Production of 5-(hydroxymethyl)-furfural from water-soluble carbohydrates and sugarcane molasses

Article abstract of DOI:10.1016/j.apcata.2017.07.049

In this work, two catalytic systems were applied for the production of 5-(hydroxymethyl)-furfural (HMF) from different water-soluble carbohydrates. These systems combined Lewis (ZnCl₂ and AlCl₃) and Br?nsted (HCl) acids in a biphasic system composed of an extraction solvent (ES) and a reaction solvent (RS). Tetrahydrofuran (THF) was used to control yield losses that are due to undesirable side reactions by promoting the continuous extraction of HMF from the aqueous phase, which was nearly saturated with sodium chloride (NaCl_aq). HMF production was enhanced when high temperatures (180?°C) and high ES/RS ratios (10) were employed in short reaction times (60?min). The combination of Lewis and Br?nsted acids and the use of the THF/NaCl_aq biphasic system proved to be the right choice for HMF production. AlCl₃/HCl was best for glucose dehydration while ZnCl₂/HCl gave higher HMF yields from sucrose and sugarcane molasses (65.6 and 49.6%, respectively). Compared to synthetic mixture containing the same amount of sugars, no matrix interference was observed during dehydration of sugarcane molasses. ZnCl₂/HCl humins were primarily composed of furan derivatives, whereas AlCl₃/HCl humins had higher molecular mass and benzofurans content, which were a result of the higher acidity of this reaction system.

Full text of DOI:10.1016/j.apcata.2017.07.049

Accepted Manuscript
Title: PRODUCTION OF
5-(HYDROXYMETHYL)-FURFURAL FROM
WATER-SOLUBLE CARBOHYDRATES AND
SUGARCANE MOLASSES
Authors: Gustavo Rodrigues Gomes, Daniel S. Rampon, Luiz
Pereira Ramos
PII:
S0926-860X(17)30369-1
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.07.049
APCATA 16354
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
21-3-2017
8-7-2017
31-7-2017
Please cite this article as: Gustavo Rodrigues Gomes, Daniel S.Rampon, Luiz Pereira
Ramos, PRODUCTION OF 5-(HYDROXYMETHYL)-FURFURAL FROM WATER-
SOLUBLE CARBOHYDRATES AND SUGARCANE MOLASSES, Applied
Catalysis A, Generalhttp://dx.doi.org/10.1016/j.apcata.2017.07.049
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

PRODUCTION OF 5-(HYDROXYMETHYL)-FURFURAL FROM WATER-
SOLUBLE CARBOHYDRATES AND SUGARCANE MOLASSES
Gustavo Rodrigues Gomesa, Daniel S. Ramponb and Luiz Pereira Ramos*a,c
a Research Center in Applied Chemistry (CEPESQ), Department of Chemistry, Federal
University of Paraná - UFPR, P. O. Box 19032 – Curitiba, PR – 81531-990 – Brazil
b
Laboratory of Polymers and Molecular Catalysis (LAPOCA), Department of
Chemistry, Federal University of Paraná - UFPR, P. O. Box 19032 – Curitiba, PR –
81531-990 – Brazil
c
INCT Energy & Environment (INCT E&A), Federal University of Paraná, Curitiba,
Brazil
* Email: luiz.ramos@ufpr.br
Graphical Abstract
Highlights
1. ZnCl2/HCl and AlCl3/HCl were used in a biphasic system to produce HMF from
water-soluble carbohydrates
2. No matrix interference was observed in the dehydration of sugarcane molasses
3. Acid dehydration in aqueous medium produced levulinic and formic acids
selectively
4. Both catalytic systems produced humins with a furanic backbone

Abstract
In this work, two catalytic systems were applied for the production of 5-
(hydroxymethyl)-furfural (HMF) from different water-soluble carbohydrates. These
systems combined Lewis (ZnCl2 and AlCl3) and Brønsted (HCl) acids in a biphasic
system composed of an extraction solvent (ES) and a reaction solvent (RS).
Tetrahydrofuran (THF) was used to control yield losses that are due to undesirable side
reactions by promoting the continuous extraction of HMF from the aqueous phase,
which was nearly saturated with sodium chloride (NaClaq). HMF production was
enhanced when high temperatures (180 °C) and high ES/RS ratios (10) were employed
in short reaction times (60 min). The combination of Lewis and Brønsted acids and the
use of the THF/NaClaq biphasic system proved to be the right choice for HMF
production. AlCl3/HCl was best for glucose dehydration while ZnCl2/HCl gave higher
HMF yields from sucrose and sugarcane molasses (65.6 and 49.6%, respectively).
Compared to synthetic mixture containing the same amount of sugars, no matrix
interference was observed during dehydration of sugarcane molasses. ZnCl2/HCl
humins were primarily composed of furan derivatives, whereas AlCl3/HCl humins had
higher molecular mass and benzofurans content, which were a result of the higher
acidity of this reaction system.
Keywords: sucrose • sugarcane molasses • acid-catalyzed dehydration •
hydroxymethylfurfural • biphasic systems
1. Introduction
The conversion of renewable feedstocks into fuel, chemicals and
materials plays an important role in the development of sustainable chemical
processes and this has been generally referred to as the biorefinery concept.
According to the National Renewable Energy Laboratory (NREL), biorefinery is
a facility that integrates equipment and processes for biomass conversion into
fuels, energy and chemicals for a variety of industrial applications. Likewise, the
International Energy Agency (IEA) defines biorefinery as the sustainable
processing of biomass into a spectrum of products to be marketed as food,
chemicals and supplies [1].
Sugarcane is a suitable example of biomass for biorefinery applications
[2]. The sugarcane processing for sugar (sucrose) and ethanol generates a

