Kinetics of monosaccharide conversion in the presence of homogeneous Bronsted acids

Article abstract of DOI:10.1016/j.apcata.2012.10.013

For the utilization of biomass in the fuel and polymer industry, a wide variety of catalysts have been studied for their activity either on the dehydration reaction in particular or on the monosaccharide degradations in general. Yet, systematic data outlining the effects of acidic features is not available and a common framework for catalytic activity comparison is missing that is needed for rational catalyst design. The current work aimed to provide insight about the effect of the nature of the acid and initial acidity on degradation kinetics of C₅ and C₆ carbohydrates and thereby built a platform allowing activity comparison. Mineral and organic acids ranging in acidic strength, hydrochloric, sulfuric, phosphoric, maleic, and propane sulfonic acid, were tested for their activity in the degradation of xylose, fructose and glucose at two different pH values; 1.5 and 3.6. In the presence of weak homogeneous acids, glucose undergoes degradation with drastically different activation energies at pH of 1.5 than at pH 3.6. Such a difference does not occur in the presence of strong homogenous acids. On the other hand, xylose and fructose undergo degradation with similar activation energies of approximately 140 kJ/mol regardless of the pH or nature of the homogeneous acid. A common framework that compiles the catalytic activities and outlines the differences potentially related to underlying mechanism changes provides the basis required for rational heterogeneous catalyst design.

Full text of DOI:10.1016/j.apcata.2012.10.013

GModel
APCATA-13924; No. of Pages6
ARTICLE IN PRESS
Applied Catalysis A: General xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Kinetics of monosaccharide conversion in the presence of
homogeneous Bronsted acids
Basak Cinlar, Tianfu Wang, Brent H. Shanks∗
Chemical and Biological Engineering Department, Iowa State University, Ames, IA 50011, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 July 2012
Received in revised form 8 October 2012
Accepted 21 October 2012
Available online xxx
For the utilization of biomass in the fuel and polymer industry, a wide variety of catalysts have been
studied for their activity either on the dehydration reaction in particular or on the monosaccharide
degradations in general. Yet, systematic data outlining the effects of acidic features is not available and
a common framework for catalytic activity comparison is missing that is needed for rational catalyst
design. The current work aimed to provide insight about the effect of the nature of the acid and initial
acidity on degradation kinetics of C5 and C6 carbohydrates and thereby built a platform allowing activity
comparison. Mineral and organic acids ranging in acidic strength, hydrochloric, sulfuric, phosphoric,
maleic, and propane sulfonic acid, were tested for their activity in the degradation of xylose, fructose
and glucose at two different pH values; 1.5 and 3.6. In the presence of weak homogeneous acids, glucose
undergoes degradation with drastically different activation energies at pH of 1.5 than at pH 3.6. Such a
difference does not occur in the presence of strong homogenous acids. On the other hand, xylose and
fructose undergo degradation with similar activation energies of approximately 140 kJ/mol regardless of
the pH or nature of the homogeneous acid. A common framework that compiles the catalytic activities
and outlines the differences potentially related to underlying mechanism changes provides the basis
required for rational heterogeneous catalyst design.
Keywords:
Glucose
Carbohydrates
Dehydration
Acid catalysis
Sugar degradation
© 2012 Published by Elsevier B.V.
1. Introduction
pH levels, the possible effect of different acidic group features can
be investigated for the hydrothermal degradation of glucose and
Dehydration of monosaccharides to furanic compounds has
regained considerable interest in the past decade either as a
reaction that could provide valuable products for the renewable
polymer and fuel industry or as one of the degradation reactions
accompanying polysaccharide hydrolysis [1]. Studies considering
dehydration as a side reaction during saccharification focus on how
to minimize its occurrence. In contrast, dehydration reaction stud-
ies have the opposite focus of how to achieve higher furanic yields
at high monosaccharide conversions. In relation to both objectives
a wide variety of catalysts ranging from traditional strong mineral
acids to novel ionic liquids immobilized on silica supports have
been examined [2,3]. However, few studies report systematic data
that compare catalytic performance using a common framework,
in which parameters like the acid strength, catalyst concentra-
tion, and reaction temperature are systematically varied in order
to explore more general information about the intrinsic kinetics
and possible mechanisms for the reaction system. In this regard,
using both mineral and organic acids of varying strength at different
xylose, which are the most common products of biomass hydroly-
sis. Additional knowledge can be gained by comparing their kinetics
to the well-understood fructose degradation kinetics. These results
can be used to build a common framework for catalytic activity
comparison for different acidic moieties.
