Kinetic and mechanistic study of glucose isomerization using homogeneous organic br?nsted base catalysts in water

Article abstract of DOI:10.1021/acscatal.5b00316

The isomerization of glucose to fructose represents a key intermediate step in the conversion of cellulosic biomass to fuels and renewable platform chemicals, namely, 5-hydroxymethyl furfural (HMF), 2,5-furandicarboxylic acid (FDCA), and levulinic acid (LA). Although both Lewis acids and Br?nsted bases catalyze this reaction, the base-catalyzed pathway received significantly less attention due to its lower selectivity to fructose and the poor yields achieved (<10%). However, we recently demonstrated that homogeneous organic Br?nsted bases present a similar performance (～31% yield) as Sn-containing beta zeolite, a reference catalyst for this reaction. Herein, we report on the first extensive kinetic and mechanistic study on the organic Br?nsted base-catalyzed isomerization of glucose to fructose. Specifically, we combine kinetic experiments performed over a broad range of conditions (temperature: 80-120 °C; pH 9.5-11.5; reactant: glucose, fructose) with isotopic studies and in situ ¹H NMR spectroscopy. Pathways leading to isomerization and degradation of the monosaccharides have been identified through careful experimentation and comparison with previously published data. Kinetic isotope effect experiments were carried out with labeled glucose to validate the rate-limiting step. The ex situ characterization of the reaction products was confirmed using in situ ¹H NMR studies. It is shown that unimolecular (thermal) and bimolecular (alkaline) degradation of fructose can be minimized independently by carefully controlling the reaction conditions. Fructose was produced with 32% yield and 64% selectivity within 7 min.

Full text of DOI:10.1021/acscatal.5b00316

Research Article
pubs.acs.org/acscatalysis
Kinetic and Mechanistic Study of Glucose Isomerization Using
Homogeneous Organic Brønsted Base Catalysts in Water
Jack M. Carraher, Chelsea N. Fleitman, and Jean-Philippe Tessonnier*
Department of Chemical and Biological Engineering, Iowa State University, Ames, Iowa 50011, United States
NSF Engineering Research Center for Biorenewable Chemicals (CBiRC), Ames, Iowa 50011, United States
S
* Supporting Information
ABSTRACT: The isomerization of glucose to fructose
represents a key intermediate step in the conversion of
cellulosic biomass to fuels and renewable platform chemicals,
namely, 5-hydroxymethyl furfural (HMF), 2,5-furandicarbox-
ylic acid (FDCA), and levulinic acid (LA). Although both
Lewis acids and Brønsted bases catalyze this reaction, the base-
catalyzed pathway received significantly less attention due to
its lower selectivity to fructose and the poor yields achieved
(<10%). However, we recently demonstrated that homoge-
neous organic Brønsted bases present a similar performance
(∼31% yield) as Sn-containing beta zeolite, a reference catalyst
for this reaction. Herein, we report on the first extensive
kinetic and mechanistic study on the organic Brønsted base-catalyzed isomerization of glucose to fructose. Specifically, we
combine kinetic experiments performed over a broad range of conditions (temperature: 80−120 °C; pH 9.5−11.5; reactant:
glucose, fructose) with isotopic studies and in situ 1H NMR spectroscopy. Pathways leading to isomerization and degradation of
the monosaccharides have been identified through careful experimentation and comparison with previously published data.
Kinetic isotope effect experiments were carried out with labeled glucose to validate the rate-limiting step. The ex situ
characterization of the reaction products was confirmed using in situ 1H NMR studies. It is shown that unimolecular (thermal)
and bimolecular (alkaline) degradation of fructose can be minimized independently by carefully controlling the reaction
conditions. Fructose was produced with 32% yield and 64% selectivity within 7 min.
KEYWORDS: glucose, fructose, base-catalyzed isomerization, Lobry de Bruyn−Alberda van Ekenstein, mechanism, kinetics,
proton transfer
should meet the following criteria: (i) guarantee a maximum
level of isomerization with a minimum of reaction byproducts,
(ii) be environmentally safe and nontoxic, (iii) be synthesized
at a large scale and low cost, (iv) be easily separated from the
product stream, and (v) be reusable without significant loss in
performance.8
Zhao et al. reviewed a broad range of homogeneous and
heterogeneous chemical catalysts active for the isomerization of
glucose to fructose, including aqueous inorganic hydroxides,
organic bases, alkali cation-exchanged zeolites, hydrotalcites,
anion exchange resins, lanthanide salts, tin-containing beta-
zeolite (Sn-β), and others.9 Many catalysts offer a high
selectivity to fructose but often at the expense of low glucose
conversion. Additionally, leaching of caustic species, difficult
and/or expensive catalyst syntheses and/or regeneration, as
well as costly separation from product streams and waste
disposal are frequently a concern.9 Lewis acids, and in particular
Sn-β, emerged as promising carbohydrate isomerization
INTRODUCTION
■
The isomerization of glucose to fructose represents an
important industrial conversion for food applications (high
fructose corn syrup) and is currently regarded as a key
intermediate step in the production of platform chemicals from
biorenewables, namely, 5-hydroxymethylfurfual (HMF), 2,5-
furandicarboxylic acid (FDCA), and levulinic acid (LA).1−3
These biobased building blocks are notably central to the
production of plastics, green solvents, lubricants, and valeric
biofuels.1−3 The industrial state-of-the-art for the large-scale
production of high fructose corn syrup currently utilizes
immobilized enzymes (D-xylose isomerase) as a biocatalyst.
4−6 Fructose yields of 42% are obtained by this method,6 which
operates near thermodynamic equilibrium.7 However, the
enzymatic system suffers from a relatively short half-life6 and
requires the careful control of pH, temperature, feedstock
purity, and flow rate to prevent any irreversible catalyst
deactivation through microbial growth and/or thermal
degradation.4−6 There is, therefore, an ongoing search for
robust high performance chemical catalysts that could replace
costly enzymes. To be considered for both food applications
and the production of renewable chemicals, these new catalysts
Received: February 13, 2015
Revised: April 14, 2015
© XXXX American Chemical Society
3162
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
Scheme 1. Glucose to Fructose Isomerization Pathways Proposed for Sn-β: Intramolecular Hydride Shift at Framework Sn
(Right) or by Deprotonation at α-C and Proton Transfer through an Enediol Intermediate at Extra-Framework SnO2 in
a
Hydrophobic Pores (Left)
a
Adapted from ref 18.
catalysts.2,10−17 Sn-β isomerizes glucose to fructose with high
selectivity (51%) and yields of up to 31%.10 The reacted
solution typically contains ∼84% of carbohydrates (glucose,
fructose, mannose) and ∼16% of carboxylic acids and other
byproducts formed through retro-aldol reactions of sugars.11,13
Several studies aimed at elucidating the mechanisms involved in
the Sn-catalyzed isomerization reaction have been carried
out.11−13,18,19 Framework tetrahedral Sn and extraframework
octahedral Sn species (SnO2) were both shown to catalyze the
isomerization of glucose to fructose.12 However, NMR
investigations using isotope labeled glucose (gluc-C2-D)
demonstrated that while framework Sn catalyzes the intra-
molecular hydride shift from the C-2 to C-1 position,
extraframework SnO2 located within the hydrophobic pores
of the zeolite promotes the isomerization through base-
catalyzed proton transfer (Scheme 1).12
the degradation of monosaccharides under alkaline conditions,
which is known to produce over 50 different byproducts
(mostly sugar derived acids).25,38,39 However, we recently
achieved high fructose yield (32%) from glucose using organic
amine catalysts.27 To the best of our knowledge, only one other
group has reported high yields of fructose using an organic
basic catalyst in an aqueous environment.40 We proposed that
isomerization likely occurs via the Lobry de Bruyn−Alberda van
Ekenstein (LdB−AvE) mechanism, first reported in
1895.30,31,41 In this mechanism, the C-2 proton on the acyclic
form of glucose is abstracted by a Brønsted base, resulting in
the formation of an enediol intermediate, and followed by
hydrogen transfer from O-2 to O-1 and protonation of C-1
(Scheme 2).
