The effect of different Br?nsted acids on the hydrothermal conversion of fructose to HMF

Article abstract of DOI:10.1039/c8gc00435h

The hydrothermal dehydration of fructose to 5-hydroxymethylfurfural (HMF), a promising platform chemical from renewable resources, is commonly known to be Br?nsted acid catalysed. However, it is not clear whether the acids used as catalysts may have an additional effect on the reaction, besides the donation of protons. In this work, we studied the effect of different Br?nsted acids on the hydrothermal conversion of fructose to HMF. In particular, phosphoric acid, if present in a high concentration, leads to a significantly stronger acceleration of the reaction than would be expected from the pH value, calculated under hydrothermal conditions. Acetic acid, on the other hand, seems to evoke an alternative reaction mechanism. The maximal HMF yield however is essentially unaffected by the pH value or the type of acid used.

Full text of DOI:10.1039/c8gc00435h

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: P. Körner, D. Jung
and A. Kruse, Green Chem., 2018, DOI: 10.1039/C8GC00435H.
Volume 18 Number
7
7
April 2016 Pages 1821–2242
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Cutting-edge research for a greener sustainable future
www.rsc.org/greenchem
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1463-9262
CRITICAL REVIEW
G. Chatel et al.
Heterogeneous catalytic oxidation for lignin valorization into valuable
chemicals: what results? What limitations? What trends?
rsc.li/green-chem

Page 1 of 12
PleaseGdr oe en no Ct ha de jmu si st t mr yargins
View Article Online
DOI: 10.1039/C8GC00435H
Green Chemistry
Paper
Effect of different Brønsted acids on the hydrothermal conversion
of fructose to HMF
Received 00th January 20xx,
Accepted 00th January 20xx
a
a
a
Paul Körner,* Dennis Jung and Andrea Kruse
DOI: 10.1039/x0xx00000x
The hydrothermal dehydration of fructose to 5-hydroxymethylfurfural (HMF), a promising platform chemical from
renewable resources, is commonly known to be Brønsted acid catalysed. Despite, it was not clear, if the acids used as
catalyst may have an additional effect on the reaction, besides the donation of protons. In this work, we studied the effect
of different Brønsted acids on the hydrothermal conversion of fructose to HMF. Especially phosphoric acid, if present in a
high concentration, leads to a significantly stronger acceleration of the reaction than it would be expected from the pH
value, calculated at hydrothermal conditions. Acetic acid, on the other hand, seems to evoke an alternative reaction
mechanism. The maximal HMF yield however is essentially unaffected by the pH value or the type of acid.
www.rsc.org/
2
–6
starting material.
Also, various solvents, such as
Introduction
As fossil resources are becoming scarce and have a
significant environmental and climate impact, mankind is
aprotic solvents or ionic liquids, and various catalysts,
such as Brønsted and Lewis acids, have been applied.
6–10
It could be demonstrated that the pH value as well as the
3
acid itself play a major role in the production of HMF .
looking
for
renewable
substitutes.
The comparison among different catalysts is difficult, if
reaction rate constants at reaction conditions were not
5
ꢀhydroxymethylfurfural (HMF) is a promising platform
chemical obtained by the dehydration of biomassꢀderived
hexoses, such as glucose or fructose. It can be further
processed into intermediates of bioꢀbased polymers or
fuels.
Subcritical water has the benefits of being cheap,
environmentally friendly, nonꢀflammable and nonꢀtoxic
which make it a promising solvent for industrialꢀscale
HMF production. Hydrothermal conversion processes of
hexoses however suffer from relatively low yields and
selectivities due to side and subsequent reactions forming
3
,11,9
calculated
. It is impossible to work out a conclusion
on the actual impact of the catalyst, namely the
acceleration of the reaction, if no kinetic analysis is done.
What has not been investigated so far, is a possible effect
of the type of Brønsted acid on the hydrothermal
dehydration, i.e.: Does it matter, if, for instance, sulfuric
or phosphoric acid is used as a catalyst, besides that they
are proton donators of different strength? For this
purpose, we studied the Brønsted acid catalysed
hydrothermal conversion of fructose to HMF at different
reaction times using phosphoric acid, sulfuric acid,
hydrochloric acid, nitric acid, citric acid, glycolic acid
and acetic acid as Brønsted catalysts. In some cases,
identical initial pH values were used, in other cases it
was varied. The reaction rate constants for every acid and
initial pH value are calculated. By means of the reaction
rate constants a comparison of the different acidic
catalyst is possible.
1
humins, levulinic acid (LA) and other organic acids.
Besides,
a
comprehensive understanding of the
underlying kinetics is crucial for the development of an
industrial process for the production of HMF.
Within many other studies, the dehydration reaction has
been studied intensively using different monoꢀ, diꢀ,
oligoꢀ polysaccharides and, occasionally, biomass as
The data obtained allows an estimation of the maximal
yield, reaction time and selectivity as function of reaction
conditions.
In addition, a deeper understanding of the role of
Brønsted acids shall be created. It has to be pointed out
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 2 of 12
View Article Online
DOI: 10.1039/C8GC00435H
Paper
Green Chemistry
here, that the properties of water, namely the ionic maximal selectivity is achieved at roughly 60% fructose
product changes with temperatures. As consequence, the conversion (i.e. 1ꢀC = 40%) and the maximal HMF yield
acidity of acids in subcritical water is different from between 70% and 80% fructose conversion. A divergent
1
2,13
ambient water. This effect differs for different acids.
behaviour, however, was observed for concentrated citric
acid (pH 1.4) and acetic acid, for which initially a
significantly higher and lower selectivity, respectively,
was found.
Results
Experimental yields, conversion and selectivities Figure 1, right, shows the CYS plot of the reaction of
fructose to LA, catalysed by various acids. The path that
In principle,
a faster fructose conversion, earlier
the reaction takes though the plot depends significantly
on the kind of acid used and the pH value, respectively.
A general observation is that only few LA is formed until
a fructose conversion of 40%. From between 70% and
maximum HMF yields, faster HMF degradation and
higher LA yields are observed with decreasing pH. The
selectivity of the HMF formation initially increases in
most cases.