Saccharification, the hydrolysis of polysaccharides into
monosaccharides, is
a key first step for the valorization of
agricultural lignocellulosic waste products [4,5]. While cellulose
is composed of only glucose, the hydrolysis of hemicellulosic
materials also produces xylose and arabinose as well as C6 sugars,
e.g., glucose, galactose and mannose. Utilization of traditional
mineral acids, such as hydrochloric and sulfuric acids, readily
causes degradation of the monosaccharides under hydrolysis
conditions, among the degradation reactions is the dehydration
of C5 and C6 sugars into furfural and 5-hydroxymethylfurfural
(HMF). Additional degradation reactions occurring in conjunction
with the dehydration reaction are condensation, fragmentation
and polymerization reactions that result in, but are not limited
to glyceraldehyde, glycolaldehyde, pyruvaldehyde, formalde-
hyde, levoglucosan, levulinic acid, acetic acid and formic acid [6].
Side reactions accompanying cellulose hydrolysis and glucose
dehydration are summarized in Scheme 1.
∗
Corresponding author. Tel.: +1 515 294 1895; fax: +1 515 294 2689.
E-mail address: bshanks@iastate.edu (B.H. Shanks).
0926-860X/$ – see front matter © 2012 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.apcata.2012.10.013
Please cite this article in press as: B. Cinlar, et al., Appl. Catal. A: Gen. (2012), http://dx.doi.org/10.1016/j.apcata.2012.10.013

GModel
APCATA-13924; No. of Pages6
ARTICLE IN PRESS
2
B. Cinlar et al. / Applied Catalysis A: General xxx (2012) xxx–xxx
has not been proposed with any other homogeneous acid. In fact,
in many studies [8,12,14,15] both hydrolysis and the accompanying
monosaccharide degradation reaction were reported to be directly
proportional to the number of available protons regardless of the
type of acid used. Glucose conversion with the dicarboxylic organic
acids resulted in rates that were independent of the pH, but in the
presence of sulfuric acid there was a direct relationship between
the measured reaction rate and pH. Similarly, Bobleter et al. [14]
reported pH-dependent reaction regimes for glucose degradation
when sulfuric acid was used, but in the presence of acetic acid
such dependency was not observed. Difference in the rates due
to the choice of catalyst was also reflected in the activation ener-
gies and this difference in activation energies has been attributed
to different mechanisms, although details about the mechanism
change were not provided [5]. Similar studies with xylose also
resulted in different activation energies for sulfuric acid and
maleic acid [17].
Several mechanisms have been proposed for glucose dehy-
dration and degradation. According to one hypothesis, glucose
dehydration occurs from the acyclic form via a 1,2-enediol inter-
mediate, which is also the intermediate in the fructose dehydration
[18]. The slower dehydration rates of glucose relative to fructose
have been explained by the lower concentration of the open-chain
form for glucose due to its higher ring stability. Another hypoth-
esis suggests that glucose isomerizes to fructose first via hydride
shift and subsequently follows the same degradation mechanism
as with fructose. Donald et al. [19] suggested hydrogen transfer
from C2 to C1 during the conversion of aldoses to furfural deriva-
tives which can account for 28% of the reaction proceeding via the
ketose. The active forms for dehydration were proposed to be the
furanose structure for C6 sugars and pyranose structure for the C5.
The higher ratio of the xylopyranose form was used to explain the
slower reaction for glucose relative to xylose, but proposed mech-
anism does not predict any change in the relative rates according
to type of the acid catalyst used.
The discrepancy in the relative rates can be extended further.
While comparable HMF yields were obtained from glucose and
fructose at 175–390 ◦C in the presence of 10−2 M mineral acids with
or without an organic extracting phase, a high yield of HMF was
only obtained from fructose at 85–90 ◦C in the presence of >0.25 M
strong mineral acids [3]. Under these conditions, the HMF yields
from glucose or other aldohexoses were found to be very small.
Whether it is the high temperatures and (or) low proton concen-
tration that caused the change in the relative rate trends remained
unresolved.
In this study, the effect of the nature of the acid and initial pH
value on degradation kinetics were explored by examining organic
and mineral acids of different strength for their activity on glu-
cose, fructose and xylose at two different proton concentrations,
pH 1.5 and 3.6. Sulfuric acid and hydrochloric acid were chosen as
the strong mineral acids, phosphoric acid as the weak mineral acid,
and maleic acid and 1-propylsulfonic acid as the weak organic acids.
By studying the effects of these acids under the same conditions the
effect of proton concentration was deconvoluted from the effect of
acidic strength. Furthermore, by calculating the activation energies
for a temperature range of 145–175 ◦C, the observed activities could
be related to the proposed mechanisms.
Scheme 1. Reaction scheme for cellulose hydrolysis and glucose degradation reac-
tions.
