2O can be hydrolyzed into chromium species and
hydrochloric acid (HCl), so we first checked the role of HCl. In a sam-
ple containing 3.7 mg of CrCl3·6H2O in 1 mL H2O, the pH value was
at ∼2.7, which indicates that the aqueous solution (e.g. reactions in
Fig. 1) was acidic before reaction took place. After heating at 110 ◦C
for 1 h, the pH of the above solution dropped to ∼2.1. Elevating
temperature could accelerate the hydrolysis of CrCl3·6H2O, there-
fore producing more HCl and leading to a lower pH. For comparison,
aqueous solution with pH of 2.7 or 2.1 was prepared by diluting con-
centrated HCl solution and used both as the solvent and catalyst for
glucose isomerization. As seen in Table 1 (Entries 2 and 3), little glu-
cose conversion could be detected, clearly indicating that the active
component for the isomerization of glucose is chromium species
but not HCl. The Fischer projection of glucose can be considered
as one glycerol molecule attached to one glyceraldehyde molecule.
Therefore, glycerol and glyceraldehyde were employed as additives
for understanding the interaction between glucose and chromium
species. As seen in Entries 4 and 5, when glycerol was added, the
glucose conversion and fructose yield showed a negligible decrease
(∼1%) compared with Entry 1 (Table 1), indicating that glycerol
did not influence the reaction much. After glyceraldehyde was
added, the isomerization of glucose was significantly suppressed
with the conversion of 3.8%. However, glyceraldehyde underwent
effective isomerization into 1,3-dihydroxyacetone, demonstrating
that chromium species could interact with the end aldehyde group
of glyceraldehyde as other metal ions [57,58]. As reported, only
about 0.002% of glucose existed in the chain form in water at ∼30 ◦C
[59], with very limited end aldehyde groups, and most aldehyde
groups of glucose existed in the hemiacetal form (the ring form).
Glucose must undergo a ring open process before interacting with
chromium species. In contrast, glyceraldehyde can provide more
Finally, we employed UV–vis spectrophotometer to monitor the
isomerization of glucose in water. The UV–vis analyses were not
carried out in DMSO solution, because more HMF would be pro-
duced therein, and the intense absorbance of HMF overshadows the
absorbance of CrCl3·6H2O, as will be discussed below. As illustrated
in Fig. 6, after CrCl3·6H2O was dissolved in H2O at room tempera-
ture, two weak absorption peaks appeared at ∼417 and ∼582.5 nm,
respectively. They could be attributed to the d–d transitions of Cr
complexes. For example, Cr3+ could coordinate with H2O molecules
to form complexes [43,44]. The molar absorptivity of the bands
caused by d–d transitions is relatively low [62], which explains
why CrCl3·6H2O solution showed the weak absorption peaks. Glu-
cose does not show signal in the UV–vis analysis. When glucose
was added, a blue shift of about 2 nm was observed for the peak
of 582.5 nm, possibly a result of Cr complexes formation between
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
6
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
Table 1
Results for the control experiments.
Entry
Feedstock
Catalyst
Additive
Glucose conversion (%)
Fructose yield (%)
1a
2
Glucose
Glucose
Glucose
Glucose
Glucose
Glyceraldehyde
Glyceraldehyde
CrCl3·6H2O
pH = 2.7b
None
None
None
Glycerol
Glyceraldehyde
None
12.0
10.5
0
0
Trace
Trace
11.8
3
pH = 2.1c
4d
5e
6
CrCl3·6H2O
CrCl3·6H2O
pH = 2.1c
CrCl3·6H2O
9.5
3.8/78.8f
7.6f
1.8/60.2g
0.6g
58.3g
7a
None
74.0f
Reaction conditions: Feedstock (50 mg of glucose or 25 mg of glyceraldehyde), additive and catalyst were added in 1 mL H2O and heated at 110 ◦C for 0.5 h.
a
3.7 mg of CrCl3·6H2O (5 mol% to glucose or glyceraldehyde) was added.
b
1 ml of HCl aqueous solution with a pH of 2.7 was used both as the solvent and catalyst.
c
1 ml of HCl aqueous solution with a pH of 2.1 was used both as the solvent and catalyst.
d
25 mg of glucose and equimolar glycerol (12.8 mg of glycerol) were added in 1 ml of H2O with 3.7 mg of CrCl3·6H2O (5 mol% to the total of glucose and glycerol).
e
25 mg of glucose and equimolar glyceraldehyde (12.5 mg of glyceraldehyde) were added in 1 ml of H2O with 3.7 mg of CrCl3·6H2O (5 mol% to the total of glucose and
glyceraldehyde).
f
Glyceraldehyde conversion.