large number of by-products such as cane bagasse from the juice extraction
step, filter cake from juice filtration and molasses after crystallization of sucrose
from the concentrated juice. Cane molasses are low cost feedstocks that
contain high concentrations of soluble sugars (mostly sucrose) with small
amounts of fructose and glucose along with other minor components such as
phenolic compounds, organic acids and salts [3]. With these properties and
composition, cane molasses can be considered a valuable feedstock for
biorefinery applications through their chemical or biochemical conversion to
platform chemicals and high value-added compounds.
5-(Hydroxymethyl)-furfural (HMF) is considered one of the top 10 most
valuable chemicals that can be produced from plant biomass and this is due to
its wide range of applications in polymers, food and fuel chemistry[3,4]. Also,
HMF is a versatile chemical intermediate that can be converted to a variety of
interesting molecules such as 2,5-hydroxymethylfuran (DHMF), 5-
(hydroxymethyl)-furanoic acid (HMFA) [5], 2,5-furandicarboxylic acid [6],
levulinic acid [7], 2,5-dimethylfuran [8] and 2,5-diformylfuran [9], among others.
The most widely accepted mechanism for the production of furans from
hexoses such as glucose first encompasses its catalytic α-hydroxy-carbonyl
isomerization to produce fructose, which can be readily converted to HMF
through a series of dehydration reactions [10,11]. The endeavour for achieving
good yields under acceptable reaction conditions has shown that a Lewis acid
mediated isomerization followed by dehydration in the presence of a strong
Brønsted acid works best for the production of these furan compounds [12,13]
while the use of basic catalysts typically leads to several undesirable side
reactions [14]. However, earlier studies have also evaluated the use of both
mineral and organic acids acting alone, metallic salts and heterogeneous
catalysts with high Brønsted acid strength [11].
Carbohydrate dehydration has been studied in several reaction media
such as aqueous, organic, biphasic and non-conventional systems such as
ionic liquids and sub- or supercritical fluids [15]. Wang et al. (2013) [16]
employed a 67% ZnCl2 aqueous solution for glucosamine dehydration at 120 °C
for 90 min to obtain an HMF yield of 21.9%. However, a singular advance in
these reaction systems was based on the use of a two-phase environment
involving a reaction solvent (aqueous phase) and an extraction solvent (organic

phase) because the in situ extraction of HMF avoids undesirable side reactions
such as rehydration to levulinic and formic acids and condensation to dark
insoluble solid residues generally referred to as humins [17]. However, despite
the good selectivity of biphasic systems for carbohydrate dehydration, due to
the good reactivity of aldehyde group of this furan compound the formation of
humins has also been reported in organic solvents [18].
Appropriate extraction solvents for carbohydrate dehydration must have
a high partition coefficient for furan compounds, meaning that, once formed,
these reactive species will be readily transferred to the organic phase. For
water-miscible organic solvents, an electrolyte such as sodium chloride (NaCl)
must be added to induce the formation of a biphasic system [19,20].
Yang et al. (2012) [21] demonstrated that AlCl3 is a promising catalyst
for HMF production from glucose in a THF/NaClaq biphasic system under
microwave irradiation. HMF yields of 65% were achieved. Pagán-Torres et al.
(2012) [22] reached a 62% HMF yield from 5 wt.% glucose in a biphasic system
containing H2O/2-butylphenol and AlCl3/HCl at 170 °C and pH 2.5 using a
glucose/Lewis acid ratio of 67. Ma et al. (2015) [23] obtained HMF yields of
5.5% and 30.3% by glucose and sucrose dehydration, respectively, using a
H2O/MIBK biphasic system at high temperatures. At these same conditions,
fructose dehydration resulted in HMF yields of 73.6%, which was readily
attributed to the higher reactivity of ketohexoses compared to aldohexoses.
Solid acids containing both Lewis and Brønsted acid sites have also been
applied as heterogeneous catalysts for carbohydrate dehydration. Jain et al.
(2015) [24] employed mesoporous zirconium phosphate for sucrose
dehydration in a biphasic system of water and diglyme. In this case, a 53%
HMF yield was achieved at the best reaction conditions.
In this work, HMF was successfully produced from glucose, sucrose
and sugarcane molasses in a THF/NaClaq biphasic system using two catalytic
systems in which both Lewis and Brønsted acids were combined (ZnCl2/HCl
and AlCl3/HCl). To the best of our knowledge, this is the first time that
sugarcane molasses is used as feedstock for HMF production in a biphasic
system, therefore revealing probable matrix effects that could partially
compromise the efficiency of carbohydrate dehydration. Besides, humins that
were formed under these reaction conditions were isolated and characterized