Many mechanistic studies have been undertaken to
investigate one or multiple elementary steps involved in the
base-catalyzed isomerization. These studies have focused on
every aspect from ring opening20,22,42−46 to overall saccharide
degradation33,38,47−53 to the nature of intermediates involved in
both isomerization and degradation. However, many results
were found to be inconsistent, and the elementary steps
Brønsted bases represent another important class of catalysts
for the isomerization of glucose to fructose.9,20−32 Isomer-
ization under strongly alkaline conditions, though, resulted in
very poor selectivity to fructose. Fructose yields below
10%9,10,18,20,21,26,30,33−38 were reported and were likely due to
3163
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
round-bottom flask with olive shape stir bar. The mother
solution was purged with Ar for an additional 20 min once
dissolution of the sugar was complete. The pH was then
measured (Mettler Toledo SevenMulti, InLab Routine Pro
electrode) and adjusted to the desired initial pH by titration
with triethylamine (≥99%, Sigma-Aldrich) using a syringe. The
flask and the pH electrode were fitted with rubber septa to
minimize introduction of atmospheric CO2 which could cause
deviations in pH due to formation of carbonic acid during the
measurement. Eight 5.0 mL aliquots of the mother solution
were then transferred via syringe to Ar-filled 5 mL thick-walled
glass reactors (Chemglass Life Sciences) containing a V-shaped
stir bar. The reactor vessels were then placed in the preheated
1 L oil bath (digital stirring hot plate IKA RCT equipped with
PT1000 thermocouple in the oil bath). No more than eight
vials were placed in the oil bath simultaneously to minimize the
drop in temperature (∼5 °C). Preliminary tests performed with
an oil bath at 100 °C showed that the reaction mixture reaches
the target temperature within ∼120 s. The reactors were
removed at set time intervals, and the reaction was quenched
using an ice bath. All the reactors were later stored in a freezer
until analysis. Analogous experiments were carried out with
NaOH (99.8%, Fisher), piperidine (≥99%, Sigma-Aldrich), and
pyrrolidine (≥99.5%, Sigma-Aldrich) for comparison. Deu-
terated glucose (C2-D) was purchased from Cambridge
Isotope Laboratories, Inc. and used without any further
treatment for the isotope exchange and kinetic isotope effect
experiments. D2O was supplied by Sigma.
Scheme 2. Lobry de Bruyn−Alberda van Ekenstein (LdB−
AvE) Mechanism for the Base-Catalyzed Isomerization of
Glucose to Fructose
involved in the reaction have been a source of debate for over a
century. The nature of the intermediate, for instance, has been
called into question several times due to discrepancies between
results obtained by different groups when using isotopically
labeled sugars and/or solvents.20,31,32,54,55 Although many of
these investigations provide important clues about the
mechanism of action, the overwhelming majority of these
studies were performed under conditions that favor the
formation of various byproducts, which probably explains the
observed discrepancies.30,31 The nature of the intermediate(s),
the effect of pH and temperature on their formation, and the
mechanisms involved in the degradation of the sugars are yet to
be identified. Actually, true equilibrium concentrations of
glucose and fructose have never been achieved with Lewis acid
and Brønsted base catalysts due to the onset of degradation
reactions.10 Understanding the chemistry of carbohydrates as a
function of temperature and pH is critical to identify the
constraints that need to be considered for future catalyst design.
We report here on the first kinetic and mechanistic study for
selective organic Brønsted base-catalyzed glucose−fructose
isomerization. We recently demonstrated that organic Brønsted
bases, specifically tertiary amines, catalyze the isomerization of
glucose to fructose with high selectivity (63%) and yield (32%),
a performance similar to the state-of-the-art Sn-β catalyst (51%
selectivity, 31% yield).27 We identified triethylamine (TEA) as
an excellent candidate for further study based on its
performance and lack of contribution to Maillard browning.
The present work aims to develop an understanding of glucose
isomerization and glucose/fructose degradation based on
kinetics, kinetic isotope effect experiments, analysis of product
For analysis, the frozen samples were brought to room
temperature, diluted using appropriate amounts of 50:50
acetonitrile/water (HPLC grade, Fisher) containing 0.2 vol %
TEA, and analyzed with a Waters Acquity H-Class ultra-
performance liquid chromatograph (UPLC) equipped with
photodiode array (PDA) and evaporative light scattering (ELS)
detectors by methods analogous to those described previously
(Supporting Information).27 The instrument was calibrated
before analyzing each set of samples (typically each day).
Kinetic traces were constructed and analyzed with Origin Pro
9.1 software (OriginLabs).
Yields are reported after 15 min (ca. 6.5 apparent half-lives at
100 °C relative glucose consumption) or after three apparent
half-lives when studying temperature effects. Standard initial
reaction conditions were 0.52 M glucose or fructose, pHo 11
1
(12
1 mol % TEA), and 100 °C. The effect of varying
distributions, and in situ H NMR studies. The broad range of
[sugar]o, pHo (mol % TEA), or temperature was determined
while holding the other two variables constant at the above-
described conditions. Byproduct yields are defined as the
difference between the initial sugar concentration and the total
sugar (glucose, fructose, and mannose) after 15 min (or three
apparent half-lives when temperature was varied).
reaction conditions studied (temperature, number of active
sites, glucose concentration, pH, ionic strength) allows us to
propose for the first time a model that captures most of the key
parameters: the kinetics of both glucose and fructose
conversion, the degradation of fructose through unimolecular
and bimolecular pathways, the effect of temperature and pH on
the degradation pathways, the buffering effects of both sugars,
and the effect of atmospheric CO2. Reaction kinetics and
kinetic isotope effect experiments point to the rate-limiting
step. The calculated rate constants and thermodynamic
parameters provide key information on conditions required to
achieve high fructose yields.
1H NMR spectra were collected with a Bruker 600 MHz
NMR (AVIII600). In order to study the rate of glucose
conversion, a time-arrayed data set was acquired in which a
single scan 1D proton spectrum was obtained at an interval of
5 s over the course of 1500 s. The temperature was regulated to
353.4 K during the acquisition.
RESULTS
MATERIALS AND METHODS
■
■
In a typical experiment, a 50.0 mL mother solution containing
0.56 0.02 M (10 wt %) D-glucose (≥99.5%, Sigma-Aldrich)
or D-fructose (≥99%, Sigma-Aldrich) was prepared with Ar-
purged DI water (Barnstead Nanopure Infinity Pyrogen Free
UF Laboratory Water System Model D8981) in a 100 mL
We previously demonstrated that primary, secondary, and
tertiary amines catalyze the isomerization of glucose to fructose
with high yields, ranging between 15 and 32% depending on
the chemical structure and pKa of the base.27 Primary and
secondary amines were the least selective due to the formation
3164
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
of colored byproducts through the Maillard reaction, a
nonenzymatic browning reaction well-known in food scien-
ces.56,57 UV−vis and H NMR spectra confirmed that tertiary
1
amines do not react and are not consumed by the Maillard
reaction, in good agreement with the mechanism for this
reaction.58 Triethylamine was found to be the most promising
homogeneous catalyst based on its high performance,
availability, low toxicity, and low cost.27,59,60 In addition,
various triethylamine derivatives are commercially available,
which offers many possibilities for the synthesis of the
corresponding heterogeneous catalysts.