1,3-Dihydroxyacetone yield.
g
glucose and chromium species. After the solution of CrCl3·6H2O
and glucose was heated at 110 ◦C for different time, the peak at
∼284 nm was enhanced remarkably, indicating the formation of
more HMF. For the chromium species, both two peaks showed
further blue shift. The original peak at ∼417 nm slightly shifted
to about ∼414 nm. Sample (g) (Fig. 6) apparently showed a more
remarkable blue shift to ∼402 nm, but it may be because the HMF
peak was so intensive that the peak at ∼414 nm was overshad-
owed to a certain degree. The original peak of ∼582.5 nm showed
a blue shift of 10.5 nm after the solution was heated at 110 ◦C for
60 min, implying more complexes or coordination between glu-
cose/fructose and chromium species. It was reported that more
HCl may influence the UV absorbance of Cr complexes in water
[44]. This possibility was also checked. As mentioned above, after
the used CrCl3·6H2O solution was heated at 110 ◦C for 60 min, its
pH value decreased from ∼2.7 to ∼2.1, which means more HCl
was produced. Based on this point, sample (a) (Fig. S7) was heated
at 110 ◦C for 30 min and 60 min, respectively, it was then mea-
sured by UV–vis spectrophotometer. As shown in Fig. S7, sample (b)
heated for 30 min showed peaks at ∼415 and ∼581 nm, while sam-
ple (c) heated for 60 min showed peaks at ∼415.5 and ∼581.5 nm,
indicating that more HCl barely affected the absorbance of
chromium species in this work. It can be seen that the absorbance
intensities of peaks 1 and 2 increased when the CrCl3·6H2O solu-
tion with glucose was heated at 110 ◦C for longer time (Fig. 6). In
fact, when the CrCl3·6H2O solution without glucose was heated at
110 ◦C for 30 min and 60 min (samples b and c, Fig. S7), respec-
tively, the intensities also increased, implying the heating process
may influence the Cr complexes in water. However, when glucose
was added and heated at 110 ◦C for 30 min and 60 min, not only
that the absorbance intensity further increased, but also the sam-
ples showed the blue shifts, which demonstrates that chromium
species interacted with glucose/fructose.
Because most blue shifts for the UV absorbance of CrCl3·6H2O
solution were only observed at conditions at which a reaction
occurred, we then checked if a particular product could cause the
blue shift through its interaction with the chromium species. After
reacting glucose with CrCl3·6H2O (5 mol%, with respect to glucose)
at 110 ◦C for 30 min, the detectable products were fructose (10.5%)
and little cellobiose (0.4%). Then cellobiose and fructose were used
as feedstock and heated with CrCl3·6H2O (5 mol%, with respect
to feedstock) at 110 ◦C for 30 min. As illustrated in Fig. S8, the
sample with cellobiose gave two absorbance peaks at ∼415.5 and
∼579.5 nm. No obvious blue shift for the absorbance of CrCl3·6H2O
was observed at such high concentration of cellobiose, indicat-
ing that the little cellobiose (a yield of 0.4%) formed in the above
reaction was not the major factor that caused the blue shift. Inter-
estingly, the sample with fructose as feedstock showed remarkable
blue shift to ∼564 nm, while the original peak at ∼417 nm disap-
peared. When fructose was used, more HMF could be produced
than glucose, and the intense HMF peak could overshadow the orig-
inal peak at ∼417 nm, which explains why the peak disappeared.
In fact, after HMF solution with CrCl3·6H2O (5 mol%, with respect
to HMF) was heated at 110 ◦C for 30 min, the peaks 1 and 2 of
CrCl3·6H2O both were overlapped as shown in Fig. S8. To get evi-
dent absorbance of CrCl3·6H2O with HMF, the amount of HMF was
reduced. For example, when HMF (1.75 mg, 5 mol% to the amount
of HMF as above) with CrCl3·6H2O (1:1 molar ratio to HMF) was
heated at 110 ◦C for 30 min, the absorption bands of CrCl3·6H2O
were at ∼412 and ∼579 nm (Fig. S8), indicating that HMF was not
the major factor for the blue shift either. As reported, the isomeriza-
tion of glucose into fructose is reversible [47]. Moreover, Choudhary
and co-workers recently reported that the isomerization of xylose
into xylulose followed a reversible intra-hydride transfer mecha-
nism in the presence of CrCl3·6H2O, where chromium species could
both interact with xylose and xylulose [44]. According to that, as
illustrated in Fig. 7, it is proposed that the chromium species could
also coordinate with both glucose and fructose, showing that the
isomerization is reversible, so the remarkable blue shift with fruc-
tose herein can be ascribed to the interaction between chromium
Fig. 6. UV–vis spectra of CrCl3·6H2O in H2O. (a) RT, 5 min; (b) with glucose, RT, 5 min;
(c) with glucose, 110 ◦C, 2 min; (d) with glucose, 110 ◦C, 8 min; (e) with glucose,
110 ◦C, 15 min; (f) with glucose, 110 ◦C, 30 min; (g) with glucose, 110 ◦C, 60 min.