by analytical pyrolysis coupled to mass spectrometry (py-GC/MS) and by
alkaline hydrolysis followed by mass spectrometry.
2. Experimental
2.1. Materials
Glucose of analytical grade and HClaq 37% were purchased from Vetec
(Duque de Caxias, RJ). AlCl3 (99%) was obtained from Sigma-Aldrich Brazil
(Jurubatuba, SP), while ZnCl2 (96%), acetic acid (99.7%) and NaCl (99.7%)
came from Neon (Suzano, SP). THF, acetonitrile and H2SO4 of HPLC grade
were purchased from Tedia (Rio de Janeiro, RJ) and sucrose of analytical
grade was acquired from F. Maia Indústria e Comércio (Belo Horizonte, MG).
Sugarcane molasses were kindly provided by Novozymes Latin America
(Araucária, PR) from batches that were collected in a large-scale sugar mill.
This mill is located in the Western region of the São Paulo State (Brazil).
2.2. Dehydration in a biphasic system
Dehydration experiments were performed in a Büchi Glas Uster®
stainless steel reactor (Uster, Switzerland), model miniclave drive, that was
equipped with mechanical stirring and both temperature and pressure controls.
The actual operating pressure corresponded to the vapor pressure of the
reaction mixture. All experiments were carried out with 0.75 mmol of
carbohydrate (0.135 g glucose; 0,257 g sucrose), whereas for sugarcane
molasses, the sample mass was 0.150 g (dry basis).
Different temperatures and ratios between extraction and reaction
solvents (ES/RS) were applied in this study. The temperature levels were 140,
160 and 180°C. The biphasic system was composed of 3 mL deionized water
and THF at three different ES/RS ratios: 5 (15 mL), 7.5 (22.5 mL) and 10 (30
mL). The amount of NaCl required to form a biphasic system was 1.05 g and
mechanical stirring was set at 1500 rpm [21]. The kinetics of glucose
dehydration was profiled in both catalytic systems (ZnCl2/HCl and AlCl3/HCl) by
withdrawing reaction aliquots at every 5 min. The applied temperature and

ES/RS ratio were those that performed best in the experimental assessments
described above.
HMF yields were calculated in relation to the corresponding theoretical
yield, which was based on the stoichiometric amount of HMF that could have
been produced from the total carbohydrate content of the starting material.
Equations 1 and 2 show the calculation for glucose and sucrose, respectively,
where HMFexp is the amount of HMF that was determined experimentally, mGlc
and mSuc are the initial mass of glucose (Glc) and sucrose (Suc), and 0.70 and
0.74 are the stoichiometric factors for glucose and sucrose conversion to HMF,
respectively.
HMF
(g)
exp
( )
% =
HMF
× 100
× 100
(1)
(2)
Glc
( )
g × 0.70
m
Glc
HMF
(g)
exp
( )
% =
HMF
Suc
( )
g × 0.74
m
Suc
HMF yields from sugarcane molasses were also calculated in relation to
the theoretical yield. However, the contribution of each carbohydrate type that
was found in the matrix (glucose, Glc; fructose, Fru; sucrose, Suc) had to be
taken into account, along with their corresponding conversion factors as
described in Equation 3.
HMF
(g)
experimental
(
)
Y
% =
× 100
(3)
HMF
( )
(
)
(
)
( )
m
g ×{[ Suc % ×0.74]+[Glc % ×0.70]+[Fru % ×0.70]}
molasses
2.3 Chromatographic analysis
Carbohydrate and organic acid analyses were carried out by high
performance liquid chromatography (HPLC) using a Shimadzu workstation that
was equipped with two detection systems: diode array UV spectrophotometry
(UV-DAD) and differential refractometry (RID). The chromatographic column
was an Agilent Hi-Plex-H that was operated at 65 °C using H2SO4 5 mmol L-1 as
mobile phase in a flow rate of 0.6 mL min-1. HMF was analyzed in Metrohm 882
Compact IC Plus HPLC, which was equipped with a dual wavelength ultraviolet
detector. The chromatographic column was a Kinetex 5μ EVO C18 100 A
(Phenomenex, 150 x 4.6 mm) that was eluted with a mobile phase containing

88% water, 11% acetonitrile and 1% acetic acid in a flow rate of 0.8 mL min-1. In
both chromatographic systems, quantification was carried out by external
calibration using the following concentration ranges: 0.5 to 5.0 g L-1 for glucose
and fructose, 2.0 to 20 mg L-1 for HMF, 0.03 a 1.50 g L-1 for levulinic acid and
0.01 to 0.50 g L-1 for formic acid.
2.4. Humins characterization
Analytical pyrolysis was carried out with solid samples (humins) that
were derived from glucose dehydration in THF/NaClaq using both catalytic
systems, ZnCl2/HCl and AlCl3/HCl. The py-GC/MS analyses were carried out in
a Thermo (Waltham, MA USA) workstation using a DB-5MS column (30 m x 0.2
mm; 0.25 µm). The pyrolysis temperature was set at 400 °C. The temperature
in the injection system was 230 °C and the sample split ratio was 1:100. The
total analysis time reached 60 min using the following temperature program:
isocratic at 50 °C for 2 min, heating at 4 °C min-1 until 280 °C and isocratic at
280°C until the end of the run. The ionization energy was set in 70 eV, the mass
analyzer was an ion trap type and the total ion current (TIC) was used to build
the GC-MS chromatograms, which were analyzed with the AMDIS software
(NIST) using a mass range of 50 to 650 m/z. All detected structures were
identified according to the NIST (National Institute of Standards) MS library. For
quantitative analysis by area normalization, all unidentified peaks corresponding
to less than 0.8% of the total chromatogram area were regarded as background
noise and removed from the py-GC/MS reports. Elution bands that were derived
from column bleeding were also ignored.
Alkaline hydrolysis of humins that were derived from glucose
dehydration was carried out in 0.5 mol L-1 NaOH at 100°C for 2 h [25]. After
hydrolysis, the oligomeric fraction was passed through a SPE C18 cartridge that
was further eluted with methanol. The cartridge eluate was analyzed by direct
infusion in a Varian 320 MS Spectrometer (Palo Alto, USA). The ionization
energy was set in 70 eV and the scan range was established between 70 and
1280 m/z.
3. Results and discussion

