215
,
be determined. In addition, by calculating the activation energies
between 150 and 170 ◦C, a relationship between observed activities
and reaction mechanism hypothesis can be correlated. Mechanis-
tic studies have been performed by Davis and co-workers on the
Sn-beta type solid Lewis acid for glucose conversion in water and
they concluded glucose/fructose isomerization proceeded through
a C2 position hydride shift mechanism [13]. While Choudhary et al.
performed kinetic isotope effects of labeled glucose catalyzed by
CrCl3, and AlCl3 under aqueous conditions [32], we also examined
the kinetic isotope effect with glucose molecules labeled at the 2-H
position under biphasic conditions relevant to glucose dehydration.
In3+, Sn4+, La3+, Dy3+, Yb3+, Zn2+, Ge4+, Cr3+, Cr2+, Cu2+, have been
reported as active catalysts for converting C5 or C6 aldose sugars
(glucose or xylose) to furanics in an aqueous media [17–25], organic
solvents [26], or ionic liquids [15,27]. Most of these reports have
similarly focused on integrating the Lewis acid properties of the
metal as the coordination center for glucose isomerization, and the
protons resulting from the metal hydrolysis as a Brønsted acid to
catalyze the dehydration reaction in order to achieve high yields
of the furanic compounds. For example, a HMF yield of over 60%
was achieved by using a combination of AlCl3 and HCl in a one-pot
biphasic system as reported by two research groups [17,23]. How-
ever, to the best of our knowledge, there is no study in the literature
to systematically tune the reaction conditions such as Lewis acidic
metal type, pH values of the reaction media, temperature, etc. As
a result, a comprehensive examination of intrinsic kinetic proper-
ties including activation energies (Ea) and reaction rates so as to
understand the Lewis acidic metal catalysts would provide further
insight into how their Lewis acidic characteristics affect the glucose
dehydration results.
2. Materials and methods
2.1. Preparation of reaction solution
Two types of Lewis acid salt solutions were prepared for
reaction testing: water-compatible lanthanide metal salts and
water-sensitive (hydrolysable) salts (e.g. AlCl3). For the lanthanide
metal salts, 25 mM LaCl3, DyCl3, or YbCl3 (Aldrich) aqueous solu-
tions were made by adding the appropriate amount of LnCl3 into
nanopure water (18 Mꢀ cm) while adjusting the pH to the targeted
value through the addition of HCl. Similarly, 5 mM AlCl3 or 25 mM
InCl3 or GaCl3 (Aldrich) aqueous solutions were prepared by the
same protocol. A glass electrode (6.0233.100, Metrohm) in combi-
nation with a Metrohm 798 MPT Titrino automatic titrator was used
to measure the pH value of the solutions. The as-made solutions
were then saturated with NaCl for further use. Glucose and fruc-
tose were purchased from Fisher Scientific as ACS Certified grade
and used as received.
It is generally known that when metal chlorides are contacted
with water, depending on their hydrolysis constant values, they
can be classified as either “water-compatible” or “water-sensitive”
categories. For example, “water-compatible” Lewis acid metal
cations refer to those associated with small hydrolysis constant
values, resulting in a limited extent of hydrolysis. Kobayashi and
Manabe [28] extensively studied “water-compatible” Lewis acids
represented by lanthanide trifluoromethanesulfonates, Ln(OTf)3
for facilitating various organic transformations such as C C and
C
O bond formation. The Lewis activity of Ln(OTf)3 was ascribed
to the high pKh values of the Ln cations of between 7.6 and 8.5 lead-
ing to their existence as coordinately unsaturated Lewis acidic aqua
ions, which were believed to be the catalytically active species. Also,
exchange rate constants for the substitution of inner-sphere water
ligands or the water exchange rate constant (WERC) have been sug-
gested to play a role in controlling the Lewis acidity of metal ions
as AlCl3, are inactive in aqueous solutions, due to their favorable
formation of mononuclear and polynuclear species in the presence
of water molecules. Studies have shown that the speciation of these
metal cations in water is highly dependent on the concentration of
the metal cation and the solution pH [29,30]. In this regard, metal
salts such as AlCl3 can work as efficacious, water-compatible Lewis
acids as well as metal salts if used at a suitable pH.
sugars to HMF and concluded that strategies combining Lewis acid-
catalyzed isomerization and Brønsted acid-catalyzed dehydration
steps to convert glucose to HMF in either a biphasic system or in
ionic liquid “one-pot” reactor configuration are quite promising
[31]. Although many reports have focused on the efficacy of such
systems on the overall conversion and selectivity, there have been
limited studies performed to examine the kinetic profiles for the
isomerization/dehydration steps. Better kinetic information can
provide insights into the optimal ratio for the two catalytic func-
tions as well as process and reactor conditions to optimize HMF
production. Additionally, since many of the Lewis acids readily
hydrolyze in water, which can either be the solvent or be pro-
duced during reaction, studies that can distinguish Brønsted and
Lewis acidity and their effect on the catalytic process are neces-
sary. In the current study, the effect of the nature of the Lewis
acidic metal salts and different initial solution pH values ranging
from 2.5 to 5.5 on the conversion kinetics of glucose were explored
by examining Al, Ga, and In salts, which are hydrolysable Lewis
acids, and La, Dy, and Yb salts, which are water-compatible Lewis
acids. By using the same initial pH conditions, the intrinsic effect of
Lewis acid strength on the glucose dehydration kinetic profile could
2.2. Biphasic reaction conditions
The biphasic glucose and fructose dehydration reaction systems
were performed in 10 ml thick-walled glass reactors (Alltech). An
oil bath was used to heat the reactor rapidly to the desired reac-
tion temperature (170 ◦C), with temperature and stirring being
controlled by an Isotemp digital stirring hot plate (Fisher Scien-
tific).
In a typical experiment, 1.5 g of an aqueous Lewis acid solu-
tion containing 5 wt% of glucose or fructose and 3.0 g of organic
extracting solvent, sec-butyl phenol (SBP) were added to the All-
tech reactor. A triangular stir bar (Fisher Scientific) was also added
to allow for adequate agitation. The reactor was then sealed with
PTFE liner covered caps (Fisher Scientific) and immersed in the oil
bath at 170 ◦C and stirred at 350 rpm. At specified times, the reac-
tors were removed from the oil bath and then cooled rapidly in
an ice bath to quench the reaction. For all the reactions, significant
amounts of soluble and insoluble brownish humins were formed
at production rates that were difficult to quantify. As with previous
literature reported, the humin formation precluded a closed carbon
balance analysis and thus the only product analyzed was HMF.
2.3. pH effect on Lewis acid salt activities
Experiments were performed to study the pH effect on the activ-
ity of the different Lewis acid metal salts. YbCl3 (25 mM) was chosen
to represent the water-compatible Lewis acids and AlCl3 (5 mM)
the water-sensitive Lewis acids. In either case, aqueous solutions
were made and the pH was adjusted to 2.5, 3.5, 4.5 or 5.5. The
same experimental procedure as described in Section 2.2 was con-
ducted with each of the solutions using a reaction temperature of
170 ◦C. Considering the relatively small size of the reaction vial,
heat transfer was assumed to be fast and as such time delay was
not considered for the kinetic analysis.