The role of metal halides in enhancing the dehydration of xylose to furfural

Article abstract of DOI:10.1002/cctc.201402842

The dehydration of xylose yields furfural, a product of considerable value as both a commodity chemical and a platform for producing a variety of fuels. When xylose is dehydrated in aqueous solution in the presence of a Bronsted acid catalyst, humins are formed via complex side processes that ultimately result in a loss in the yield of furfural. Such degradative processes can be minimized via the insitu extraction of furfural into an organic solvent. The partitioning of furfural from water into a given extracting solvent can be enhanced by the addition of salt to the aqueous phase, a process that increases the thermodynamic activity of furfural in water. Although the thermodynamics of using salts to improve liquid-liquid extraction are well studied, their impact on the kinetics of xylose dehydration catalyzed by a Bronsted acid are not. The aim of the present study was to understand how metal halide salts affect the mechanism and kinetics of xylose dehydration in aqueous solution. We found that the rate of xylose consumption is affected by both the nature of the salt cation and anion, increasing in the order no salt+++ and no salt---. Furfural selectivity increases similarly with respect to metal cations, but in the order no salt--- for halide anions. Multinuclear NMR was used to identify the interactions of cations and anions with xylose and to develop a model for explaining xylose-metal halide and water-metal halide interactions. The results of these experiments coupled with ¹⁸O-labeling experiments indicate that xylose dehydration is initiated by protonation at the C1OH and C2OH sites, with halide anions acting to stabilize critical intermediates. The means by which metal halides affect the formation of humins was also investigated, and the role of cations and anions in affecting the selectivity to humins is discussed. Get your kicks from kinetics: The effect of metal halides on the mechanism and kinetics of xylose dehydration in aqueous solution have been investigated. We found that both the rate of xylose consumption and furfural selectivity are affected by the nature of the salt cation and anion pairing.

Full text of DOI:10.1002/cctc.201402842

DOI: 10.1002/cctc.201402842
Full Papers
The Role of Metal Halides in Enhancing the Dehydration of
Xylose to Furfural
Kristopher R. Enslow[a, b] and Alexis T. Bell*[a, b]
The dehydration of xylose yields furfural, a product of consid-
erable value as both a commodity chemical and a platform for
producing a variety of fuels. When xylose is dehydrated in
aqueous solution in the presence of a Brønsted acid catalyst,
humins are formed via complex side processes that ultimately
result in a loss in the yield of furfural. Such degradative pro-
cesses can be minimized via the in situ extraction of furfural
into an organic solvent. The partitioning of furfural from water
into a given extracting solvent can be enhanced by the addi-
tion of salt to the aqueous phase, a process that increases the
thermodynamic activity of furfural in water. Although the ther-
modynamics of using salts to improve liquid–liquid extraction
are well studied, their impact on the kinetics of xylose dehy-
dration catalyzed by a Brønsted acid are not. The aim of the
present study was to understand how metal halide salts affect
the mechanism and kinetics of xylose dehydration in aqueous
solution. We found that the rate of xylose consumption is af-
fected by both the nature of the salt cation and anion, increas-
ing in the order no salt<K+ <Na+ <Li+ and no salt<Clꢀ <
Brꢀ <Iꢀ. Furfural selectivity increases similarly with respect to
metal cations, but in the order no salt<Iꢀ <Brꢀ <Clꢀ for halide
anions. Multinuclear NMR was used to identify the interactions
of cations and anions with xylose and to develop a model for
explaining xylose-metal halide and water-metal halide interac-
tions. The results of these experiments coupled with 18O-label-
ing experiments indicate that xylose dehydration is initiated by
protonation at the C1OH and C2OH sites, with halide anions
acting to stabilize critical intermediates. The means by which
metal halides affect the formation of humins was also investi-
gated, and the role of cations and anions in affecting the selec-
tivity to humins is discussed.
Introduction
Furfural is a high value chemical that can be used to produce
a wide range of commodity chemicals[1,2] and fuels.[3–6] At pres-
ent, the industrial production of furfural is accomplished
through the Brønsted acid-catalyzed hydrolysis of agricultural
waste, a process that has a product selectivity of less than
60%.[2,7] Furfural can also be sourced from any biomass materi-
al containing hemicellulose, a biopolymer which can be hydro-
lyzed to its principle component, xylose.[8] The subsequent de-
hydration of xylose produces furfural, however, this reaction is
accompanied by complex side reactions leading to humins, in-
tractable condensation products that can greatly reduce furfu-
ral selectivity. One approach for minimizing the loss of furfural
is in situ extraction. Recent studies have shown that furfural se-
lectivities in excess of 85% can be achieved by extracting the
furfural formed by Brønsted acid-catalyzed dehydration of
xylose using a biphasic system consisting of water and 4-meth-
ylpentan-2-one (MIBK)[9] or water/DMSO (1:1 v/v) and MIBK/2-
butanol (7:3 v/v).[10] The increased furfural selectivity is attribut-
ed to the removal of furfural from the aqueous phase, thereby
minimizing the possibility for furfural to undergo self-resinifica-
tion or react with xylose to form humins. It has also been re-
ported that the addition of salt, such as metal halides, to the
aqueous phase enhances the extraction of furfural from the
aqueous phase into the organic phase as a result of the salt
lowering the activity of water, consequently increasing the ac-
tivity coefficient of the furfural.[11]
A question seldom addressed in previous studies is how dis-
solved salt affects the rate of xylose dehydration and/or the
rate of humins formation in an aqueous phase. To the best of
our knowledge, this subject has received very little attention.
What is known is that dissolved metal halides enhance the
yield of furfural obtained through the aqueous phase dehydra-
tion of xylose.[12] Notably, this work was performed with low
substrate concentrations and high temperature (35 mm xylose,
2008C), conditions for which entropic effects are expected to
diminish condensation reactions that lead to the formation of
humins.[2] The only explanation given for the observed effect
of salt addition is that it enhances the formation of 1,2-enediol,
an intermediate in the mechanism of xylose dehydration.[12]
Thus, the effects of dissolved salts on the rates of xylose dehy-
dration and humins formation are little understood.
[a] K. R. Enslow, Dr. A. T. Bell
Energy Biosciences Institute
Energy Biosciences Building, 2151 Berkeley Way
Berkeley, CA 94704-1011 (USA)
E-mail: alexbell@berkeley.edu
[b] K. R. Enslow, Dr. A. T. Bell
Department of Chemical and Biomolecular Engineering
University of California
Berkeley, CA 94720-1462 (USA)
The present study was undertaken with the aim of develop-
ing a more thorough understanding of the role of metal hal-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201402842.