Initially, the efficiency of glucose dehydration was evaluated in the
THF/NaClaq biphasic system by employing the ZnCl2/HCl catalyst under several
temperatures and ES/RS ratios. According to Román-Leshkov and Dumesic
(2009) [20], THF fulfills the solvent requirements for proper furans extraction in
aqueous biphasic systems. However, THF is partially miscible in water so it is
necessary to add a simple electrolyte such NaCl to induce the formation of a
biphasic system. Table 1 depicts the influence of temperature and ES/RS ratio
on glucose dehydration with ZnCl2/HCl. The main products were fructose, HMF,
and a black residue that is attributed to the formation of humins [18].
Satisfactory HMF yields were obtained at higher temperatures with full
carbohydrate conversion being achieved at 180 °C (entries 5 and 6). Also, HMF
production was improved at higher ES/RS ratios (entries 1 and 2, 3 and 4, and
5 and 6) even at the highest temperature levels of 180 °C (entry 6). Fructose
was observed in the reaction media only at the lowest reaction temperature
(entries 1 to 4), suggesting that an efficient aldose−ketose isomerization is the
first step for the subsequent conversion of a six- to a five-membered ring [26].
Also, higher temperatures accelerated the rate of unwanted side reactions such
as condensation involving carbohydrates, furan compounds and organic acids
to produce humins (entries 5 and 6) [27]. However, the effective extraction of
furan compounds at high ES/RS ratios reduced the extent of such side
reactions because furans became less exposed to the more reactive aqueous
phase.
Table 1
A comparative study involving both Lewis and Brønsted acid types
acting separately or in combination was carried out to investigate their influence
in the catalytic dehydration of glucose (Fig. 1). The combined catalytic systems
resulted in good HMF yields, which corresponded to 48.8 ± 0.9% for ZnCl2/HCl
and to 48.6 ± 0.9% for AlCl3/HCl.
Fig. 1

Glucose dehydration with AlCl3 alone gave also acceptable HMF yields.
AlCl3 dissociates in presence of water to produce an hexaaquo complex (4) that
is further hydrolyzed to release the corresponding Brønsted acid in the reaction
medium (5 and 6) and these events enhanced the production of HMF when
AlCl3 was applied as the reaction catalyst [28].
Screening of different reaction times for glucose dehydration revealed
that both ZnCl2/HCl and AlCl3/HCl catalytic systems reached their maximum
HMF yields after 60 and 55 min, which corresponded to 48.8 ± 0.9% and 54.3 ±
1.0% of the theoretical yield, respectively (Fig. 2). At longer reaction times, the
HMF yields decreased in both catalytic systems due to the formation of humins
and this is readily attributed to the excessive exposure of HMF and other
carbohydrate derivatives to the catalyst acid strength. Such yield losses were
more pronounced with AlCl3/HCl because this catalytic system is more acidic
than ZnCl2/HCl.
Fig. 2
To demonstrate the advantages of using a biphasic system, HMF yields
in THF/NaClaq were compared to those derived from reactions that were carried
out in the absence of an organic solvent (Table 2). As already expected, the
best performances were achieved in THF/NaClaq with both ZnCl2/HCl and
AlCl3/HCl because HMF was continuously extracted from the acidic aqueous
phase, thereby minimizing its further conversion to humins (by aldol
condensation) and both levulinic and formic acids (by rehydration). By contrast,
when the reaction was carried out in aqueous system, the latter two organic
acids were formed predominantly and selectively. Also, ZnCl2/HCl was slightly
better than AlCl3/HCl for the production of both levulinic and formic acids. This
observation is interesting because levulinic acid is an important value-added

chemical intermediate and its formation could be easily directed by a simple
change in the reaction solvent.
Table 2
Table 3 shows the effect of catalyst loading on the conversion of
glucose to HMF in THF/NaClaq using both AlCl3/HCl and ZnCl2/HCl catalytic
systems. Three different catalyst loadings were evaluated using the same molar
ratio between both Lewis and Brønsted acid types. Both catalytic systems were
relatively equivalent for glucose dehydration, with the highest HMF yields being
obtained by adding 0.2 mmol (27 mol%) of the acid catalysts. These HMF yields
corresponded to 62.7 ± 0.7% for ZnCl2/HCl (entry 2) and 66.9 ± 1.5% for
AlCl3/HCl (entry 5).
Table 3
The pH of the catalytic solutions was measured to evaluate its
correlation with HMF yields. In all experiments described in Table 3, the
AlCl3/HCl system always exhibited the lowest pH values (entries 4-6) and this
was attributed to the intrinsic properties of this harder Lewis acid, since the
Brønsted/Lewis acid ratio was always the same [28]. The best HMF yields were
reached using 0.2 mmol (27 mol%) of catalyst at intermediate pH values of 1.36
for ZnCl2/HCl and 1.12 for AlCl3/HCl (entries 2 and 5). A further increase in the
catalyst loading to 0.3 mmol (40 mol%) had negligible impact on HMF formation
using both catalytic systems (entries 3 and 6). By contrast, the HMF yield
decreased at lower pH values because this promoted the aldol
addition/condensation of HMF, which is the primary reaction for humins
formation [29]. In order to investigate the effect of catalytic systems applied on
glucose dehydration, experiments were also carried out in absence of any
added catalyst. The HMF yield that was obtained from glucose by thermal
conversion was 9.7 ± 1.6%, which is much lower compared to the performance
of both catalytic systems.
Compared to glucose, ZnCl2/HCl and AlCl3/HCl had considerably
different catalytic performances in sucrose dehydration (Table 3). The highest

