Efficient green catalysis for the conversion of fructose to levulinic acid

Article abstract of DOI:10.1016/j.apcata.2017.03.016

Highly efficient and selective production of levulinic acid has been achieved from D-fructose in the presence of polystyrene-based sulphonic acid resin catalyst, Dowex 50?×?8-100, at mild reaction conditions of 120?°C, over 24?h in a 50:50 mixture of water/GVL resulting in 72?mol% yield under optimized reaction conditions. Optimization of the effect of reaction temperature, time, pressure, catalyst to substrate ratio, fructose concentration and solvent was performed. Various polystyrene-based sulfonic acid resins were also investigated for quantitative production of LA from 5-hydroxymethylfurfural (5-HMF) in pure water. Catalyst recycling was carried out up to 6 cycles. Significant mechanistic information was obtained for the formation of “humins”, which are the primary cause of catalyst fouling, by the identification of soluble by-products and polymerization presursors using Q-Tof mass spectrometry based on accurate masses.

Full text of DOI:10.1016/j.apcata.2017.03.016

Accepted Manuscript
Title: Efficient Green Catalysis for the Conversion of Fructose
to Levulinic Acid
Authors: Indira Thapa, Brian Mullen, Ammar Saleem, Cora
Leibig, R. Tom Baker, Javier B. Giorgi
PII:
S0926-860X(17)30115-1
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.03.016
APCATA 16175
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
16-11-2016
31-1-2017
15-3-2017
Please cite this article as: Indira Thapa, Brian Mullen, Ammar Saleem,
Cora Leibig, R.Tom Baker, Javier B.Giorgi, Efficient Green Catalysis
for the Conversion of Fructose to Levulinic Acid, Applied Catalysis A,
Generalhttp://dx.doi.org/10.1016/j.apcata.2017.03.016
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.

Efficient Green Catalysis for the Conversion of Fructose to Levulinic Acid
Indira Thapaa, Brian Mullenb, Ammar Saleemc, Cora Leibigb, R. Tom Bakera, Javier B. Giorgia*
aCentre for Catalysis Research and Innovation, Department of Chemistry and Biomolecular
Sciences, University of Ottawa, Ottawa, Ontario, Canada. K1N 6N5.
bSegetis, Inc. 680 Mendelssohn Ave N, Golden Valley MN 55427, US
cLaboratory for the Analysis of Natural and Synthetic Environmental Toxins, Centre for
Advanced Research in Environmental Genomics, University of Ottawa, Ottawa, Ontario
E-mail: jgiorgi@uottawa.ca
Fax: +1(613)562-5170; Tel: +1(613)562-5800x6037
Corresponding author: Javier B. Giorgi, jgiorgi@uottawa.ca
Graphical abstract for TOC
1

Highlights
Demonstration of fructose to levulinic acid with very high activity and selectivity using Dowex 50x8-100
resin.
We obtained an efficient system capable to produce the transformation under benign conditions.
This work provides more flexibility than dual catalyst systems reported previously, behaving as a
switchable catalyst for LA and HMF production depending on the nature of solvent.
Mechanistic insight provided on the formation of by-products leading to humins, the primary catalyst
fouling mechanism for this reaction.
Abstract
Highly efficient and selective production of levulinic acid has been achieved from D-fructose in
the presence of polystyrene-based sulphonic acid resin catalyst, Dowex 50x8-100, at mild
reaction conditions of 120 oC, over 24 h in a 50:50 mixture of water/GVL resulting in 72 mol%
yield under optimized reaction conditions. Optimization of the effect of reaction temperature,
time, pressure, catalyst to substrate ratio, fructose concentration and solvent was performed.
Various polystyrene-based sulfonic acid resins were also investigated for quantitative production
of LA from 5-hydroxymethylfurfural (5-HMF) in pure water. Catalyst recycling was carried out
up to 6 cycles. Significant mechanistic information was obtained for the formation of “humins”,
which are the primary cause of catalyst fouling, by the identification of soluble by-products and
polymerization presursors using Q-Tof mass spectrometry based on accurate masses.
Keywords: Levulinic acid, hydroxymethylfurfural, fructose, Dowex, green catalysis
2

1. Introduction
The continuity and reconversion of the current energy supply are major challenges
increasingly faced by modern civilization. The present massive exploitation of energy resources
and the consequent negative impact on the environment is also a growing major concern. Fossil
fuels such as coal, natural gas, and petroleum provide about 90% of the energy used today by
developed countries. Renewable and sustainable sources are the only alternative to this rising
global problem. Biomass,[1] one of the largest renewable resources, is composed of cellulose,
hemicellulose and lignin. Cellulose is composed in turn of carbohydrates. Carbohydrate research
receives special attention from leaders in the field, as it forms the largest natural source of
carbon.[2] An alternative way to use carbohydrates as feedstocks is their dehydration into furans
and levulinic acid. Levulinic acid (LA) and 5-hydroxymethylfurfural (HMF) are among the most
desirable biomass-derived platform chemicals, polymer precursors and feedstocks that can be
efficiently converted into fuel additives and other valuable chemicals.[1] Levulinic acid is
categorized by the US Department of Energy as one of the top 12 building block chemicals
among the 300 potential biorefinery products identified.[3–5]
The presence of reactive carbonyl functional groups make LA a very attractive precursor
for a variety of value-added chemical products. Ethyl/methyl-levulinate can be used as diesel
additives, methyl-tetrahydrofuran (MTHF) and γ-valerolactone (GVL), obtained from the
hydrogenation of levulinic acid, can be directly used as solvents and gasoline blenders.[6–11]
LA is generally obtained from the hexose sugars of cellulose through a consecutive dehydration
and rehydration process in the presence of acid catalysts according to the mechanism proposed
by Horvat et al.[12–19] The bottleneck in the reaction mechanism is the formation and stability
of the intermediate HMF under the specified reaction conditions in an aqueous medium.[16,20–
22] Conventionally, LA has been produced with homogeneous catalysts such as mineral acids
(HCl, H2SO4, H3PO4 etc.) due to their ready availability, easy applicability, short reaction times
and high catalytic activity.[23–28] However, the catalyst recovery process is energy intensive
and waste disposal is costly due to the environmental impact exerted by these mineral acids.
Alternatively, development of efficient heterogeneous catalysts is an active field of
research to minimize the problems posed by homogeneous catalysts. Easy separation from
3