Including dehydration, the monosaccharide conversion reac-
tions can all be acid catalyzed and have very similar activation
energies [7]. This fact further creates a challenge for selective
monosaccharide dehydration and has prevented economically
viable HMF production from glucose. In the pursuit for sustain-
able chemicals, HMF and furfural production from biorenewable
resources has regained interest, because they are classified as
potentially important building blocks for the polymer industry
[1,8]. A wide range of novel catalysts have been recently reported
that is attempting to achieve high yields of furanic compounds
[9–11]. Also, a considerable amount of past literature investigated
monosaccharide conversion, in particular dehydration, but few
links have been established between the more recent studies with
heterogeneous catalysts to these prior studies with homogeneous
acids [3,6,8,12]. The use of different reaction conditions and lack
of evaluation of intrinsic kinetic values such as activation ener-
gies and turnover numbers accounts for the absence of such links.
Therefore, a more systematic evaluation of the reaction system
will provide valuable insight about how acidic features affect the
reaction results.
Parameters thought to play a key role in the dehydration reac-
tion performance are temperature, nature of the solvent, and the
type and concentration of the acid [4,13–16]. The effect of different
acids on dehydration has been demonstrated by Dumesic and co-
workers where they examined HMF formation from fructose in the
presence of different mineral acids [16]. In experiments at pH 1.5 in
a DMSO–water mixture with an extracting phase, phosphoric acid
gave the highest selectivity at 170 ◦C followed by hydrochloric acid.
Sulfuric acid yielded the lowest selectivity under their conditions.
Accordingly, it was concluded that selectivity was affected by the
choice of the acid. As these selectivity values were reported at dif-
ferent conversions there might have been a convolution between
the conversion and the effect of the acid itself.
Another study comparing the effect of acid catalyst choice on
monosaccharide degradation examined organic acids, including
dicarboxylic acids, and compared their performance to that of sul-
furic acid [5]. In that study, which explored cellobiose hydrolysis
and degradation of the resulting glucose, maleic acid was active for
hydrolysis and acetic acid was not under the reaction conditions
employed. The maleic acid activity was postulated to be related
to its being a diacid. Interestingly, maleic acid did not show any
appreciable glucose degradation activity under the hydrolysis con-
ditions employed, so it was a catalyst with higher selectivity for
glycosidic bond hydrolysis relative to sulfuric acid. The ability to
selectively hydrolyze cellulosic compounds without degradation
2. Experimental
Hydrochloric acid (12 N, Fisher Chemicals), sulfuric acid (18 N,
Fisher Chemicals), phosphoric acid (85%, Fisher Chemicals), maleic
acid (99%, Acros), and 1-propanesulfonic acid (99%, Acros) were
used as purchased. The reactants, d-fructose (Fisher Chemicals),
d-xylose (Acros) and ␣-d-glucose (99%, Acros), were also used as
Please cite this article in press as: B. Cinlar, et al., Appl. Catal. A: Gen. (2012), http://dx.doi.org/10.1016/j.apcata.2012.10.013

GModel
APCATA-13924; No. of Pages6
ARTICLE IN PRESS
B. Cinlar et al. / Applied Catalysis A: General xxx (2012) xxx–xxx
3
100
90
80
70
60
50
40
30
20
10
0
purchased without further purification. For the kinetic experiments
aqueous solutions of hydrochloric acid, sulfuric acid, phosphoric
acid, maleic acid, or 1-propanesulfonic acid at two different pH
values, 1.5 and 3.6 were prepared. The pH values of the solu-
tions were measured at ambient temperature using a combined
glass electrode (6.0233.100, Metrohm) attached to a Metrohm 798
MPT Titrino automatic titrator. Before each measurement the elec-
trode was calibrated using three standard buffers at 4.00, 7.00
and 10.00.
c
a
b
The kinetic experiments were performed at 145, 160 and 175 ◦C
in a 250 ml stirred batch reactor (Parr Instruments) equipped with
a glass liner. Temperature control was maintained with a PID
controller attached to a heating jacket and cooling coil. The cor-
responding sugar was added to the acidic solution in the amount
needed for a final concentration of 0.11 M before the solution was
charged to the reaction vessel. Time zero for the reaction was taken
as the time when the desired reaction temperature was reached. As
some sugar conversion occurred during the heating procedure, the
reactor was sampled at this zero time and the resulting conver-
sion was determined relative to this “starting” concentration. The
reactor was pressurized with 300 psig nitrogen to ensure a con-
densed phase. A stirring rate was determined and selected to be
sufficiently high to overcome mass transfer limitation. Samples col-
lected during the kinetic runs were filtered through a 0.2 ␮m nylon
filter (Cobert Assoc.) and pH was back-adjusted to 6–7 by addition
of 4.0 M NaOH prior to HPLC analysis.