8
0% when normally the maximal HMF yield is achieved
the LA formation increases, but very differently. In HCl
and HNO solution high LA yields are achieved even
before fructose is converted completely, while in H SO
In order to depict product yield (Y), educt conversion (C)
and selectivity (S) of the reaction within one diagram, a
ternary plot was designed which we will subsequently
term a CYS plot. It can be read as follows: on the left
axis the amount of unreacted educt is plotted (1ꢀC), on
the right axis the yield of product (Y) and on the bottom
axis the yield of byꢀ and degradation products (CꢀY), or,
in other words, what is deficit to 100%. The selectivity
can be found by drawing a line from the data point to the
lower left edge of the triangle and identifying the
intersection point with the right axis. Thus, a point laying
exactly on the bottom axis would represent a selectivity
of 0%, while a point on the left axis would represent
3
2
4
solution, selectivity and LA yield are lower. Still,
incomplete fructose conversion does not hinder high LA
yields. In H PO , pH 1.0 also very high LA yields are
3 4
achieved and the reaction proceeds considerably faster.
However, the LA yield does not exceed 30% before
nearly all fructose is converted. Interestingly, the curves
of H PO , pH 2.0 and citric acid progress similarly. In
3 4
glycolic and acetic acid only low LA yields are achieved,
even after a long reaction time.
1
00% selectivity.
Within figure 1, left, the CYS plot of the reaction of Reaction rate constants
fructose to HMF, catalysed by various acids, is depicted.
The measurement values were used to obtain the reaction
rate constants by kinetic modelling as described in the
method section. Within figure 2 k , k and k +k are
plotted against the proton concentration as determined at
For most acids investigated in our work the reaction
follows the same path, even though the reaction rates
differ drastically in some cases. Generally speaking, the
1
2
3
4
Figure 1
left: CYS plot of the reaction of fructose to HMF, catalysed by various acids. right: CYS plot of the reaction of fructose to LA, catalysed by various acids
2
| J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 3 of 12
PleaseGdr oe en no Ct ha de jmu si st t mr yargins
View Article Online
DOI: 10.1039/C8GC00435H
Green Chemistry
Paper
Figure 2
3 4
Plotting of the reaction rate constants against [H]RT. In the plotting of k +k , the values of nitric and hydrochloric acid are excepted from the fitting in order to
achieve a good linear correlation
room temperature ([H]RT) yielding a more or less linear The k in phosphoric acid at pH 1.0 (k (H PO , 1.0) is
1
1
3
4
1
trend. However, certain deviations must be observed: substantially higher than all other k values, while
k
1
(H
linear correlation is achieved, provided that the values for Ymax (tYmax
nitric and hydrochloric acid are excepted from the fitting,
as they deviate strongly downwards. In order to described within the method section. The results are
determine the effect of the actual proton concentration depicted in table 1. Interestingly,
under hydrothermal reaction conditions, [H]140°C was similar for each acid amounting 43% and 52%,
2
SO
4
, 1.1) is unexpectedly small. For k
3
+k
4
the best maximal HMF yield (
, the fructose conversion and the selectivity at
Ymax, respectively) for each acid as
Ymax), the reaction time to achieve
)
Y
max
(CYmax and S
Y
max and SYmax are very
calculated as described in the methods section.
Plotting the reaction rate constants against [H]140°C yields
a different picture as shown in figure 3.
respectively, while ꢀꢁꢂꢃꢄ varies more strongly between
5% and 95%. ꢅꢁꢂꢃꢄ in principle increases with
7
ꢆ
increasing pH, however fluctuates strongly even for
identical initial pH values.
The linear correlation of k and k as function of the
1
2
calculated proton concentration is better than as function
Table 1
Ymax, tYmax, Cymax and SYmax calculated from the modelled k values by using
of the initial pH value. However, the values for the analytical solution for Frc and HMF.
Acid
pHRT
Ymax [%]
t
Ymax[min]
C
[
Ymax
S
Ymax
phosphoric acid at high proton concentration have to be
excepted from the fitting, as they deviate significantly
upwards. For k +k no linear correlation can be identified
%]
[%]
51
Phosphoric
Sulfuric
1
45
40
41
40
42
42
43
44
13
32
87
83
75
76
78
88
77
88
3
4
1.1
1.2
1.2
1.4
1.8
2
48
55
53
54
48
56
50
in the plotting against [H]140°C
.
Hydrochloric
Nitric
22
19
Theoretical maximum HMF yield
Citric
53
The analytical solutions of the kinetic model for [Frc]
and [HMF] were used to calculate the theoretical
Glycolic
Citric
224
167
431
Phosphoric
2
Figure 3
Plotting of the reaction rate constants against [H]140°C. In the plotting of k
1
and k
2
, the value of phosphoric acid at pH 1.0 ([H]140°C = 0.047) is excepted in
order to achieve a good linear correlation.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 4 of 12
View Article Online
DOI: 10.1039/C8GC00435H
Paper
Green Chemistry
Acetic
2
48
243
95
51
± 2.3%
± 6.4%
± 2.6%
average
42.6
82.9
51.5
performing an extrapolation of temperature and H
concentration.
Jiang et al. investigated the decomposition of glucose to
2 4
SO
Discussion
10
Comparison of the kinetic data with literature
values
HMF and LA at 140°C, catalysed by FeCl . The kinetic
3
data they provide bases on the assumption that there is no
A comparison with literature values seems appropriate
even though it is difficult, because the reaction
conditions often are not directly comparable. Besides, it
must be taken into consideration that reaction rate
constants usually are strongly temperatureꢀdepending.
This can be a reason for alleged deviations from
literature values, as different systems for temperature
determination are used.
direct conversion of glucose to HMF, but that it is
isomerised to fructose first. Hence, their data is in
principle way more comparable with ours as the rate
constant for fructose conversion is available. FeCl of
3
course is a Lewis acid which however is partially
hydrolysed in water yielding HCl. Thus, depending on
the FeCl
3
concentration the initial pH in their study was
linear
1
.06 and 1.74, respectively. Assuming
a
The pH values noted here have to be understood as initial
pH values.
dependency of k on [H], we can calculate the k values for
5
pH 1.2. Although, except for k these calculated values
2
Fachri et al. investigated the HMF formation from inulin
are substantially higher than our findings for HCl, pH
at 170°C and 30 min, catalysed by different acids at their
concentration of 0.006 M. They found higher HMF
yields for stronger acids, which however is due to the
fact that stronger acids cause a lower pH than weaker
acids at the same concentration. Thus, the reason for the
observed effect are different pH values and not
necessarily the type of acid. Lower pH values cause an
accelerated HMF formation as well as a faster HMF
degradation. This is why it is difficult to interpret
different HMF yields, if the reaction was studied at only
one reaction time and one acid concentration.
1
.2, indicating that the different iron species formed from
FeCl in water have an additional effect on the reaction
3
apart from the isomerisation of glucose to fructose,
which is in complete accordance with the conclusion of
the authors.