ChemCatChem 0000, 00, 0 – 0
1
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ides in the dehydration of xylose to furfural in aqueous solu-
tion. Multinuclear NMR spectroscopy was used to probe both
the interactions of metal cations and halide anions with water
and xylose. These studies were complemented by experiments
utilizing 13C labeling of the xylose anomeric carbon to probe
isomeric distributions and 18O labeling of the xylose ring
oxygen to identify reaction pathway populations. The results
of these investigations have led to a mechanistic understand-
ing of the means by which metal halides alter the kinetics of
Brønsted acid-catalyzed dehydration of xylose in aqueous solu-
tion as well as the influence metal halides have on the forma-
tion of humins.
a Brønsted acid. This hypothesis was confirmed by observing
the reaction of an aqueous solution of furfural, containing
375 mm of furfural and 200 mm HCl, at 1408C. After 2 h of re-
action, the concentration of furfural decreased by 24% due to
resinification (self-coupling), a process reported to occur readi-
ly at temperatures below 2008C.[2] Furfural can also undergo
condensation with anhydro-xylose intermediates via dioxolane-
like bridging structures. Consistent with this proposal, it was
observed (see Figure 2) that the addition of furfural at the start
Results and Discussions
Xylose dehydration in water
Shown in Figure 1 is the temporal evolution of products
formed during HCl-catalyzed dehydration of xylose in water at
1408C. Over the course of the reaction, furfural is observed as
the primary identifiable product. The difference between the
Figure 2. Dehydration of xylose (750 mm) at 1408C catalyzed by HCl
(200 mm) in water with and without an initial concentration of furfural
([F]0 =400 mm). Furfural yields shown only represent furfural produced di-
rectly from xylose, such that in the experiment starting with xylose and fur-
fural, furfural yield=[total furfural]ꢀ[F]0.
of xylose dehydration reduces the furfural produced from
xylose (from 26 mol% to 9 mol%). However, direct condensa-
tion of furfural with xylose is unlikely, since the rate at which
xylose is consumed is unchanged by the initial furfural concen-
tration. These two observations indicate that a reaction occurs
between furfural and an intermediate along the pathway from
xylose to furfural. Scheme 1 shows a proposed pathway for the
dehydration of xylose to furfural and for the loss of xylose and
furfural via condensation reactions. The mechanisms by which
the latter two processes might occur are given in Scheme 2
and 3.
Figure 1. Dehydration of xylose (750 mm) at 1408C catalyzed by HCl
(200 mm) in water. Humins as presented represent the molar sum of soluble
and insoluble degradation products.
consumption of xylose and the formation of furfural is attribut-
ed to the formation of humins—a combination of soluble (but
not identifiable) products contributing to the discoloration of
the reacting solution and a dark brown-to-black precipitate.
During 600 min of reaction 92% of the xylose is consumed.
The furfural yield increases initially, reaches a maximum of
30% after 200 min, and then slowly declines. By contrast, the
yield of humins increases monotonically, and after 600 min
constitutes the principal product of xylose dehydration.
At early reaction times (<10 min), xylose is consumed at
a rate that exceeds the rate of furfural formation, suggesting
the loss of xylose due to self-condensation or reaction with
anhydro-xylose intermediates in addition to its dehydration to
furfural. The decline in furfural yield for reaction times longer
than that needed to reach the maximum yield suggests that
furfural undergoes secondary reaction in the presence of
Scheme 1. A representative reaction pathway for xylose dehydration and
degradation catalyzed by a Brønsted acid.
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
2
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 2. A representative reaction pathway for furfural resinification via the coupling of two furfural molecules.
Scheme 3. A representative reaction pathway for furfural condensation with a xylose intermediate (mono-dehydrated at the C2 position).[2]
The kinetics of xylose consumption and furfural and humins
formation can be described by the following set of ordinary
differential equations [Eqs. (1)–(4)], which are based on
Scheme 1:
Xylose dehydration in the presence of metal halides
The effects of metal halides on the Brønsted acid-catalyzed de-
hydration of xylose in an aqueous solution were investigated
using the chloride, bromide, and iodide salts of the monova-
lent metals lithium, sodium, and potassium. Each of these ex-
periments was performed at 1408C in a 5m metal halide solu-
tion containing 50 mm HCl. The results of these experiments
are presented in Figures 3–5.
d½Xꢁ
¼ ꢀðk1 þ k4½IꢁÞ½Xꢁ½Hþꢁ
ð1Þ
ð2Þ
ð3Þ
ð4Þ
dt
d½Iꢁ
¼ ðk1½Xꢁ ꢀ k2½Iꢁ ꢀ k3½Iꢁ½Fꢁ ꢀ k4½Iꢁ½XꢁÞ½Hþꢁ
¼ ðk2½Iꢁ ꢀ k3½Iꢁ½Fꢁ ꢀ k5½FꢁÞ½Hþꢁ
dt
Analysis of Figures 3A–5A reveals that xylose consumption
occurs much more rapidly in brine than in water, the rate de-
pending on the nature of both the cation and anion. For
a given anion, the rate of xylose consumption increases in the
following order: no salt<K+ <Na+ <Li+. The initial rate of fur-
fural formation increases in the same order, as can be seen
from the data presented in Figures 3A–5B. In each case, the
yield of furfural reaches a maximum and then declines with
further reaction time, the time at which the maximum occurs
decreasing as initial rate of furfural formation increases.