0
30
60
90
0
30
60
90
0
30 60 90 120
time (min)
Fig. 1. Sugar (glucose (a), xylose (b), fructose (c)) conversions at 175 ◦C, pH 3.6 in the
presence of HCl (ꢀ), H2SO4 (ꢁ), H3PO4 (×), maleic acid (*) and 1-propane sulfonic
acid (ꢂ).
monosaccharide degradation or dehydration reactions were con-
ducted in the presence of strong mineral acids with the highest
molarity being 50 mM, which for sulfuric acid corresponds to a
pH value of approximately 1.5. In contrast, more recent studies
with heterogeneous catalysts were performed at higher pH values.
Unfortunately, the exact proton concentrations in the presence of
heterogeneous catalysts such as ion-exchangers and zeolites were
typically not reported. The pH values were reported for organic acid
functionalized mesoporous silica when used for cellobiose hydrol-
ysis and they range from a pH of 3.0 to 4.5 [20]. Therefore, studying
the proton activity at pH values of 1.5 and 3.6 bracketed a significant
portion of the literature and enabled a bridge between the results
using either homogeneous acids or heterogeneous acids. Values of
the pH of less than 1.5 were not explored due to the difficulty of
achieving those values with the weak acids.
The samples were analyzed with a Hi-Plex H+ column (Poly-
mer Lab.) at 65 ◦C on a Waters HPLC system equipped with a
Waters 2414 Refractive Index detector (RID) and Waters 996 Pho-
toiodide Array Detector (PAD). The mobile phase was a 10 mM
sulfuric acid aqueous solution flowing at 0.6 ml/min. The RID peak
areas and intensities were used to estimate the sugar concentra-
tions to avoid peak convolutions in the presence of weak acids,
dehydration product concentrations were determined using both
the PAD peak area at 280 nm (conc. <0.2 wt%) and the RID peak
areas (conc. >0.2 wt%) to maintain a linear relationship of the peak
area or intensity with the concentration. The glucose concentra-
tions in the samples were also confirmed with a Biorad Glucose
Analyzer.
A similar objective was followed when choosing the tempera-
ture range to investigate. Although some mechanistic studies were
carried out at much higher temperatures such as 390 ◦C [21], the
bulk of the studies were performed within the range of 145–175 ◦C.
Particularly in the case of glucose, reaction rates at 145 ◦C are rather
difficult to determine as the error bars become large relative to the
change in reaction rates if not very carefully analyzed. Also at the
lower rates of reaction, the kinetics can be modified by the for-
mation of acids during the reaction. These acids will self-catalyze
the reaction and cause higher rates. At higher temperatures, and
hence higher decomposition rates, the formation of these acids
was reported to be insignificant [12]. Overall, comparing the activa-
tion energies is a more reliable way for understanding the effect of
acidity rather than comparing the relative rates at a single reaction
temperature.
Shown in Figs. 1 and 2 are the rates of monosaccharide conver-
sion at 175 ◦C and 160 ◦C, respectively, when using an initial pH
of 3.6. As discussed in the Section 2, the conversion values were
adjusted to zero corresponding to time zero, which was the time
that the desired reaction temperature was reached. This adjust-
ment allowed a more direct comparison of the conversion rates
amongst the different monosaccharides. In reality, the reactants
were charged to the reactor initially and some conversion occurred
during the heat-up period. This initial loss was determined to be
less than 10% and was insignificant compared to the total conver-
sion during the time of the reaction run at the designated reaction
temperature. For all of the runs, significant amounts of soluble and
3. Results and discussion
When using Bronsted acids, the rates of dehydration and degra-
dation for monosaccharides are generally proportional to the
number of available protons (or hydronium ions). Therefore, the
performance of different acids can best be achieved by measuring
their catalytic activities at equal proton concentrations. As such
the current study compared the activities of the various acids,
hydrochloric, sulfuric, phosphoric, maleic and 1-propanesulfuric
acid, at the same pH values rather than at the same molar acid
amounts. The pH or available proton concentration at elevated tem-
perature for the studied acids were calculated and the results are
summarized in the supplemental material. Using a pH of 3.6 at room
temperature as an example, the predicted pH values at 160 ◦C are
3.66 and 3.87 for sulfuric acid and phosphoric acid, respectively.
And as such its potential effect on the rates was found to be within
the experimental error margin for the reaction results. Therefore
the change in the pH of the solution due to the increase of the pKa or
activity coefficient was neglected and the results were interpreted
as if the final proton concentrations in the experimental runs at
the different temperatures were essentially equal to their ambient
temperature values.
The two different proton concentrations, pH 1.5 and 3.6,
were chosen to cover most of the pH range used in the litera-
ture. Commonly, kinetic and mechanistic studies regarding the
Please cite this article in press as: B. Cinlar, et al., Appl. Catal. A: Gen. (2012), http://dx.doi.org/10.1016/j.apcata.2012.10.013

GModel
APCATA-13924; No. of Pages6
ARTICLE IN PRESS
4
B. Cinlar et al. / Applied Catalysis A: General xxx (2012) xxx–xxx
Table 1
Apparent activation energies for glucose, fructose and xylose.