The experimental setup was further improved since our
preliminary experiments. Various protocols were tested in order
to increase the heating rate, lower the impact of atmospheric
CO2, and other possible artifacts. Aluminum heating blocks,
which are commonly used for reactions with 5−10 mL reactors,
were found to have a good heat conductivity but a relatively low
heat capacity. Temperature dropped by approximately 10 °C
when placing more than five vials at a time. Therefore, a
conventional 1 L oil bath was preferred to perform kinetic
studies as temperature dropped by only ∼5 °C and the reactors
reached the target reaction temperature within ∼120 s. The
conversion of glucose during heating is relatively negligible at
pHo 10.5 or below; however, at pHo 11.5, nearly 17%
conversion of glucose is observed by the first point of data
collection (Figure 1). The experimental protocol was also
improved in order to prevent any artifact due to atmospheric
CO2. Carbonic acid significantly impacts both selectivity and
conversion (Supporting Information, Figure S1), and its effect
varies depending on the base and pH of the reaction mixture.
Therefore, all the reactions performed for the present work
were carried out under Ar. Finally, small differences in glucose
and fructose concentrations were measured when storing the
reacted solutions in a fridge for more than 24 h, thus indicating
that the reaction proceeds even at low temperature. However,
the composition did not change when storing the reactors in a
freezer, even for an extended period of time (1−2 weeks).
Isomerization from Glucose: Kinetics. The first few
[Glu]t data points were used to construct linear ln([Glu]t/
[Glu]o) versus t plots (eq 1) and observed first-order rate
constants, kobs, were obtained from the slope. The pH
decreased during this time period but typically by less than
0.3 units. The observed first-order rate constant was found to
be 9.3 0.9 × 10−4 s−1 at pHo 11.0 and showed no dependence
on initial glucose concentration (Table 1), consistent with the
kinetic traces previously reported under similar conditions.40
The zero-order dependence of kobs on [Glu] indicates that the
process is unimolecular with respect to glucose, likely ring
opening. kobs showed an upward trend with pHo (Figure 2).
The apparent pH dependence on reaction kinetics suggests that
ring opening most likely occurs from a deprotonated form of
ring glucose.20,24 As the reaction approached completion, the
pH decreased dramatically (typically by 2−4 units). This likely
results from the known degradation of sugars under alkaline
conditions yielding acid products,25,38,39 and as alluded to in
our previous work,27 is likely the leading factor in reversible
catalyst poisoning. Finally, kobs varied substantially with
temperature (Table 2). The plot of ln(kobs/T) versus 1/T
shown in Figure 3 is linear and fits eq 2, where kB is the
Boltzmann constant, h is Planck’s constant, R is the ideal gas
law constant, and T is temperature in Kelvin. The slope and
intercept provide activation parameters for glucose conversion
Figure 1. Plots of glucose (a) and fructose (b) concentrations as a
function of time. Each data point corresponds to a single experiment.
Reactions were carried out at 100 °C in degassed aqueous solutions of
■
●
▲
▼
0.52 M glucose with 0.03 , 1.29 red , 3.87 blue , 12.24 pink ,
◆
and 108.68 green
mol % TEA relative to glucose.
respectively (Table 3). Kinetic traces obtained from monitoring
the formation of fructose were much more complicated, and
they indicate fructose formation and decay on the same time
scale (Figure 1). At the highest pHo studied, fructose yields
were observed to reach a maximum and then decay. The
product distributions obtained from the above-described
reactions are complicated by this degradation, and are discussed
in detail in the next section, vide infra.
⎛
⎜
⎝
⎞
⎟
⎠
[Glu]t
[Glu]
ln
= −k × t
obs
(1)
o
; where [Glu]t = glucose concentration at time t, [Glu]o = initial
glucose concentration, kobs = observed rate constant, and t =
time in seconds.
⎛
⎜
⎝
⎞
⎟
⎠
⎛
⎜
⎞
⎟
kB
h
k
T
ΔS‡
R
ΔH‡
RT
ln
= ln
+
−
⎝
⎠
ΔH‡ and ΔS‡ of 58
8 kJ/mol and −144
30 J/molK,
(2)
3165
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
Table 1. Effects of Glucose Concentration on Kinetics and
a
Catalytic Performance
b
c
d
e
entry [Glu]o (M) 104 × kobs (s−1
)
pHf XGlu
YFru
YMan
SFru
1
2
3
4
5
6
7
8
9
0.288
0.390
0.521
8.4 0.1
8.9 0.1
9.1
9.9
9.5
8.7
9.3
9.4
9.9
9.5
9.9
58
56
57
52
49
55
56
59
64
25
26
28
0
5
7
4
4
43
46
49
44
49
49
47
46
33
10
14
15
1
2
1
f
h
0.523
23
24
g
h
0.523
0.523
i
j
10.9 0.2
9.4 0.2
8.6 0.2
10.7 0.2
27
-
0.644
0.746
2.900
27
27
21
6
7
6
a
Standard reaction conditions: 100 °C, 0.52 M glucose, 12 1 mol %
TEA relative to glucose (pHo 11.0), reactors quenched after 2, 3, 4, 5,
7, 10, 12, and 15 min. Conversion and yields reported to 1% error.
Rate constants were determined by fitting eq 1 to the data.
Conversion, yields, selectivity, and pHf reported at 15 min reaction
b
c
d
e
time. Glucose conversion. Fructose yield. Mannose yield. Fructose
f
g
selectivity. 7.5 mol % piperidine (pHo 11.0). 9.4 mol % pyrrolidine
(pHo 11.0). Relatively low fructose yields and larger rate constants
h
Figure 3. Effect of temperature on rate constants for glucose
consumption, unimolecular fructose consumption, and bimolecular
fructose consumption. kfructose unimolecular and kfructose bimolecular (for
reactions 3 and 4) were determined from the y-intercept and slope,
respectively, in the plot of kobs vs [Fru]ave shown in Figure 5.
are likely due to participation of secondary amines in bimolecular
i
Maillard browning (glu + R2NH).27 12 mol % TEA with 0.1 M NaCl
j
(ionic strength ca. 0.24 M). Difficult to distinguish from baseline in
UPLC.
Table 3. Activation Parameters for Glucose Isomerization
a
and Fructose Degradation
a
e
reaction
ΔH‡ (kJ/mol)
58
63 11
11
ΔS‡ (J/molK)
Ea100 C (kJ/mol)
b
Gluunimol
8
−144 30
−134 32
−274 45
61
66
14
c
Fruunimol
d
Frubimol
7
a
Standard reaction conditions: 0.52 M glucose, 12 mol % TEA (pHo
b
11.0). Isomerization from glucose to fructose. kobs obtained from eq 1
for glucose conversion. Isomerization from fructose to glucose. kuni
obtained from y-intercept in Figure 5. kbimol obtained from the slope
c
d
e
in Figure 5. Activation energies were calculated from temperature and
ΔH‡ = Ea − RT.
; where k = kobs obtained from eq 1, T = temperature in Kelvin,
kB = Boltzmann constant, h = Planck’s constant, R = ideal gas
constant, ΔS‡ = entropy of activation, and ΔH‡ = enthalpy of
activation.
Figure 2. Conversion, selectivity, or yield (left axis) and kobs (right
axis) as a function of initial pH for glucose isomerization in the
presence of TEA. Reactions were carried out at 100 °C in degassed
aqueous solutions containing 0.52 M glucose and 0.03−108.68 mol %
TEA relative to glucose.