Initial rate data for xylose conversion and furfural formation
is summarized in Table 1 for all metal halide solutions. The ini-
tial rates of both xylose consumption and furfural formation in-
crease in the order no salt<Clꢀ <Brꢀ <Iꢀ. However, it is evi-
dent from Figures 3–5 that for longer reaction times the orders
of Brꢀ and Iꢀ are reversed. This change in order is attributable
to the slow oxidation of iodide anions to molecular iodine in
aqueous systems at elevated temperatures via a sequence of
reactions that sum to the following stoichiometric reaction
[Eq. (5)]:[13]
d½Fꢁ
dt
d½Hꢁ
¼ ðk3½Fꢁ½Iꢁ þ k4½Xꢁ½Iꢁ þ k5½FꢁÞ½Hþꢁ
dt
Here [X], [I], [F], and [H] are the concentrations of xylose, in-
termediate product, furfural, and humins, respectively. As the
intermediates involved in the conversion of xylose-to-furfural
were not observed, we can invoke the pseudo-steady-state as-
sumption and set d[I]/dtꢂ0. Values for k1, k2, k3, k4, and k5
were determined by least squares minimization of the residuals
between the predicted concentrations of X, F, and H obtained
by solving Equations (1)–(4), and the experimental data shown
in Figure 1. The best-fit values obtained in this manner are k1 =
2.9(ꢃ0.6)ꢁ10ꢀ4 Lmolꢀ1 sꢀ1, k2 =7.5(ꢃ0.4)ꢁ10ꢀ4 Lmolꢀ1 sꢀ1, k3 =
2.15(ꢃ0.08)ꢁ10ꢀ3 L2 molꢀ2 sꢀ1, k4 =7.5(ꢃ0.6)ꢁ10ꢀ4 L2 molꢀ2 sꢀ1,
and k5 =1.0(ꢃ0.1)ꢁ10ꢀ4 Lmolꢀ1 sꢀ1. These results demonstrate
that the rate coefficient for the consumption of intermediates
produced from xylose is over twice as large as that for their
production, and that humins are formed through condensation
of furfural and intermediates involved in xylose dehydration at
a rate that is more than twice as fast as the rate of condensa-
tion between xylose and the intermediates involved in xylose
dehydration.
4 Iꢀ þ O2 þ 4 Hþ ! 2 I2 þ 2 H2O
ð5Þ
The occurrence of this reaction leads to a loss of both iodide
anions and protons from solution and the formation of iodine.
Evidence for this process was observed as a darkening of the
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
3
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 3. A) dehydration of xylose (750 mm) at 1408C catalyzed by HCl
(50 mm) in various 5m metal chloride (aq) solutions. B) Resultant furfural for-
mation.
Figure 4. A) dehydration of xylose (750 mm) at 1408C catalyzed by HCl
(50 mm) in various 5m metal bromide (aq) solutions. B) Resultant furfural for-
mation.
Table 1. Initial rate of xylose (750 mm) dehydration and furfural forma-
tion at 1408C catalyzed by HCl (50 mm) in various 5m metal halide (aq)
solutions.[a]
greater roles in degradation reactions than do chloride anions.
Notably, these selectivities are lower than those observed in
Figures 3–5 due to an induction period that exists in the for-
mation of furfural at short reaction times.
Salt
r0, Xylose
[mmollꢀ1 sꢀ1
r0, Furfural
[mmollꢀ1 sꢀ1
Furfural
]
]
selectivity [%]
None
LiCl
LiBr
LiI
NaCl
NaBr
NaI
KCl
KBr
KI
19.1
156.3
277.7
356.4
91.3
118.7
148.7
89.6
5.8
69.3
75.1
86.8
38.6
41.1
50.8
35.0
41.2
48.8
30.4
44.3
27.0
24.4
42.3
34.6
34.2
39.1
38.3
38.6
Effects of metal halides on the activity of water
The data presented in Figure 3–5 and Table 1 clearly indicate
that the presence of metal halides strongly affects rate of
xylose dehydration, as well as the formation of furfural and
humins. To understand how halide salts affect these processes,
we examined the effect of salt composition on the thermody-
namic activity of water and the manner in which salt cations
and anions interact with dissolved xylose.
107.6
126.3
[a] Includes furfural selectivity data for each salt. Data taken over the
course of the first 5 min of each reaction.
The activity of water was determined for 5m metal halide
solutions using data taken from the literature.[14] The results,
presented in Table 2, show that metal halides reduce the activi-
ty of water in the order no salt>Clꢀ >Brꢀ >Iꢀ for a given
cation and no salt>K+ >Na+ >Li+ for a given anion. Compari-
son of these trends with the observed trend in the rate of
xylose consumption reveals that the rate of xylose dehydration
increases with decreasing water activity. This correlation can
be attributed to the ability of metal halides cations and anions
reaction solution and a reduction in the rate of xylose dehydra-
tion. Further details of this process are discussed in the Sup-
porting Information (Figure S2).
Table 1 also lists the selectivity toward furfural for each salt.
In general, the furfural selectivity increases in the order Iꢀ ꢂ
Brꢀ <Clꢀ, indicating that bromide and iodide anions may play
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
4
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Na+, 5 for Li+, and to a lesser degree 1 for Clꢀ/Brꢀ and 0 for
Iꢀ).[16] In the presence of strong kosmotropic species, water
molecules are less able to interact with other solutes, and
hence the activity coefficient of these solutes (such as xylose)
increases, enabling them to interact more effectively with
metal halide ions and protons. The latter effect is thought to
enhance the extent of xylose protonation, as discussed in
more detail below.
Effects of metal halides on the distribution of xylose con-
formers
Xylose in its crystalline form is present almost exclusively as a-
xylopyranose; however upon dissolution in water it isomerizes
to an equilibrium distribution of the five conformers shown in
Scheme 4.[17] The distribution of xylose conformers was deter-
Figure 5. A) dehydration of xylose (750 mm) at 1408C catalyzed by HCl
(50 mm) in various 5m metal iodide (aq) solutions. B) Resultant furfural for-
mation.
Table 2. Molalities, salt activity coefficients, osmotic coefficients, and
water activities of various 5m metal halide (aq) solutions at 298 K.[13]
Scheme 4. The five isomers of xylose found after dissolving crystalline xylose
in water.
Salt
Salt molality
Salt activity
coefficient
Osmotic
coefficient
Water
activity
[kgmolꢀ1
]
None
LiCl
LiBr
LiI
NaCl
NaBr
NaI
KCl
KBr
KI
–
–
1
1
mined by 13C NMR (see Figure S3 in the Supporting Informa-
tion for peak assignments). As shown in Table 3, in the absence
of acid and salt, only four of the five isomers are evident, the
principal forms being b-xylopyranose (63.8%) and a-xylopyra-
nose (35.6%). In 5m salt solutions the ratio a-xylopyranose/b-
xylopyranose increases slightly in the order K+ <Na+ <Li+ for
a given halide anion and in the order Clꢀ <Brꢀ <Iꢀ for a given
cation. Although lithium and sodium halide solutions increase
the a-xylopyranose/b-xylopyranose ratio as compared to the
case where no salt is added, potassium halide solutions pro-
mote marginally lower a-xylopyranose/b-xylopyranose ratios.
Overall, the observed change in the ratio of a-pyranose/b-pyra-
nose isomers of xylose in the presence of metal halides is
smaller than that reported for isomers of ribose and galac-
tose.[18,19] The larger effect of salts on these latter sugars is at-
tributable to the presence of pyranose conformers containing
axial-equatorial-axial (ax.-eq.-ax.) orientations of hydroxyl
groups, for which the tri-oxygen core strongly attracts cations.