20
Glucose
Ea (kJ/mol)
Literature values
Ea (kJ/mol)
pH 1.5 pH 3.6
Hydrochloric acid
Sulfuric acid
Phosphoric acid
Maleic acid
138
142
145
138
137
138
82
78
83
1.5 M sulfuric acid [4]
50 mM sulfuric acid [8]
50 mM maleic acid [5]
20% SBA-15-BuCOOH [20]
15% SBA–SO3H [20]
137
118
73
79
75
Propanesulfonic acid 148
10
Fructose
Hydrochloric acid
Sulfuric acid
Phosphoric acid
Maleic acid
132
139
146
138
136
128
144
142
50 mM sulfuric acid [4]
1.5 M sulfuric acid [8]
138
136
Xylose
Hydrochloric acid
Sulfuric acid
Phosphoric acid
Maleic acid
133
135
128
138
136
138
126
150
127
50 mM sulfuric acid [4]
1.5 M sulfuric acid [8]
Hydrothermolysis [3]
50 mM maleic acid [8]
15% SBA–SO3H [20]
134
134
137
204
150
0
Propanesulfonic acid 150
0
30
60
90
120
time (min)
fructose, the selectivity followed the decreasing order of H3PO4,
H2SO4 and HCl, which also follows the trend of the basic strength
of the anion. Therefore, it could be speculated that complexion
of the dissociated anions with an intermediate leading to dehy-
dration products might have occurred. Such a complex formation
Fig. 2. Glucose conversions at 160 ◦C and pH 3.6 in the presence of HCl (ꢀ), H2SO4
(ꢁ), H3PO4 (×), maleic acid (*) and 1-propane sulfonic acid (ꢂ).
insoluble humin formation were observed, but their production
rate was not quantified. The formation of the insoluble humins pre-
cluded performing closed carbon balances on the samples. The only
reaction product that was quantified was the yield of the furanic
compounds.
of HMF or its intermediate with the Cl− or SO4
was proposed
2−
previously when the glucose dehydration was examined in the
presence of magnesium and aluminum salts [22]. In that study,
saturated solutions of the salts were used in the absence of acidic
protons and the different activity of chloride and sulfate ions were
explained by the sulfates and chlorides constituting classes of salts,
which could exert different effects on glucose reactivity. Dehydra-
tion to HMF was proposed to proceed via participation of aquo- and
hydroxyl-complexes leading to an extensively hydrogen-bonded
species intermediate in the presence of sulfate ions, whereas an
exocyclic CH2OH elimination originating from so called “anions
guiding rail” was responsible for the furan ring formation in the
presence of chloride ions [23]. The conditions used in this study,
e.g., much lower concentration of anions and the presence of
acidic protons, did not lead to a discernable glucose conversion
difference between HCl and H2SO4. Solutions of sulfuric acid con-
tain predominantly bisulfate ions rather than the sulfate ions and
these bisulphate ions are less strong complexation agents. There-
fore, a better explanation for the higher selectivity with H2SO4
than HCl is the complexation of Cl− anions with the HMF prod-
uct itself causing its further conversion and thereby decreasing the
yield [23].
As can be seen in Fig. 1 no differences in conversion rates were
observed amongst the homogeneous acids at pH 3.6 and 175 ◦C.
However, when the temperature was 160 ◦C at pH 3.6 the strong
acids gave significantly higher conversion rates than the weak
acids. A similar trend was also observed when the reaction tem-
perature was 145 ◦C and the pH value 3.6 with the conversions in
the presence of maleic acid, phosphoric acid and 1-propanesulfonic
acid being similar to each other, whereas the conversions when
using hydrochloric acid or sulfuric acid were the same and at a
higher level. The results for pH 1.5 are not shown as no significant
difference was observed in the sugar conversion rates for any of
the acids examined. This result at pH 1.5 was in agreement with
the chemistry of the monosaccharide conversion being performed
via hydronium ions such that no effect of the dissociated anions
could be observed.
Interestingly, the different acids when tested at 160 ◦C and pH
1.5 were found to give different selectivity toward dehydration
products for each of the sugars as shown in Fig. 3. For glucose and
In the Bronsted acid catalyzed conversion of monosaccharides,
first order reaction kinetics based on the sugar concentration has
been widely used in modeling studies [5,8]. Shown in Table 1 is the
apparent activation energies calculated for the acids using initial
reaction rate data and the Arrhenius relationship (Eq. (1)) while
assuming a first order reaction rate expression.