HMF yields were produced with 0.2 mmol (27 mol%) of ZnCl2/HCl (65.6 ± 0.4%,
entry 2), showing values 17% higher than those obtained with AlCl3/HCl (entry
5). This was a remarkable change in catalytic activity because AlCl3/HCl was
the best catalyst for glucose dehydration. Sucrose is a non-reducing
disaccharide whose glycosidic linkage is relatively weak and its faster acid
hydrolysis facilitates the direct formation of the fructofuranosyl cation, which is
readily converted to HMF by tautomerization followed by dehydration (Scheme
1) [30]. However, probably due to its rapid built up in the AlCl3/HCl more acidic
reaction medium, the longer exposure of this HMF moiety to acid-catalyzed
aldol condensation caused a significant drop in its final yield [31,32].
Scheme 1
Since sucrose is hydrolyzed to equivalent amounts of glucose and
fructose, one additional experiment was carried out to assess the effect of
hydrolysis on sucrose dehydration. For this, an equimolar mixture of glucose
and fructose (0.135 g each, which is equivalent to 0.257 g of sucrose) was
dehydrated to HMF under the same reaction conditions using both AlCl3/HCl
and ZnCl2/HCl in a THF/NaClaq biphasic system (Fig. 3). The HMF yields from
glucose plus fructose were lower than those from sucrose alone, reaching 53.3
± 0.4% for ZnCl2/HCl and 46.0 ± 0.9% for AlCl3/HCl compared to entries 2 (65.6
± 0.4%) and 4 (54.4 ± 1.4%) of Table 3, respectively. The glycosidic bond of
sucrose can be easily hydrolyzed by either Brønsted or Lewis acid catalysts to
form directly the fructofuranosyl cation. By contrast, free fructose has to be
converted to this HMF precursor and this is not so easily achieved under these
reaction conditions [11,30].
Fig. 3
Sugarcane molasses are interesting materials for chemical conversion
because of their high carbohydrate content and low cost. This matrix contains
62.4 wt.% sucrose, 8.5 wt.% glucose and 4.5 wt.% fructose, in addition to other
minor components such as phenolic compounds, organic acids and inorganic

salt [33]. Hence, it would be extremely relevant to know if the conditions used in
this study were also adequate to produce HMF from this widely available
industrial residue. For this purpose, cane molasses were dehydrated using both
AlCl3/HCl and ZnCl2/HCl in THF/NaClaq and the results were compared with
those obtained from a synthetic mixture containing the same concentration of
sucrose, glucose and fructose (Fig. 3).
Compared to sucrose alone, lower HMF yields were obtained from
sugarcane molasses (49.6 ± 0.4% for ZnCl2/HCl and 43.2 ± 1.0% for AlCl3/HCl)
because part of the available sugars were found as glucose and fructose and
these are less stable under the reaction conditions. However, identical HMF
yields were derived from a synthetic mixture that contained the same
carbohydrate composition as cane molasses. Therefore, other minor
components of this complex matrix such as proteins, organic acids and
inorganic salts had no influence on HMF yields as well as on the activity of both
catalytic systems employed in this study.
It is well known that the acid-catalyzed carbohydrate dehydration may
lead to the accumulation of undesired reaction by-products that are derived
from HMF by two different pathways, one leading to humins and another to
levulinic and formic acids. Humins formation depends on the process conditions
as well as on the nature of the acid catalysts used for conversion. Hence, the
characterization of these reaction by-products is an interesting way to elucidate
the pathway of such catalytic conversion processes [34].
Humins derived from glucose in THF/NaClaq using both ZnCl2/HCl and
AlCl3/HCl catalytic systems were analyzed by py-GC/MS at 400 °C (Fig. 4) and
the qualitative analysis of the pyrolysate was carried out by comparison with the
NIST MS library [34]. After subtraction of the background elution bands, the
remaining peaks accounted for 90% of the total chromatogram area. Both
ZnCl2/HCl and AlCl3/HCl humins had similar py-GC/MS profiles, with 19
components identified on the basis of their mass spectra (Table 4). Furan
derivatives responded for nearly 90% of the list of identified organic
compounds, led by 2-methylfuran (2), 2,5-dimethylfuran (3), 2-acetylfuran (9),
one unidentified alkylfuran derivative (5), furan (1), 5-methylfurfural (12) and 5-
methyl-2-furanone (7), in this order (Table 4). The furan-based structure of the
isolated humins was revealed by the predominance of these and other furan

derivatives in the py-GC/MS profiles, as well as by the low contribution of
benzene derivatives and HMF rehydration by-products such as levulinic acid
and 2,5-dioxo-6-hydroxy hexanal (DHH) [27].
Fig. 4
Phenolic componds such as m-cresol (14) and 2,4-di-(tert-butyl)-phenol
(19), benzofurans such as 2-methylbenzofuran (16), 7-methylbenzofuran (15)
and 4,7-dimethylbenzofuran (18), and ketones such as 2-cyclohexen-1-one (8),
3-methyl-2-cyclohexen-1-one (11) and allylacetone (10) were also observed by
analytical pyrolysis in lower amounts, particularly in the py-GC/MS profile of
AlCl3/HCl humins. Benzofurans are derived from the Diels-Alder reaction and
dehydration of the active furanyl intermediates and olefin fragments [35,36].
Unsaturated ketones come from the pyrolysis of carbohydrates [37] and
phenols have been observed already as a result of humins pyrolysis [38]. The
presence of 5-methyl-2-furanone in both py-CG/MS profiles was probably arisen
from levulinic acid moieties that were physically entrapped in humin clusters
[39].
Wang et al. (2016) [34] investigated the analytical pyrolysis of humins
that were derived from glucose dehydration in dilute sulfuric acid at high
temperatures (600°C). A wide product distribution was observed including a
range of benzofurans, phenols and reactive furan derivatives. Similar
compounds were observed in the py-GC-MS analysis of both ZnCl2/HCl and
AlCl3/HCl humins but phenol and benzofuran derivatives were present in a
much lower extent, meaning that the pathways of humins formation were the
same but the use of lower temperatures and acid strengths in a biphasic system
produced humins with a more characteristic furanic backbone.
Table 4
Humins were also characterized by mass spectrometry after alkaline
hydrolysis. Van Zandvoort et al. (2015) [25] developed this approach to
investigate the structure and chemical composition of highly recalcitrant humins
that were produced by acid-catalysed dehydration of glucose. Alkaline