5.593
6.467
7.888
5.595
5.800
6.198
5.927
6.173
6.635
2.404
4.785
7.001
0.939
1.224
1.804
0.781
0.652
0.734
1.748
2.093
3.132
1.238
1.368
1.854
1.030
1.048
1.097
0.703
0.614
0.411
0.779
0.751
0.661
0.803
0.792
0.769
to interact preferentially with water molecules, thereby reduc-
ing the interactions of water molecules with themselves or
with other solute species. Such an interaction interrupts the
hydrogen bond network within water, promoting a more
stable restructuring of water around the metal halides.[15,16]
This “kosmotropic” effect increases with ionic radius and/or
charge, and can be estimated by the number of water mole-
cules that hydrate each free ion (for example, 4 for K+ and
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
5
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Table 3. Xylose isomer distribution (%) of 1-13C-xylose (750 mm) in D2O
(no salt), D2O with 50 mm HCl (no salt w/50 mm HCl), and various 5m
metal halide (aq–D2O) solutions as determined by measuring the 13C NMR
signals for the C1 carbons of individual xylose isomers.[a]
Xylopyranose
Xylofuranose
Acyclic xylose
Salt
b-
a-
b-
a-
None
None w/50 mM HCl
63.8
64.4
63.1
60.3
57.2
63.0
62.6
62.0
67.5
66.9
66.8
34.2
35.6
36.9
39.7
42.8
37.0
37.4
38.0
32.5
33.1
33.2
1.2
–
–
–
–
–
–
–
–
0.8
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
LiCl
LiBr
LiI
NaCl
NaBr
NaI
KCl
KBr
KI
–
–
–
–
[a] Dashes in place of numbers indicate that the isomer was not identified
in solution. Data was recorded at room temperature after allowing 24 h
to achieve an equilibrium distribution.
By contrast, the hydroxyl groups in b-xylopyranose are in eq.-
eq.-eq. orientation and those in a-xylopyranose are in ax.-eq.-
eq. orientation, and hence are less able to interact with cations.
The distribution of xylose conformers reported in Table 3 was
determined at room temperature. Previous studies have shown
that the ratio of a-xylopyranose/b-xylopyranose in aqueous
solutions increases with increasing temperature and, therefore,
it is expected that this ratio will be higher at reaction tempera-
ture (1408C) than at room temperature.[17]
Figure 6. Dependence of chemical shift (Dd) for each carbon of the a-xylo-
pyranose A) and b-xylopyranose B) forms on halide anion for the sodium
salts as determined by 13C NMR analysis of 750 mm xylose in D2O at 298 K in
the presence of varying 5m sodium halides.
Evidence for metal halide cation and anion interactions with
xylose
1
13C and H NMR were used to observe the interactions of metal
ciate with the positive end of the CꢀOꢀH dipole. The location
of these anions is presumed to be close to the nearest hydrox-
yl proton given the steric barrier of such bulky ions nestling
between the attached carbon and proton of the hydroxyl
group. The cation–oxygen interaction withdraws electron den-
sity from carbon atoms of xylose, whereas the anion-proton in-
teraction provides electron density. Given the charge density
disparity between the cations and anions, the influence of the
cations on shielding of carbon atoms is expected to dominate
over the influence of anions, resulting in positive changes in
carbon chemical shifts due to electron deshielding. However,
as the anion associated with a given cation becomes larger,
the anion becomes increasingly less effective in shielding the
CꢀOꢀH group and the net deshielding effect of the cation–
anion interaction with the sugar increases. This is precisely
what is generally observed in Table S1 (Supporting Informa-
tion)—as the charge density of the anion decreases, the mag-
nitude of DdCn increases, such that the change in chemical
shift increases with respect to anion in the following manner:
Clꢀ <Brꢀ <Iꢀ.
cations and halide anions with molecules of xylose dissolved in
water. Evidence for these interactions was given by shifts in
the positions of the 13C NMR peaks or shifts in the 1H NMR
peaks for protons bonded directly to carbon atoms (proton
chemical shifts for COH groups were not observed, owing to
deuterium exchange with the D2O solvent used to preserve
the chemical nature of solution interactions). The observed
changes in chemical shift (DdCn and DdCnH) are presented in
Tables S1 and S2 (in the Supporting Information). An example
of the 13C shifts observed for sodium halides is shown in Fig-
ure 6A and 6B for a- and b-xylopyranose, respectively. With
a few exceptions DdCn is positive for all carbons of both xylose
conformers regardless of the metal halide composition, indicat-
ing the occurrence of electron deshielding at the carbon
atoms. Each carbon is covalently bound to an oxygen atom
that can interact with the metal cations in solution. These in-
teractions withdraw electron density from the oxygen and in
turn from the attached carbons, resulting in electron deshield-
ing of the carbon nucleus. The metal cations interacting with
the oxygen atoms of CꢀOꢀH groups are associated with the
negative end of the dipole moment presented by these
groups, whereas charge-compensating halide anions can asso-
The interactions of metal cations with the ring oxygen atom
have a qualitatively different effect from that observed for in-
teractions of metal cation with CꢀOꢀH groups. While the
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
6
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
direct interaction between cation and ring oxygen is observed,
as evidenced by positive values of DdC5 in all cases, the C5-
O5-C1 dipole is aligned such that the positive end extends
within the ring itself, prohibiting interaction with large halide
anions. Consequently, the anion interacts with the cation on
the periphery of the xylose molecule and consequently elec-
tron shielding of the C5 carbon atoms does not occur, and
only the effects of the cation interaction with the ring oxygen
are observed. This explains why the DdCn is greater at the C5
position than at any other position. Furthermore, DdC5 increas-
es with anion polarizability (Clꢀ <Brꢀ <Iꢀ), owing to the greater
charge separation between a given cation and a more polariza-
ble anion that affords a stronger dipole-ion interaction be-
tween the cation and ring oxygen. This effect, coupled with
resonance effects from the nearby C4 carbon, can explain the
trend in increasing values of DdC5 with respect to anion size.