products and reusability are among the most important advantages of heterogeneous catalysts.
The search for a solid acid catalyst possessing high stability and accessible strong acid sites
which is economically amenable for LA production from carbohydrates is an ongoing process
among researchers in the field.[29,30] Jow and coworkers have used LZY zeolite for the
dehydration of fructose to LA at high temperature (383 to 433K) over 8-15 h to yield 40%
selectivity.[31] Similar results were observed with Y-Zeolite from glucose dehydration to LA
under similar reaction conditions.[32,33] Increasing the reaction temperature to 453K in an
autoclave with ZRP zeolite containing different SiO2/Al2O3 ratio did not improve the LA yield
(36%) from glucose.[34] Recently, Dumesic and co-workers presented promising results of LA
production (>70%) in monophasic THF-H2O solvent with furfural alcohol using H-ZSM-5 as
catalyst.[35] Following this research Requies and co-workers published the improved LA yield
(77%) from furfural alcohol using ZSM-5 under an optimized reaction conditions.[36] Lourvanij
and Rorrer studied a pillared clay as a solid acid catalyst for the dehydration of glucose with
typical LA yields of 20%.[37] Wang and co-workers used sulfated TiO2 for cellulose
dehydration to LA, produced 32% yield (38% of theoretical) under their optimum
2-
conditions.[38] Recently, Chen and co-workers reported use of a solid superacid (S2O8 /ZrO2-
SiO2-Sm2O3), at high temperature (473K) which yielded 23% of LA from rice straw.[39]
Commercially available ion exchange resins have been used in various industrially
important chemical transformations such as alkylation, trans-alkylation, isomerization,
oligomerization, acylation, esterification and nitration.[40] However, only limited studies have
explored the potential of ion exchange resins as catalysts for HMF and LA formation.[21,41–44]
Acidic ion-exchange resins (Dowex MSC, Amberlysts), studied by Schraufnagel and Rase for
the formation of HMF and LA, afforded moderate yields (24 mol%) after 24 h reaction time at
100 ℃.[45] Improved yields were reported by Sanborn and co-workers with a solid acid cation
exchange resin, Amberlyst 35, which produced 41% and 62% of LA from fructose and high-
fructose corn syrup at 423K in an autoclave over 4.5 and 18 h, respectively.[29] Smith, and co-
workers made use of Dowex 50x8-100 for the production of HMF from fructose in acetone-
water by a microwave heating technique.[41] At 2 wt% fructose concentration the HMF yield
was 73.4%, which was reduced to 54.3% when the fructose concentration was increased to 20
wt%, likely due to the formation of soluble and insoluble polymers (humins) at higher initial
4

fructose concentrations.[41] Ebitani and co-workers reported the synthesis of LA from fructose
using Amberlyst 15 at 393K obtaining 52% yield in water.[6][7–11] Amberlyst 70 was used as a
solid acid catalyst for hydrothermal decomposition of cellulose to LA by the group of Huber,
however, the LA selectivity was low (28% of theoretical).[46] An encouraging discovery was
reported by the group of Dumesic for conversion of cellulose to LA with Amberlyst 70 in a
mixture of 90% GVL and 10% water which produced 69% LA.[47] In recent studies they have
reported the use of Amberlyst-70 for the formation of HMF from fructose and glucose in GVL
and in THF which produced 80% and 25-32% HMF selectivity at 90% conversion.[48][49]
Recent studies have been published by the group of Horvath to use GVL as a green solvent to
convert Fructose, Glucose and HMF into LA in presence of H2SO4 catalyst.[50] Other solvent
systems have been reported previously for LA formation such as fluorinated solvents,[51,52]
acetone, water, supercritical carbon dioxide[53–55] and ionic liquids.[56–59] A recent
publication by the group of Li and co-workers have also highlighted the importance of solvent
polarity on the LA yield as well as polymer formation.[59]
In this work, we take advantage of high throughput facilities to explore a selection of
sulfonated polystyrene catalysts under multiple conditions of pressure, temperature, solvent and
reactant concentration. Among these polymeric acid catalysts, Dowex 50x8-100, cation
exchange resin with strongly acidic sulfonic groups has shown the most relevant activity for the
synthesis of LA from fructose in pure water and in selected co-solvents. We have also
demonstrated that very high activity and selectivity for LA could be achieved using Dowex
50x8-100 resin under optimized reaction conditions. We have explored GVL as a solvent which
was reported as the solvent of choice for solubilising char or humins formed during cellulose
decomposition and consequently improved biomass conversion.[47,60–62] In addition to being
derived from biomass, GVL is also an excellent solvent for the one-pot downstream process to
upgrade LA to GVL as demonstrated by Dumesic.[47] Mechanistic insight, derived from
identification of side-products formed during the dehydration of fructose to levulinic acid, is
discussed and reusability of the catalysts is explored.
2. Experimental
5

Catalysts and reactants were purchased from Sigma-Aldrich, Dow, or Rohm and Haas as
listed and used as supplied: Amberlyst® 15 cation exchange resin (Alfa Aesar), AmberlystTM 35
wet resin (Rohm and Haas), Amberlyst® BD20 (Rohm and Haas), AmberlystTM 70 (Rohm and
Haas), Dowex® 50x8 (hydrogen form, strongly acidic, 50-100 mesh, Sigma-Aldrich), Dowex®
50x8 (Hydrogen form, 50-100 mesh, Sigma-Aldrich), Dowex® 50x8 (hydrogen form, strongly
acidic, 20-50 mesh, Sigma-Aldrich), Dowex®MarathonTM MSC (hydrogen form, Sigma-Aldrich),
gamma-valerolactone (GVL, Sigma-Aldrich; 99%), water (LC-MS grade, Fisher Scientific,
Oakville, ON), D-fructose (Sigma-Aldrich; >99%), tetrahydrofuran, acetonitrile, methanol (LC-
MS grade, Fisher Scientific, Oakville, ON, Canada), trifluoroacetic acid (EMD Chemicals Inc.),
sulphuric acid (Sigma-Aldrich; 99.9%), 5-(hydroxymethyl) furfural (5-HMF, Sigma-Aldrich;
>99%), levulinic acid (Sigma-Aldrich; 99%), formic acid (Sigma-Aldrich; >95%) and used
directly without further purification.
Fructose conversion was monitored by disappearance of the reactant and appearance of
LA, formic acid (FA) and HMF products for a range of catalysts and reaction conditions as
described below. In all cases, we can define the conversion for the reaction based on the
disappearance of fructose (equation 1) and the yield of each product based on its own
quantification (equation 2). When necessary, selectivity can be determined by the ratio of yield
for a particular product to the conversion of fructose.
[
]
[
]
푓푟푢푐푡표푠푒 − 푓푟푢푐푡표푠푒
푖
퐶표푛푣푒푟푠푖표푛 (푚표푙%) =
푓 ∗ 100
(1)
(2)
[
]
푖
푓푟푢푐푡표푠푒
[
]
푠
푝푟표푑푢푐푡
푦푖푒푙푑 푝푟표푑푢푐푡 (푚표푙%) =
∗ 100
]
푖
[
푓푟푢푐푡표푠푒
Where, [fructose]i = molar concentration of initial fructose (reactant), [fructose]f = molar
concentration of fructose in the product; [product]s = molar concentration of species in the
reaction product.
2.1 Conversion of fructose to LA, FA and HMF using sulfonic acid-based heterogeneous
catalysts
2.1.1 Method A: Reactions were carried out using a Symyx High Throughput Module (Symyx
Discovery Tools). In a typical experiment, fructose (0.28 mmol, 0.05 g, 9 wt%) was added to a
6