ꢀ
ꢁ
−Ea
k = A ∗ exp
(1)
RT
As shown in Table 1, the apparent activation energy values were
found to be quite similar for all of the acids with the exception
of the weaker acids under pH 3.6 conditions. For this case, the
apparent activation energy for glucose was found to be signifi-
cantly lower than the commonly reported values for strong mineral
acids. A similar low apparent activation energy value was reported
in the presence of organic acid functionalized mesoporous sil-
ica previously [20] as well as with maleic acid [5]. In the later
study, it was proposed that dehydration using maleic acid led
to a different dehydration mechanism, which was attributed to
Fig. 3. HMF/furfural selectivity after 30 min at 160 ◦C in the presence of different
acids at pH 1.5 for xylose (ꢃ), glucose ( ), fructose ( ) and at pH 3.6 for glucose
(ꢁ).
Please cite this article in press as: B. Cinlar, et al., Appl. Catal. A: Gen. (2012), http://dx.doi.org/10.1016/j.apcata.2012.10.013

GModel
APCATA-13924; No. of Pages6
ARTICLE IN PRESS
B. Cinlar et al. / Applied Catalysis A: General xxx (2012) xxx–xxx
5
Table 2
represented by acyclic route in Fig. S1 [3]. As the open-chain form
of the sugar was responsible for the formation of the 1,2-enediol
intermediate, the lower activity of glucose was explained as being
due to the lower ratio of open-chain form than with fructose. For
this mechanism, the rate-limiting step was proposed to be the pro-
tonation of the pyranose form and subsequent ring opening [3].
An alternative mechanism proposed that glucose first isomer-
ized to fructose and then followed the same conversion pathway
as fructose [12]. The rate-determining step for this mechanism was
proposed to be the isomerization step, which would lead to the
glucose conversion having a different apparent activation energy
than fructose would. Glucose can be isomerized to fructose via base
or acid/base catalysis with base catalysis being more effective for
the isomerization. In the presence of both an acid and base, a con-
certed push–pull mechanism was suggested, that is first driven by
the attack of the base [18]. In the presence of the base only, the
isomerization appeared to occur via a hydride shift from C2 to C1
and as such the formation of an 1,2-enediol intermediate was not
observed [18].
Pre-exponential rate constants for glucose.
a
Glucose
A (min−1
)
A0 (min−1
)
pH 1.5
pH 3.6
pH 1.5
pH 3.6
Hydrochloric acid
Sulfuric acid
Phosphoric acid
Maleic acid
5.46E14
7.18E14
8.15E14
8.26E14
6.96E14
4.40E12
5.78E12
3.57E04
3.52E05
4.26E05
1.73E16
2.27E16
2.58E16
2.61E16
2.20E16
1.75E16
2.30E16
1.42E08
1.40E09
1.70E09
Propanesulfonic acid
a
Modified Saeman equation constants with the assumption of m equal to 1.
maleic acid being a diacid. While more work is needed to decipher
the exact reaction mechanism, the kinetic results in the current
study may indirectly suggest that there is potentially a different
dehydration mechanism associated with weaker acids when a
pH value of 3.6 is used, but that this result was not correlated
with whether a mono- or diacid was used. The possibility of
different dehydration mechanisms was previously proposed by
Antal et al. [3], but the reaction conditions causing the differ-
ent mechanisms to become predominant were not systematically
evaluated.
The change in the apparent activation energy was also mani-
fested in the preexponential factors, which were determined from
the Arrhenius plot and are shown for glucose in Table 2. Saeman
[24] modified the classical Arrhenius equation to predict the effect
of the proton concentration on cellulose hydrolysis and glucose
degradation, where the acid concentration was expressed sepa-
rately from the pre-exponential factor. This relationship has been
widely accepted and was further generalized by replacing the acid
concentration with the proton concentration to produce a modified
Saeman equation (Eq. (2)) [5].
Simulation studies of the xylose and glucose degradation path-
ways [7,25,26] suggested that different mechanisms can dominate
depending on the reaction conditions, such as the acidity, pres-
ence of co-solvents, or the temperature. However, for glucose the
protonation of the glucose molecule was still found to be the rate
limiting step. The strongest proton affinity was found for the C2 OH
group and only the protonation of this group led to the formation of
HMF either in simulations corresponding to vacuum or with explicit
water molecules. The simulations suggested that the protonation
of the C3 OH group might or might not take place depending on
the acidity. At lower acidities, the proton at C3 usually transferred
back to the water molecule, while at higher acidities this protona-
tion might lead to degradation [25]. For the latter case, ring opening
was observed following protonation. In contrast, protonation at C2
resulted in the formation of 2,5-anhydride ketose intermediates
directly via a hydride shift [25]. Subsequent elimination of water
from the 2,5-anhydride intermediate would then readily occur,
which is similar to fructose degradation, as represented by the
cyclic route shown in Fig. S1.
ꢀ
ꢁ
−Ea
k = A0 ∗ [H+]m ∗ exp
(2)
RT
In this equation the exponential term m was defined to account
for the effects realized when different acid catalysts or reaction
conditions were used. To be able to determine the constant m, the
proton concentration at the reaction temperature has to be known.