Conditions: 3.7 mg of CrCl3·6H2O with or without glucose (50 mg) was added in
1 mL H2O and stirred at different temperatures for different time. After reaction, the
solution was diluted by 2 mL H2O and analyzed by UV–vis spectrophotometer.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
7
Fig. 7. Proposed reaction pathway in the chromium-catalyzed glucose-to-fructose isomerization.
system was estimated to be 58.6 kJ mol−1. Control experiments, 13
C
species and fructose at C1 and C2 positions. The reason why fruc-
tose gave more remarkable blue shift may be due to the fact that
more fructose (0.8%) exists in the open chain form in water at
∼30 ◦C, a more favor structural anomer to form coordination bond-
ing with the chromium species. In comparison, the concentration
of the open chain form of glucose (0.002%) is only 1/400 of that of
fructose [59].
NMR and UV–vis spectroscopic analyses provided some mechanis-
tic insights into the CrCl3·6H2O catalyzed isomerization of glucose.
Chromium species played a role as the catalyst in the isomeriza-
tion by coordinating with end aldehyde group of the chain form of
glucose, while the HCl produced by the hydrolysis of CrCl3·6H2O
did not work. 13C NMR spectra confirm that the mutarotation of
glucose occurred during the reaction process, implying the forma-
tion of the chain form of glucose with end aldehyde group, which
is important for the catalyzed isomerization reaction. The findings
reported herein offer not only chemistry toward the Cr-catalyzed
isomerization of glucose, but also new opportunities for designing
appropriate aqueous system for sugar conversion.
The reaction of glyceraldehyde was investigated by UV–vis
analysis as well (as shown in Fig. S9). Similarly, after reacting
for some time, the absorption peak of chromium species showed
more remarkable blue shift. For example, the blue shift reached
∼19 nm from 582.5 to 563.5 nm after 30 min reaction, which indi-
cates that chromium species could have a stronger interaction
with glyceraldehyde than with glucose. The results demonstrate
that glyceraldehyde was more active, which was consistent with
the data in Table 1 (Entry 5). In addition, when glycerol and 1,3-
dihydroxyacetone were used as feedstock with CrCl3·6H2O (5 mol%
with respect to glycerol or 1,3-dihydroxyacetone) and heated at
110 ◦C for 30 min, respectively, the sample with glycerol gave two
peaks at ∼415 and ∼581 nm, further indicating that the OH groups
of glycerol were not the major factor for the blue shift; the sample
with 1,3-dihydroxyacetone gave two peaks at ∼400 and ∼563.5 nm.
The peak shifted to ∼400 nm is because it was overshadowed by
the intense absorption peak of 1,3-dihydroxyacetone, and the blue
shift to ∼563.5 nm is ascribed to the interaction between chromium
species and 1,3-dihydroxyacetone at C1 and C2 positions as those
for fructose.
Acknowledgements
This work was supported by the China Postdoctoral Science
Foundation (2013M530952), National Natural Science Foundation
of China (21306186), and the Chinese Government “Thousand Tal-
ent” program funding.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.cattod.
2014.02.038.
References
4. Conclusions
[1] A. Corma, S. Iborra, A. Velty, Chemical Reviews 107 (2007) 2411–2502.
[2] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933–1937.
[3] C. Sievers, I. Musin, T. Marzialetti, M.B.V. Olarte, P.K. Agrawal, C.W. Jones, Chem-
SusChem 2 (2009) 665–671.
[4] C. Li, Z.K. Zhao, A. Wang, M. Zheng, T. Zhang, Carbohydrate Research 345 (2010)
1846–1850.
[5] T.S. Hansen, J. Mielby, A. Riisager, Green Chemistry 13 (2011) 109–114.
[6] L. Lai, Y. Zhang, ChemSusChem 4 (2011) 1745–1748.
[7] J. Lu, Y. Yan, Y. Zhang, Y. Tang, RSC Advances 2 (2012) 7652–7655.
[8] T. Thananatthanachon, T.B. Rauchfuss, ChemSusChem 3 (2010) 1139–1141.