mixture of GVL: H2O (50:50; 0.5 mL) in 1 mL glass vials equipped with magnetic stir bars and
loaded on the Core Module. To the resulting solution was added sulfonic acid-based solid acid
catalysts (0.05 g) in each of the corresponding vials placed in a 96-well plate. Reaction mixtures
were heated at temperatures ranging from 50 to 150 oC in different plates containing each of
these reaction mixtures with continuous stirring (estimated as 500 rpm) for 24 h. After the
reaction, 100 µL of the supernatant from each of the reaction vials were diluted with 900 µL of
LC-MS grade water followed by filtration through a 0.2 micron filter. The reactant fructose and
product amounts were determined by high performance liquid chromatography equipped with an
Aminex HPX-87H column (Bio-Rad Laboratory) equipped with DAD and ELS detectors.
2.1.2 Method B: Reactions were carried out in a thick-walled glass pressure reactor. Fructose
(0.28 mmol, 0.4g, 9 wt%) was added to a mixture of GVL: H2O (50:50; 4 mL) in the glass tube
and shaken well until a homogeneous solution was formed. To the resulting solution was added
Dowex 50x8-100 catalyst (0.4g) and the mixture heated at 120 oC with 600 rpm stirring for 24 h.
The reaction products were analyzed as in Method A.
2.1.3 Pressure Experiments: Reactions were carried out as in Method A above. Reaction
mixtures were pressurized with N2 gas from 100 to 500 psi for different plates and heated at 120
oC with continuous stirring for 24 h. The reaction products were analyzed as in Method A.
2.1.4 Catalyst Regeneration: Catalyst regeneration was done following a literature procedure.22
The used catalyst was filtered after each run and stirred in 10 mL of 30% H2O2 for overnight.
The solution was then filtered and the residual catalyst on the filter funnel was washed with
deionized water (5 mL x 3) followed by methanol (5 mL x 1). The catalyst beads were then dried
in an oven at 90 oC overnight before use for the next run.
2.1.5 HPLC DAD-ELSD Analysis: Separation and detection of target compounds were
developed on an HPLC-DAD-ELSD (Agilent Technologies, Santa Clara, CA, USA). The system
consisted of an Agilent 1100 series liquid chromatograph equipped with a diode array detector
and a 1200 series ELSD. The separations were performed on an Aminex HPX-87H column 300
mm × 7.8 mm (Bio-Rad, Hercules, CA, USA). Optimized mobile phase composition was: 0.1%
trifluoroacetic acid, 2% acetonitrile, 97.9% water delivered in 50 min. Column thermostat was
7

kept at 50 °C and flow rate was maintained at 0.6 mL/min. The column was equilibrated for 90
min before DAD was set at 210 nm and 270 nm (band width 4, reference off). ELSD conditions
were: temperature 30 °C, gain 7, pressure 3.4 bar. Quantification of the target compounds was
achieved using calibration curves based on the area under the peak using linear fits that bracket
the response of target peaks in the samples. In HPLC analysis, the typical relative standard error
was 1-2% for multiple injections from the same samples and 2-5% for replicate samples.
2.1.6 UPLC-Q-TOF Analysis: Accurate mass analysis was performed on a Xevo G2 Q-TOF
(Waters Milford, MT, USA). The system consisted of an Acquity UPLC chromatograph
equipped with a Xevo G2 Q-TOF. Separations were performed on an Acquity BEH C18, 1.7 μm
particle size, 2.1 × 50 mm column connected with a VanGuard Pre-column 2.1 mm× 5 mm
(Waters, Milford, MT, USA). The mobile phase was composed of water + 0.1% formic acid, B-
acetonitrile + 0.1% formic acid (Fisher Optima LC-MS). Flow rate, 0.8 mL/min, column
temperature, 50 °C, sample temperature 10 °C. Mobile phase composition, 100% A, isocratic in
0.5 min, 0.5-5 min 0-60% B, 5-5.5 min 60-100% B, 5.6-6 min 50-0% B, seal wash 5 min (total
run time 6 min). Sample injection conditions: 1 uL injection followed by a strong wash 200 μL
(90% acetonitrile + 10% water) and weak wash 600 uL (10% acetonitrile + 90% water). Q-TOF
analysis conditions: MassLynx software, MSe ESI+ mode, lock mass LeucineEnkephalin12C
556.2615, source temperature 120 °C, desolvation temperature 400 °C, Cone gas (N2) flow 50
L/hr, desolvation gas (N2) flow 1200 L/hr. MSe conditions, mass range 100-1500 Da, F1: CE
6V, F2CER 20-50V, Cone voltage 35V, Scan time 0.08 sec. Instrument calibration, 50-1000 Da
sodium formate.
3. Results and discussion
According to the reaction stoichiometry (Scheme 1), one molecule of fructose converts
into one molecule each of formic and levulinic acids.[12] However, for both homogeneous and
heterogeneous experiments, it is found that the FA yield is always higher than that of LA.
According to Deng, the intermediate formed after formation of FA is responsible for low LA
yield as it may lead to side reactions.[63] On the other hand, Zeng and co-workers reported that
the difference between FA and LA yield is a property of the specific catalyst which may lead to
8

decomposition of LA.[34] In either case, the difference between FA and LA confirms the
formation of by-products.
The problem in obtaining a high yield of LA (and HMF if only the dehydration reaction
is pertinent) arises from losses due to polymerization. The polymerization products are ill
defined and are believed to originate either from self-condensation of HMF or LA, or from their
cross-condensation reactions, resulting in brown soluble polymers and insoluble brown to black
‘humins’.[42,44] To increase the yield of LA, then, formation of polymerization products must
be tightly controlled.
3.1 Exploration of reaction space
As there are many variables that govern the conversion of fructose to levulinic acid, we
have taken a rapid throughput approach to explore the reaction space. In these experiments, vial
volume and stirring speeds are fixed by the instrument and a reaction time can be chosen for
comparison (typically 24 hr). The influence of the various components (substrate to catalyst
ratio, temperature, etc) is analyzed independently leading to an overall optimization of reactants
and reaction conditions. Under optimized conditions, mass transfer limitations are not applicable
as demonstrated by subsequent kinetic experiments.
3.1.1 Influence of substrate to catalyst ratio
Substrate to catalyst ratio is an important parameter to be examined for obtaining the
optimal yield of products. Ebitani and co-workers reported that a fructose to catalyst ratio of 1:1
was optimal for dehydration of fructose to LA using Amberlyst15 in pure water.[6–11] We have
examined the effect of catalyst ratio and fructose concentration with various polystyrene-based
sulfonic acid resin catalysts such as Amberlyst15, Amberlyst35, Amberlyst70, AmberlystBD20,
Dowex MSC and Dowex 50x8-100 in pure water. Amberlyst 35 has the highest nominal proton
exchange capacity as provided by the manufacturer, and Amberlyst 15 showed the highest value
as measure by a NaCl/NaOH titration method.[64] A full list of the exchange capacities can be
found in the supplementary information (Table ESI 1).
9