2
Swift et al. studied the HMF conversion and production
from fructose in HCl at pH values between 0.7 and 1.6
and temperatures ranging from 70 to 150°C. The kinetic
model they take as a basis is more complex than ours in
such a way that pH independent rate constants are
determined by assuming that each conversion is a
bimolecular reaction depending on both, the
concentration of the educt and of the protons. For the
conversion of fructose to HMF they take into
consideration that an intermediate is formed of which
consumption, not formation, is the rateꢀlimiting step.
However, they also assume that only fructoꢀfuranose can
directly be converted into this intermediate, while the
partition between openꢀchain, pyranose and furanose
fructose depends on the temperature. Attempts of
calculating the rate constants for pH 1.2 and 140°C from
their data in order to compare them with k(HCl, pH 1.2)
yields values that are more or less in a range with ours in
1
4
Weiqi & Shubin studied the LA formation from glucose
in a mixture of H PO and CrCl that according to our
3
4
3
calculation should have an initial pH of around 2 at
temperatures ranging from 150 to 180°C. By
extrapolating the Arrhenius graph with their data in order
to estimate the rate constants at 140°C, we find that k
and k are within the same dimension of k(H PO , pH
.0), but k and k are significantly larger. Obviously,
1
3
3
4
2
2
4
CrCl increases especially the rehydration rate, which is
3
also a basic outcome of the cited study.
The LA formation from glucose was also studied by
1
5
Chang et al. at 170°C and 190°C, respectively, by using
%, 3% or 5% sulfuric acid as catalyst. They provide an
1
the case of k
Swift and coꢀworkers, a divergent understanding of what
and k represent is applied. According to their
1 2 3 4
and k , but not k and k . In the work of
empirical model to predict k as a function of the H SO4
2
concentration and the temperature. However, if we apply
k
3
4
our data (H
for k2 lays within the range of our findings, while
predicted k and k are several orders of magnitude
2 4
SO pH 1.1, 140°C), only the predicted value
hypothesis, humins may be formed from HMF via one
path and from fructose via two paths of which one is
accompanied with the formation of formic acid. In our
work, k and k represent the formation of undefined side
1
3
smaller than the experimental values. Obviously, a
comparison of glucose and fructose conversion is not
fruitful under these conditions. At that the risks of
3
4
and degradation products which certainly comprise
humins, but not exclusively. The mechanism of humin
4
| J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 5 of 12
PleaseGdr oe en no Ct ha de jmu si st t mr yargins
View Article Online
DOI: 10.1039/C8GC00435H
Green Chemistry
Paper
formation from fructose or HMF, respectively, has not Which can be inserted into equ. 4:
been fully understood, yet. Our data suggests that at least
ꢑ
ꢇꢈꢉꢊ
ꢇꢅ
ꢌ
(7)
ꢍꢐ
the overall formation of side and degradation products is
not or not exclusively Brønsted acid catalysed. This may
be an additional reason why an implication of the proton
concentration into the determination of k may lead to
divergent results.
ꢋ ꢌꢑ
ꢈꢎꢊꢈꢒꢊ
ꢍꢔ
ꢑ
ꢓ ꢌꢑ
ꢌ
ꢍꢔ
ꢍꢖ
Equating equ. 1 and 7 and rearranging them
x
after k gives:
ꢑ
ꢑ
ꢌ ꢌ
(8)
ꢍꢔ ꢍꢐ
ꢌꢍ ꢋ
ꢈꢒꢊ
ꢑ
ꢓ ꢌꢑ
ꢌ
ꢍꢔ
ꢍꢖ
Dependency of the rate constants on the type of
acid
When plotting the reaction rate constants against the From equ. 8 we learn that
ꢌ
ꢍ
depends linearly on [H]. So,
proton concentration as measured at room temperature if we assume that [H] is the only catalyst in our system
[H]RT), a more or less linear trend can be identified for and there is no effect by the type of acid, we should be
, k , and to some extent for k +k
, as it can be seen in able to identify one linear trend when plotting k against
figure 2. This would be expected for Brønsted acid [H] for all acids.
catalysed reactions, as our kinetic model bases on the According to this reflection, our findings confirm that
assumption that the reaction of an educt E to a product P especially the HMF formation from fructose (k ) and the
) are Brønsted acid catalysed.
(
k
1
2
3
4
x
1
is of first order with respect to E.
HMF rehydration to LA (k
2
Furthermore, the linear correlation of the plot improves,
if not the proton concentration at room temperature, but
at reaction temperature is considered. This suggests that
it makes sense to determine the latter one, in particular if
different acids shall be compared. From the same data we
can conclude that the overall formation of side and
E → P
ꢇꢈꢉꢊ
ꢇꢅ
ꢋ ꢌ ꢈꢎꢊ
(
1)
ꢍ
Thus, it neglects that fact that this reaction, at least in
some cases, can be homogeneously catalysed by protons
3 4
degradation products (k +k ), which comprise humins,
acetic acid, formic acid, furfural and others, either is not
Brønsted acid catalysed or the catalytic effect is strongly
superimposed by other mechanisms, as no linear trend
x
which makes k a function of [H].
However, if we assume the common kinetic model for
homogeneously catalysed reactions as:
can be identified in the k
3 4
+k plot (figure 3).
Although, considering the k and k plot in figure 3,
1
2
certain deviations from the linear trend are observed.
Especially, k (H PO , pH 1.0) and k (H PO , pH 1.0)
deviate substantially upwards, which suggests that the
type of acid can have an additional effect on these
reactions.
In order to discuss the effect on the acid type on the
dehydration reaction, we distinguish between three cases:
A) acid as coꢀcatalyst; B) acid as coꢀsolvent; C) acid as
coꢀreagent.
(
R : formation of a protonated intermediate EH; R :
1
2
1
3
4
2
3
4
formation of P) and furthermore assume quasiꢀ
stationarity for [EH], then we can draft three ordinary
differential equations (equ. 2ꢀ4):
ꢇꢈꢎꢊ
ꢇꢅ
(2)
(3)
(4)
ꢑ
ꢑ
ꢋ ꢏꢌ ꢈꢎꢊꢈꢒꢊ ꢓ ꢌ ꢈꢎꢒꢊ
ꢍꢐ
ꢍꢔ
ꢇꢈꢎꢒꢊ
ꢇꢅ
ꢋ ꢌꢑ ꢈꢎꢊꢈꢒꢊ ꢏ ꢕꢌꢑ ꢓ ꢌ ꢍꢑ ꢖꢗꢈꢎꢒꢊ ꢋ 0
ꢍꢐ
ꢍꢔ
Case A comprises catalytic effects of the acid molecule
or the corresponding anion, besides the donation of
protons, that could be the formation of intermediate
complexes with the substrate, the transfer of electrons or
protons etc. For instance, sulfuric acid is a common
catalyst in esterification reactions as it does not only
provide protons, but also withdraws the formed water
molecules by complexing them. For transesterification
reactions it has been shown that the counter anion of the
acid catalyst (which usually is sulfuric acid) participates
ꢇꢈꢉꢊ
ꢑ
ꢋ ꢌ ꢈꢎꢒꢊ
ꢍꢔ
ꢇꢅ
Equ. 3 can be formed into:
ꢑ
ꢑ
ꢑ
ꢌ ꢈꢎꢊꢈꢒꢊ ꢋ ꢕꢌ ꢓ ꢌ ꢗꢈꢎꢒꢊ
(5)
(6)
ꢍꢐ
ꢍꢔ
ꢍꢖ
Transposing equ. 5 leads to:
ꢑ
ꢌ
ꢍꢐ
16
ꢈꢎꢒꢊ ꢋ
ꢈꢎꢊꢈꢒꢊ
in the transition state, at least if the solvent is nonpolar.