However, the pH value changes according to reaction temperature
and, therefore, the y-intercept of the fitted Arrhenius plot includes
an average effect of these temperature dependent pH values. As
the determination of constant m was not the intent of the current
study, the pre-exponential factors were determined using the con-
ventional Arrhenius plot and not the modified Saeman equation
and as such the proton concentration would be implicitly included
in those pre-exponential values. This approach is reasonable for
the current study as the different acids were compared under the
same pH condition. As can be seen in Table 2, the values for the
pre-exponential factors were quite similar for all of the acids at the
pH 1.5 conditions. However, at pH 3.6 there was 5–6 orders of mag-
nitude difference in the value between the weak acids and strong
acids. To provide a rough comparison of the pre-exponential factors
for different pH values, the pre-exponential factors were calculated
assuming that the factor m in Saeman equation was equal to 1,
which is a commonly used value [4]. These new pre-exponential
factors based on the Saeman equation are also listed in Table 2.
When the effect of proton concentration was deconvoluted from
the pre-exponential rate constant, similar values were obtained at
different pH values in the presence of strong acids. However, the
values found for the weak acids still gave a large difference between
those at pH 1.5 and pH 3.6.
A mechanistic study of fructose dehydration to HMF at 250 ◦C
with 50 mM sulfuric acid suggested that HMF formation occurred
via
a fructofuranosyl-cationic intermediate rather than a 3-
deoxyhexosulose intermediate resulting from an enediol reaction
[6]. Only under weak acidic conditions was some formation of HMF
from fructose observed via 3-deoxyhexuloses, but at rates rela-
tively slower than with the fructofuranosyl-cationic intermediate
mechanism. A fructofuranosyl-cationic intermediate mechanism,
which is shown in Fig. S2, was originally proposed by Antal
et al. [6] was further confirmed in later experimental stud-
ies at varying acidic conditions and temperatures. In our study,
the observed apparent activation energy for fructose conversion
was in good agreement with previously reported values and
did not change according to pKa of the acid catalyst or the pH
of the reaction solution. Therefore, the activation energy data
from the current kinetics results do not conflict the proposed
mechanism.
Further insight can be provided by the xylose degradation mech-
anisms, since xylose is an aldose like glucose but it has a higher ratio
of the ␤-furanose form at equilibrium than with fructose. Mecha-
nistic studies on furfural formation from xylose revealed that at
high temperature in the acidic regime xylose is initially present in
three different forms [27]. The open-chain form of xylose has been
suggested to be responsible for fragmentation product formation
and the xylopyranose form leads to furfural via 2,5-anhydride inter-
mediates, while the furanose form is stable [3]. The ring-opening
isomerization was reported to be relatively low at 250 ◦C in the
presence of 10−3 to 10−2 M H2SO4 and 3-deoxyglycosuloses were
In one of the proposed mechanisms for dehydration of glucose
to HMF, the activated transition state for glucose conversion was
suggested to involve an 1,2-enediol intermediate that then could
dehydrate via further enolization and formation of hexosuloses as
Please cite this article in press as: B. Cinlar, et al., Appl. Catal. A: Gen. (2012), http://dx.doi.org/10.1016/j.apcata.2012.10.013

GModel
APCATA-13924; No. of Pages6
ARTICLE IN PRESS
6
B. Cinlar et al. / Applied Catalysis A: General xxx (2012) xxx–xxx
important intermediates only at low temperatures during base cat-
alyzed reactions. These different postulated pathways for xylose
decomposition are summarized in Fig. S3. A mechanism change
did not appear to be observed in our study either due to change
in acidic strength or pH. Our values were consistent with most
literature values. The exception is a report in which an apparent
change in mechanism was observed for xylose degradation dur-
ing hemicellulose hydrolysis in the presence of maleic acid [17].
It is possible that if we had explored broader pH and temperature
regimes that we might have observed such a change.
Acknowledgement
This material is based upon work supported by the National
Science Foundation under Award No. EEC-0813570.
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.apcata.2012.10.013.
The results for glucose conversion with the different acids and
pH levels can be interpreted by considering the different proposed
pathways for xylose and the apparent HMF formation mechanism
from fructose. Within this context, it seems most likely that at lower
pH values glucose first isomerizes to fructose and then decomposes
like fructose. Whereas, at the pH 3.6 value in the presence of weak
acids the glucose decomposes directly via the 1,2-enediol interme-
diate although it is also possible that the cyclic route (see Fig. S1)
could be followed at the higher pH value.
References
[1] J.N. Chheda, Y. Roman-Leshkov, J.A. Dumesic, Green Chem. 9 (2007) 342–350.
[2] C.Z. Li, Z.H. Zhang, Z.B.K. Zhao, Tetrahedron Lett. 50 (2009) 5403–5405.