Notably, for a given metal halide, DdCn is always greater for
a-xylopyranose than for b-xylopyranose at the C1 position,
whereas the opposite is true at the C2 position. This trend is
indicative of a stronger salt–sugar interaction at the C1OH
group when xylose is present in the alpha form, whereas the
interaction at the C2OH is strongest when xylose is present in
the beta form. These differences have implications for the reac-
tion mechanism, as discussed below.
number of Clꢀ anions interacting with xylose using the follow-
ing correlation [Eq. (6)]:[20]
½xyloseꢁ% MWNaCl
Ddobs ¼ Ddfree þ ðDdbound ꢀ DdfreeÞN
ð6Þ
MWxylose
½NaClꢁ%
For which: Ddobs =observed peak width of chloride ion,
Ddfree =peak width of chloride unbound to xylose, Ddbound
=
peak width of chloride bound to xylose, N=ratio of chloride
ions to xylose molecules, [NaCl]% =weight% of NaCl in solu-
tion, [xylose]% =weight% of xylose in solution, MWNaCl =molec-
ular weight of NaCl, 58.44 gmolꢀ1, and MWxylose =molecular
weight of xylose, 150.13 gmolꢀ1.
By fitting Equation (6) to a plot of the measured values of
the chloride ion peak width versus the weight percent xylose
it was possible to determine a value of N. A value of Nꢂ3.8 is
determined, suggesting that every molecule of xylose interacts
with ꢂ4 chloride anions or that each hydroxyl group interacts
with one chloride anion.
18O Isotopic labeling experiments
The NMR results presented in Table 2 show that the presence
of lithium and sodium halides can cause a small increase in the
ratio of a-xylopyranose relative to b-xylopyranose, and as
noted earlier this ratio is expected to increase further when
the temperature is raised from room temperature to the reac-
tion temperature (1408C). However, these changes do not
appear to be sufficiently large to account for the significant ef-
fects of metal halides on the rates of xylose consumption and
furfural formation presented in Figures 3–5. Therefore, it is
more likely that these effects are attributable to specific inter-
actions of metal cations and halide anions with molecules of
xylose. Of particular note in this connection is the observation
that cation and anion interactions are stronger with a-xylopyr-
anose than with b-xylopyranose at the C1 position, whereas
the opposite is true at the C2 position.
The trends in DdCnH shown in Table S2 (Supporting Informa-
tion) follows the patterns observed for DdCn. For all protons,
DdCnH is positive and increases with respect to anion for
a given cation in the following order: Clꢀ <Brꢀ <Iꢀ.
The interactions of chloride anions with xylose were probed
by acquiring 35Cl NMR spectra as a function of xylose concen-
tration. Figure 7 shows that as the concentration of xylose in-
creases, the width and the height of the chloride ion peak in-
crease. These data can be used to determine the average
To further probe how metal halides affect the dehydration
of a-xylopyranose and b-xylopyranose, experiments were per-
formed with xylose labeled with 18O-labeled at the ring oxygen
position (d-[5-18O]xylose]) both in the presence and absence of
NaCl. At the end of reaction, furfural was extracted from the
aqueous solution using toluene and then analyzed by GCMS
to determine the location of 18O in the furfural. The relative
population of peaks at m/z=69 versus 67 were used to deter-
mine whether 18O in the xylose ring remained within the furan
ring or transferred to the carbonyl group, respectively. In the
absence of metal halides, 69% of the furfural produced con-
tained 18O within the furan ring and the remaining 31% con-
tained 18O in the carbonyl group. When the reaction was per-
formed in 5m NaCl, 52% of the furfural produced contained
18O in the furan ring and 48% contained 18O in the carbonyl
group. The results of these experiments suggest that the inter-
actions of Na+ and Clꢀ with xylose influence the mechanism
and kinetics of xylose dehydration. The interpretation of these
results is discussed in the next section.
Figure 7. 1D 35Cl NMR of the Clꢀ peak of 5m NaCl as a function of varying
xylose concentration from 0–10 wt% in D2O at 298 K.
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
7
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Mechanism of xylose dehydra-
tion
Aqueous-phase dehydration of
xylose to furfural catalyzed by
a Brønsted acid can occur in one
of two ways, initiated either by
ring-opening to form the acyclic
isomer of xylose followed by de-
hydration or by direct dehydra-
tion of either a- or b-xylopyra-
nose. Experimental results have
suggested that the Brønsted
acid-catalyzed dehydration of
xylose proceeds via the acyclic
pathway,[21] however, theoretical
studies of xylose dehydration
suggest that pyranose ring
opening followed by dehydra-
tion of the resulting aldose is
less favorable energetically than
direct dehydration of the pyra-
nose ring followed by intramo-
lecular rearrangement.[22] This
study also shows that dehydra-
tion initiated at the C2OH posi-
tion of xylopyranose is favored
over that at the C1OH position,
and that dehydration initiated at
either C3OH or C4OH leads to
fragmentation products rather
than furfural. Hence, under the
conditions investigated in this
work, Brønsted acid-catalyzed
dehydration of xylose in water
with or without metal halides is
presumed to proceed via either
C1OH or C2OH dehydration,
since fragmentation products
were not observed.
Scheme 5. The involvement of halide anions in stabilizing the carbocation formed during the initial dehydration
of xylose at the C2 position.
Scheme 6. The involvement of halide anions in stabilizing the carbocation formed during the initial dehydration
When xylose dehydration is
performed in the presence of
metal halides, the metal cations
of xylose at the C1 position.
can interact with the hydroxyl groups of the ring oxygen of
xylose thereby increasing the CꢀOH and C-OꢀC bond lengths
and lowering the energy necessary to break such CꢀO bonds.
The use of a metal cation with strong kosmotropic character
more effectively interrupts the xylose solvation shell, sterically
hindering rehydration of the carbocation formed after dehy-
dration. Moreover, halide anions can interact with the electro-
negative regions of xylose and stabilize any carbocations
formed. The proposed influence of halide anions on Brønsted
acid dehydration of xylose initiated at the C2OH and C1OH po-
sitions is presented in Schemes 5 and 6, respectively.
lowed by the loss of water and the formation of a carbocation
at the C2 position, which has been established in previous
studies to be the rate-limiting step.[13] The carbocation is a reac-
tive intermediate that can undergo rearrangement to form 2,5-
anhydroxylose, which then further dehydrates to form furfural.
The presence of a halide anion near the C2 carbon of the car-
bocation intermediate can aid in its stabilization and reduce
the extent to which rehydration occurs at that position. As the
halide nucleophilicity increases, so does the ability of the
halide to stabilize the carbocations in protic solvents, i.e., Iꢀ >
Brꢀ >Clꢀ.[13] This trend parallels that seen for the effect of
halide composition on the rate of xylose dehydration (see Fig-
ures 3–5).
Scheme 5 shows the dehydration pathway proposed for
xylose dehydration initiated by reversible protonation at the
C2OH position of either a- or b-xylopyranose. This step is fol-
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
8
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 7. The dehydration of 2,5-anhydroxylose at the C3 position (top) and C4 position (bottom), respectively.