ꢑ
ꢓ ꢌꢑ
ꢌ
ꢍꢔ
ꢍꢖ
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 6 of 12
View Article Online
DOI: 10.1039/C8GC00435H
Paper
Green Chemistry
Case B comprises the possibility of the acid used to substance to the reaction medium may not only change
change the properties of the solvent and hence to act as a the amount of protons, but also their activity or
coꢀsolvent. The effects of organic solvents in biomass reactivity, respectively, which can have a significant
1
7
conversion reactions have been reviewed by Shuai and effect on the overall reaction rate.
1
7
Luterbacher . Within this paragraph we will focus on the Another effect can take place with respect to the reactant,
possible solvent effects of the acid on the catalyst, the i.e. the substrate which is fructose in our work. It is
reactants and the transition state. It should however be commonly known that hexoses appear in five different
taken into consideration that solvent effects of the acid isomers that exist in an equilibrium: an
probably only appear in the case that high concentrations furanose, an ꢀ and a ꢀpyranose as well as an openꢀchain
of these acids are used as we did for H PO pH 1.0 form. The ratio of these isomers does not only depend on
1.5 M), citric acid pH 1.4 (2.1 M), glycolic acid pH 1.8 the temperature, but also on the solvent.
αꢀ and a βꢀ
α
β
3
4
18,19
(
(
Swift and coꢀ
1.7 M) and acetic acid pH 2.0 (5.7 M). Although, only workers determined the effect of neglecting the isomer
PO pH 1.0 strongly deviating reaction rate distribution on the apparent fructose dehydration rate and
2
for H
3
4
constants have been identified, while the other systems found a deviation by a factor of two. Hence, the isomer
rather stand out by a different reaction behaviour in terms distribution is a relevant factor and the possible effect of
of intermediate selectivity and yield, which we will the acid used on the isomer distribution should be taken
discuss in next section.
into consideration.
While a possible coꢀcatalytic effect of the acid used or The most obvious effect of the solvent on the reaction of
especially the counterion should not be excluded, there is a molecule arises from the fact that it always surrounds
no doubt that the formation of HMF from fructose is the molecule. In doing so, it affects the mobility of the
catalysed by protons. In this context it is intriguing to molecule, the likeliness of protonation etc. In other
note that the standard Gibbs free energy of the protons words, the solvent has an influence on the transition state
1
7
depends on the solvent. Thus, adding a coꢀsolventꢀlike with various consequences. Insofar, the coꢀsolvent
20,21,22, 23
Figure 4
Different hypotheses about the HMF formation from the intermediate Frcf+
6
| J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 7 of 12
PleaseGdr oe en no Ct ha de jmu si st t mr yargins
View Article Online
DOI: 10.1039/C8GC00435H
Green Chemistry
Paper
effect is comparable with a coꢀcatalytic effect. The the substrate or intermediates that could either be deadꢀ
difference consists in that a coꢀcatalyst is active at a low ends (or lead to deadꢀends) of the reaction which would
concentration due to a very specific interaction with the in particular affect the selectivity of the reaction, or that
substrate, while a coꢀsolvent must be present in a high could be further intermediates of which formation can
amount.
affect both, the selectivity and the velocity of the
Case C comprises the possibility of the acid molecule or reaction.
the corresponding anion to form additional products with
H+
2
-H O
Frcf+
Frcf
HMF
-X-
-
-
-, H
2 4
PO
-,
X : Cl , HSO
4
X-
-
3
CH COO etc.
FrcX
+
Figure 5
Possible formation of intermediate fructose derivatives by nucleophilic attack on Frcf
About the exact mechanism of the HMF formation from contradicts the assumption that the first dehydration is a
fructose no consensus has been reached, yet. Antal and classic E reaction. It also implicates that the formation
1
2
0
+
coꢀworkers somehow ended the discussion whether it of Frcf is comparatively fast, which gives rise to the
proceeds via an openꢀchain or a cyclic path. Since then, it consideration that it may feature an attractive target for
is generally accepted that the first step comprises the nucleophilic attacks as summarized in figure 5.
abstraction of the hydroxyl group from C2 of fructoꢀ Of course, in an aqueous medium the most important
furanose (Frcf) leading to a fructoꢀfuranosyl cation nucleophile is water and, to some extent, hydroxide ions.
+
+
ꢀ
(
Frcf ), which in the end yields HMF without any The recombination of Frcf with water or OH ions,
intermittent ring opening. About the further reaction of respectively, would simply yield fructose back. However,
+
Frcf different hypotheses exist, that are condensed in it is also imaginable that the acid used or especially the
figure 4.
corresponding anion acts as nucleophile yielding
One assumption is that a subsequent abstraction of the fructoseꢀ2ꢀphosphate or fructoseꢀ2ꢀsulfate, for instance.
2
4
proton from C1 takes place leading to an enol (FrcE) Phan and coꢀworkers studied the hydrolysis of methyl
which in principle features an E reaction. FrcE is then glucopyranoside under nonꢀhydrothermal conditions.