hydrolysis is important because humins that are hardly soluble in organic
solvents can be solubilized completely after alkaline oligomerization [40].
The mass spectra of humins that were synthesized from glucose using
ZnCl2/HCl in THF/NaClaq showed a maximum m/z of 613 and 3 fragmentation
signals of m/z of 475, 301 and 129. The first mass loss was equivalent to an
m/z of 138 and the two subsequent mass losses corresponded to an m/z of
173. On the other hand, the mass spectra of humins derived from the AlCl3/HCl
system showed a maximum m/z of 1240, which is twice bigger than that of
ZnCl2/HCl. In addition, there were several fragments whose mass losses were
around an m/z of 158. Based on these data, it is clear that a more acidic
catalytic system such as AlCl3/HCl contributed more to condensation reactions
than ZnCl2/HCl, leading to the formation of humins of higher molecular mass.
Also, a much larger number of stable fragments were observed in the mass
spectrum of AlCl3/HCl humins.
4. Conclusion
The combination of Lewis and Brønsted acids was a good way to
improve the production of HMF from simple sugars such as glucose and
sucrose. Also, the use of a biphasic system containing THF reduced the extent
of undesirable side reactions. HMF production was enhanced when high
temperatures (180 °C) and extraction/reaction solvent ratios (ES/RS) were
used. The best HMF yields from glucose were 62.7 ± 0.7% for ZnCl2/HCl and
66.9 ± 1.5% for AlCl3/HCl. For saccharinic matrices, ZnCl2/HCl performed better
than AlCl3/HCl. The best HMF yields from sucrose were 65.6 ± 0.4% for
ZnCl2/HCl and 54.4 ± 1.4% for AlCl3/HCl. Good HMF yields were also obtained
from sugarcane molasses and no matrix interference was observed in
dehydration reactions, indicating that this feedstock has a great potential for
HMF production. The analytical pyrolysis of both ZnCl2/HCl and AlCl3/HCl
humins revealed that they are formed mostly by furan derivatives, especially by
2-methylfuran. Also, the alkali oligomerization of AlCl3/HCl humins produced
m/z fragments twice as big as those derived from ZnCl2/HCl humins, showing
that the catalyst with higher acid strength contributed to condensation reactions
even in the presence of a suitable biphasic system.

5. Acknowledgements
The authors are grateful to CNPq (grants 551404/2010-8 and
311554/2011-3) and to the INCT in Energy and Environment for providing
financial support to carry out this study, as well as to Novozymes Latin America
(Araucária, PR, Brazil) for donating a large batch of fresh sugarcane molasses.
Also, the authors wish to thank CAPES for granting an M.Sc. scholarship to
G.R.G. and Fabricio Augusto Hansel from Embrapa-Florestas (Colombo, PR,
Brazil) for the py-GC/MS analyses of humins.
6. References
[1] S. K. Maity, Renewable Sustainable Energy Rev. 43 (2015) 1427-1445.
[2] J. Liu, J. Basnayake, P.A. Jackson, X. Chen, J. Zhao, P. Zhao, L. Yang, Y.
Bai, H. Xia, F. Zan, W. Qin, K. Yang, L. Yao, L. Zhao, J Zhu, P. Lakshmanan, X.
Zhao, Y. Fan, Field Crop Res. 196 (2016) 418-425.
[3] D. (D. J.) Liu, E. Y-X. Chen, ACS Catal. 4 (2014) 1302-1310.
[4] P. Mäki-Arvela, E. Salminen, T. Riittonen, P. Virtanen, N. Kumar, J-P.
Mikkola, Int. J. Chem. Eng. 2012 (2012) 1-10.
[5] S. Subbiah, S. P. Simeonov, J. M. S. S. Esperença, L. P. N. Rebelo, C. A.
M. Afonso, Green Chem. 15 (2013) 2849-2853.
[6] S. Siankevich, G. Savoglidis, Z. Fei, G. Laurenczy, D. T.L. Alexander, N.
Yan, P.J. Dyson, J. Catal. 2014, 315, 67-74.
[7] N. A. S. Ramli, N. A. S. Amin, Appl. Catal., B 163 (2015) 487-498.
[8] L. Hu, L. Lin, S. Liu, Ind. Eng. Chem. Res. 53 (2014) 9969-9978.
[9] G. A. Halliday, R. J. Young, V. V. Grushin, Org. Lett. 11 (2003) 2003-2005.
[10] H. Rasmussen, H. R. Sørensen, A. S. Mayer, Carbohydr. Res. 385 (2014)
45-57.