Scheme 6 illustrates the proposed pathway for the xylose
dehydration initiated by reversible protonation of the C1OH
position of a-xylopyranose. As mentioned previously, dehydra-
tion of the b-xylopyranose conformer at the C2 position is
more energetically favorable than at the C1 position, however,
the a-xylopyranose conformer presents an anomeric hydroxyl
group in the axial position. Hydroxyl groups in such positions
are more easily protonated and removed as water than their
equatorial counterparts. Therefore, both C1OH and C2OH initi-
ated dehydration pathways are possible for a-xylopyranose.
However, after the initial C1OH dehydration and the formation
of the oxocarbinium cation, the next step requires the oxygen
of the C2OH group to attack the C5 carbon. This step is ex-
pected to have a large energetic barrier because the equatorial
C2OH must rotate around the plane of the xylose molecule
and attack a carbon without formal charge. The addition of
metal halides reduces this energetic barrier by not only elon-
gating the C2-OH bond length allowing for increased mobility
(via the metal cation-OH interaction), but also by increasing
the partial positive charge on the carbon due to electron de-
shielding provided by the metal halide. Therefore, the salt–
sugar interaction facilitates the production of furfural through
the C1OH pathway by not only increasing the starting concen-
tration of a-xylopyranose, but also by decreasing the barrier
for ring reformation following the initial dehydration.
The results obtained from 18O labeling experiments are fully
consistent with the observed effects of metal halides on the
pathways for xylose dehydration. In contrast to published re-
sults utilizing xylose-1-14C that suggest the C1 carbon remains
nearly completely at the aldehydic carbon of furfural,[21] 18O la-
beling experiments indicate that the presence of metal halides
can influence the final location of the C1 carbon and ring
oxygen. Dehydration by means of protonation of the C1OH
group, which results in the 18O-labeled ring oxygen appearing
in the furfural carbonyl group, occurs more rapidly than dehy-
dration initiated through protonation of the C2OH group,
which results in the 18O-labeled ring oxygen remaining in the
ring oxygen of furfural, when metal halides are present.
(2,5-anhydroxylose intermediate), furfural is formed via elimina-
tion of the hydroxyl groups at the C3 and C4 positions, as
shown in Scheme 7. The ring structure is maintained through-
out these final dehydrations because oxolane is very stable[23]
and particularly resistant to attack by hydrolytic agents.[24]
The carbocationic intermediates produced during the dehy-
dration of xylose can participate in condensation processes
leading to humins. In the presence of metal halides the selec-
tivity to furfural at initial reaction times decreased with increas-
ing anion size, indicating that the ratio of the rates of conden-
sation to dehydration increases with respect to anion in the
order Clꢀ <Brꢀ <Iꢀ. As larger anions stabilize carbocation inter-
mediates, they form longer, lower energy bonds with the
carbon atoms that results in a less polar charge distribution
between anion and carbocation. This effect lowers the partial
positive charge on the carbon making it a weaker Brønsted
acid that is less capable of participating in elimination reac-
tions.[13] Together with the fact that larger anions are better
leaving groups improves the probability that substitution
occurs at the carbocation. Hence, although large anions pro-
mote the rates of xylose consumption and furfural production,
such anions may also enhance the rate of promote more nu-
cleophilic substitution and, consequently, lower selectivity to
furfural.
Condensation reactions leading to humins can occur via
a number of pathways involving either furfural or xylose and
an intermediate involved in the dehydration of xylose. At the
beginning of the Results and Discussion we showed that con-
densation involving furfural is preferred kinetically. Scheme 3 il-
lustrates a possible mechanism for this process. In this exam-
ple, xylose is substituted at the C2OH position by the nucleo-
philic carbonyl of furfural, followed by a nucleophilic attack on
the carbonyl carbon by the C1OH oxygen, creating a furfural-
xylose condensation product with a dioxolane bridge. The con-
densation of two furfural molecules (Scheme 2) can also be
aided by the presence of metal halides. Furan resonance struc-
tures forming oxocarbenium ions can be stabilized by halide
anions, allowing for the nucleophilic attack of a furfural C5
carbon on the carbonyl carbon of another furfural to form
a di-furanyl resinification product.
After primary dehydration at either hydroxyl group followed
by ring rearrangement to form the 5-membered oxolane ring
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
9
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Improving overall furfural yields
found to increase the dehydration of xylose initiated by proto-
nation of C1OH versus C2OH groups. These findings demon-
strate that alkali metal halides dissolved in water can be used
to enhance the rate of xylose dehydration and the selectivity
towards furfural versus humins formation. By coupling these
effects with the extraction of furfural into an organic phase,
high yields of furfural can be achieved.
To this point, we have only discussed the effects of salt addi-
tion on the rate of xylose dehydration in aqueous solution and
on the competing formation of humins. However, as noted in
the Introduction, the yield of furfural can be enhanced by
in situ extraction of furfural into a water-immiscible phase. For
this reason, we end our discussion of the effects of salt addi-
tion by briefly examining the influence of NaCl on the dehy-
dration of xylose performed in a two-phase water/toluene
system. We note that the partition coefficient of furfural be-
tween a toluene phase and an aqueous phase (1:1 by volume)
increases from 9 to 20 when 5m NaCl is added to the aqueous
phase (measured at 508C). Dehydrating xylose in such a bipha-
sic system also improves the selectivity to furfural from 54 to
72% when 5m NaCl (aq) was added to the aqueous phase (see
Figure S6 in the Supporting Information). Thus, the addition of
metal halides to a biphasic reaction system has the thermody-
namic benefit of improving the efficacy of the extracting sol-
vent relative to water
Experimental Section
Materials
d-xylose (99%, Sigma–Aldrich), d-[5-18O]-xylose (90 atom% 18O,
Omicron Biochemicals, Inc.), and 2-furaldehyde (furfural, 99%,
Sigma–Aldrich) were used as reagents and standards (for quantifi-
cation). d-xylose-1-13C (99 atom% 13C, Sigma–Aldrich) was used in
1H and 13C NMR studies.