1
either isomerised to the corresponding aldehyde (FrcA) They found that the hydrolysis rate constant is
20
prior to the dehydration of the hydroxyl group from C3 , significantly larger in a HBr system compared with a
or the dehydration from C3 takes place first, resulting HCl or SO system, even though the proton
subsequently in the formation of the aldehyde concentration should have been identically large in all
H
2
4
2
1,22
function
. Another assumption consists in a hydride three systems or even larger in the H SO system. They
2 4
+
ꢀ
shift taking place from C1 to C2 in Frcf yielding attributed this finding to a direct participation of the Br
23
somehow a protonated FrcA that is further dehydrated . ion in the reaction. When the solvent was modified by
In an E reaction, the abstraction of the leaving group adding 74% and 82% 1,4ꢀdioxane, respectively, the
1
usually is the rateꢀlimiting step, while the subsequent reaction rate constants increased considerably, also in the
2
deprotonation is fast. Swift et al. studied the reaction of case of the HCl system, which suggests that the
unlabelled fructose and fructose that is deuterated at C1. nucleophilicity of chloride and bromide increases with
They found that both, the fructose conversion and the increasing 1,4ꢀdioxane content allowing them for a direct
HMF formation proceed considerably more slowly in the participation in the reaction. In the frame of a subsequent
2
5
case of deuterated fructose due to the kinetic isotope work , Phan et al. confirmed their assumption by
effect. This finding suggests that the abstraction of the demonstrating that the addition of KCl and in particular
proton or hydride from C1 is rateꢀlimiting which KBr to the respective system causes a further increase of
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 8 of 12
View Article Online
DOI: 10.1039/C8GC00435H
Paper
Green Chemistry
the reaction rate constant. It also increased in the case the enormously high k
1
(H
3
PO
4
, pH 1.0) compared with
that KHSO was added to the H SO system, but the the k values in other systems even at higher proton
4
2
4
1
authors attribute this to the higher proton activity rather concentration. Possibly, fructoseꢀ2ꢀphosphate was
than to a direct participation of the anion. However, the formed in a significant amount, as large amounts of
ꢀ
formation of sulfated sugars is not in principle devious,
as such compounds are formed to a large extent, if
H
H
3
3
PO
PO
4
4
/H
2
PO
4
were available (the concentration of
at pH 1.0 is roughly 1.5 M), which however is
carbohydrates are treated with concentrated sulfuric less stable than fructose itself. This results in a higher
2
6
+
acid .
Besides,
numerous
naturally
sulfated quasiꢀstationary concentration of Frcf and thus in a
polysaccharides such as agar, carrageenan and higher HMF formation rate. By contrast, fructoseꢀ2ꢀ
chondroitin sulfate are known. Also, sugar phosphates chloride and fructoseꢀ2ꢀsulfate were not formed in a
are common natural compounds whereby DNA is a significant amount within the respective systems due to
prominent example. Phosphorylases produce glucoseꢀ1ꢀ the low concentration of chloride (0.063 M) and
phosphate from glucans via phosphorolysis. If chitin is hydrogensufate (0.060 M), respectively. This renders
treated with concentrated phosphoric acid,
N
ꢀ
these systems much more close to an exclusively
acetylglucosamineꢀ1ꢀphosphate is formed from an Brønsted acid catalysed system.
2
7
intermediate carbocation. This reaction can even be
used to determine the degree of ꢀacetylation in
N
2
8
chitin/chitosan. Overall, it seems permissible to assume Selectivity of the HMF and LA formation from
fructose
that the acid used in the hydrothermal treatment of
fructose is able to form fructose derivatives of the FrcX
type. The contribution of especially halide ions in the
fructose dehydration as both, nucleophile attacking on C2
and base removing the proton from C1 in order to form
FrcfE, is also proposed in other works, even though nonꢀ
The yield and the selectivity of HMF and also of LA
formation strongly depend on the reaction time which is
influenced by the reaction temperature, the solvent and
the catalyst. They also depend on the kind of
2
9,30
carbohydrate used as educt
, though within the frame
6
,8
aqueous solvents were used in those.
of this work we focus on the conversion of fructose.
Generally speaking, a short reaction time leads to low
HMF and LA yields while the selectivity with regard to
HMF might be relatively high. This is attributed to a low
fructose conversion. Conversely, a long reaction time can
also be the reason for low HMF yields and selectivities
due to decomposition or polymerisation of the substance.
In return, the LA yield increases. This ambivalent
behaviour of HMF renders it difficult to classify the
effect of certain reaction conditions, if only few reaction
times were studied.
The formation of FrcX ought to be understood as a SN1
reaction, which consists of two partial steps: the
dissociation of the leaving group from the molecule
+
yielding a carbocation (Frcf in this case), followed by
the attack of a nucleophile. Normally, the first step is
rateꢀlimiting. Of course, the counterions of strong or
medium strong acids are weak nucleophiles in aqueous
media. However, because the second step of a SN1
reaction is much faster than the first one, the
nucleophilicity of X, does not affect the formation of RX,
+
3
provided that X is available in a sufficient amount. Frcf
Asghari & Yoshida investigated the effect of various
will literally react with the first nucleophile that comes
along. Although, the nucleophilicity of X affects the
stability of RX with regard to the reꢀdissociation, as a
acids at different pH values at 240°C and a residence
time of 120 s on the formation of degradation products
from fructose in water. Even though their results are not
easily comparable with those in our work, as they did not
only employ a higher reaction temperature, but especially
do not provide any reaction rate constants, they were able
to show that the HMF yield differs for different acids
even at same room temperature pH. We plotted a
selection of their data against the initial proton
concentration in figure 6, left, where it can be seen that
there are apparently large differences in HMF and LA
yields within one pH value. However, if we calculate the
actual proton concentrations at 240°C, similarly as we
did it with our own data, and plot the yields against them,
a more reasonable picture appears. As it can be seen in
weak nucleophile is a good leaving group. As k depends
x
on [EH] (as we discussed a few pages ago), which in this
+
case is identical with [Frcf ], a change in the quasiꢀ
+
stationary concentration of [Frcf ] has an impact on the
HMF formation rate. We assume that the dissociation of
+
ꢀ
FrcX into Frcf and X is fast, but its velocity depends on
the stability of FrcX. Furthermore, we assume that the
formation of FrcX is even faster, but its rate depends on
the concentration of X. Eventually, we assume that the
+
reaction of Frcf towards HMF is slow: then the
+
intermediate concentration of [Frcf ] and hence the HMF
formation rate increases with increasing amount of X and
decreasing stability of FrcX. This could be the reason for
8
| J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 9 of 12
PleaseGdr oe en no Ct ha de jmu si st t mr yargins
View Article Online
DOI: 10.1039/C8GC00435H
Green Chemistry
Paper
figure 6, right, the highest HMF yields are achieved at
low proton concentration when only little degradation to former assumption would be in accordance with our
LA has taken place. With increasing proton finding of a much larger k (H PO , pH 1.0).