All of the catalysts examined showed a similar trend of activity and selectivity in terms of
catalyst loading with the substrate (fructose) concentration. Among these catalysts, Dowex 50x8-
100 showed the highest LA yield and selectivity (Table 1). Figure 1 shows that high catalyst
loading (fructose:catalyst = 1:1 and 1:2, weight ratio) affords highest reactivity and selectivity
towards LA formation (58 mol% at 99 mol% fructose conversion). For catalyst loadings below
1:1, such as 1:0.5 and 1:0.2, the LA yield gradually decreased as previously reported using
Amberlyst15.[6–11] As will be shown below, this suggests that sufficient catalyst is required to
prevent the accumulation of intermediates that will lead to polymerization and other by-products.
Further reducing the catalyst concentration to one tenth of the fructose concentration decreased
the LA yield to 28 mol% at 76 mol% of fructose conversion (Figure 1). However, at such high
ratios, the 5-HMF intermediate appears in substantial quantities and reaction is not complete
after 24 h. The ratio influences the required reaction time, as expected from basic kinetics and
shown below. With these results in hand, we considered a weight ratio of 1:1 as optimal for the
Dowex 50x8-100 resin and we have therefore used this ratio for most subsequent experiments.
3.1.2 Influence of Solvent
Solvent plays a vital role in the series of dehydration and rehydration events. Various
polystyrene-based sulfonic acid resin catalysts were chosen for producing LA from fructose
using pure water or using GVL, THF, acetonitrile (ACN), methanol (MeOH) and ethanol (EtOH)
as co-solvent in water. The reaction was first carried out at 0.5 mL scale using the Symyx High
Throughput instrument and was then scaled up to 4 mL and 20 mL. Using pure water as the
solvent, it is unavoidable to produce humins (solid products) with all the solid acid catalysts
investigated (Table 1). In pure water, most catalysts produced low LA yields as observed
previously.[6–11,47,65] The highest yield, 58 mol%, was observed with Dowex 50x8-100
catalyst under the specified reaction conditions (Table 1) which is however comparatively higher
than that previously reported using Amberlyst 15 in water.[6–11] Under the same reaction
conditions, in water/GVL solvent (50:50), only traces of solid products were observed with at
least 10-15 mol% increased LA yield compared to that in pure water. This behaviour could be
attributed to the swelling ability of resins in the presence of GVL, allowing increased diffusion
10

of reactants through the pores and consequently increasing the catalytic activity of the resins as
noted by Dumesic in the case of Amberlyst 70.[47] An alternative explanation is that the ability
of GVL to dissolve the reaction side-products prevents catalyst fouling, thereby allowing higher
yields.
An increased LA yield in water/GVL system was reported by Dumesic and co-workers
during cellulose degradation to LA.[47] Further increasing the amount of GVL to 70 and 90 wt%
resulted in lower LA yield but increased the solubility of undesirable solid products (Table 2).
The decrease in GVL concentration to 30 and 10 wt% did not significantly improve the LA yield
rather it increased solid products indicating that the appropriate amount of water/GVL mixture is
critical to improve LA yield and suppress humins formation.
Dowex 50x8-100 (a strongly acidic hydrogen form) resin produced the highest amount of
LA (72 mol%) in 24 hrs at 120 oC among the other resins under the specified reaction conditions
at 99.5% fructose conversion. All the Amberlyst catalysts investigated showed at least 10 mol%
increased LA yield with reduced solid side-products in water/GVL (50:50) mixture compared to
pure water, producing ~60 mol% LA at 99% fructose conversion (Table 1). A blank experiment
was performed in the absence of catalyst in 50:50 water/GVL mixtures under the specified
reaction conditions, which showed formation of 7 mol% of HMF as a result of the thermal
decomposition of fructose (Table 1). Minimal conversion of fructose (20 mol%) and no detection
of FA and LA during the HPLC-DAD-ELSD analysis suggests formation of a small amount of
by-products along with 5-HMF upon thermal decomposition of fructose. This finding also
suggests that GVL does not undergo decomposition to form LA under the specified reaction
conditions.
The highest catalytic activity shown by the Dowex 50x8-100 strongly acidic hydrogen
form resin prompted us to further investigate this catalyst in different solvents. With 50:50
water/THF mixture, LA yield was comparable (72 mol%) with that of water/GVL (50:50) at 98
mol% fructose conversion. Dumesic and co-worker recently reported the high yield of LA
production (>70%) with furfural alcohol in THF-H2O using H-ZSM-5 as catalyst.[35,36]
Literature precedent describes that Sn-W mixed oxide and Sn-Beta/HCl catalyst systems
produced high HMF yields in water/THF solvent at 393 K.[66][67] Interestingly, it was observed
11

that a 50:50 mixture of water/acetonitrile produced a higher yield of HMF (66 mol%) at 99
mol% fructose conversion with only 15 mol% LA using Dowex 50x8-100 (Table 1). The Dowex
50x8-100 thus behaves as a switchable catalyst for LA and HMF production depending on the
nature of solvent. The preference for HMF in water/ACN solvent may be related to the ability of
acetonitrile to stabilize HMF. This would be consistent with previously reported results that polar
aprotic solvents like DMSO favour HMF by strongly binding with its hydroxyl and carbonyl
functions during fructose dehydration.[68] With these results in hand, we considered the Dowex
50x8-100 resin as the best resin catalysts in water/GVL (50:50) mixture and used this solid acid
catalyst for further investigations.
The catalytic behaviour of Dowex 50x8-100 in water/alcohol solvent was also
investigated for fructose dehydration. In 50:50 mixtures of water/methanol and water/ethanol
solvent systems only 28 mol% and 36 mol% of LA was observed at 99 mol% fructose
conversion (Table 1). Comparatively low LA yield with both water/methanol and water/ethanol
mixtures was attributed to the esterification of LA as well as acetal and ether formation of HMF
promoted by alcohol.[21,69–71] Under our reaction conditions, the use of methanol produced
methyl levulinate (see electronic supplementary information, ESI, Fig. 1) and ethanol produced
ethyl levulinate (ESI, Fig. 2a) as well as 5-ethoxymethylfurfural (ESI, Fig. 2b) as detected by
GC-MS. Although these derivatives of LA and HMF are valuable both for commercial use and
mechanistic insight, further optimization and quantification of the products were not undertaken
in this work.
3.1.3 Influence of fructose concentration
The absolute concentration of fructose is another important parameter affecting the LA
yield. From the initial screening reaction with 1 to 50 wt% of fructose examined using Dowex
50x8-100 in the water/GVL solvent system, it was found that increasing fructose concentration
led to a gradual decrease in LA product yield (Figure 2). When the fructose concentration was as
low as 1 wt%, all of the fructose was converted, resulting in 92 mol% of FA and 75 mol% of LA.
Since there was no 5-HMF measurable, the significant difference between FA and LA could
suggest the preferential decomposition of LA for formation of side-products or a secondary
mechanism for formation of FA at 1 wt% fructose compared to the 5 and 9 wt% of fructose. At
12