Isomerization from Glucose: Products. Variations in
pHo (9.5−11.5) when heating a 0.52 M glucose solution to 100
°C resulted in a linear increase in glucose consumption (Figure
2). Likewise, yields of mannose increased linearly with pHo to a
maximum of 9%. Fructose yields, however, showed a peculiar
trend in that they rose steadily from 6% to 30% between pH 9.5
and 10.7, remained constant (30%) between pH 10.7 and 11.3,
and then decreased to 23% at pH 11.5. The maximum yield of
fructose obtained in this work was 32% (100 °C, pHo 11.3, 7
min). Increasing the initial glucose concentration from 0.29 to
0.75 M had little to no effect on glucose conversion or fructose
yield (Table 1). However, increasing glucose concentrations to
2.9 M (45 wt %) resulted in a decrease in fructose yield to only
21%. Factors like fructose degradation and buffering effects of
the sugars must be considered when analyzing the product
distribution of these very high [Glu] reactions. The multiple
degradation pathways described below in conjunction with the
long residence times (that is, the system being run in batch as
opposed to flow reactors) likely work together to diminish
yields at high [Glu].
Table 2. Effects of Temperature on Kinetics and Catalytic
a
Performance
b
c
d
e
temp (°C)
104 × kobs (s−1
)
XGlu
YFru
YMan
SFru
60
80
1.6 0.2
5.6 0.5
26
52
51
56
14
24
28
19
2
6
6
52
45
56
35
100
120
26
39
1
6
f
-
a
Standard reaction conditions: 0.52 M glucose, 12 mol % TEA relative
to glucose (pHo 11.0). Rate constants were determined by fitting eq 1
to the data. Conversion and yields reported to 1% error. Conversion,
yields, and selectivity are reported after three apparent half-lives.
b
c
d
e
Glucose conversion. Fructose yield. Mannose yield. Fructose
f
selectivity. Difficult to distinguish due to noisy baseline in UPLC;
however, earlier results suggest 7% mannose.27
3166
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
Isomerization from Fructose: Kinetics. Study of the
reverse reaction (i.e., isomerization of fructose to glucose)
resulted in similar trends (Figure 4). Like glucose, the first few
Figure 5. Plot of kobs as a function of average fructose concentration:
[Fru]ave = 1/2([Fru]o + [Fru]240 s). All reactions were carried out with
an initial pH of 11.0 0.1.
Table 4. Temperature Dependence of Rate Constants
Obtained from Fructose Experiments
temp (°C)
104 × kunimol (s−1
)
104 × kbimol (M−1 s−1
)
80
100
120
3
19
29
1
2
5
9
8
3
2
15 10
ratio increases as OH− is consumed. Therefore, the rate
constant is expected to scale with [Fru]: kobs = kunimol + kbimol
×
[Fru]ave. The activation parameters for these two reactions are
shown in Table 3. Studies beginning with fructose were also
attempted at 60 °C, but kobs contained substantial error.
Additionally, these reaction conditions were not studied in
depth due to half-lives on the order of hours.
Fructosering → Products
kunimol
kbimol
(3)
(4)
Fructosering + X → Products
Figure 4. Plots of fructose (a) and glucose (b) concentrations as a
function of time. Each data point corresponds to a single experiment.
Reactions were carried out at 100 °C in degassed aqueous solutions of
Isomerization from Fructose: Products. The product
distributions obtained when beginning the isomerization from
fructose (Figure S2) were more complicated likely due to the
multiple reaction pathways observed kinetically (eqs 3, 4).
Variations in temperature resulted in no discernible trend with
regard to fructose conversion, Table 5 entries 3, 6, and 11 or
entries 4, 8, and 12, but glucose yields clearly decrease with
increasing temperature. The decrease in glucose yield at higher
temperatures is most likely due to reactions other than
isomerization of linear forms of the sugars, given the majority
of kobs is made up of the unimolecular process (Figure 5). In
■
●
▲
▼
0.52 M fructose with 0.03 , 1.14 red , 2.67 blue , 9.67 pink ,
◆
and 67.98 green
mol % TEA relative to fructose.
data points for fructose consumption fit eq 1, and kobs increased
with pHo (Figure S2). However, kobs for fructose consumption
had a linear dependence on [Fru]ave (Figure 5). At the highest
pHo (11.5), kinetic traces monitoring formation of glucose
reached a maximum and clearly decayed (Figure 4). Although
similar to isomerization from glucose, this behavior was far less
pronounced when fructose was the sugar initially present. The
rate constants obtained from the reaction series in Figure 5 are
shown in Table 4 and are indicative of unimolecular (y-
intercept) and bimolecular (slope) reactions (eqs 3, 4). Rate
constants obtained in the presence of NaOH in place of TEA
were similar, therefore, reactant X in eq 4 is believed to be
OH−, vide infra. Additionally, under the conditions of these
experiments, fructose is considered the excess reagent and OH−
the limiting reagent, as initially [Fru]/[OH−] = 520, and this
contrast, at 80 °C, where [Fru] has a significant effect on kobs
,
the increase in glucose yields with [Fru] could suggest that at
least some of the bimolecular pathway ultimately leads to
glucose formation. The effect is also observed at 100 °C, albeit
less pronounced, as kobs at 0.5 M [Fru] is approximately 4:1
unimolecular/bimolecular.
Isomerization from Glucose: Kinetic Isotope Effects.
Isomerization (pHo 10.9, 100 °C) of glucose isotopically
labeled at the C-2 position (glucose-C2-D) yielded kobs = 4.9
0.6 × 10−4 s−1. The corresponding kinetic isotope effect, kie =
3167
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
a
Table 5. Effects of Temperature and Concentration on Fructose Conversion at Three Apparent Half-Lives
b
c
d
e
f
entry
[Fru]ave (M)
temp (°C)
103 × kobs (s−1
)
XFru
YGlu
YMan
SGlu
1
2
0.260
0.460
0.490
0.697
0.233
80
80
0.53 0.1
0.55 0.1
0.65 0.1
0.90 0.1
2.0 0.1
2.0 0.2
2.3 0.1
2.4 0.1
2.4 0.1
3.3 0.6
3.2 0.5
3.8 0.6
22
46
46
50
58
49
62
47
48
52
40
45
10
24
24
28
21
20
18
24
25
18
16
17
4
7
7
3
6
8
52
51
56
36
41
29
51
52
34
40
38
3
80
4
80
5
100
100
100
100
100
120
120
120
g
h
6
0.391
-
7
0.425
0.616
0.714
0.204
0.379
0.525
3
9
2
5
5
5
8
9
10
11
12
a
Standard reaction conditions: 12 1 mol % TEA (pHo 11.0), vary [Fru]o and temperature. Rate constants were determined by fitting eq 1 to the
b
c
d
e
data. Product data reported after three apparent half-lives. [Fru]ave = 1/2([Fru]o + [Fru]∞). Fructose conversion. Glucose yield. Mannose yield.
f
g
h
Glucose selectivity. 10 mol % NaOH (pHo = 11.05). Difficult to distinguish from baseline in UPLC.
1
Figure 6. H NMR spectra of glucose (top) and fructose (bottom) in the presence of TEA, 600 MHz, D2O. TEA signals are omitted for clarity.