3 4
H PO is either faster or more selective. At least the
1
3
4
concentration, the HMF yield decreases while the LA The fructose dehydration has been studied in various
yield increases. Still, there are significant differences in reaction systems. In a purely aqueous system, i.e. without
the HMF yield even at similar proton concentration: way catalyst, HMF yields up to 56% were achieved with a
better yields are achieved with H
3
PO
4
than with selectivity of 78%, when 4.5 wt% fructose were
HCl/H
2
SO
4
, which suggest that the HMF formation in
3
Figure 6 Data from Asghari & Yoshida ; HMF (blue) and LA (orange) yield plotted against [H]
31
dehydrated at 175°C for 1.5 h. With 11 wt% fructose at behaviour of acetic acid in the dehydration of fructose
00°C for 30 min 51% HMF were achieved with a can also be concluded from our data.
2
3
2
selectivity of 57%. The presence of Brønsted acid The utilisation of heterogeneous catalysts in aqueous
catalysts seemingly tends to decrease the reaction time, media leads to a large range of yields and selectivities,
which however is not necessarily favourable for the HMF depending on the reaction conditions. However, they
yield, because the HMF rehydration which gives LA is usually are not larger than with homogeneous catalysts.
also acidꢀcatalysed. In 1 mM H SO4 9 wt% fructose In nonꢀaqueous, especially aprotic solvents significantly
2
yielded 23% HMF with a selectivity of 25% after 5 min better yields and selectivities towards HMF are achieved,
1
1
at 200°C. From 4.5 wt% fructose in 16 mol% H
40°C and 4 mol% H PO at 260°C, respectively, 40% absence of water, among other reasons.
HMF were produced. At 95°C, 9 wt% fructose in In order to find an appropriate basis for the comparison
00 mol% HCl gave 26% and 30% HMF with a of HMF yield and selectivity in dependence on the
selectivity of 57% and 48% after 16 and 24 min, catalyst, The analytical solutions of the kinetic model for
3 4
PO at which can be attributed to the missing rehydration in the
1
2
3
4
4
4
3
3
respectively. An interesting effect can be stated for the [Frc] and [HMF]
use of formic and acetic acid as catalyst, because maximal HMF yield (
comparably high yields were achieved with 56% and
, the fructose conversion and the selectivity at
8% at a selectivity of 61% and 64% after 10 and 20 Ymaxꢆ
CYmaxꢗꢆ and SYmax, respectively) for each acid as
min, respectively, at 200°C and a catalyst loading of 100 described within the method section. As it can be seen in
ꢆ
were used to calculate the theoretical
Y
max), the reaction time to achieve
Y
maxꢆ Ymax
ꢕt ꢗ
5
ꢕ
3
2
wt%.
table 1, the highest Ymax is achieved in the acetic acid
In one case, even a selectivity of 100% is reported for the system while the H SO systems gives the lowest yield.
2
4
conversion of 4.5 wt% fructose in formic acid pH 2.7 at Although, the differences in Ymax as well as in SYmax
3
1
1
75°C after 45 min. The yield was 56%. Possibly, the between different systems are not significant. In this
presence of acetic and formic acid, which are degradation
products, prevents side and subsequent reactions by
shifting the equilibrium state. At least, a divergent
context, it has to be stressed that
higher pH values. It is only a matter of time when it is
achieved which is an interesting finding. Swift and coꢀ
Ymax is not lower for
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 9
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 10 of 12
View Article Online
DOI: 10.1039/C8GC00435H
Paper
workers found that
Green Chemistry
2
Y
max depends on the reaction the LA yield is low, even at high fructose conversion and
temperature as it was higher for higher temperatures. after long reaction times. Interestingly, this is not
Obviously, an analogue effect does not take place in case reflected in the selectivity towards HMF that is not
of the initial pH value (at least, if it is 2 or below).ꢆtYmax higher compared with the other systems. Obviously, the
expectedly increases with increasing pH value, but low rehydration rate is compensated by higher sideꢀ
fluctuates strongly even for identical pH values. Also, reaction rates. In the case of phosphoric and citric acid,
C
Ymax
ꢆ
differs for different systems such as in the acetic the LA yield does not exceed 30% before nearly all
acid system it is possible to aim at a nearly complete fructose is converted which makes these acids more
fructose conversion in order to achieve
max, while in the suitable catalyst for processes that strive selectively for
Y
HCl system a conversion exceeding 75% lowers the HMF.
HMF yield. With regard to a technical process, the latter The curve progressions denoted here correlate perfectly
case is of course not favoured as it either means that a with the ratio of LA formation to fructose conversion
significant percentage of the substrate is dissipated or a
recycling is required.
(
i.e.ꢆꢌ
ꢔ
⁄ꢕꢌ
ꢐ
ꢓ ꢌ
ꢖ
ꢗ
) which is 0.54 and 0.53 for HCl and
SO and 0.10 and 0.01
HNO
3
, respectively, 0.31 for H
2
4
The discussion of yield, conversion and selectivity as a for glycolic and acetic acid, respectively. For phosphoric
function of the reaction time is often hindered, because and citric acid, it fluctuates between 0.19 and 0.25. It is
multiple diagrams must be used to depict the relation. In interesting to note that, even though k
order to facilitate the representation of all four and (H PO pH 1.0) have been found to be
parameters, we developed the CYS plot which is a single exceptionally high, the ꢓ ꢌ ratio is similar to
⁄ꢕꢌ
diagram for an entire reaction. From the representation of that of citric acid and more diluted H PO
the reaction of fructose to HMF within a CYS plot A final important observation is that the selectivity
figure 1) we learn that for most acids investigated in our towards HMF initially increased, especially at the higher
1 3 4
(H PO , ,pH 1.0)
k
2
3
4
,
ꢌ
ꢔ
ꢗ
ꢖ
ꢐ
.
4
3
(
work the reaction follows an identical path, even though pH values investigated. This might be unexpected when
the reaction rates differ drastically in some cases. The considering that the HMF formation always has to
divergent behaviour of concentrated citric acid (pH 1.4) compete with the HMF degradation and polymerisation.
2
and acetic acid indicates that either a different reaction In addition, it contradicts the outcome of other works.
mechanism takes place in these solvents, or the relations Although, it can be attributed to the assumption that an
of the elementary reaction rates differ. Especially, the intermediate is formed from fructose of which
behaviour of the acetic acid system might be explainable consumption, not formation, is rateꢀlimiting in the HMF
with hypotheses proposed previously: First, as
a
formation. Hence, the initial fructose consumption is
degradation product acetic acid may shift the equilibrium faster than the initial HMF formation, which is reflected
state preventing the degradation of fructose to other in a lower selectivity at the beginning of the reaction.
products than HMF. Second, it may act as a coꢀsolvent However, this kind of behaviour is possibly not observed
influencing the transition states, for instance, as it is at all reaction conditions and in particular not, if the
present in a high amount (5.7 M). Third, acetate may reaction proceeds fast due to a low pH value and/or a
contribute to the dehydration of fructose by forming high temperature. Also, it requires that the intermediate
2ꢀO
ꢀacetlyfructose or other intermediates which, in the does not occur in a quasiꢀstationary state, but that its
beginning of the reaction, causes a high fructose concentration depends on the amount of fructose or
conversion without HMF formation what is indeed fructose derivative, respectively.
observed in our experiments. Later, these intermediates
may react more selectively towards HMF, which is why
in the end the highest yields were obtained with the
Experimental
acetic acid system.