50 wt% fructose, however, only 59 mol% of LA was observed at >98% fructose conversion with
13 mol% HMF remaining. These observations led to an exploration of the reaction kinetics.
Nevertheless, from the experimental data in Figure 2, 9 wt% of fructose concentration is optimal
for our reaction conditions.
3.1.4 Kinetic optimization
Kinetic experiments were performed on lab scale (20 mL) with refluxing conditions
under nitrogen for 48 h (as opposed to the high throughput experiments that used 0.5 mL volume
for 24 h with the Symyx instrument). During the kinetic experiment, aliquots were taken at
regular time intervals and each aliquot was analysed for composition. The total molar
concentration, fructose to catalyst ratio, solvent mixture (water/GVL 50:50) and temperature
were kept the same in both experiments. Figure 3 shows selected HPLC traces at various time
intervals and the corresponding reaction profile as a function of time for the primary components
of the reaction for fructose conversion to LA using Dowex 50x8-100 (H+ form) resin. Fructose
conversion increases rapidly whereas LA and FA yield increased slowly as the reaction
progressed. The conversion of fructose goes hand in hand with the generation of the HMF
intermediate which reaches a maximum (63-64 mol%) around 4-6 h of reaction. At about 8 h
more than 99 mol% of fructose has been converted under the reaction conditions. After 33 h,
HMF gradually decreased to 9 mol%, at which time LA yield reached a maximum of 78 mol%.
At longer reaction times up to 48 h, no significant improvement in LA yield was observed. This
reaction profile is consistent with Scheme 1.
Figure 3a shows the appearance of additional components after approximately 5 hrs of
reaction (marked as * in the HPLC traces). Analysis of these components by accurate mass
spectrometry reveals condensation products A-C (Scheme 2) and higher molecular weights
consistent with polymerization. Compound A (ESI, Fig 3a) has been reported previously.[72]
Compound B (ESI, Fig 3b) is a prime example of side reactions between products and the HMF
intermediate and could be prevented under conditions where the products are removed from the
reaction vessel. On the other hand, compound C (ESI, Fig 3c) shows the onset of polymerization
of the HMF intermediate, which likely leads to the insoluble polymers that are traditionally
observed in this reaction.[12,13,15–19,38,42,72,73] As an example, Lund et al. reported this
13

HMF self-polymerization using IR spectroscopy. However, using accurate mass UPCL-Q-TOF
analysis we were able to identify the products of cross aldol condensation/addition between
HMF and LA, as well as the self-aldol condensation/addition of HMF and/or fructose. Further
mechanistic studies are underway.
The reaction temperature and residence time had a large effect on the fructose conversion
and LA yield. To better understand the correlation between reaction rates as a function of
temperature, further experiments were carried out at different times and temperatures using
Dowex 50x8-100 catalyst (Table 3). When the reaction temperature was 80 oC, fructose
conversion was 96 mol% and LA yield was only 54 mol% with the presence of 20 mol% of
HMF even after 120 h. This suggests that at lower temperature hydration of HMF to LA was
slower even if the reaction was allowed to take place for 5 days. When the temperature was
increased to 100 oC, fructose conversion and LA yield increased to 99 and 65 mol%, respectively
after 72 h. Conversion of fructose and LA yield reached 99.5 and 72 mol%, respectively, after 24
h at 120 oC. In the latter reaction only 0.2 mol% HMF remained. As the temperature was
increased to 160 oC, fructose conversion and LA yield reached 99.4 and 68 mol% respectively
after just 3 h. At 200 oC, the rate of fructose conversion was much higher reaching the maximum
(100 mol%) with correspondingly high LA yield, 76 mol% after 1 h (Table 3). Therefore, higher
temperature with shorter residence time is beneficial for high catalytic activity. However,
considering the stability of the strong acid sulfonated copolymer resin which might not be stable
for repeated runs at temperatures above 150 oC,17b,20 we chose our optimal reaction temperature
and time as 120 oC and 24 h in this study.
The reactions clearly take place in the liquid phase. Nevertheless, for consistency the
effect of pressure over the liquid was briefly explored. The effect of pressure (nitrogen) on the
conversion of fructose to LA was examined at 100 to 500 psi nitrogen pressure at 120 oC over 24
h under our optimal reaction conditions. At lower pressures of 100 and 200 psi there was a
minimal effect on catalytic activity of the Dowex-50x8-100 resin on LA yields (70 and 72 mol%
respectively). At 300 psi there was a slight decrease in LA yield (69 mol%) and slight increase in
FA yield (87 mol%). Further decrease in LA yield (66 mol%) was observed at 500 psi with 2
14

mol% HMF remaining unconverted. However, fructose conversion remained constant (> 99
mol%) at all pressures. These results suggest that higher pressures are not beneficial for LA yield
with this catalyst under our reaction conditions.
3.2 Focus on the hydration step: HMF as starting feed
Using HMF as a starting material, all of the polystyrene-based sulfonic acid catalysts
showed high yield of LA formation in pure water (Table 4). Dowex 50x8-100 and Dowex-MSC
(macro-reticular) resins showed 90 mol% of LA yield under the specified reaction conditions.
The highest LA yield (92 mol%) was observed with Amberlyst BD20, whereas Amberlyst15 and
Amberlyst35 gave 86 and 83 mol% of LA respectively. The lowest catalytic activity (63 mol%)
was shown by Amberlyst70, which exists in the chloride form resin rather than the proton form,
suggesting the importance of having a proton (strongly acidic property) to assist the hydration of
HMF to LA.
All of these polystyrene based sulfonic acid catalysts showed high activity of LA production by
direct hydration of HMF in pure water under our optimized conditions compared to fructose as
the initial feed. This also suggests the ability of these heterogeneous catalysts to stabilize HMF
under our reaction conditions and also the reduction of by-product formation. In this vein,
Nikolakis and Vlachos studied the H-BEA-18 Zeolite and indicated the preferential adsorption of
HMF on the zeolite surface over sugars and thereby produced higher LA yield from HMF
compared to fructose under mildly acidic conditions.[74]
3.3 Catalyst recycling
The reusability of the catalyst is an important parameter to be considered when using
heterogeneous catalysts in the production of LA at commercial scale.[34,66,67] In figure 4a, the
catalyst was recovered after each run and treated with 5 mL of water (3X), followed by 5 mL of
acetone (2X) to remove residual organic compounds and then dried in an oven overnight at 80
oC.[12,13,16,18,19,73,75] The reused catalyst, Dowex 50x8-100, showed a progressive decrease
in catalytic selectivity, affording just 37 mol% LA after the 5th run, although high fructose
conversion (> 99 mol%) was maintained (Figure 4a). This decrease in selectivity is attributed to
15

formation of residuals such as humins deposited onto the catalyst surface, thereby clogging the
active sites for reaction.
To remove humins and other side-products that reduced LA yield, we examined the
regeneration of the catalyst after every run. The catalytic activity was recovered by washing the
used catalyst overnight with 30% aqueous H2O2 at room temperature followed by washing with
methanol and water (figure 4b).[47] After regeneration, catalytic selectivity was revived but a
small decrease in LA yield persisted, from 61 mol% to 55 mol% after 6 runs (Figure 4b).
Fructose conversion was > 99 mol% in every run. Interestingly, the FA yield was slightly
increased up to the 4th cycle. The difference in LA and FA yield, however, suggests formation of
other by-products. In addition, the regaining of catalytic selectivity after regeneration suggests
that catalyst deactivation is not due to a real degradation of the catalyst, but rather to clogging of
the catalyst active sites due to humins formation.[47] Elemental analysis (C, H, S) of used
Dowex resin after fructose dehydration to HMF performed by Smith and co-workers, showed an
increased C (4.6%) and H (0.94%) and decreased S content by 1.3%.20 The authors’
interpretation was a likely increase in carbon content due to humins, although slight catalyst
degradation by loss of sulfonic groups could not be ruled out. As such, it was thought that even
at 150 oC for 10 minutes of microwave irradiation, the Dowex resin decomposition was
minimal.[41] These results would indicate that under our reaction conditions (lower temperature)
the catalysts remains stable and any observed decrease in performance is due to clogging.
The difference in LA yield shown in Figure 4 and Table 1 was due to the different
conditions that were used in the reactions. Catalyst recycling reactions were carried out in a
pressure glass reactor in 4 mL scale. In this reaction condition since the reaction volume was
larger, hydration of HMF to LA was not complete after 24 h, consequently, the LA yield shown
in Figure 4 is less than the that shown in Table 1, where the reaction was carried out in 0.5 mL
scale with the Symyx/FreeSlate high-throughput instrument.
3.3.1 Influence of particle size
In this work, we have explored a selection of sulfonated polystyrene catalysts. Of those,
Dowex 50x8-100 afforded the highest LA yields. Next, we explored the various particle sizes
available for commercial samples of this catalyst. Table 5 shows that both larger and smaller
16