3.8, indicates that deprotonation of the C-2 position plays a
substantial role in the in the conversion of the linear form of
glucose. In addition, evidence for competitive reactions from
linear glucose is easily seen when considering the product yields
from glucose-C2-D relative to glucose−C2-H. Glucose
conversion, fructose yields, and fructose selectivity (36%, 7%,
and 20%, respectively) increased when replacing deuterated
glucose-C2-D by glucose-C2-H (47%, 30%, and 64%,
respectively). Mannose yields remained relatively unchanged
at 5% (glu-D) and 6% (glu-H). Reactions with glu-H carried
out in D2O were identical in both kinetics and product
distribution to those in H2O, as indicated by the solvent kinetic
isotope effect kie = 1. The solvent is therefore not expected to
play a kinetically significant role as a proton donor (e.g., for the
protonation of the saccharides) during isomerization.
by integration of the following regions: fructose (2 H δ 4.25−
4.35) and glucose (0.6 H δ 3.5−3.65) relative to TEA (9H δ
1.09) signals, used here as an internal standard. The time-
resolved evolution of these signals and the in situ kinetic traces
for the reaction of 0.5 M glucose and 12 mol % TEA (pDo =
11.0) at 82 °C are shown in Figures 7 and 8. The shifts in peak
positions observed during the first 66 s (Figure 7) are due to
the increase in temperature from 22 to 82 °C. Interestingly, the
onset of the fructose signal is slightly shifted compared to the
glucose consumption (Figures 7 and 8). This shift could be due
to integration errors arising from the signal/noise ratio of a
broad fructose signal particularly obvious at low concentration.
1H NMR spectra acquired at room temperature before and
after reaction (Figure S3) only revealed peaks characteristic of
the carbohydrates. Byproducts such as carboxylic acids were not
observed directly, most probably because the concentration of
each byproduct was below the detection limit of the technique.
Room-temperature spectra acquired before and after reaction
suggested a glucose conversion and fructose yield on the order
of 23% and 17%, respectively. UPLC analysis of the same
solution revealed 29% glucose conversion with 18% yield of
fructose, thus confirming that NMR is a reliable technique to
Isomerization from Glucose: Products Formation and
1
Kinetics by In Situ H NMR. Glucose and fructose present
several distinctive peaks in 1H NMR spectra, which can be used
to monitor changes in concentrations of both sugars during
reaction (Figure 6). Reactions were carried out in J. Young
tubes that were sealed under ambient temperature and
pressure. The concentrations of the sugars were determined
3168
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
Figure 7. 1H NMR in situ kinetics of 0.52 M glucose, 12 mol % TEA relative to glucose (pDo = 11.0), 82 °C, 600 MHz, D2O. Bitmap of 1H NMR
(horizontal) and time (vertical) from 0 (bottom) to 1500 s (∼3 apparent half-lives, top). Blue signals at 4.00−4.05, 4.13−4.15, and 4.25−4.35 ppm
correspond to fructose formation.
further kinetic studies were done by NMR. Of interest,
however, was (i) the validation that the same products are
detected by NMR and UPLC and (ii) the possibility to use
variations in chemical shifts of TEA signals as an in situ pH
probe (Figure S4).
DISCUSSION
■
General Considerations. Due to limitations in exper-
imental design and the complexity of the system, the observed
rate constants obtained from kinetic traces monitoring the
consumption of either glucose or fructose are at the very least
composite rate constants for the reversible process (eq 5) and
degradation reactions outlined in Scheme 4. Even at the lowest
pHo studied, ca. 7% of glucose was converted to products by
the time the first data point was collect (120 s). By this time,
both forward and reverse reactions were active and contributed
to the observed rate constant. We were therefore unable to
study the forward and reverse reactions independently.
Additionally, considerations must be made for decomposition
pathways and changes in pH under reaction conditions.
1
Figure 8. In situ H NMR kinetic traces for 0.52 M glucose and 12
mol % TEA relative to glucose, 600 MHz, D2O, 82 °C. Sugar
concentrations (glucose: red ; fructose: blue ) were exper-
imentally determined relative to TEA signals. The solid black lines are
kforward
GlucoseHooooooooIFructose
kreverse
fits to equations for exponential decay.
quantify carbohydrates. However, these values are on the order
expected for pDo ∼ 9.5−10.2. Analogous experiments carried
out at pH 10.2 in 5 mL reaction vials and heated in the oil bath
yielded a glucose conversion of 25% and fructose yield of 18%
(5)
The thermodynamic maximum yield of fructose, based on
glucose−fructose equilibrium, is 57.4% at 100 °C (does not
take into account the formation of mannose).7 However, both
our previous27 and current work indicate glucose conversions
far beyond that (75%) and fructose yields well below 57%
(32%). Clearly, glucose is either diverted from the pathway that
generates fructose and/or fructose is consumed under reaction
conditions. Both are implicated by the kinetic traces in Figures
1 and 4, and the fructose bimolecular reaction observed in
Figure 5. Fructose decomposes more readily than glucose,61
likely due to the relative stability of the pyranose versus
furanose forms of the sugars. The effect of fructose
decomposition is particularly pronounced at high pHo.
1
after 8 min at 100 °C. It appears that while H NMR is a
suitable method for determination of glucose conversion and
fructose yield at room temperature, the use of TEA as the
internal standard provides erroneous results due to its volatility
at 82 °C and the significant difference in head space between 5
mL reaction vials (∼0.5 mL head space) and a J. Young tube
(2−3 mL head space). As a result [TEA] in solution is lower at
elevated temperatures in a J. Young tube relative to the 5 mL
reaction vials. This causes a decrease in pH and a false standard
against which sugar signals are compared. For this reason and
the necessity of remaining below the boiling point of D2O, no
3169
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
a
Scheme 3. Proposed Mechanism for Isomerization of Glucose to Fructose
a
Mannose was omitted for clarity.
a
Scheme 4. Decomposition Pathways in Base-Catalyzed Isomerization of Glucose to Fructose
a
Mannose was omitted for clarity.
Reactive Species Involved in the Isomerization
Reaction. We proposed in our previous work27 a reaction
mechanism consistent with the kinetic model summarized by
Kooyman et al.62 in which isomerization occurs after
deprotonation of an acyclic form of the sugar. Here, the lack
of effect that initial glucose concentration has on kobs (Table 1)
supports our previous assessment that glucose conversion is a
unimolecular process (Scheme 3). The pH dependence of kobs
(Figure 2) suggests ring opening from glucose anions and
requires expression of [Glu−] as a function of [OH−] and
[Glu]total in the rate law to take this effect into account (eq 6
and Supporting Information). This is consistent with many
previous works carried out under high concentrations of
inorganic bases and having no dependence on [OH−].20,24
Under such extreme conditions, [OH−] ≫ KbGlu and the first
term in eq 6 simplifies to −k1[Glu]total. Under our experimental
conditions, a distinction cannot be made between deprotona-
tion of acyclic glucose followed by intramolecular proton
transfer proposed by de Wit et al.20 (Scheme 3) or proton
abstraction by an external base (Scheme 2) due to the multiple
pathways by which intermediates may react. As a result, the two
reaction pathways are kinetically difficult to distinguish without
going to the extremes of high [OH−]. However, at such
extremes, the side reactions will likely become too over-
whelming to make clear distinctions. A strong argument can be
made in favor of the intramolecular pathway, as will be
discussed in more detail below. Additionally, rate constants and
product distributions obtained with piperidine, pyrrolidine, and
NaOH at iso-pHo (11.0) were similar to TEA, except the
contribution from bimolecular Maillard browning for secondary
amines. The similarity in kobs between NaOH and TEA
indicates that conversion of glucose is not dependent on the
structure and/or pKa of the external base; the basic species
involved in the reaction are the hydroxide ions (OH−)
generated in situ through interaction between the amine and
water. Similar observations have been made with tetramethyl-
guanidine and NaOH.40 Product distributions and kobs for the
reverse reaction are similar for TEA and NaOH, Table 5,
entries 6−7. This also supports that OH− is the unidentified
reactant “X” in eq 4 and that the bimolecular reaction likely
describes the alkaline degradation of fructose (e.g., the
formation of carboxylic acids from fructose though nucleophilic
attack of OH−).33,52,53
d[Glu−]
= ‐k1[Glu−] + k−1[Glu−open
]
dt
[OH−][Glu]total
= − k
+ k−1[Glu−open
]
1 KbGlu + [OH−]
(6)
d[Glu−open
]
= − (k−1 + k2 + k5 + k6[OH−])[Glu−open
+ k1[Glu−] + k−2[ED]
]
dt
(7)
3170
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
opening22 ultimately returning to glucose. Finally, the fact that
all of these reactions (except for decomposition) are reversible
d[ED]
dt
= − (k−2 + k3 + k7)[ED] + k2[Glu−open] + k−3[Fru−open
]
(8)
and have similar forward and reverse rate constants (Koverall
k
=
forward/kreverse = 1.348 at 100 °C)7 is an additional hindrance to
d[Fru−open
]
= − (k−3 + k4 + k8 + k9[OH−])[Fru−open] + k3[ED]
high fructose yields. In light of this analysis, it is easy to
understand why this reaction system is so challenging and why
most works reported low selectivities to fructose. Clearly,
temperature and pH have significant effects on the reaction
pathways and therefore yields. Of equal importance is the
residence time of the sugars under these reaction conditions as
repeated cycling between glucose and fructose results in
degradation through the intermediates, thus consumption of
the carbohydrates. This can be easily seen when calculating the
carbon balance (taking only the sugars into account) as a
function of time, as it decreased from 92% to 84% when
extending the reaction time from 5 to 15 min. During that 10
min time span, fructose yields only increased by an additional
1−2% at most. Interestingly, the carbon balance at 30 min
calculated for TEA (79%) was similar to the value obtained for
Sn-β (84%), thus indicating that similar levels of sugar
degradations can be expected for Lewis acids.