Lithium chloride (LiCl, BioXtra, ꢄ99.0%–titration), lithium bromide
(LiBr, anhydrous, Redi-Dri, ReagentPlus, ꢄ99%), lithium iodide (LiI,
99.9%–trace metals basis), sodium bromide (NaBr, puriss., 99–
100.5%), sodium iodide (NaI, anhydrous, Redi-Dri, ACS reagent,
ꢄ99.5%), potassium chloride (KCl, approx. 99%), potassium bro-
mide (KBr, BioXtra, ꢄ99.0%), potassium iodide (KI, puriss.,
ꢄ99.5%), magnesium chloride (MgCl2, anhydrous, ꢄ98%), and cal-
cium chloride (CaCl2, anhydrous, Redi-Dri, ꢄ97%) were purchased
from Sigma–Aldrich; sodium chloride (NaCl, ꢄ99.0%) was pur-
chased from Fisher Scientific. Sulfuric acid (H2SO4, 98%, Sigma–Al-
drich) and hydrochloric acid (HCl, 37% v/v, Fisher Scientific) were
used as Brønsted acid catalysts. All materials were used as pur-
chased, without further purification or modification.
Notably, adjustments to reaction conditions can result in
similar enhancements in selectivity. For example, lowering the
metal halide concentration from 5m reduces the difference be-
tween rate of xylose conversion and that of furfural produc-
tion, resulting in increased selectivity (see Figure S1 of Sup-
porting Information). The furfural selectivity can also be in-
creased by increasing the temperature, which decreases the
extent of side reactions and thereby increases the furfural se-
lectivity, and at 2008C approaches 100%.[2]
Experimental approach
Conclusions
All experiments were performed in 10 mL glass vials (from Sigma–
Aldrich via supplier Supelco) sealed with 20 mm aluminum-crimp-
ed PTFE septa and heated using a silicon oil bath to maintain con-
stant reaction temperature and stirring rate. In a representative
xylose dehydration experiment, xylose was dissolved in 5m brine
(i.e. 5m NaCl in nanopure water), to which HCl was added to
create a solution (750 mm xylose, 50 mm H+). A 4 mL aliquot of
this solution was sealed into a 10 mL glass vial. The vial was then
placed in a silicone oil bath heated to 1408C and stirred at
600 rpm. Reaction time began after 1 min had transpired (control
experiments determined this time to be necessary for the reactor
to reach reaction temperature). Upon completion of the reaction,
the sample was removed and quenched in an ice bath. An internal
standard (1 mL of 75 mgmLꢀ1 1,6-hexanediol in water) was added
and the sample centrifuged to remove all water insoluble particu-
lates. A portion of the reaction mixture (500 mL) was diluted in
a 1:1 ratio with nanopure water and taken for HPLC analysis. For
reactions involving an additional organic phase, the aqueous
phase volume was reduced to 1 mL and toluene (4 mL) was added
prior to sealing the reaction vial. At reaction end, the organic
phase was separated via centrifuging. The aqueous phase was
treated as above and analyzed via HPLC, whereas a different inter-
nal standard (1 mL of 5 mgmLꢀ1 guaiacol in toluene) was added to
the organic phase prior to GC/MS analysis.
This study has shown that the addition of metal halides in-
creases the rate of xylose dehydration in an aqueous phase
catalyzed by a Brønsted acid. The enhancement in the dehy-
dration rate is a function of both cation and anion composition
and increases in the order no salt<K+ <Na+ <Li+ and no salt
<Clꢀ <Brꢀ <Iꢀ. The selectivity to furfural is also affected by
salt composition, increasing in the order K+ <Na+ <Li+ and
Iꢀ <Brꢀ <Clꢀ (as observed in initial rate studies). At 1408C, the
maximum furfural selectivity was 44% when the reaction was
performed in a 5m LiCl solution. The effects of salt on the
rates of xylose dehydration and humins formation are complex
and involve interactions of the salt cations and anions with
both water and xylose. Metal cations disrupt the solvation of
xylose by water and interact with the hydroxyl groups and
ring oxygen atoms of xylose, leading to a weakening of CꢀO
bonds. Halide anions interact to a lesser degree with water
than do cations, but do interact with the positive end of hy-
droxyl groups on xylose, such that on average 4Clꢀ ions inter-
act with each xylose molecule. These interactions stabilize cat-
ionic intermediates formed during the dehydration of the
sugar, the degree of stabilization increasing with the nucleo-
philicity of the anion. Significantly, more nucleophilic anions
also enhance reactions involving these intermediates to form
humins, and consequently furfural selectivity decreases in the
order Clꢀ <Brꢀ <Iꢀ. The addition of metal halides was also
The influence of metal halides on the dehydration of xylose and
the formation of furfural was performed using a 5m solution of the
metal halide. The choice of this concentration was based on pre-
liminary experiments (see Figure S1 in Supporting Information) per-
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
10
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
formed with NaCl solutions ranging in concentration from 0 to 5m.
The natural log of the initial reaction rate versus the salt concentra-
tion revealed three distinct regimes: 1) at low concentrations
(500 mm and under), rate of xylose dehydration was roughly zero-
order, 2) at intermediate concentrations (500 mm to 3.5m), the rate
was roughly first-order, and 3) at high concentrations (3.5m to
5m), the rate was roughly zero-order. To maximize initial xylose
conversion/furfural production rates, the highest salt concentration
that could be achieved for all salts with respect to room-tempera-
ture solubility limits was found to be 5m. Only salts containing Clꢀ,
Brꢀ, and Iꢀ were used. Fluoride (Fꢀ) salts were purposely excluded
in order to avoid a reduction in the pH of the reaction solution
due to the reaction of proton (from the hydrochloric acid Brønsted
acid catalyst, pKa =ꢀ8) with the more basic fluoride anions. Conse-
quently, only halides of equal or lesser basicity than Clꢀ were ex-
plored.
Keywords: anions · carbohydrates · cations · kinetics · salt
effect
[1] T. Werpy, G. Petersen, Report No. NREL/TP-510-35523, Top Value Added
Chemicals from Biomass: Vol. 1—Results of Screening for Potential Candi-
dates from Sugars and Synthesis Gas, National Renewable Energy Labo-
ratory, Golden, CO, 2004.
[2] K. J. Zeitsch, Sugar Ser. 2000, 13, 1–353.
[3] S. Bayan, E. Beati, Chim. Ind. 1941, 23, 432.
[4] a) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007,
46, 7164; Angew. Chem. 2007, 119, 7298; b) G. A. Tompsett, N. Li, G. W.
Huber, Thermochemical Processing of Biomass: Conversion into Fuels,
Chemicals and Power (Ed.: Robert C. Brown), Wiley, Chichester, 2011,
pp. 223–279.
[5] B. Madhesan, E. R. Sacia, A. T. Bell, ChemSusChem 2014, 7, 1078.
[6] a) G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Science 2003,
300, 2075; b) G. W. Huber, J. A. Dumesic, Catal. Today 2006, 111, 119.