[9] N. Jiang, R. Huang, W. Qi, R. Su, Z. He, Bioenergy Research 5 (2012) 380–386.
[10] X. Tong, M. Li, N. Yan, Y. Ma, P.J. Dyson, Y. Li, Catalysis Today 175 (2011)
524–527.
[11] F. Wang, A. Shi, X. Qin, C. Liu, W. Dong, Carbohydrate Research 346 (2011)
982–985.
[12] N. Mittal, G.M. Nisola, W. Chung, Tetrahedron Letters 53 (2012) 3149–3155.
[13] Z. Zhang, B. Liu, Z. Zhao, Carbohydrate Polymers 88 (2012) 891–895.
[14] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith Jr., Green Chemistry 11 (2009)
1327–1331.
[15] X. Qi, H. Guo, L. Li, Industrial & Engineering Chemistry Research 50 (2011)
7985–7989.
[16] X. Guo, Q. Cao, Y. Jiang, J. Guan, X. Wang, X. Mu, Carbohydrate Research 351
In this work, the results obtained in aqueous reaction medium
demonstrate that the CrCl3·6H2O catalyst is effective for the isom-
erization of glucose to fructose as the main primary product and
the plot of fructose yield versus glucose conversion follows a gen-
eral curve independent of reaction variables, such as reaction time,
temperature, CrCl3·6H2O loading, halide ion and initial glucose con-
centration. With glucose loading of 5 wt% in the aqueous solvent,
fructose production showed a general linear increase at lower glu-
cose conversion (e.g. below ∼30%), while it gradually deviated from
the general curve at higher glucose conversion (e.g. over ∼30%),
accompanied with more secondary and further products, such as
HMF. To change the product distribution from glucose isomeriza-
tion, changing the solvent system is the most effective approach.
For example, when DMSO was employed as a second solvent in
this work, it was found that tuning the composition of DMSO and
H2O effectively changed the behavior of fructose production and
the distribution of products from glucose. In DMSO/H2O (v/v) of
8/2, higher yield of HMF and moderate yield of fructose could be
obtained while circumventing the formation of undesired prod-
uct, such as cellobiose. The Ea of glucose conversion in the studied
(2012) 35–41.
[17] F.H. Richter, K. Pupovac, R. Palkovits, F. Schu¨ th, ACS Catalysis 3 (2013) 123–127.
[18] R. Alamillo, A.J. Crisci, J.M.R. Gallo, S.L. Scott, J.A. Dumesic, Angewandte Chemie
International Edition 52 (2013) 10349–10351.
[19] A.S. Amarasekara, L.D. Williams, C.C. Ebede, Carbohydrate Research 343 (2008)
3021–3024.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038

G Model
CATTOD-8931; No. of Pages8
ARTICLE IN PRESS
8
S. Jia et al. / Catalysis Today xxx (2014) xxx–xxx
[20] K. Shimizu, R. Uozumi, A. Satsuma, Catalysis Communications 10 (2009)
1849–1853.
[41] E. Nikolla, Y. Román-Leshkov, M. Moliner, M.E. Davis, ACS Catalysis 1 (2011)
408–410.
[21] J.B. Binder, R.T. Raines, Journal of the American Chemical Society 131 (2009)
1979–1985.
[42] V. Choudhary, A.B. Pinar, S.I. Sandler, D.G. Vlachos, R.F. Lobo, ACS Catalysis 1
(2011) 1724–1728.
[22] T.S. Hansen, J.M. Woodley, A. Riisager, Carbohydrate Research 344 (2009)
2568–2572.
[23] A.I. Torres, M. Tsapatsis, P. Daoutidis, Computers & Chemical Engineering 42
(2012) 130–137.
[24] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597–1600.
[25] G. Yong, Y. Zhang, J.Y. Ying, Angewandte Chemie International Edition 47 (2008)
9345–9348.
[26] C. Li, Z. Zhang, Z.K. Zhao, Tetrahedron Letters 50 (2009) 5403–5405.
[27] Y. Su, H.M. Brown, X. Huang, X. Zhou, J.E. Amonette, Z.C. Zhang, Applied Catalysis
A: General 361 (2009) 117–122.
[28] E.A. Pidko, V. Degirmenci, R.A. Santen, E.J.M. Hensen, Angewandte Chemie
International Edition 49 (2010) 2530–2534.
[29] Z. Yuan, C. Xu, S. Cheng, M. Leitch, Carbohydrate Research 346 (2011)
2019–2023.
[30] Y. Zhang, E.A. Pidko, E.J.M. Hensen, Chemistry – A European Journal 17 (2011)
5281–5288.