Intramolecular vs Bimolecular Deprotonation.
Although the results indicate that deprotonation occurs from
acyclic forms of the sugars, it is not obvious which species is
responsible for accepting the protonfor example, intra-
molecular proton transfer (deprotonation of C-2, protonation
of O-5; reaction 2 of Scheme 3) or bimolecular deprotonation
with OH− followed by reprotonation from H2O (Scheme 2).
The results obtained for glu-C2-H in D2O are somewhat
ambiguous. If deprotonation of acyclic forms of the sugars is a
bimolecular reaction, and reprotonation occurs from the
solvent, then a solvent kie = 1 could suggest that either (i)
reprotonation of the enediol is post-rate-limiting step (as one
would expect to observe a kinetic isotope effect from the
solvent) or (ii) deprotonation of the enediol is an intra-
molecular process and, therefore, does not involve the solvent.
Unfortunately, no such distinction can be made from these
results. However, considering the extreme conditions (high
[OH−]) in regards to the above proposed reaction mechanism,
one would expect to see the following. If OH− is responsible for
deprotonation of acyclic sugar as shown in Scheme 2, then k2
and k−3 should exhibit first-order dependence on [OH−].
Therefore, in the limit of high [OH−], the term k2 [OH−]/(k−1
+ k2 [OH−] + k5 + k6 [OH−]) simplifies to k2/(k2 + k6). As a
result, fructose yields would be expected to plateau. To the
contrary, an intramolecular deprotonation event (having no
dependence on [OH−]) would yield k2/(k6 [OH−]) at high
[OH−]. Infinitely high [OH−] would favor direct decom-
position (k6) without formation of enediol intermediate and
therefore no formation of fructose. Admittedly, these treat-
ments are oversimplified. They do not take into account
competing reactions for other intermediates or bimolecular
decay of fructose, but the low yields reported in the literature at
high [OH−] anecdotally support intramolecular proton
abstraction. Additionally, a marked decrease in fructose yields
and increase in degradation products were reported during
base-catalyzed isomerization as the ratio of NaOH to glucose
was increased from 0.12 to 1.00.40 On the other hand, strictly
invoking intramolecular proton transfer (reversible proton
transfer exclusively between C-2 and O-5) does not allow for
the incorporation or loss of deuterium or tritium observed in
some of the isotopically labeled experiments.20,23,24,29,32,54 Even
the large negative entropy of activation for the unimolecular
dt
+ k−4[Fru−]
(9)
d[Fru−]
dt
= − (k−4 + k10/KbFru)[Fru−] + k4[Fru−open
]
(10)
Isotope Effects and Competitive Reactions. Although
earlier isotope labeling experiments report a variety of results
that are somewhat ambiguous when compared,20,23,24,29,32,54
the results in this investigation offer a great deal of insight into
the mechanisms presented in Schemes 3 and 4. The effect of
replacing H with D at the C-2 position in glucose is substantial
(kH/kD = 3.8). The following conclusions can be drawn: (i)
deprotonation of acyclic glucose occurs prior to the rate
limiting step, evidenced by a kie of 3.8 observed in the kinetics.
(ii) The significant decrease in glucose conversion and fructose
yield upon retarding enediol formation (Scheme 4, k2) indicates
that a competition exists between deprotonation (k2) and
decomposition of the acyclic form (Scheme 4, k5, k6). (iii) This
competition strongly favors formation of the enediol
intermediate when isomerizing proteo-glucose (glu-C2-H),
which is supported by the relative product distributions after
short reaction times. Fructose selectivity of 78% was obtained
from glu-C2-H (pHo 10.9, 100 °C, 3 min) compared to only
45% from glu-C2-D under the same conditions. Additionally,
fructose selectivities up to 91% were observed after 2 min at
100 °C, pHo = 11.5. The final mechanistic detail that the
labeling experiments reveal is (iv) one of the pathways by
which acidic byproducts are formed is decomposition from
acyclic glucose. The changes in pH after three apparent half-
lives with glu-C2-H and glu-C2-D were 1.5 units and 2.0 units,
respectively. The more rapid decrease in pH not only resulted
in lower fructose yields but also lower glucose conversion (47%
glu-H vs 36% glu-D).
The retardation effect on reaction 2 in Scheme 4 resulting
from deuteration of glucose most easily explains the changes in
product distribution when considering the fate of acyclic
glucose. If the formation of the enediol intermediate is the only
pathway that leads to fructose, then the probability of its
formation can be expressed as k2/(k−1 + k2 + k5 + k6 [OH−]).
Once glucose has undergone ring opening to the acyclic form,
this intermediate is consumed by one of four reactions: (i) ring
closing (k−1), (ii) deprotonation to the enediol (k2), (iii)
thermal decomposition (k5), and (iv) nucleophilic attack by
OH− (k6 [OH−]). All of these reactions are supported by the
isotopically labeled experiment, pH effects, or literature studies
on decomposition25,33,38,39,52,53 and ring opening/clos-
ing.20,22,42−46 Once the enediol intermediate is generated, a
similar competition takes place: (i) rearrangement to acyclic
fructose (k3), (ii) reformation of acyclic glucose (k−2), or (iii)
decomposition (likely through β-scission, k7).20 The probability
of successful formation of acyclic fructose is therefore k3/(k−2
+
k3 + k7). Competition for the resulting acyclic fructose now
begins and essentially mirrors that of glucose. The probability
of successful ring closing to fructose anion and eventually
neutral fructose is k−3/(k−3 + k4 + k8 + k9 [OH−]); however,
unlike glucose (which is relatively stable in its neutral ring
form), fructose can undergo a bimolecular reaction with OH−
which may lead to either decomposition or bimolecular ring
3171
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
process calculated in this work could either suggest bimolecular
deprotonation from acyclic sugars or significant solvent
rearrangement as a result of ring cleavage. Perhaps the
inconsistencies observed in the earlier isotope labeling
experiments20,23,24,29,32,54 point to parallel pathways by which
the fraction of deprotonation/reprotonation by either the
intramolecular reaction or by OH− is dependent on temper-
ature (thus varying the ratio of the elementary rate constants
for each reaction {kintra/kOH}) and [OH−] thereby varying the
When one considers (i) the lack of dependence that fructose
yields have on pHo in the range 10.7 to 11.3, (ii) no conversion
expected below pHo 9,28 and (iii) the observed decrease in pH
with time, it becomes evident that although the equilibrium
constants are independent of pH, the degree of reactivity in the
system (e.g., glucose conversion) is pH-dependent. As a result,
under more basic conditions larger quantities of glucose can be
converted to products before the formation of acidic by-
products overwhelms the combined buffering capacity of the
glucose, fructose, and TEA in solution. Simultaneously, a high
pH environment enhances degradation pathways of ring
fructose and acyclic forms of both sugars, thus stifling yields.