[11] R-J. van Putten, J. C. van der Waal, E. de Jong, C. B. Rasrendra, H. J.
Heeres, J. G. de Vries, Chem. Rev. 113 (2013) 1499−1597.
[12] J. B. Binder, R. T. Raines, J. Am. Chem. Soc. 131 (2009) 1979-1985.
[13] V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N. S.
Marinkovic, A. I. Frenkel, S. I. Sandler, D. G. Vlachos, J. Am. Chem. Soc. 135
(2013) 3997-4006.
[14] B. Y. Yang, R. Montgomery, Carbohydr. Res. 280 (1996) 27-45.
[15] S. P. Teong, G. Yi, Y. Zhang, Green Chem. 16 (2014) 2015-2026.
[16] Y. Wang, C. M. Pedersen, T. Deng, Y. Qiao, X. Hou, Bioresour. Technol.
143 (2013) 384-390.
[17] B. Saha, M. M. Abu-Omar, Green Chem. 16 (2014) 24-38.
[18] X. Zhang, B. B. Hewetson, N. S. Mosier, Energy Fuels 29 (2015) 2387-
2393.
[19] F. N. D. C. Gomes, L. R. Pereira, N. F. P. Ribeiro, M. M. V. M. Souza, Braz.
J. Chem. Eng. 32 (2015) 119-126.
[20] Y. Román-Leshkov, J. A. Dumesic, Top. Catal. 52 (2009) 52, 297-303.
[21] Y. Yang, C. Hu, M. M. Abu-Omar, Green Chem. 14 (2012) 509-513.
[22] Y. J. Pagꢀn-Torres, T. Wang, J. M. R. Gallo, B. H. Shanks, J. A. Dumesic,
ACS Catal. 2 (2012) 930-934.
[23] H. Ma, F. Wang, Y. Yu, L. Wang, X. Li, Ind. Eng. Chem. Res. 54 (2015)
2657-2666.
[24] A. Jain, A. M. Shore, S. C. Jonnalagadda, K. V. Ramanujachary, A.
Mugweru, Appl. Catal., A 489 (2015) 72-76.
[25] I. van Zandvoort, E. R. H. van Eck, P. de Peinder, H. J. Heeres, P. C. A.
Bruijnincx, B. M. Weckhuysen, ACS Sustainable Chem. Eng. 3 (2015) 533−543.
[26] Y. Yang, C. Hu, M. M. Abu-Omar, J. Mol. Catal. A: Chem. 376 (2013) 98-
102.

[27] I. van Zandvoort, Y. Wang, C. B. Rasrendra, E. R. H. van Eck, P. C. A.
Bruijnincx, H. J. Heeres, B. M. Weckhuysen, ChemSusChem 6 (2013) 1745-
1758.
[28] F. Fringuelli, F. Pizzo, L. Vaccaro, J. Org. Chem. 66 (2001) 4719-4722.
[29] S. Wang, H. Lin, J. Chen, Y. Zhao, B. Ru, K. Qiu, J. Zhou, RSC Adv. 5
(2015) 84014-84021.
[30] C. P. Locas, V. A. Yaylayan, J. Agric. Food Chem. 56 (2008) 6717–6723.
[31] A. S. Amarasekara, L. D. Williams, C. C. Ebede, Carbohydr. Res. 343
(2008) 3021-3024.
[32] H. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science 316 (2007) 1597-
1600.
[33] M. Sharma, S. N. Patel, K. Lata, U. Singh, M. Krishania, R. S. Sangwan, S.
P. Singh, Bioresour. Technol. 219 (2016) 311-318.
[34] S. Wang, H. Lin, Y. Zhao, J. Chen, J. Zhou, J. Anal. Appl. Pyrolysis 118
(2016) 259-266.
[35] Y. T. Cheng, G. W. Huber, ACS Catal. 1 (2011) 611-628.
[36] S. Vaitheeswaran, S. K. Green, P. Dauenhauer, S. M. Auerbach, ACS
Catal. 3 (2013) 2012-2019.
[37] S. Wang, X. Guo, K. Wang, Z. Luo, J. Anal. Appl. Pyrolysis 91 (2011) 183-
189.
[38] T. M. C. Hoang, E. R. H. van Eck, W. P. Bula, J. G. E. Gardeniers, L.
Leffertsa, K. Seshan, Green Chem. 17 (2015) 959-972.
[39] N. Baccile, G. Laurent, F. Babonneau, F. Fayon, M-M. Titirici, M. Antonietti,
J. Phys. Chem. C 113 (2009) 9644–9654.
[40] G. Tsilomelekis, M. J. Orella, Z. Lin, Z. Cheng, W. Zheng, V. Nikolakis, D.
G. Vlachos, Green Chem. 18 (2016) 1983-1993.

Figure Captions
Fig. 1. Comparative study involving the use of Lewis and Brønsted acids in
glucose dehydration. Reactions were carried out in THF/NaClaq at 180 °C for 60
min using 0.75 mmol glucose with aqueous HCl pH 1.9 (0.3 mmol or 25 µL
HClaq 37%) and 0.98 (3.6 mmol or 300 µL HClaq 37%) as well as with 0.3 mmol
ZnCl2 (0.043 g) or AlCl3 (0.073 g) with and without 0.3 mmol HCl.
Fig. 2. Glucose dehydration profiles for (A) ZnCl2/HCl (B) AlCl3/HCl in
THF/NaClaq. Reactions were carried out at 180 °C employing 0.75 mmol
glucose plus 0.3 mmol ZnCl2 (0.043 g) or AlCl3 (0.073 g) containing 0.3 mmol
HCl (or 25 µL HClaq 37%).
Fig. 3. HMF production from sugarcane molasses and their main carbohydrate
components using ZnCl2/HCl (0.029 g ZnCl2 plus 15 µL HClaq 37%) and
AlCl3/HCl (0.024 g AlCl3 and 10 µL HClaq 37%) in THF/NaClaq. Reactions were
carried out at 180 °C for 60 (ZnCl2/HCl) and 55 min (AlCl3/HCl) using (A)
sucrose (0.257 g); (B) glucose and fructose (0.135 g each); (C) cane molasses
(0.150 g) and (D) synthetic media containing sucrose (0.094 g), glucose (0.013
g) and fructose (0.006 g).
Fig. 4. Py-GC-MS analysis of humins that were derived from glucose using (A)
ZnCl2/HCl and (B) AlCl3/HCl in a THF/NaClaq biphasic system. Reactions were
carried out at 180 °C for 60 min using 0.75 mmol glucose plus 0.3 mmol ZnCl2
(0.043 g) or AlCl3 (0.073 g) containing 0.3 mmol HCl (or 25 µL HClaq 37%).
Scheme 1. Formation of the fructofuranosyl cation from sucrose followed by its
conversion to HMF.