particle sizes led to decreased LA yields relative to the 100 mesh (149 micron) catalyst. This
observation suggests that both surface area and accessibility of acid sites are contributing to high
LA yields. This information could be useful for designing new green catalysts containing higher
surface area and more accessible acidic moieties.[76,77]
4. Conclusions
We have demonstrated the use of Dowex 50x8-100 resin as an efficient heterogeneous
catalyst for the production of LA in yields up to 77 mol% under optimized reaction conditions.
After regeneration, the catalyst can be reused with only nominal decrease in reactivity and
selectivity as demonstrated here for five reaction cycles. Some of the soluble side-products were
identified as aldol and cross-aldol condensation products of LA and HMF by using Q-TOF mass
spectrometry based on accurate mass.
Our study provides some important insight for future directions. First this work provides
more flexibility than dual catalyst systems reported previously in which one catalyst affords high
HMF yield from fructose in the first stage and the second catalyst yields quantitative LA from
HMF in the second stage. Secondly, in order to minimize polymerization products and increase
the LA or HMF yield, catalyst residence time is critical. One way to achieve this goal is to design
a flow reactor that would allow for the collection of LA as it is formed. This is one of the main
advantages of using heterogeneous catalysts which not only have the capability for catalyst reuse
but also may control to some extent, the aldol chemistry that would otherwise result in larger
oligomeric products (humins).
Acknowledgements
Financial assistance for this work was provided by the Natural Sciences and Engineering
Research Council of Canada (NSERC) and Segetis, Inc. The authors are grateful for ongoing
support of the Centre for Catalysis Research and Innovation (CCRI) by the University of Ottawa,
the Canada Foundation for Innovation and the Ontario Ministry of Economic Development and
17

Innovation. Discussions with CCRI high-throughput facility manager Roxanne Clément are also
greatly appreciated.
References:
[1] D.M. Alonso, J.Q. Bond, J.A. Dumesic, GREEN Chem. 12 (2010) 1493–1513.
[2] Y. Roman-Leshkov, J.N. Chheda, J.A. Dumesic, Science (80-. ). 312 (2006) 1933–1937.
[3] G. Werpy, T. and Petersen, T. Werpy, G. Petersen, Other Inf. PBD 1 Aug 2004 (2004)
Medium: ED; Size: 76 pp. pages.
[4] J.J. Bozell, G.R. Petersen, GREEN Chem. 12 (2010) 539–554.
[5] L.E. Manzer, in:, Bozell, JJ and Patel, MK (Ed.), Feed. Futur. RENEWABLES Prod.
Chem. Mater., 2006, pp. 40–51.
[6] P.A. Son, S. Nishimura, K. Ebitani, React. Kinet. Mech. Catal. 106 (2012) 185–192.
[7] D.C. Elliot, J.G. Frye, Hydrogenated 5-Carbon Compound and Method of Making,
US005883266A, n.d.
[8] M.A. Lake, S.W. Burton, W.C. Fuller, R. Sasser, M.E. Lindstrom, J.T. Wheless,
Oroduction of Levulinic Acid and Levulinate Esters from Biomass, US 2010/0312006 A1,
2010.
[9] A. Demolis, N. Essayem, F. Rataboul, ACS Sustain. Chem. Eng. 2 (2014) 1338–1352.
[10] L.E. Manzer, Production of 5-Methylbutyrolactone from Levulinic Acid, 6617464B2,
2003.
[11] M. Chia, J.A. Dumesic, Chem. Commun. 47 (2011) 12233–12235.
[12] J. HORVAT, B. KLAIC, B. METELKO, V. SUNJIC, TETRAHEDRON Lett. 26 (1985)
2111–2114.
[13] T. Runge, C. Zhang, Ind. Eng. Chem. Res. 51 (2012) 3265–3270.
[14] J.A. Dumesic, S.G. Wettstein, D.M. Alonso, E.I. Guerbuz, Solute Enhanced Production of
Gamma Valerolactone from Aqueous Solutions of Levulinic Acid, 20120302766A1,
2012.
[15] 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.
[16] D.J. Hayes, S. Fitzpatrick, M.H.B. Hayes, J.R.H. Ross, in:, Biorefineries-Industrial
Process. Prod., Wiley-VCH Verlag GmbH, Weinheim, Germany, n.d., pp. 139–164.
[17] B. Girisuta, L.P.B.M. Janssen, H.J. Heeres, Ind. Eng. Chem. Res. 46 (2007) 1696–1708.
[18] S.W. Fitzpatrick, Production of Levulinic Acid from Carbohydrate Containing Materials,
5608105, n.d.
[19] S.W. Fitzpatrick, Lignocellulose Degradation to Furfural and Levulinic Acid, 4897497,
n.d.
[20] Y. Zhang, J. Wang, J. Ren, X. Liu, X. Li, Y. Xia, G. Lu, Y. Wang, Catal. Sci. Technol. 2
(2012) 2485–2491.
[21] 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.
[22] W.A. Farone, J.E. Cuzens, Method for the Production of Levulinic Acid and Its
Derivatives, 6054611, n.d.
18