An ideal system is one in which glucose is introduced to a high
pH (∼11.3) environment at high temperature for a short
period of time so that each molecule participates in the
isomerization cycle only one time. As a result, losses due to
degradation of acyclic forms of the sugars will be minimized. In
addition, fructose should be removed from the reaction system
immediately upon formation to avoid the observed bimolecular
reaction.
ratio of the apparent rate constants for each step (kintra
/
{kOH[OH−]}).
Activation Parameters. Because the rate constants
measured for unimolecular fructose reaction and the glucose
reaction are a composite of forward and reverse rate constants
for the same isomerization, they are expected to be similar.
Indeed, the plot of ln(k/T) versus 1/T shown in Figure 3
reveals statistically identical activation parameters.
Large negative values of ΔS‡ are typically observed as a result
of bimolecular reactions (kbimol ΔS‡ = −274 J/molK). However,
other results clearly show that the isomerization is a
unimolecular reaction with respect to the monosaccharide.
Therefore, the large negative ΔS‡ (ca. −140 J/molK) observed
for the unimolecular pathways here may be due to a portion of
the intermediates reacting with OH− and/or substantial
rearrangement of solvent molecules to accommodate the
intermediates. Calculated activation energies for ring opening
of monosaccharides vary substantially upon the inclusion of a
single water molecule.46,63
ASSOCIATED CONTENT
* Supporting Information
The following file is available free of charge on the ACS
■
S
Publications website at DOI: 10.1021/acscatal.5b00316.
Detailed UPLC methods, amine screening results with
and without the exclusion of CO2, catalyst performance
for fructose to glucose isomerization as a function of pH,
and NMR spectra (PDF)
Analogies and Differences with Lewis Acids. The
catalytic performance of this isomerization system is similar to
Sn- β in several ways: turnover rates (0.018 0.01 molFru s−1
AUTHOR INFORMATION
Corresponding Author
*E-mail: tesso@iastate.edu. Phone +1 515-294-4595.
at 100 °C) versus Sn-β (0.027 molFru s−1 mol at 100
−1
−1
■
mol
TEA
Sn
°C),12 mass balance (both ∼84% at prolonged reaction times),
fructose yields (32% TEA vs 31% Sn-β), and selectivity (64%
TEA vs 51% Sn-β). Both systems produce sugar derived
carboxylic acids which play a role in catalyst deactivation, but
unlike Sn-β,16 the deactivation of the TEA system is a result of
reversible protonation of the base. Additionally, the activation
energy for Sn-β (93 kJ/mol) at 100 °C is substantially higher
than the base-catalyzed route measured in this work (61 kJ/
mol).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This material is based upon work supported in part by Iowa
State University and the National Science Foundation Grant
Numbers EEC-0813570 and EPSC-1101284. The authors
would like to thank Mr. Chi (Alex) Liu and Mrs. Sarah
Curry for their assistance in experimental preparation and
insightful discussions, as well as Dr. Sarah Cady (ISU Chemical
Instrumentation Facility) for training and assistance pertaining
to the AVIII-600 results included in this publication.
CONCLUSIONS
■
In light of the disadvantages of enzyme-catalyzed glucose
isomerization (cost, sensitivity to feed purity, irreversible
deactivation, etc.), many researchers have put considerable
efforts in developing and understanding novel isomerization
catalysts. Early studies of Brønsted base-catalyzed glucose
isomerization reactions typically resulted in dismally low yields
of fructose. However, the majority of these studies were carried
out under strongly alkaline conditions. The reaction mechanism
under these conditions has been a point of great interest and
debate since its initial discovery over a century ago. Although a
great deal of progress has been made in understanding the
chemistry of base-catalyzed glucose isomerization, high fructose
yields have only recently been reported.27,40 This work
investigated the kinetics and product distribution under high
fructose yield conditions with consideration to the substantial
amount of the work already done in this area in order to better
understand the base-catalyzed mechanism and lay the
foundation for the synthesis and implementation of novel
base catalysts.
REFERENCES
■
547.
(1) Climent, M. J.; Corma, A.; Iborra, S. Green Chem. 2014, 16, 516−
(2) Wang, T.; Nolte, M. W.; Shanks, B. H. Green Chem. 2014, 16,
548−572.
(3) van Putten, R.-J.; van der Waal, J. C.; de Jong, E.; Rasrendra, C.
B.; Heeres, H. J.; de Vries, J. G. Chem. Rev. 2013, 113, 1499−1597.
(4) Bhosale, S. H.; Rao, M. B.; Deshpande, V. P. Microbiol. Rev. 1996,
60, 280−300.
(5) Misset, O. In Handbook of Food Enzymology; Whitaker, J. R.,
Voragen, A. G. J., Wong, D. W. S., Eds.; Marcel Dekker: New York,
2003; pp 1057−1077.
(6) Buchholz, K.; Seibel, J. Carbohydr. Res. 2008, 343, 1966−1979.
(7) Tewari, Y. B.; Goldberg, R. N. Appl. Biochem. Biotechnol. 1985,
11, 17−24.
(8) Aider, M.; Halleux, D. d. Trends Food Sci. Technol. 2007, 18, 356−
364.
3172
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173

ACS Catalysis
Research Article
(9) Zhao, S.; Guo, X.; Bai, P.; Lv, L. Asian J. Chem. 2014, 26, 4537−
(43) Fab
THEOCHEM 1994, 307, 171−178.
́
ian
́
, P.; Asbot
́
h, B.; Nar
́
ay-Szabo,
́
G. J. Mol. Struct.:
4543.
(44) Grimshaw, C. E.; Whistler, R. L.; Cleland, W. W. J. Am. Chem.
Soc. 1979, 101, 1521−1532.
́
(10) Moliner, M.; Roman-Leshkov, Y.; Davis, M. E. Proc. Natl. Acad.
Sci. U. S. A. 2010, 107, 6164−6168.
(45) Los, J. M.; Simpson, L. B.; Wiesner, K. J. Am. Chem. Soc. 1956,
78, 1564−1568.
(11) Bermejo-Deval, R.; Assary, R. S.; Nikolla, E.; Moliner, M.;
́
Roman-Leshkov, Y.; Hwang, S.-J.; Palsdottir, A.; Silverman, D.; Lobo,
(46) Assary, R. S.; Curtiss, L. A. Energy Fuels 2011, 26, 1344−1352.
(47) de Bruijn, J. M.; Kieboom, A. P. G.; van Bekkum, H. J.
Carbohydr. Chem. 1986, 5, 561−9.
R. F.; Curtiss, L. A.; Davis, M. E. Proc. Natl. Acad. Sci. U. S. A. 2012,
109, 9727−9732.
(12) Bermejo-Deval, R.; Gounder, R.; Davis, M. E. ACS Catal. 2012,
2, 2705−2713.
(48) de Bruijn, J. M.; Kieboom, A. P. G.; van Bekkum, H. Recl. Trav.
Chim. Pays-Bas 1987, 106, 35−43.