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Scheme 1.

Table 1. Influence of temperature and solvent ratio in glucose dehydration using
ZnCl2/HCl as the reaction catalyst in a THF/NaClaq biphasic system.
Conversion
(mol %)
Yield (mol %)
Solid
Entrya T (°C) ES/RSb
residue (wt %)
Fructose
HMF
1
2
3
4
5
6
140
140
160
160
180
180
5
73.3
70.4
85.2
79.3
100
11.9
9.0
1.5
5.9
ndc
nd
2.4
8.5
12.5
10.7
20.1
16.1
36.4
33.7
10
7.5
7.5
5
18.9
21.0
33.1
48.8
10
100
[a]Reactions were carried for 60 min in the THF/NaClaq biphasic system using 0.75 mmol
glucose (0.135 g) plus 0.3 mmol ZnCl2 (0.043 g) containing 0.3 mmol HCl (or 25 µL HClaq
37%);[b] ES/RS= extraction solvent/reaction solvent; [c] nd = not detected.

Table 2. Comparative study of glucose dehydration using both ZnCl2/HCl and
AlCl3/HCl catalyst in biphasic and aqueous systems.
Biphasic system[a,c]
Aqueous system[b,c]
Compounds
(mol%)
ZnCl2/HCl
AlCl3/HCl
54.3 ± 1.0
nd
ZnCl2/HCl
AlCl3/HCl
0.06 ± 0.01
41.1 ± 0.1
36.0 ± 0.8
HMF
48.8 ± 0.9
nd[d]
0.26 ± 0.02
46.8 ± 0.7
40.4 ± 0.2
Levulinic acid
Formic acid
nd
nd
[a] Binary system composed of THF and NaClaq (30 mL THF plus 1.05 g NaCl in 3 mL H2O) plus
0.75 mmol glucose and the indicated amount of catalyst;
[b] Aqueous system containing water (30 mL) plus 0.75 mmol glucose and the indicated amount
of catalyst;
[c] Reactions were carried at 180 °C using 0.3 mmol ZnCl2 (0.043 g) or AlCl3 (0.073 g) containing
0.3 mmol HCl (or 25 µL HClaq 37%); the reaction time was 60 min for ZnCl2/HCl and 55 min for
AlCl3/HCl;
[d] nd, not detected

Table 3. Influence of different catalyst loadings in both glucose and sucrose
dehydration using ZnCl2/HCl and AlCl3/HCl in a THF/NaClaq biphasic system.
Catalyst loadingb
(mmol)
HMF/glucose
(mol %)
HMF/sucrose
(mol (%)
Entrya
pH
ZnCl2/HCl
59.6 ± 0.2
62.7 ± 0.7
48.8 ± 0.9
AlCl3/HCl
64.6 ± 0.4
66.9 ± 1.5
54.3 ± 1.0
1
2
3
0.1 (13 mol%)
0.2 (27 mol%)
0.3 (40 mol%)
60.6 ± 0.6
65.6 ± 0.4
60.0 ± 0.7
1.43 ± 0.04
1.36 ± 0.05
1.18 ± 0.04
4
5
6
0.1 (13 mol%)
0.2 (27 mol%)
0.3 (40 mol%)
54.4± 1.4
48.6 ± 0.3
49.6 ± 0.4
1.40 ± 0.06
1.12 ± 0.04
0.98 ± 0.04
[a]
Reactions were carried for 60 min in the THF/NaClaq biphasic system using 0.75 mmol
glucose (0.135 g) or sucrose (0.257 g) plus the indicated amount of catalyst; the reaction time
was 60 min for ZnCl2/HCl and 55 min for AlCl3/HCl;
[b] 0.1 mmol: 0.014 ZnCl2 or 0.024 g AlCl3 with 10 µL HClaq 37%; 0.2 mmol: 0.029 ZnCl2 or 0.049
g AlCl3 with 15 µL HClaq 37%; 0.3 mmol: 0.043 ZnCl2 or 0.072 g AlCl3 with 25 µL HClaq 37%.

Table 4. Py-GC/MS analyses of ZnCl2/HCl and AlCl3/HCl humins
Area (%)
#
Compound
m/z
ZnCl2/HCl
AlCl3/HCl
8.0
1
2
furan
68
53; 82
96
8.6
24.7
12.1
2.1
2-methylfuran
21.0
10.3
2.5
3
2,5-dimethylfuran
5-methyl-2-ethylfuran
unidentified alkyllfuran
2-allylfuran
4
95; 110
95
5
9.5
12.4
1.7
6
79; 108
55; 98
68; 96
95; 110
55; 98
110
0.82
9.10
3.5
7
5-methyl-2-furanone
2-cyclohexen-1-one
2-acetylfuran
7.6
8
3.8
9
11.9
3.2
11.6
3.6
10
11
12
13
14
15
16
17
18
19
allylacetone
3-methyl-2-cycloexen-1-one
5-methylfurfural
0.83
7.4
1.1
110
6.4
5-methyl-2-acetylfuran
m-cresol
109; 124
108
1.8
3.3
0.93
nd
1.4
7-methylbenzofuran
2-methylbenzofuran
3-(cycloexen-1-yl)-furan
4,7-dimethylbenzofuran
2,4-di-tert-butylphenol
132
1.1
132
1.5
1.2
148
nd
0.85
2.1
146
0.94
1.1
191
nd
[a] nd, not detected