[23] J.Y. Cha, M.A. Hanna, Ind. Crops Prod. 16 (2002) 109–118.
[24] Q. Fang, M.A. Hanna, Bioresour. Technol. 81 (2002) 187–192.
[25] N.S. Mosier, C.M. Ladisch, M.R. Ladisch, Biotechnol. Bioeng. 79 (2002) 610–618.
[26] B. Girisuta, L. Janssen, H.J. Heeres, Chem. Eng. Res. Des. 84 (2006) 339–349.
[27] C. Chang, P. Cen, X. Ma, Bioresour. Technol. 98 (2007) 1448–1453.
[28] L. Yan, N. Yang, H. Pang, B. Liao, CLEAN-SOIL AIR WATER 36 (2008) 158–163.
[29] A.J. Sanborn, Preparation of Levulinic Acid from Fructose Using an Acid Catalyst, a
Polyethylene Glycol and an End-Capped Polyethylene Glycol, 2233477A1, 2013.
[30] B.D. Mullen, D.J. Yontz, C.M. Leibig, Process to Prepare Levulinic Acid,
PCT/US2014/025581, 2014.
[31] J. JOW, G.L. RORRER, M.C. HAWLEY, D.T.A. LAMPORT, BIOMASS 14 (1987)
185–194.
[32] K. Lourvanij, G.L. Rorrer, J. Chem. Technol. Biotechnol. 69 (1997) 35–44.
[33] K. LOURVANIJ, G.L. RORRER, Ind. Eng. Chem. Res. 32 (1993) 11–19.
[34] W. Zeng, D. Cheng, H. Zhang, F. Chen, X. Zhan, React. Kinet. Mech. Catal. 100 (2010)
377–384.
[35] M.A. Mellmer, J.M.R. Gallo, D.M. Alonso, J.A. Dumesic, ACS Catal. 5 (2015) 3354–
3359.
[36] I. Guzman, A. Heras, M.B. Guemez, A. Iriondo, J.F. Cambra, J. Requies, Ind. Eng. Chem.
Res. 55 (2016) 5139–5144.
[37] K. LOURVANIJ, G.L. RORRER, Appl. Catal. A-GENERAL 109 (1994) 147–165.
[38] P. Wang, S. Zhan, H. Yu, in:, Liang, J and Wang, L (Ed.), Adv. Ecol. Environ. Funct.
Mater. ION Ind., 2010, pp. 183–187.
[39] C. Hongzhang, Y. Bin, J. Shengying, Bioresour. Technol. 102 (2011) 3568–3570.
[40] M.A. Harmer, Q. Sun, Appl. Catal. A-GENERAL 221 (2001) 45–62.
[41] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith Jr., GREEN Chem. 10 (2008) 799–805.
[42] B.F.M. KUSTER, STARCH-STARKE 42 (1990) 314–321.
[43] V. Heguaburu, J. Franco, L. Reina, C. Tabarez, G. Moyna, P. Moyna, Catal. Commun. 27
(2012) 88–91.
[44] X. Hu, S. Kadarwati, S. Wang, Y. Song, M.D.M. Hasan, C.-Z. Li, FUEL Process.
Technol. 137 (2015) 212–219.
[45] R.A. SCHRAUFNAGEL, H.F. RASE, Ind. Eng. Chem. Prod. Res. Dev. 14 (1975) 40–44.
[46] R. Weingarten, W.C. Conner Jr., G.W. Huber, ENERGY Environ. Sci. 5 (2012) 7559–
7574.
[47] D.M. Alonso, J.M.R. Gallo, M.A. Mellmer, S.G. Wettstein, J.A. Dumesic, Catal. Sci.
Technol. 3 (2013) 927–931.
[48] E.I. Guerbuez, J.M.R. Gallo, D.M. Alonso, S.G. Wettstein, W.Y. Lim, J.A. Dumesic,
Angew. CHEMIE-INTERNATIONAL Ed. 52 (2013) 1270–1274.
[49] J.M.R. Gallo, D.M. Alonso, M.A. Mellmer, J.A. Dumesic, GREEN Chem. 15 (2013) 85–
90.
[50] L. Qi, Y.F. Mui, S.W. Lo, M.Y. Lui, G.R. Akien, I.T. Horvath, ACS Catal. 4 (2014)
1470–1477.
[51] H. Heeres, R. Handana, D. Chunai, C.B. Rasrendra, B. Girisuta, H.J. Heeres, GREEN
Chem. 11 (2009) 1247–1255.
[52] C. Chun, M. Xiaojian, C. Peilin, CHINESE J. Chem. Eng. 14 (2006) 708–712.
19

[53] A. Boisen, T.B. Christensen, W. Fu, Y.Y. Gorbanev, T.S. Hansen, J.S. Jensen, S.K.
Klitgaard, S. Pedersen, A. Riisager, T. Stahlberg, J.M. Woodley, Chem. Eng. Res. Des. 87
(2009) 1318–1327.
[54] M. Bicker, J. Hirth, H. Vogel, GREEN Chem. 5 (2003) 280–284.
[55] W. Sangarunlert, P. Piumsomboon, S. Ngamprasertsith, KOREAN J. Chem. Eng. 24
(2007) 936–941.
[56] S. Hu, Z. Zhang, Y. Zhou, B. Han, H. Fan, W. Li, J. Song, Y. Xie, GREEN Chem. 10
(2008) 1280–1283.
[57] C. Moreau, A. Finiels, L. Vanoye, J. Mol. Catal. A-CHEMICAL 253 (2006) 165–169.
[58] H. Ren, Y. Zhou, L. Liu, Bioresour. Technol. 129 (2013) 616–619.
[59] X. Hu, S. Wang, R.J.M. Westerhof, L. Wu, Y. Song, D. Dong, C.-Z. Li, FUEL 141 (2015)
56–63.
[60] I.T. Horvath, H. Mehdi, V. Fabos, L. Boda, L.T. Mika, GREEN Chem. 10 (2008) 238–
242.
[61] L. Qi, I.T. Horvath, ACS Catal. 2 (2012) 2247–2249.
[62] D. Fegyverneki, L. Orha, G. Lang, I.T. Horvath, Tetrahedron 66 (2010) 1078–1081.
[63] L. Deng, J. Li, D.-M. Lai, Y. Fu, Q.-X. Guo, Angew. CHEMIE-INTERNATIONAL Ed.
48 (2009) 6529–6532.
[64] P.F. Siril, H. Griffiths, D. Brown, J. Mol. Catal. A Chem. 279 (2008) 63–68.
[65] J. Hegner, K.C. Pereira, B. DeBoef, B.L. Lucht, TETRAHEDRON Lett. 51 (2010) 2356–
2358.
[66] E. Nikolla, Y. Roman-Leshkov, M. Moliner, M.E. Davis, ACS Catal. 1 (2011) 408–410.
[67] K. Yamaguchi, T. Sakurada, Y. Ogasawara, N. Mizuno, Chem. Lett. 40 (2011) 542–543.
[68] G. Tsilomelekis, T.R. Josephson, V. Nikolakis, S. Caratzoulas, ChemSusChem 7 (2014)
117–126.
[69] J. Liu, Y. Tang, K. Wu, C. Bi, Q. Cui, Carbohydr. Res. 350 (2012) 20–24.
[70] X. Hu, C.-Z. Li, GREEN Chem. 13 (2011) 1676–1679.
[71] G.-Z. Xu, C. Chang, W.-N. Zhu, B. Li, X.-J. Ma, F.-G. Du, Chem. Pap. 67 (2013) 1355–
1363.
[72] G.R. Akien, L. Qi, I.T. Horvath, Chem. Commun. 48 (2012) 5850–5852.
[73] D.M. Alonso, S.G. Wettstein, M.A. Mellmer, E.I. Gurbuz, J.A. Dumesic, ENERGY
Environ. Sci. 6 (2013) 76–80.
[74] J.S. Kruger, V. Choudhary, V. Nikolakis, D.G. Vlachos, ACS Catal. 3 (2013) 1279–1291.
[75] B.M. Choudary, S. Madhi, N. Chowdari, M.L. Kantam, B. Sreedhar, J. Am. Chem. Soc.
124 (2002) 14127–14136.
[76] J. Potvin, E. Sorlien, J. Hegner, B. DeBoef, B.L. Lucht, TETRAHEDRON Lett. 52 (2011)
5891–5893.
[77] H.N. Lim, M.A. Yarmo, N.M. Huang, P.S. Khiew, W.S. Chiu, J Phys Sci 20 (2009) 23–
36.
20