(13) Bermejo-Deval, R.; Orazov, M.; Gounder, R.; Hwang, S.-J.;
Davis, M. E. ACS Catal. 2014, 4, 2288−2297.
(49) de Bruijn, J. M.; Kieboom, A. P. G.; van Bekkum, H. Starch/
Staerke 1987, 39, 23−8.
(14) Caratzoulas, S.; Davis, M. E.; Gorte, R. J.; Gounder, R.; Lobo, R.
F.; Nikolakis, V.; Sandler, S. I.; Snyder, M. A.; Tsapatsis, M.; Vlachos,
D. G. J. Phys. Chem. C 2014, 118, 22815−22833.
(15) Li, Y.-P.; Head-Gordon, M.; Bell, A. T. ACS Catal. 2014, 4,
1537−1545.
(50) de Bruijn, J. M.; Kieboom, A. P. G.; van Bekkum, H.; van der
Poel, P. W.; de Visser, N. H. M.; de Schutter, M. A. M. Int. Sugar J.
1987, 89, 190−195.
(51) de Bruijn, J. M.; Touwslager, F.; Kieboom, A. P. G.; van
Bekkum, H. Starch/Staerke 1987, 39, 49−52.
(16) Rajabbeigi, N.; Torres, A. I.; Lew, C. M.; Elyassi, B.; Ren, L.;
Wang, Z.; Je Cho, H.; Fan, W.; Daoutidis, P.; Tsapatsis, M. Chem. Eng.
Sci. 2014, 116, 235−242.
(52) Eggleston, G.; Vercellotti, J. R. J. Carbohydr. Chem. 2000, 19,
1305−1318.
(53) Shaw, P. E.; Tatum, J. H.; Berry, R. E. J. Agric. Food Chem. 1968,
16, 979−982.
(17) Choudhary, V.; Pinar, A. B.; Lobo, R. F.; Vlachos, D. G.;
Sandler, S. I. ChemSusChem 2013, 6, 2369−2376.
(54) Fodor, G.; Sachetto, J.-P. Tetrahedron Lett. 1968, 9, 401−403.
(55) Bothner-By, A. A.; Gibbs, M. J. Am. Chem. Soc. 1950, 72, 4805−
4805.
́
(18) Roman-Leshkov, Y.; Moliner, M.; Labinger, J. A.; Davis, M. E.
Angew. Chem., Int. Ed. 2010, 49, 8954−8957.
(19) Rai, N.; Caratzoulas, S.; Vlachos, D. G. ACS Catal. 2013, 3,
2294−2298.
(56) Corzo-Martínez, M.; Corzo, N.; Villamiel, M.; del Castillo, M.
D. In Food Biochemistry and Food Processing, 2nd ed.; Simpson, B. K.,
Nollet, L. M. L., Toldra, F., Benjakul, S., Paliyath, G., Hui, Y. H., Eds.;
John Wiley & Sons: Hoboken, NJ, 2012; pp 56−83.
(57) Kim, J.-S.; Lee, Y.-S. Food Chem. 2008, 108, 582−592.
(58) Wirth, D. D.; Baertschi, S. W.; Johnson, R. A.; Maple, S. R.;
Miller, M. S.; Hallenbeck, D. K.; Gregg, S. M. J. Pharm. Sci. 1998, 87,
31−39.
(59) Sixty-ninth meeting of the Joint FAO/WHO Expert Committee
on Food Additives (JECFA). Safety Evaluation of Certain Food
Additives; World Health Organization: Geneva, 2009; p 155.
(60) U.S. Department of Health and Human Services. National
Toxicology Program - Triethylamine. http://ntp.niehs.nih.gov/go/
12383 (accessed July 28, 2014).
(61) Shallenberger, R. S.; Mattick, L. R. Food Chem. 1983, 12, 159−
165.
(62) Kooyman, C.; Vellenga, K.; De Wilt, H. G. J. Carbohydr. Res.
1977, 54, 33−44.
(63) Lewis, B. E.; Choytun, N.; Schramm, V. L.; Bennet, A. J. J. Am.
Chem. Soc. 2006, 128, 5049−5058.
(20) de Wit, G.; Kieboom, A. P. G.; van Bekkum, H. Carbohydr. Res.
1979, 74, 157−175.
(21) Despax, S.; Estrine, B.; Hoffmann, N.; Le Bras, J.; Marinkovic,
S.; Muzart, J. Catal. Commun. 2013, 39, 35−38.
(22) Goux, W. J. J. Am. Chem. Soc. 1985, 107, 4320−4327.
(23) Isbell, H. S.; Frush, H. L.; Wade, C. W. R.; Hunter, C. E.
Carbohydr. Res. 1969, 9, 163−175.
(24) Isbell, H. S.; Linek, K.; Hepner, K. E., Jr Carbohydr. Res. 1971,
19, 319−327.
(25) Knill, C. J.; Kennedy, J. F. Carbohydr. Polym. 2003, 51, 281−
300.
(26) Lima, S.; Dias, A. S.; Lin, Z.; Brandao, P.; Ferreira, P.; Pillinger,
̃
M.; Rocha, J.; Calvino-Casilda, V.; Valente, A. A. Appl. Catal., A 2008,
339, 21−27.
(27) Liu, C.; Carraher, J. M.; Swedberg, J. L.; Herndon, C. R.;
Fleitman, C. N.; Tessonnier, J.-P. ACS Catal. 2014, 4, 4295−4298.
(28) Simonov, A. N.; Pestunova, O. P.; Matvienko, L. G.; Parmon, V.
N. Russ. Chem. Bull. 2005, 54, 1967−1972.
(29) Sowden, J. C.; Schaffer, R. J. Am. Chem. Soc. 1952, 74, 505−507.
(30) Speck Jr, J. C. In Advances in Carbohydrate Chemistry; Melville,
L. W., Ed.; Academic Press: New York, 1958; Vol. 13, pp 63−103.
(31) Angyal, S. J. In Glycoscience; Stutz, A., Ed.; Springer: Berlin,
̈
2001; Vol. 215, pp 1−14.
(32) Topper, Y. J.; Stetten, D. J. Biol. Chem. 1951, 189, 191−202.
(33) de Bruijn, J. M.; Kieboom, A. P. G.; van Bekkum, H. Recl. Trav.
Chim. Pays-Bas 1986, 105, 176−183.
(34) Lecomte, J.; Finiels, A.; Moreau, C. Starch/Staerke 2002, 54,
75−79.
(35) Moreau, C.; Durand, R.; Roux, A.; Tichit, D. Appl. Catal., A
2000, 193, 257−264.
(36) Roman
1580.
́
-Leshkov, Y.; Davis, M. E. ACS Catal. 2011, 1, 1566−
(37) Souza, R. O. L.; Fabiano, D. P.; Feche, C.; Rataboul, F.;
Cardoso, D.; Essayem, N. Catal. Today 2012, 195, 114−119.
(38) Yang, B. Y.; Montgomery, R. Carbohydr. Res. 1996, 280, 27−45.
́
̌
(39) Novotny, O.; Cejpek, K.; Velísek, J. Czech J. Food Sci. 2008, 26,
117−131.
(40) Jia, S.; Liu, M.; Gong, Y.; Feng, J.; Guo, X. Acta Pet. Sin., Pet.
Process. Sect. 2012, 28, 940−949.
(41) de Bruyn, C. A. L.; van Ehenstein, W. A. Ber. Dtsch. Chem. Ges.
1895, 28, 3078−3082.
(42) Serianni, A. S.; Pierce, J.; Huang, S. G.; Barker, R. J. Am. Chem.
Soc. 1982, 104, 4037−4044.
3173
DOI: 10.1021/acscatal.5b00316
ACS Catal. 2015, 5, 3162−3173