[31] B. Kim, J. Jeong, D. Lee, S. Kim, H. Yoon, Y. Lee, J.K. Cho, Green Chemistry 13
(2011) 1503–1506.
[43] V. Choudhary, S.H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N.S. Marinkovic, A.I.
Frenkel, S.I. Sandler, D.G. Vlachos, Journal of the American Chemical Society
135 (2013) 3997–4006.
[44] V. Choudhary, S.I. Sandler, D.G. Vlachos, ACS Catalysis 2 (2012) 2022–2028.
[45] S. Bali, M.A. Tofanelli, R.D. Ernst, E.M. Eyring, Biomass and Bioenergy 42 (2012)
224–227.
[46] E.F. Dunn, D. Liu, E.Y.-X. Chen, Applied Catalysis A: General 460–461 (2013)
1–7.
[47] Y. Tewari, Applied Biochemistry and Biotechnology 23 (1990) 187–203.
[48] V. Choudhary, A.B. Pinar, R.F. Lobo, D.G. Vlachos, S.I. Sandler, ChemSusChem 6
(2013) 2369–2376.
[49] Y. Yia, M. Hab, J. Leea, C. Chung, Chemical Engineering Journal 180 (2012)
370–375.
[50] B.F.M. Kuster, Starch/Stärke 42 (1990) 314–321.
[51] D.B. Wong, K.P. Sokolowsky, M.I. El-Barghouthi, E.E. Fenn, C.H. Giammanco,
A.L. Sturlaugson, M.D. Fayer, Journal of Physical Chemistry B 116 (2012)
5479–5490.
[52] F. Rallc, F. Rodante, P. Silvestroni, Thermochimica Acta 1 (1970) 311–316.
[53] M. Ohara, A. Takagaki, S. Nishimura, K. Ebitani, Applied Catalysis A: General
383 (2010) 149–155.
[32] L. Hu, Y. Sun, L. Lin, Industrial & Engineering Chemistry Research 51 (2012)
1099–1104.
[33] S. Hu, Z. Zhang, J. Song, Y. Zhou, B. Han, Green Chemistry 11 (2009) 1746–1749.
[34] Y. Tian, J. Deng, T. Pan, Q. Guo, Y. Fu, Chinese Journal of Catalysis 32 (2011)
997–1002.
[54] C. Kooyman, K. Vellenga, H.G.J. Wilt, Carbohydrate Research 54 (1977) 33–44.
[55] H. Asaoka, Carbohydrate Research 137 (1985) 99–109.
[56] J. Lecomte, A. Finiels, C. Moreau, Starch/Stärke 54 (2002) 75–79.
[57] R.W. Nagorski, J.P. Richard, Journal of the American Chemical Society 123 (2001)
794–802.
[58] C.B. Rasrendra, B.A. Fachri, I.G.B.N. Makertihartha, S. Adisasmito, H.J. Heeres,
ChemSusChem 4 (2011) 768–777.
[59] S.J. Angyal, Advances in Carbohydrate Chemistry and Biochemistry 42 (1984)
15–68.
[60] Q. Ren, Y. Huang, H. Ma, F. Wang, J. Gao, J. Xu, BioResources 8 (2013) 1563–1572.
[61] S.J. Angyal, Carbohydrate Research 263 (1994) 1–11.
[62] L.E. Orgel, An Introduction to Transition-Metal Chemistry, Ligand Field Theory,
2nd ed., Methuen, London, 1966.
[35] Z. Zhang, Q. Wang, H. Xie, W. Liu, Z. Zhao, ChemSusChem 4 (2011) 131–138.
[36] W.T. Cefalu, F.B. Hu, Diabetes Care 27 (2004) 2741–2751.
[37] S. Lima, A.S. Dias, Z. Lin, P. Brandão, P. Ferreira, M. Pillinger, J. Rocha, V. Calvino-
Casilda, A.A. Valente, Applied Catalysis A: General 339 (2008) 21–27.
[38] M. Moliner, Y. Román-Leshkov, M.E. Davis, Proceedings of the National
Academy of the Sciences of the United States of America 107 (2010) 6164–6168.
[39] R.O.L. Souza, D.P. Fabiano, C. Feche, F. Rataboul, D. Cardoso, N. Essayem, Catal-
ysis Today 195 (2012) 114–119.
[40] S. Saravanamurugan, M. Paniagua, J.A. Melero, A. Riisager, Journal of the Amer-
ican Chemical Society 135 (2013) 5246–5249.
Please cite this article in press as: S. Jia, et al., Reaction media dominated product selectivity in the isomerization of glucose by chromium
trichloride: From aqueous to non-aqueous systems, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.02.038