[7] C. Moreau, R. Durand, D. Peyron, J. Duhamet, P. Rivalier, Ind. Crops Prod.
1998, 7, 95.
[8] a) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044; b) C. E.
Wyman, S. R. Decker, M. E. Himmel, J. W. Brady, C. E. Skopec, L. Viikari,
Polysaccharides: Structural Diversity and Functional Versatility, Vol. 2 (Ed.:
S. Dumitriu), Marcel Dekker, New York, 2005, pp. 995–1033.
[9] R. Weingarten, J. Cho, W. C. Conner, G. W. Huber, Green Chem. 2010, 12,
1423.
Product analysis
A Shimadzu HPLC equipped with a Phenomenex Rezex RFQ-Fast
Acid H+ column (100ꢁ7.8 mm; 0.01 N H2SO4; 1.0 mLminꢀ1; 558C)
and a refractive index detector (RID) was used to analyze all aque-
ous samples. Product quantities were determined by converting in-
tegrated HPLC peak areas into concentrations using a 7-point cali-
bration curve generated from purchased standards.
[10] J. N. Chheda, Y. Roman-Leshkov, J. A. Dumesic, Green Chem. 2007, 9,
342.
A Varian CP-3800 Gas Chromatograph equipped with a FactorFour
Capillary Column (UF-5 ms 30 m, 0.25 mm, 0.25 mm, P/N CP8944)
connected to a Varian quadrupole-mass spectrometer (MS) and
flame ionization detector (FID) was used to analyze all organic
phase samples. After product identification by mass spectrometer,
product concentrations were determined from integrated FID peak
areas using a 6-point calibration curve generated from purchased
standards.
[11] a) E. O. Eisen, J. Joffe, J. Chem. Eng. Data 1966, 11, 480; b) E. I. Gꢂrbꢂz,
S. G. Wettstein, J. A. Dumesic, ChemSusChem 2012, 5, 383; c) B. Saha,
N. S. Mosier, M. M. Abu-Omar, Advances in Plant Biology: Plants and Bio-
Energy, Vol. 4 (Eds.: M. McCann, M. S. Buckeridge, N. C. Carpita), Springer,
New York, 2013, pp. 267–276; d) Y. Romꢃn-Leshkov, J. A. Dumesic, Top.
Catal. 2009, 52, 297.
[12] a) G. Marcotullio, W. D. Jong, Green Chem. 2010, 12, 1739; b) G. Marco-
tullio, W. D. Jong, Carbohydr. Res. 2011, 346, 1291.
[13] T. W. G. Solomons, C. B. Fryhle, Organic Chemistry, 8th ed., Wiley, Hobo-
ken, NJ, 2004.
A Bruker AVQ-400 console with a Bruker 9.39 T magnet was used
1
to perform all H (400.13 MHz) and 13C (100.613 MHz) NMR experi-
ments. 1H NMR spectra were taken by using a 8.2 ms 908 pulse,
whereas 13C NMR spectra were taken using a standard gated de-
coupling pulse sequence (8.5 ms pulse width) to remove signal
splitting caused by proton coupling. d-xylose-1-13C labeled xylose
dissolved in D2O was used to improve signal quality in 13C NMR
xylose isomeric distribution studies. A Bruker AVB-400 console with
a Bruker 9.39 T magnet was used to perform all 35Cl (39.2 MHz)
NMR experiments. 35Cl-NMR spectra were taken using a 10 ms 908
pulse.
[14] a) W. J. Hamer, Y. C. Wu, J. Phys. Chem. Ref. Data 1972, 1, 1047; b) R. N.
Goldberg, R. L. Nuttall, J. Phys. Chem. Ref. Data 1978, 7, 263.
[15] a) C. D. Cappa, J. D. Smith, K. R. Wilson, B. M. Messer, M. K. Gilles, R. C.
Cohen, R. J. Saykally, J. Phys. Chem. B 2005, 109, 7046; b) J. D. Smith,
R. J. Saykally, P. L. Geissler, J. Am. Chem. Soc. 2007, 129, 13847.
[16] J. Israelachvili, Intermolecular and Surface Forces, 2nd ed., Academic
Press, London, UK, 1991.
[17] J. P. Mikkola, R. Sjoholm, T. Salmi, P. Maki-Arvela, Catal. Today 1999, 48,
73.
[18] S. J. Angyal, Adv. Carbohydr. Chem. Biochem. 1989, 47, 1.
[19] P. Ortiz, J. Fernandez-Bertran, E. Reguera, Spectrochim. Acta Part A 2005,
61, 1977.
[20] a) R. C. Remsing, R. P. Swatloski, R. D. Rogers, G. Moyna, Chem. Commun.
2006, 1271; b) J. J. Falke, R. J. Pace, S. I. Chan, J. Biol. Chem. 1984, 259,
6472; c) J. J. Falke, K. J. Kanes, S. I. Chan, J. Biol. Chem. 2005, 260, 9545.
[21] B. Danon, G. Marcotullio, W. de Jong, Green Chem. 2014, 16, 39.
[22] M. R. Nimlos, X. Qian, M. Davis, M. E. Himmel, D. K. Johnson, J. Phys.
Chem. A 2006, 110, 11824.
Reagent and product yields are reported as molar percentages rel-
ative to initial molar concentrations of xylose (i.e. furfural yields=
moles furfural/initial moles xylose). All reported yields were typical-
ly reproducible to within a +/- 5% relative error (based upon the
calculation of one standard deviation).
[23] X. H. Qian, M. R. Nimlos, D. K. Johnson, M. E. Himmel, Appl. Biochem. Bio-
technol. 2005, 124, 989.
Acknowledgements
[24] M. J. Antal, T. Leesomboon, W. S. Mok, G. N. Richards, Carbohydr. Res.
1991, 217, 71.
This work was supported by the Energy Biosciences Institute
funded by BP. All NMR work was performed using equipment
funded by NSF grant CHE-0130862. Special thanks to research as-
sistance provided by Daniel Tjandra and David Doan.
Received: October 20, 2014
Published online on && &&, 0000
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
11
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
K. R. Enslow, A. T. Bell*
Get your kicks from kinetics: The
effect of metal halides on the mecha-
nism and kinetics of xylose dehydration
in aqueous solution have been investi-
gated. We found that both the rate of
xylose consumption and furfural selec-
tivity are affected by the nature of the
salt cation and anion pairing.
&& – &&
The Role of Metal Halides in
Enhancing the Dehydration of Xylose
to Furfural
&
ChemCatChem 0000, 00, 0 – 0
www.chemcatchem.org
12
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!