Material
Considering the CYS plot of the reaction of fructose to
LA (figure 1), we find three types of courses: In the case
of HCl, HNO and H SO , high LA yields are already
3 2 4
achieved even before all fructose is converted into HMF.
In the consequence, the maximal HMF yield should be
lower for these acids which in principle is true according
Fructose, phosphoric acid, sulfuric acid, hydrochloric
acid, nitric acid, citric acid, glycolic acid, acetic acid,
methanol, acetonitrile and trifluoroacetic acid were
purchased from VWR. The analytical standards levulinic
acid and 5ꢀHydroxymethylfurural were purchased from
VWR and Sigma Aldrich, respectively.
to
Ymax we calculated for them (table 1). On the other
Procedure
hand, the selectivity towards LA is high preventing sideꢀ
reactions of HMF. In the case of acetic and glycolic acid,
1
0 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 11 of 12
PleaseGdr oe en no Ct ha de jmu si st t mr yargins
View Article Online
DOI: 10.1039/C8GC00435H
Green Chemistry
00 g/l fructose are dissolved in the respective aqueous ꢇꢈꢟ
Paper
1
ꢐ
ꢓ ꢟ
ꢇꢅ
ꢊ
ꢔ
ꢋ ꢌ ꢈꢘꢙꢚꢊ ꢓ ꢌ ꢈꢒꢛꢘꢊ
(3)
ꢖ
ꢜ
acid. The initial pH is determined with a pH probe.
0.6 ml of the resulting solution are filled into a
stainlessꢀsteal microautoclave (total volume ≈ 15 ml) and
heated up to 140°C in a modified GC oven. Reaction
1
With
as the undefined degradation products from fructose and
HMF. ꢈFrcꢊ,ꢆ ꢈHMFꢊ,ꢆ ꢈLAꢊ and ꢈR ꢆ ꢓꢆ R are the molar
concentrations (mol/l) of the substances.
1 2 3 4 1 2
k ,ꢆk ,ꢆk ,ꢆk the reaction rate constants and R ꢆꢓꢆR
1
2
ꢊ
R R
time (t ) counts from reaching 139.0°C (i.e. t = 0 min),
the preheating time is around 15 min. After the respective
reaction time, the microautoclave is removed from the
oven and quickly cooled down in a water bath. The
product solution is analysed by HPLC.
The calculation of the reaction rate constants has been
done with MATLAB R2015b by numerical integration of
equations 9ꢀ12 with ODE45 followed by an optimization
with the lsqnonlin tool.
Analysis
HPLC analysis was carried out in
Prominence System comprising a refractive index
detector.
a Shimadzu
Determination of yield, conversion and
selectivity
For the determination of the fructose content, a YMCꢀ Product yield
Y, educt conversion C and selectivity of the
Pack Polyamine II column (250 x 4.6 mm I.D.) was used. reaction are calculated using equ. 13ꢀ15 for both, the
S
The measurements were run isocraticly with 78:22 experimental and the theoretical data.
acetonitrile:water containing 10 mM ammonium formiate
as the eluent at a flow rate of 1.0 ml/min and an oven
temperature of 35°C.
ꢚ ꢢ
ꢡ
ꢣ
ꢠ ꢋꢆ
(4)
ꢚ ꢢ
ꢣ
ꢡ
For the determination of the HMF and LA content, a
Luna C18 column (150 x 4.6 mm I.D.) was used. The
measurements were run isocraticly with 20:80
methanol:water containing 0.1% trifluoroacetic acid as
the eluent at a flow rate of 0.5 ml/min and an oven ꢥ ꢋ
temperature of 25°C.
ꢚꢣꢤ ꢏ ꢚꢣ
ꢚꢣꢤ
ꢀ
ꢋ
(5)
(6)
ꢠ
ꢀ
Prior to each measurement, the samples were filtered
through a syringe filter. A dilution was not necessary.
The injection volume was 30 µl per analysis.
For the calculation we used the molar concentration
instead of the amount of substance, as in principle there
is no volume change during the reaction on the one hand
and the exact determination of the amount of substance
would be more errorꢀprone on the other hand.
Basing on the kinetic model applied (scheme 1), the
analytical solutions for ꢈFrcꢊ and ꢈHMFꢊ have been
identified (equ. 16 and 17).
Kinetic model
The reaction pathways in scheme 1 were used to set up a
kinetic model of the reaction network (equ. 9ꢀ12).
ꢈꢘꢙꢚꢊ ꢋ ꢈꢘꢙꢚꢊꢤꢦꢧꢕꢨ
ꢪꢨ
ꢫ
ꢗꢬ
ꢩ
(7)
ꢌꢐꢈꢘꢙꢚꢊ
ꢤ
ꢕ
ꢧꢨ
ꢧꢨ
ꢫ
ꢪꢨ
ꢯ
ꢪꢨ ꢬ
ꢰꢗ
ꢈꢒꢛꢘꢊ ꢋ ꢭ
ꢮꢦ
ꢩ
ꢏ 1ꢱ
ꢏꢌ ꢏ ꢌ ꢓ ꢌ ꢓ ꢌ
ꢜ
ꢐ
ꢖ
ꢔ
(
8)
ꢯ ꢰ
ꢪꢨ
ꢧꢕꢨ ꢗꢬ
Scheme 1
Proposed reaction pathways for fructose conversion to HMF and
ꢓ ꢈꢒꢛꢘꢊ ꢲ ꢦ
ꢤ
levulinic acid. R1 and R2 are unknown decomposition products.