100
80
60
40
20
0
100
80
60
40
20
0
(1:2)
(1:1)
(1:0.5)
(1:0.2)
(1:0.1)
Fructose to catalyst ratio
Figure 1: Effect of catalyst concentration with respect to fructose on fructose dehydration to LA in water.
Reaction conditions: fructose (9 wt%, 0.28 mmol, 0.05 g), H2O (0.5 mL), 120 oC, 24 h. Fructose
conversion, Formic acid yield, Levulinic acid yield, 5-HMF yield.
100
80
60
40
20
0
100
80
60
40
20
0
1
5
9
20
30
40
50
Fructose Concentration (Wt %)
Figure 2: Effect of absolute fructose concentration on fructose conversion to LA. Reaction
conditions: fructose:catalyst weight ratio (1:1), H2O/GVL (50:50; 0.5 mL), 120 oC, 24 h. Fructose
conversion, Formic acid yield, Levulinic acid yield, 5-HMF yield.
21

Figure 3: a) Formation of the reaction products and by-products as a function of time. HPLC
chromatograms monitored by UV absorption at 210 nm. unknown by-products, F = Formic acid, L=
* =
Levulinic acid (more products were being formed as a function of time), G= GVL, H= HMF (peaks are
observed to grow to a maximum and then completely disappear as a function of time). b) Dehydration of
fructose to 5-HMF and hydration of HMF to LA as a function of time with Dowex 50x8-100. Fructose
conversion, Formic acid yield, Levulinic acid yield, 5-HMF yield. Reaction conditions: fructose
(9 wt%, 0.28 mmol), H2O/GVL (50:50, 20 mL), fructose:catalyst (1:1), 120 oC.
22

(a)
100
80
60
40
20
0
1 (fresh)
2
3
4
5
Recycle times
(b)
100
80
60
40
20
0
1 (fresh)
2
3
4
5
6
Recycle Times
Figure 4: (a) catalyst recycle after acetone/water wash. (b) catalyst recycle after H2O2/MeOH
regeneration. Reaction conditions: Fructose (9 wt %, 0.28 mmol), H2O/GVL (50:50, 4 mL), 24 h, 120
oC. Fructose conversion, Formic acid yield, Levulinic acid yield, 5-HMF yield.
23

HO
OH
OH
O
O
O
- 3H2O
H+
+ 2H2O
H+
H
+
H
HO
H
C
O
HO
OH
O
H
O
O
O
Formic Acid
Levulinic Acid
D-Fructose
5-HMF
Polymers and humins
Scheme 1: Reaction scheme of fructose dehydration to LA, FA and by-products.
OH
O
OH
OH
O
HO
HO
- H2O/H+
O
O
2
OH
HO
HO
O
OH
HO
OH
A
O
O
O
O
O
O
HO
H
O
O
-H2O
HO
+
OH
HO
OH
O
OH
O
O
B
O
O
O
O
HO
O
2
O
H
- H2O/H+
O
C
Scheme 2: Reaction scheme for the formation of by-products. A) α-D-fructofuranose-β-D-fructofuranose
1,2’:2’3-dianhydrate[66,67]; B) 4-acetyl-5-(-(hydroxymethyl)furan-2-yl)-dihydrofuran-2(3H)-one; C)
Bis(5-methylfurfural)ether.
24

Table 1. Comparison of LA and FA formation from fructose in pure water, (50:50) water/GVL,
water/THF and water/ACN mixtures using polystyrene based sulfonic acid resin catalysts.
Catalysts
Solvents
Fructose
conversion
(%)
Formic acid Levulinic acid 5-HMF
(mol%)
(mol%)
(mol%)
Amberlyst 15
Water
99
62
71
55
65
0.4
3.0
water/GVL*
99
99
99
99
99
99
99
99
99.5
Amberlyt 35
Water
45
64
49
65
51
63
60
77
42
63
43
62
46
59
58.0
72
0.4
5.0
0.1
6.0
0.5
15
water/GVL*
Water
Amberlyst 70
water/GVL*
Water
Amberlyst BD 20
Dowex 50x8-100
water/GVL*
Water
0.07
0.2
water/GVL*
water/THF*
water/ACN*
98
99
78
22
72
15
0.5
66.0
Water/MeOH*
Water/EtOH*
Water
99
99
99
99
20
51
57
52
69
0
28
36
49
64
0
0.1
0.1
0.6
6.8
7.0
Dowex MSC
Blank
water/GVL*
water/GVL*
Reaction conditions: Fructose (9 wt%, 0.28 mmol), fructose:catalyst (1:1), solvent volume (0.5 mL), 120
oC, 24 h; *(50:50) mixture of corresponding solvent with water.
25

Table 2: Effect of water:GVL ratio in fructose dehydration to LA using Dowex 50x8-100.
Water:GVL Fructose
Formic acid
Levulinic acid
(mol %)
5-HMF
(mol %)
Conversion (%) (mol %)
90:10
99.9
44
38
0
70:30
50:50
30:70
10:90
99.9
99.5
99.0
99.0
67
77
73
73
62
72
70
68
0.1
0.1
0.1
0.1
Reaction conditions: fructose (9 wt%, 0.28 mmol), fructose:catalyst weight ratio (1:1), H2O/GVL (0.5
mL), 120 oC, 24 h.
Table 3: Correlation between reaction rate and temperature to yield LA from fructose using Dowex 50x8-
100 (strongly acidic hydrogen form) resin.
Temperature
(oC)
Time
(h)
Fructose
conversion (%)
Formic acid
(mol%)
Levulinic acid 5-HMF
(mol%)
(mol%)
80
120
72
96
99
62
74
54
65
20
100
1.0
120
160
200
24
3
>99
>99
100
77
71
90
72
68
76
0.2
1.0
2.0
1
Reaction conditions: fructose (9 wt%, 0.28 mmol), fructose:catalyst (1:1), H2O/GVL (50:50; 0.5 mL).
26

Table 4: Production of LA from HMF in pure water using polystyrene-based sulfonic acid
catalysts.
Catalysts
Formic acid
(mol%)
Levulenic acid 5-HMF
(mol%)
(mol%)
Amberlyst 15
Amberlyst 35
Amberlyst 70*
Amberlyst BD20
Dowex 50x8-100
Dowex MSC
100
98
86
83
63
92
90
90
0
0
74
0.2
2
100
100
100
0
0
Reaction conditions: HMF (3.3 wt %, 0.21 mmol), H2O (0.5 mL), 120 oC, 24h, HMF:catalyst (1:1).
*chlorine form; rest of the catalysts were hydrogen form.
Table 5: Effect of particle sizes in LA production using Dowex-50x8 catalysts.
Catalysts
Fructose
conversion
(mol%)
Formic acid
(mol%)
Levulinic acid 5-HMF
(mol%)
(mol%)
Dowex 50x8-20
Dowex 50x8-100
Dowex 50x8-200
99.5
99.5
99.5
71
77
74
63
72
66
3.0
0.2
2.0
Reaction conditions: fructose (9 wt%, 0.28 mmol), fructose:catalyst (1:1), H2O/GVL (50:50; 0.5 mL),
120 oC, 24h.
27