[3] M.J. Antal, T. Leesomboon, W.S. Mok, G.N. Richards, Carbohydr. Res. 217 (1991)
71–85.
[4] N.S. Mosier, C.M. Ladisch, M.R. Ladisch, Biotechnol. Bioeng. 79 (2002)
610–618.
[5] N.S. Mosier, A. Sarikaya, C.M. Ladisch, M.R. Ladisch, Biotechnol. Progr. 17 (2001)
474–480.
[6] M.J. Antal, W.S.L. Mok, G.N. Richards, Carbohydr. Res. 199 (1990) 91–109.
[7] X.H. Qian, M.R. Nimlos, D.K. Johnson, M.E. Himmel, Appl. Biochem. Biotechnol.
121 (2005) 989–997.
[8] B.F.M. Kuster, Starch - Starke 42 (1990) 314–321.
[9] C. Carlini, M. Giuttari, A.M.R. Galletti, G. Sbrana, T. Armaroli, G. Busca, Appl.
Catal. A: Gen. 183 (1999) 295–302.
4. Conclusions
Understanding the degradation pathways of monosaccharides
and the effect of acidic properties on those pathways is of great
importance for optimizing both polysaccharide hydrolysis and
monosaccharide dehydration. Unfortunately, there is apparently
conflicting results presented in the literature as the comparison
between different acid catalysts is not performed in a systematic
manner in which the effects of acid strength and number of acid
sites is deconvoluted. The current study examined homogeneous
acids with a range of acid strengths and compared their reactivity
at fixed pH values. By comparing the apparent activation energies
of model monosaccharides, it was found that sugars such as fruc-
tose and xylose displayed distinctly different dehydration reaction
characteristics than glucose. For fructose and xylose, it was found
that regardless of the nature of the acid, only the H+ activity was
responsible for the degradation conversion rates, while the associ-
ated anions only led to some differences in selectivity toward HMF
and furfural. In contrast, the strength of the acid did affect glucose
degradation. In the presence of weak homogeneous acids, glucose
underwent degradation at pH values of 1.5 and 3.6 with differences
in both the apparent activation energies and the pre-exponential
factors. Such a difference did not appear to occur in the presence of
the strong homogenous acids.
[10] C. Carlini, P. Patrono, A.M.R. Galletti, G. Sbrana, Appl. Catal. A: Gen. 275 (2004)
111–118.
[11] M. Watanabe, Y. Aizawa, T. Iida, R. Nishimura, H. Inomata, Appl. Catal. A: Gen.
295 (2005) 150–156.
[12] H.E. Vandam, A.P.G. Kieboom, H. Vanbekkum, Starch
95–101.
- Starke 38 (1986)
[13] A.S. Dias, M. Pillinger, A.A. Valente, J. Catal. 229 (2005) 414–423.
[14] O. Bobleter, W. Schwald, R. Concin, H. Binder, J. Carbohydr. Chem. 5 (1986)
387–399.
[15] W. Schwald, O. Bobleter, J. Carbohydr. Chem. 8 (1989) 565–578.
[16] Y. Roman-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933–1937.
[17] Y.L. Lu, N.S. Mosier, Biotechnol. Progr. 23 (2007) 116–123.
[18] M.L. Mednick, J. Org. Chem. 27 (1962) 398–403.
[19] N. Bhandari, D.G. Macdonald, N.N. Bakhshi, Biotechnol. Bioeng. (1984)
320–327.
[20] J.A. Bootsma, B.H. Shanks, Appl. Catal. A: Gen. 327 (2007) 44–51.
[21] T.M. Aida, K. Tajima, M. Watanabe, Y. Saito, K. Kuroda, T. Nonaka, H. Hattori,
R.L. Smith, K. Arai, J. Supercrit. Fluids 42 (2007) 110–119.
[22] S.K. Tyrlik, D. Szerszen, M. Olejnik, W. Danikiewicz, Carbohydr. Res. 315 (1999)
268–272.
[23] S.K. Tyrlik, D. Szerszen, S. Szymanski, New J. Chem. 19 (1995) 1019–1021.
[24] J.F. Saeman, Ind. Eng. Chem. 37 (1945) 43–52.
[25] M.R. Nimlos, X. Qian, M. Davis, M.E. Himmel, D.K. Johnson, J. Phys. Chem. A 110
(2006) 11824–11838.
[26] X. Qian, M.R. Nimlos, M. Davis, D.K. Johnson, M.E. Himmel, Carbohydr. Res. 340
(2005) 2319–2327.
[27] E.R. Garrett, B.H. Dvorchik, J. Pharm. Sci. 58 (1969) 813–820.
Please cite this article in press as: B. Cinlar, et al., Appl. Catal. A: Gen. (2012), http://dx.doi.org/10.1016/j.apcata.2012.10.013