ꢇꢈꢘꢙꢚꢊ
ꢋ ꢏꢆꢌ ꢈꢘꢙꢚꢊ ꢏ ꢌ ꢈꢘꢙꢚꢊ
(9)
(10)
(11)
ꢐ
ꢖ
ꢇꢅ
In order to determine the theoretical maximum yield
ꢇꢈꢒꢛꢘꢊ
ꢇꢅ
(
[
Y
max) as well as
HMF]Ymax are obtained from equ. 16 and 17,
is obtained by
CYmaxꢆ and SYmax, [Frc]Ymax and
ꢋ ꢆꢌ ꢈꢘꢙꢚꢊ ꢏ ꢌ ꢈꢒꢛꢘꢊ ꢏ ꢌ ꢈꢒꢛꢘꢊ
ꢐ
ꢔ
ꢜ
ꢇꢈꢝꢞꢊ
respectively, by applying
t = tYmaxꢆ .ꢆtYmaxꢆ
ꢋ ꢌ ꢈꢒꢛꢘꢊ
ꢔ
ꢇꢅ
equating equ. 10 with zero.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 11
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 12 of 12
View Article Online
DOI: 10.1039/C8GC00435H
Paper
Green Chemistry
Computational determination of the proton
concentration at elevated temperature
A. Riisager, Chem. ꢀ A Eur. J., 2011, 17, 1456–1464.
C. Fan, H. Guan, H. Zhang, J. Wang, S. Wang and X.
Wang, Biomass and Bioenergy, 2011, 35, 2659–2665.
Y. Jiang, L. Yang, C. M. Bohn, G. Li, D. Han, N. S.
Mosier, J. T. Miller, H. I. Kenttämaa and M. M. Abuꢀ
Omar, Org. Chem. Front., 2015, 2, 1388–1396.
M. Watanabe, Y. Aizawa, T. Iida, T. M. Aida, C.
Levy, K. Sue and H. Inomata, Carbohydr. Res., 2005,
340, 1925–1930.
9
1
The determination of the proton concentration at reaction
0
34
temperature bases on a model developed by Helgeson ,
in which the pK as a function of the temperature is
a
calculated from the free energy of dissociation. Helgeson
also provides a set of parameters for different acids.
Details about these calculations can be found in the
supplementary information.
11
12
A. Kruse and E. Dinjus, J. Supercrit. Fluids, 2007,
3
9, 362–380.
A. Kruse and N. Dahmen, J. Supercrit. Fluids, 2015,
6, 36–45.
W. Weiqi and W. Shubin, Chem. Eng. J., 2017, 307,
89–398.
C. Chang, X. Ma and P. Cen, Chinese J. Chem. Eng.,
006, 14, 708–712.
L. Reyes, I. NicolásꢀVázquez, N. MoraꢀDiez and J. R.
AlvarezꢀIdaboy, J. Org. Chem., 2013, 78, 2327–2335.
L. Shuai and J. Luterbacher, ChemSusChem, 2016, 9,
133–155.
1
1
1
1
3
4
5
6
9
Conclusion
Within the frame of this work we pointed out that the
effect of Brønsted acids on the hydrothermal dehydration
of fructose to HMF exceeds the sole donation of
catalytically active protons. In fact, there are additional
effects caused by the type of acid, which can be
3
2
identified as a change in the overall reaction velocity or 17
mechanism and that become the more relevant, the
1
1
8
9
S. J. Angyal, Carbohydr. Res., 1994, 263, 1–11.
H. Kimura, M. Nakahara and N. Matubayasi, J. Phys.
Chem. A, 2013, 117, 2102–2113.
M. J. Antal, W. S. L. Mok and G. N. Richards,
Carbohydr. Res., 1990, 199, 91–109.
R. S. Assary, T. Kim, J. J. Low, J. Greeley and L. A.
Curtiss, Phys. Chem. Chem. Phys., 2012, 14, 16603.
G. Yang, E. A. Pidko and E. J. M. Hensen, J. Catal.,
2012, 295, 122–132.
higher the acid concentration is. Interestingly, the type of
acid used has however only a low impact on the maximal
HMF yield. Future investigations will target a better
understanding of the substrateꢀcatalyst interaction.
2
2
2
0
1
2
Acknowledgement
This work by Paul Körner was supported by a State
Graduate Scholarship and is part of the bioeconomy 23
graduate program BBW ForWerts.
S. Caratzoulas and D. G. Vlachos, Carbohydr. Res.,
2011, 346, 664–672.
2
2
2
2
2
2
3
4
5
6
7
8
9
0
H. D. Phan, T. Yokoyama and Y. Matsumoto, Org.
Biomol. Chem., 2012, 10, 7382.
H. D. Phan, T. Yokoyama and Y. Matsumoto, J.
Carbohydr. Chem., 2013, 32, 223–239.
K. Nagasawa, Y. Tohira, Y. Inoue and N. Tanoura,
Carbohydr. Res., 1971, 18, 95–102.
M. Vincendon, Carbohydr. Polym., 1997, 32, 233–
There are no conflicts to declare.
References
1
R.ꢀJ. van Putten, J. C. van der Waal, E. de Jong, C. B.
Rasrendra, H. J. Heeres and J. G. de Vries, Chem.
Rev., 2013, 113, 1499–1597.
2
37.
T. Wu and S. Zivanovic, Carbohydr. Polym., 2008,
3, 248–253.
7
2
T. D. Swift, C. Bagia, V. Choudhary, G. Peklaris, V.
Nikolakis and D. G. Vlachos, ACS Catal., 2014, 4,
J. B. Binder, A. V. Cefali, J. J. Blank and R. T.
Raines, Energy Environ. Sci., 2010, 3, 765.
R. J. Van Putten, J. N. M. Soetedjo, E. A. Pidko, J. C.
Van Der Waal, E. J. M. Hensen, E. De Jong and H. J.
Heeres, ChemSusChem, 2013, 6, 1681–1687.
B. F. M. Kuster and H. M. G. Temmink, Carbohydr.
Res., 1977, 54, 185–191.
Y. Li, X. Lu, L. Yuan and X. Liu, Biomass and
Bioenergy, 2009, 33, 1182–1187.
B. F. M. Kuster and H. S. van der Baan, Carbohydr.
Res., 1977, 54, 165–176.
2
59–267.
F. S. Asghari and H. Yoshida, Ind. Eng. Chem. Res.,
006, 45, 2163–2173.
V. E. Tarabanko, M. Y. U. Chernyak, I. V
Nepomnyashchiy and M. A. Smirnova, Chem.
Sustain. Dev., 2006, 14, 49–53.
3
4
2
31
32
33
34
5
6
7
8
B. A. Fachri, R. M. Abdilla, C. B. Rasrendra and H. J.
Heeres, Chem. Eng. Res. Des., 2016, 109, 65–75.
J. B. Binder and R. T. Raines, J. Am. Chem. Soc.,
2009, 131, 1979–1985.
H. C. Helgeson, J. Phys. Chem., 1967, 71, 3121–
J. Liu, Y. Tang, K. Wu, C. Bi and Q. Cui, Carbohydr.
Res., 2012, 350, 20–24.
3136.
T. Ståhlberg, S. RodriguezꢀRodriguez, P. Fristrup and
1
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins