Levulinic Acid as a Catalyst for the Production of 5-Hydroxymethylfurfural and Furfural from Lignocellulose Biomass

Article abstract of DOI:10.1002/cctc.201501105

Levulinic acid (LA) was used as a catalyst for the first time to produce 5-hydroxymethylfurfural (5-HMF) and furfural (FAL) from pinewood and eucalyptus sawdust in a mono- or biphasic solvent system. 2-Methyltetrahydrofuran was used as a co-solvent with water in different ratios and temperatures (140-200 °C). Highest yields of 5-HMF and FAL were obtained at 180 °C and 2 h reaction time; however, at 160 °C, high yields of C₆ and C₅ sugars were obtained. Both hydrolysis and dehydration steps were accelerated in the MTHF/water biphasic system compared to pure aqueous phase. In particular, 1:2 w/w ratio of MTHF/water resulted in the highest yield of 5-HMF and FAL, whereas 2:1 w/w ratio showed highest yield of C₆ and C₅ sugars. Increasing the ratio of MTHF/water resulted in a higher fraction of dehydrated products extracted into the organic phase. LA as a catalyst is beneficial because it is miscible in both the phases and the presence of LA favours the equilibrium towards 5-HMF production.

Full text of DOI:10.1002/cctc.201501105

DOI: 10.1002/cctc.201501105
Full Papers
Levulinic Acid as a Catalyst for the Production of
5-Hydroxymethylfurfural and Furfural from Lignocellulose
Biomass
Bhogeswararao Seemala, Victoria Haritos, and Akshat Tanksale*[a]
Levulinic acid (LA) was used as a catalyst for the first time to
produce 5-hydroxymethylfurfural (5-HMF) and furfural (FAL)
from pinewood and eucalyptus sawdust in a mono- or biphasic
solvent system. 2-Methyltetrahydrofuran was used as a co-sol-
vent with water in different ratios and temperatures (140–
2008C). Highest yields of 5-HMF and FAL were obtained at
1808C and 2 h reaction time; however, at 1608C, high yields of
C6 and C5 sugars were obtained. Both hydrolysis and dehydra-
tion steps were accelerated in the MTHF/water biphasic
system compared to pure aqueous phase. In particular, 1:2 w/
w ratio of MTHF/water resulted in the highest yield of 5-HMF
and FAL, whereas 2:1 w/w ratio showed highest yield of C6
and C5 sugars. Increasing the ratio of MTHF/water resulted in
a higher fraction of dehydrated products extracted into the or-
ganic phase. LA as a catalyst is beneficial because it is miscible
in both the phases and the presence of LA favours the equilib-
rium towards 5-HMF production.
Introduction
Depletion of petrochemical resources coupled with growing
concerns about climate change has prompted investigation of
lignocellulosic biomass utilisation for the production of value-
added chemicals and biofuels.[1] Lignocellulose biomass is an
attractive feedstock because it is the most abundant renew-
able carbon resource. However, pre-treatment of lignocellulose
remains one of the main challenges for the conversion of
lignocellulose into desirable products.
liquid alkanes respectively, through two-step process such as
furan ring opening followed by hydrodeoxygenation reac-
tion.[4c,5b,7] The process for the production of FAL from xylose is
already industrialised. However, the production of 5-HMF from
glucose is more difficult because this step requires isomeriza-
tion of glucose molecule into fructose followed by dehydration
of fructose into 5-HMF, which is limited by equilibrium conver-
sion at both the steps.[8] Moreover, 5-HMF may rehydrate into
levulinic acid in aqueous conditions. Many attempts have been
made for selective synthesis of 5-HMF with metal chlorides,
bases and inorganic acids, however, separation process of
those reagents are not yet clear.[9] Ionic liquids have also been
tried in the synthesis of 5-HMF and FAL with high yields, how-
ever, separation and cost of those reagents are not favourable
for commercial process.[10] Thermochemical process is efficient
in depolymerization of lignocellulose, however, at higher tem-
perature the conditions suppress product selectivity and yield.
Temperatures in excess of 2008C and acidic conditions are usu-
ally required to synthesise 5-HMF from lignocellulose but
under these conditions formation of humins through conden-
sation of C6 and C5 sugars with 5-HMF and FAL, and self-con-
densation of furan products are major limiting factors.[11]
Therefore, a process that requires minimal pre-treatment is
highly desirable.[2] Conversion of lignocellulose into furan de-
rivatives, particularly 5-hydroxymethylfurfural (5-HMF) and fur-
fural (FAL) are of high interest because they may be used as
feedstock for liquid fuels and other important chemicals such
as organic green solvents and high-volume plastics.[3] General-
ly, conversion of lignocellulose into chemicals occurs through
multiple steps: first by cleaving ester bonds between hemicel-
lulose and lignin to release cellulose chain, followed by hydrol-
ysis of hemicellulose and cellulose into monomer and further
acid-dehydration of monosaccharides into value-added chemi-
cals.[4] 5-HMF is a platform chemical with its derivatives levulin-
ic acid (LA, produced through furan ring opening), 2,5-furan di-
carboxylic acid (by oxidation), dimethylfuran (by hydrogenoly-
sis) and dihydroxymethylfuran (C=O group hydrogenation).[5]
FAL is a source of furfuryl alcohol, tetrahydrofurfurylalcohol,
methylfuran and tetrahydrofuran.[6] Both 5-HMF and FAL may
be used as alternative source for the synthesis of C6 and C5
5-HMF and FAL product selectivity can be improved by ra-
tional selection of the solvent. Dimethylsulfoxide (DMSO)
showed positive effects on increasing the final yields of 5-HMF
and also suppressed the unwanted side products because of
its high boiling point. However, the stability of DMSO in acidic
medium during the reaction is low and the separation of 5-
HMF and FAL is difficult.[12] Water-immiscible organic solvents
in the biphasic systems play an important role of extracting 5-
HMF from aqueous phase. However, most of the solvents
show poor partitioning of 5-HMF into the organic phase.[5b] In
the present work, remarkable results were obtained with 2-
[a] Dr. B. Seemala, Prof. V. Haritos, Dr. A. Tanksale
Catalysis for Green Chemicals Group
Department of Chemical Engineering
Monash University
Clayton, VIC 3800 (Australia)
Fax: (+61)3-9905-5686
E-mail: akshat.tanksale@monash.edu
ChemCatChem 2016, 8, 640 – 647
640
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
methyltetrahydrofuran (MTHF) and toluene as solvents with
water for hydrolysis–dehydration of lignocellulose. Simeó et al.
claimed that the physical properties of MTHF are similar to
those of toluene, however, MTHF can be derived from renew-
able resources (FAL to MTHF), but toluene mostly from petro-
chemicals.[13] MTHF has been also proven as a promising alter-
native solvent for environmentally benign synthesis strategies
such as the synthesis of Grignard reagents and lithium alumini-
um hydride catalysts as well as carbon–carbon (CÀC) coupling
reactions.[14] Owing to its low solubility in water, better stability,
and high boiling point compared to tetrahydrofuran (THF), it
favours easy recovery of the final products into the organic
phase.[1f,5c,15] In the present work, we report a new strategy to
produce 5-HMF and FAL by using LA as a catalyst for the first
time. LA is completely miscible in MTHF whereas it is immisci-
ble in toluene. Pinus radiata (pinewood) and Eucalyptus reg-
nans sawdust was used as feedstock in a mono- or biphasic
solvent system. MTHF and toluene were used as co-solvents
with water in different ratios at a temperature of 160–2008C.
The use of LA as a catalyst in a biphasic system is particularly
advantageous because of (i) better control of the side reactions
to achieve high yield of 5-HMF; (ii) easier extraction of desira-
ble compounds in the organic phase; (iii) higher feasibility of
the product mixture to be upgraded into chemicals and fuels;
and (iv) absence of inorganic salts and acids. LA and MTHF
both can be potentially produced from biomass, making the
processes greener.
vent was not active for the conversion of pinewood sawdust
(Table 1, experiment no.1). Cello-oligomers and hemi-mono-
mers concentrations increased with increasing water content
from 4:1 to 2:1 MTHF/water w/w ratio. The cello-oligomers
molar yield was equal to 10.5% and 22.8% in 4:1 and 2:1
MTHF/H2O ratios, respectively, based on 41.7 wt% cellulose
content in pinewood,[17] which is one of the highest yield in
just 15 min reaction time.[18] Increasing the aqueous phase
composition to 1:1 and 1:2 w/w ratios resulted in marginal de-
crease in the yield of C6 and C5 sugars (experiment nos. 5–7,
respectively). Yields of dehydration products enhanced with in-
creased water composition, with maximum concentration of 5-
HMF and FAL obtained at 1:2 ratio of MTHF/water, which was
more than a factor of five increase over their production at 4:1
ratio. Increasing the water content enhanced the polarity of
the system thereby accelerating the cleavage of glycosidic
ether bonds and enhanced hydrolysis step. Further increasing
the ratio of aqueous phase to 1:4 suppressed the 5-HMF
amounts significantly, attributed to the rehydration conversion
of 5-HMF into LA. This conversion is caused by a higher frac-
tion of 5-HMF and FAL in the aqueous phase. The percentage
of 5-HMF and FAL in the organic phase dropped from 94%
and 99% in the 4:1 MTHF/water mixture to only 5.6% and
16.3%, respectively, in the 1:4 mixture. Condensed products
can also be formed by increasing the amount of the water
media because of increase in the source of hydrogen ions in
the aqueous phase. Weingarten et al. also observed that in-
creasing the amount of water in the solvent decreased the re-
action rate.[19] Since increasing the fraction of water in the sol-
vent mixture shifted the distribution of 5-HMF and FAL into
the aqueous phase, the extraction and purification of these
products would require an additional separation process. Usu-
ally, the polar phase is used to dissolve and convert the sugars,
and the less polar or polar aprotic phase extracts the reactive
products or intermediates such as 5-HMF and FAL and pre-
vents further exposure to acidic protons in the polar phase.[19]
Although the LA catalyst is soluble in both MTHF and water,
the difference in polarity of the solvents enables different abili-
ties to transfer the hydrogen to reactants and, therefore, can
Results and Discussion
In Table 1 the effects of MTHF/water ratio in the biphasic
system (w/w ratio=4:1, 2:1, 1:1, 1:2 and 1:4) at 1808C after
15 min reaction time are shown. The maximum mass conver-
sion achieved was 45.3%. As LA is a weak acid (pKa =4.17), it is
likely to catalyse the cleavage of weak glycosidic ether bonds
and hydrogen bonds between cellulose, hemicellulose and
lignin leading to deconstruction of lignocellulose.[16] The parti-
tion coefficient of LA in MTHF/water at equilibrium is 1.08. Al-
though LA is completely miscible in MTHF, pure MTHF as a sol-
Table 1. Effect of solvent ratio on conversion of pinewood sawdust at 1808C.[a]
Exp.
no.
Solvent ratio
(MTHF/H2O w/w ratio)
Concentration
[mgLÀ1
Distribution of product in
organic phase [wt%]
X
[%]
]
Cello-oligomers
Hemi- monomers
5-HMF
FAL
5-HMF
FAL
1
2
3
4
5
6
7
8
9
only MTHF
4:1
2:1
2:1[b]
1:1
1:1[b]
1:2
1:4
230
1904
4440
3213
2196
1834
1878
1525
756
100
498
ND
56.8
168
375
307
585
353
200
692
ND
61
–
–
4.5
94.0
74.0
78.5
71.5
73.5
52.5
5.6
99.0
92.6
92.5
90.0
91.3
81.5
16.3
76.6
25.2
45.3
41.8
40.7
40.6
38.0
36.2
24.0
3350
2450
3090
2450
2530
2310
1020
158
432
288
792
417
133
596
1:1
4.6
toluene/water
[a] Reaction conditions: Pinewood=0.4 g, solvent=20 g, LA catalyst=0.1 g, reaction temperature=808C, rpm=0, reaction time=15 min. Cello-oligo-
mers=cellopentose+cellotettrose+cellotriose+cellobiose+glucose; hemi-monomers=xylose+arabinose+mannose, ND=Not detected. [b] Pine-
wood=1 g, water=25 g, MTHF=25 g, LA catalyst=0.250 g, reaction temperature=1808C, rpm=400, reaction time=15 min.
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
641
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
affect the catalytic behaviour by changing the solvent ratio.
95% of LA was found in the organic phase during the reaction
in 4:1 solvent ratio, and reducing the MTHF/water ratio to 2:1,
1:1, 1:2 and 1:4 reduced the catalyst amount in the organic
phase to 73.5, 55.9, 44.4 and 14.4%, respectively. Therefore,
the quantity of acidic protons and the polarity in the aqueous
phase increased with increasing water content, which could ef-
fectively facilitate the depolymerisation of lignocellulose.
Lower polarity of pure MTHF and 4:1 ratio of MTHF/water re-
sulted in lower yields of desirable products. A comparison of
the C6 and C5 products yields in the non-stirred and stirred re-
actors under identical operating conditions is shown in Table 1
(experiment no. 4 and 6). The mass conversion of pinewood
sawdust and sugars yield was not influenced by stirring, which
suggests that the hydrolysis reaction was kinetically controlled.
However, the final yields of 5-HMF and FAL increased two- to
three-fold with stirring, which suggests that the rate of mass
transfer of 5-HMF and FAL from aqueous phase to organic
phase was enhanced by stirring, thereby shifting the equilibri-
um of sugar monomer dehydration.
stock and LA catalyst loading was kept constant at 0.4 g and
0.1 g, respectively, whereas the total volume of the solvent
was varied from 10 g to 40 g, keeping the reaction time con-
stant at 15 min. Increasing the total solvent amount from 10 g
to 20 g resulted in an increase in the mass conversion and the
mass yield of both cello- and hemi-sugars (from 8.9 to 22.1 mg
and 4.9 to 25.9 mg, respectively). Particularly, 30 g of MTHF/
water (1:1) solvent system yielded higher amounts of both
cello-oligomers and hemi-monomers at 903 and 1520 mgLÀ1
(mass yield=27.1 and 45.6 mg, respectively). Moreover, more
than 76% and 93% of 5-HMF and FAL, respectively, were re-
covered into the organic phase. However, further increasing
the solvent amount to 40 g reduced the mass conversion and
the yield of the products. Lowering the concentration of LA-
catalyst had a negative influence on the depolymerisation of
lignocellulosic biomass. Effect of catalyst loading is shown in
Table 3. Increasing the catalyst amount showed positive effect
Table 3. Influence of amount of LA catalyst on pinewood conversion.[a]
In comparison to the MTHF/water (1:1) mixture, the toluene/
water (1:1) mixture resulted in an increased yield of both 5-
HMF and FAL by over two-fold (Table 1 experiment no. 9). Max-
imum 5-HMF molar yield of 7.4% was achieved in the toluene/
water (1:1) mixture compared to only 3.1% molar yield in the
MTHF/water (1:1) mixture after 15 min reaction time. This sug-
gests that toluene dissolved the lignin fraction of the biomass
and exposed hemicellulose and cellulose to acidic protons in
the aqueous media. Moreover, the LA catalyst is insoluble in
toluene; therefore a high concentration of acid in the aqueous
phase enhanced the dehydration reaction and also shifts the
equilibrium towards formation of 5-HMF in aqueous phase.
Since FAL has lower solubility in water as compared to 5-HMF,
76.6% of FAL was extracted into the toluene phase. However,
<5% of the 5-HMF was obtained in the toluene phase; where-
as in comparison, >70% of the 5-HMF partitioned into the
MTHF phase in the MTHF/water (1:1) mixture.
Exp.
no.
Levulinic
acid [g]
Concentration
[mgLÀ1
Mass
conv. [%]
]
C-ol
H-ol
5-HMF
FAL
1
2
3
4
NIL
1175
1646
2196
2433
400
2300
3090
3080
35
184
307
715
40
173
288
692
18.5
31.8
40.7
42.6
0.05
0.10
0.15
[a] Reaction conditions: Pinewood=0.4 g, solvent (MTHF/water, 1:1 w/
w)=20 g, reaction temperature=1808C, reaction time=15 min. Cello-
oligomers (C-ol)=cellopentose+cellotettrose+cellotriose+cellobiose+
glucose; hemi-monomers (H-ol)=xylose+arabinose+mannose. NIL=no
LA added.
on the furan products, indicating that the reaction was mass
transfer limited. 5-HMF and FAL concentrations increased by
more than a factor of two when LA-catalyst loading was in-
creased from 0.100 g to 0.150 g, whereas sugar concentrations
increased marginally. This is believed to be due to increase in
the dehydration step, which is favourable at higher concentra-
tion of acid at low reaction temperatures. Similar effect has
been observed with inorganic acids.[9d]
MTHF/water was chosen for further activity studies because
it is greener than toluene/water and 1:1 ratio was selected be-
cause it resulted in higher yields of sugar monomers and fur-
anic compounds. In Table 2 the effect of initial concentration
of feedstock and LA-catalyst in (1:1) MTHF/water solvent on
the conversion of lignocellulose at 1808C is shown. The feed-
To find out the optimum temperature, the reaction was con-
ducted at four different temperatures: 140, 160, 180 and
2008C (Table 4). As the temperature increased from 140 to
1608C, the yield of C6 sugars increased to nearly four times
with marginal increase in the yields of C5 sugars, 5-HMF and
FAL. Further increasing the temperature to 1808C resulted in
remarkable increase in 5-HMF (to 307 mgLÀ1) with reduction in
the yield of cello-oligomers. Yield of FAL and hemi-monomers
also increased remarkably. Increasing the reaction temperature
to 2008C resulted in increase in the yield of both the furanic
products with lower yields of both cello-oligomers and hemi-
monomers (Table 4, experiment no. 3 and 4). Even though the
yields of furanic compounds were higher, the overall selectivity
of the desirable products (sugars+furans) was lower at 2008C.
At this temperature the reaction mixture turned dark
(Figure 1), which may be caused by the formation of humins
Table 2. Effect of varying the amount of solvent.[a]
MTHF/H2O
[g]
Concentration
[mgLÀ1
Distribution of product in
organic phase [wt%]
X
[%]
]
C-ol H-ol 5-HMF FAL 5-HMF
FAL
5:5
885 489 778
1103 1280 897
903 1520 654
486 768 529
372 43.9
495 70.4
384 76.3
269 66.9
28.9
90.9
93.2
87.5
33.8
40.7
36.5
24.2
10:10
15:15
20:20
[a] Reaction conditions: Pinewood=0.4 g, solvent=MTHF/water, LA cata-
lyst=0.1 g, reaction temperature=1808C, reaction time=1 h. Cello-oligo-
mers (C-ol)=cellopentose+cellotettrose+cellotriose+cellobiose+glu-
cose; hemi-monomers (H-ol)=xylose+arabinose+mannose.
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
642
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
sic system, it was still limited by the hydrolysis step
at 1608C. An advantage of the biphasic system was
that more than 70% of the furan compounds were
extracted into the organic phase that resulted in
higher accessibility of aqueous acidic protons to
polymeric sugars, which favoured the depolymerisa-
tion step. Lange et al. and Cai et al. also reported
that pure aqueous system with dilute mineral acids
typically suffers from low 5-HMF yields.[20]
Table 4. Effect of temperature on pinewood conversion.[a]
Exp.
no.
Temp.
[8C]
Concentration [mgLÀ1
Hemi-monomers
]
X
[%]
Cello-oligomers
5-HMF
FAL
22.1
66
288
535
1
2
3
4
140
160
180
200
1002
3907
2196
1141
1704
1830
3090
1902
18
53
307
803
21
33
40.7
45
[a] Reaction conditions: Pinewood=0.4 g, solvent (MTHF/water, 1:1 w/w)=20 g, LA
catalyst=0.1 g, reaction time=15 min. Cello-oligomers=cellopentose+cellotet-
trose+cellotriose+cellobiose+glucose; hemi-monomers=xylose+arabinose+man-
nose.
At 1808C (Figure 3(a)), dehydration of sugar mono-
mers was favoured over hydrolysis of oligosacchar-
ides, resulting in higher concentrations of 5-HMF and
FAL (1186 and 536 mgLÀ1 at 120 min, respectively)
Figure 1. Reaction mixtures at different temperatures after reaction
a) 1408C, b) 1608C, c) 1808C, d) 2008C; pinewood=0.4 g, solvent (MTHF/
water, 1:1 w/w)=20 g, LA catalyst=0.1 g, reaction time=15 min.
as by-products. Normally, humins can form through the self-
condensation of furan products and also condensation be-
tween the 5-HMF and FAL with sugars at temperatures above
1808C in highly acidic conditions. Even though the initial reac-
tion conditions were not highly acidic at 2008C, acids could be
generated in situ by depolymerisation of lignocellulose, which
would favour the condensation reactions.[1f]
In Figure 2a the extent of the reaction as a function of time
in aqueous medium with LA catalyst at 1608C is shown. If
using pure water as a solvent, C6 sugars concentration showed
a peak at 30 min (at 1081 mgLÀ1), whereas C5 sugars concen-
tration showed
a plateau at approximately 120 min (at
2090 mgLÀ1). 5-HMF and FAL concentrations increased linearly
for the duration of study. This suggests that the rate of depoly-
merisation was the limiting step and is limited by the low tem-
perature. In comparison to the aqueous media, the MTHF/
water biphasic system with LA-catalyst resulted in higher yield
and selectivity of sugars and furans at 1608C (Figure 2b). The
depolymerisation step was promoted in the biphasic system
with subsequent hydrolysis–dehydration into C6 sugars and
furans. C6 sugars decreased from the maximum of 3957 mgLÀ1
at 15 min to approximately 2000 mgLÀ1 at 120 min, whereas
HMF yields increased linearly from 50 mgLÀ1 to 447 mgLÀ1 at
the respective times.
Figure 2. Effect of the solvent system on the conversion of pinewood saw-
dust to 5-HMF and chemicals using a) only water as solvent, b) MTHF/water
(1:1 w/w) co-solvent system; pinewood=0.4 g, solvent=20 g, LA cata-
&
*
lyst=0.1 g, reaction temperature=1608C. ( ) Cello-oligomers, ( ) Hemi-olig-
~
!
omers, ( ) 5-HMF, ( ) FAL. The bullet points are results from separate runs
and the solid lines are fitted curves to illustrate the trends.
whereas the cello-oligomers and hemi-monomers concentra-
tion reduced to 865 and 477 mgLÀ1, respectively. This results
in 5-HMF molar yield of 18.3%. The rate of dehydration reac-
tion further increased at 2008C as observed from the rapid in-
crease in concentration of 5-HMF and FAL to 1350 and
643 mgLÀ1, respectively, at 60 min (Figure 3(b)). This results in
5-HMF molar yield of 20.8%. However, at this temperature, the
C5 sugars concentration increased from 1700 mgLÀ1 at
15 min to 3150 mgLÀ1 at 60 min. Therefore it can be conclud-
ed that even though the reaction rate increased in the bipha-
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
643
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
with sugar molecules.[20b,21] Long residence time and high tem-
perature also favoured LA production. The concentration of
LA, excluding the LA catalysis amount, increased linearly at
1808C. The fraction of 5-HMF that is present in the aqueous
phase is susceptible to rehydration into LA,[21] and since this re-
action mixture contained 1:1 ratio of MTHF/water, as the con-
centration of 5-HMF increased, the concentration of LA also in-
creased because of the equilibrium. As shown in Figure 4a–c,
SEM images of the pinewood samples before and after the re-
action showed remarkable variation in the surface morphology.
In the reaction with MTHF/water solvent at 180 and 2008C, the
structure was completely disrupted and pores are created on
the cell wall, whereas smooth surface can be seen in the un-
treated sample. These pores are clearly visible on both 180
and 2008C treated residue samples and may suggest the ex-
traction of lignin by contact with organic solvent following the
reaction. Lignin is insoluble in water and soluble in MTHF and
therefore can be separated by vacuum distillation. Lignin was
recovered by this method from the reaction mixture as a semi-
solid material and tested by using FTIR spectra. The character-
istic FTIR peak of lignin at 1515 nm (CÀC aromatic stretching)
was evident (Figure 5). However, as the boiling point of LA is
high (2458C) and it was difficult to separate it from the reac-
tion mixture, some LA was recovered with lignin as seen in the
peak at 1715 for the C=O group in LA.
The yield of desirable products from various feedstock was
tested in 1:1 MTHF/water mixture at 1808C (Table 5). The con-
centration of C6 and C5 sugars obtained from eucalyptus saw-
dust was less than that from pinewood after 15 min reaction
time. The SEM images in Figure 4 also evidenced that the sur-
face of the eucalyptus was partially damaged compared to
that of pinewood. However, the concentration of FAL from eu-
calyptus sawdust was significantly higher than from pinewood.
This may be caused by the presence of larger composition of
C5 carbohydrates, in particular, xylose. Typically, eucalyptus
contains 15–17 wt% hemicellulose, 38–49 wt% cellulose and
28–31 wt% lignin[22] in comparison to pinewood, which con-
Figure 3. Effect of temperature on the conversion of pinewood saw dust to
5-HMF and chemicals at a) 180 and b) 2008C, pinewood=0.4 g, solvent
&
*
(MTHF/water, 1:1 w/w)=20 g, LA=0.1 g. ( ) Cello-oligomers, ( ) Hemi-oligo-
~
!
*
mers, ( ) 5-HMF, ( ) FAL, ( ) levulinic acid. The bullet points are the results
from separate runs and the solid lines are fitted curves to illustrate the
trends.
desirable products were lost to humins formation at longer
residence time as the reaction mixture turned to dark brown
after 120 min (Figure 1) suggesting that humins formed
through self-condensation of furans or condensation of furans
Figure 4. a) Pinewood before treatment, b) after treatment at 1808C c) after treatment at 2008C, d) eucalyptus before treatment, e) after treatment at 1808C.
Reaction conditions: feedstock=0.4 g, solvent (MTHF/water, w/w)=20 g, LA=0.1 g, reaction time=15 min.
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
644
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
however, the latter sample contains only approximately 45%
cellulose.
Moreover, owing to the recalcitrance of crystalline cellulose
depolymerisation reaction is the rate limiting step and longer
times might be needed to achieve similar yield of 5-HMF. If
using 0.4 g of glucose as a feedstock, 5-HMF was obtained as
the major product (1570 mgLÀ1) with a small quantity of LA
byproduct. Only 20% of glucose was converted into 5-HMF
and LA within 15 min reaction time at 1808C compared to
40.7% and 37% conversion of pinewood and eucalyptus saw-
dust. These results are comparable with those reported in liter-
ature.[8,9f] Even though the reaction conditions are slightly dif-
ferent from those in the literature, the amount of glucose con-
verted into 5-HMF is higher in the present case. In addition,
Zhao et al. also reported that more than 70% yield of 5-HMF
was obtained from glucose in the presence of 1-ethyl-3-meth-
ylimidazolium chloride [EMIM]Cl as a solvent with CrCl2 salt.[25]
They proposed a mechanism with the plausible formation of
À
CrCl3 from [EMIM]Cl and, further, that anions(CrCl3À) play
a role in proton transfer that could facilitating mutarotation of
glucose in [EMIM], which results in fructose. In the present
condition, LA catalyst may act as both Lewis acid and dehy-
drating agent. LA catalyst contains a carboxylic group at one
end and a carbonyl group located in between the active meth-
ylene (CH2) and the methyl group (CH3). The carboxylate and
carbonyl groups favour the formation of hydrogen bonds with
the C2-hydroxyl group of the glucose molecule, which facili-
tates keto–enol isomerisation. The carboxylic group can
donate a proton that acts like a Brønsted site, which facilitates
dehydration reaction. Recently, Luterbacher et al. reported high
yields of carbohydrate production from hard and softwood
Figure 5. FTIR spectra of lignin byproduct obtained at a) 180 and b) 2008C
reaction temperature. 1) C=O group stretching in carboxylic acid; 2) CÀC
stretching in aromatic group.
biomass-derived g-valerolactone with
a water co-solvent
system with using dilute H2SO4 as a catalyst.[1f] Although the
final concentrations of C6 and C5 sugars are low in the present
work, it achieved better yields and selectivity of 5-HMF. The
presence of the furan ring in MTHF, 5-HMF and FAL may facili-
tate the extraction and these reactive intermediates are highly
stable in the organic phase.[26] More importantly, MTHF is more
volatile than FAL and 5-HMF, therefore separation of these
products is easier by vacuum distillation process.[19,27] Further
the LA and lignin content in the organic phase can be separat-
ed by adding water to it. In addition by optimisation of reac-
tion time, volume and temperature to maximise the yields
from biomass conversion, the ratio of MTHF/water biphasic
system allowed the additional control of the extent of reaction
to achieve high yields of desirable products with good selectiv-
ity.[28] Moreover, in this work the salts and mineral acid were re-
placed with LA, which along with MTHF can be produced from
lignocellulose.
Table 5. Conversion of different feedstock to chemicals.[a]
Feedstock
Concentration [mgLÀ1
]
X [%]
C-ol
H-ol
5-HMF
FAL
LA
pinewood
eucalyptus
cellulose[b]
glucose
2196
491
3410
1258
3090
2720
459
–
307
93
555
1570
288
565
106
–
177
–
80
40.7
37.0
24.3
20.0
175
[a] Reaction condition: Feedstock=0.4 g, solvent (MTHF/water, 1:1 w/
w)=20 g, LA catalyst=0.1 g, reaction temperature=1808C, reaction
time=15 min. [b] Reaction time=1 h. Cello-oligomers (C-ol)=cellopen-
tose+cellotettrose+cellotriose+cellobiose+glucose; hemi-monomers
(H-ol)=xylose+arabinose+mannose.
tains 20.6 wt% hemicellulose, 41.7 wt% cellulose and
25.9 wt% lignin.[22b,23] Eucalyptus hemicellulose contains glu-
curonoxylan, xylan and acetyl groups, whereas pinewood
hemicellulose contains galactoglucomannans, mannose and
galactose units.[22b,24] Concentration of C6 sugars obtained from
microcrystalline cellulose was three times higher than from
pinewood sawdust after 60 min reaction, however, the dehy-
dration into 5-HMF was comparatively lower. High sugar yield
is expected from microcrystalline cellulose because the mass
of the feedstock used was same as from eucalyptus sawdust;
Conclusions
Levulinic acid (LA) was used as a catalyst for the conversion of
lignocellulose biomass into 5-hydroxymethylfurfural (5-HMF)
and furfural (FAL) in mono- and biphasic systems. LA showed
several important benefits as the catalyst. LA is miscible in
both water and MTHF solvents, therefore, it was equally parti-
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
645
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
tioned into the two phases. In the aqueous phase it depoly-
merised lignocellulose biomass followed by dehydration of
sugar monomers to produce 5-HMF and FAL. The presence of
LA in the aqueous phase maintained the equilibrium towards
5-HMF by preventing rehydration of 5-HMF into LA. Large frac-
tions of 5-HMF and FAL were recovered in the MTHF phase,
which is important for increasing the conversion of lignocellu-
lose into these products and would allow easier separation
downstream. Both hydrolysis and dehydration reactions were
accelerated in the MTHF/water biphasic system compared to
those in the pure aqueous system. Highest yields of 5-HMF
and FAL were achieved with 1:1 w/w ratio of MTHF/water in
a stirred reactor. Although highest conversion was achieved at
2008C, the optimum temperature was found to be 1808C at
which good yield and higher selectivity of 5-HMF and FAL was
achieved. Pinewood sawdust was found to yield higher
amounts of desirable products than eucalyptus sawdust. The
SEM micrographs showed clear indications of loss of material
in the treated pinewood sample, compared to the treated eu-
calyptus samples.
Yield of 5-HMF ð%Þ ¼
ꢀ
ꢁ
ð2Þ
Moles of 5-HMF
 100
Theoretical moles of glucose in the feed
The theoretical moles of glucose are based on 41.7 wt% cellulose
content in p. radiata.[17]
The distribution of 5-HMF and FAL in the organic phase was calcu-
lated according to the following equations:
Distribution of 5-HMF in organic phase ð%Þ ¼
ꢀ
ꢁ
ð3Þ
ð4Þ
Mass of 5-HMF in organic phase
Total mass of 5-HMF in the product
¼ Â100
Distribution of FAL in organic phase ð%Þ ¼
ꢀ
ꢁ
Mass of FAL in organic phase
Total mass of FAL in the product
 100
By definition, the distribution of 5-HMF and FAL in the aqueous
phase will be 100 minus the distribution in the organic phase.
Product analysis
After each experiment, both aqueous and organic phases were
sampled and analysed by using HPLC equipped with Rezex RHM-
Monosacccharide 300”7.80 mm column. Retention times and
quantitation of sugars, HMF, LA and FAL were undertaken by com-
parison of retention times and calibration curves derived from
peak areas of pure standards. Surface morphology of the pine-
wood and eucalyptus sawdust samples before and after the reac-
tion were analysed using FEI Magellan 400 FEGSEM scanning elec-
tron microscope (SEM). Lignin byproduct was characterized using
Experimental Section
Materials
The following feedstock were used in this study: pinus radiata
(pinewood, 550 mm) and eucalyptus regnans (500 mm) were ob-
tained from Pollard’s Sawdust Supplies, Victoria, Australia and Sig-
macell microcrystalline cellulose (20 mm) and d-glucose were ob-
tained from Sigma–Aldrich.
a PerkinElmer Spectrum GX FTIR with spectral resolution of 4 cmÀ1
The spectrum represents an average of 10 scans.
.
Reaction procedure
Acknowledgements
The required amount (400 mg) of feedstock was taken into
a 100 mL stainless-steel non-stirred Parr reactor and then LA (0.1 g,
Aldrich 98%) as a catalyst was added followed by the solvent
(20 g). MTHF (Sigma–Aldrich 99%), toluene (Sigma–Aldrich, 99%),
and distilled water were used to make the mixture of MTHF/water
(4:1, 2:1, 1:1, 1:2 and 1:4 w/w ratio) and toluene/water (1:1 w/w
ratio) that were used as solvents in this reaction. The reactor was
initially purged with N2 to flush out air and then pressurized with
nitrogen gas up to 10 bar. The reactor was heated to the desired
temperature (140 to 2008C) as measured by internal thermocouple
and the reaction was conducted for time periods between 15 and
120 min. After the reaction, the unreacted solid was filtered from
the reaction mixture, dried at 858C for 24 h and weighed. The or-
ganic phase was separated from the aqueous phase by using a sep-
arating funnel. Reactions were also conducted in a 300 mL Berghof
stirred reactor at a stirring speed of 400 rpm containing pinewood
sawdust (1 g), LA (0.25 g) and the solvent mixture (50 g). Mass con-
version was calculated according to the following equation:
The authors acknowledge the funding from CSIRO Flagship Col-
laboration Fund. The authors thank Ms Kathryn Waldron,
Monash Centre for Electron Microscopy, for her help with the
SEM analysis.
Keywords: acidity
·
biomass
·
biphasic catalysis
·
carbohydrates · hydrolysis
[1] a) D. M. Alonso, J. Q. Bond, J. A. Dumesic, Green Chem. 2010, 12, 1493–
1513; b) F. L. Chan, A. Tanksale, ChemCatChem 2014, 6, 2727–2739;
c) F. L. Chan, K. Umeki, A. Tanksale, ChemCatChem 2015, 7, 1329–1337;
d) M. J. Climent, A. Corma, S. Iborra, Green Chem. 2011, 13, 520–540;
e) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098;
f) J. S. Luterbacher, J. M. Rand, D. M. Alonso, J. Han, J. T. Youngquist, C. T.
Maravelias, B. F. Pfleger, J. A. Dumesic, Science 2014, 343, 277–280;
g) L. R. Lynd, J. H. Cushman, R. J. Nichols, C. E. Wyman, Science 1991,
251, 1318–1323.
ꢀ
ꢁ
[2] a) N. R. Savadekar, S. T. Mhaske, Carbohydr. Polym. 2012, 89, 146–151;
b) V. K. Thakur, M. K. Thakur, Carbohydr. Polym. 2014, 109, 102–117.
[3] C. H. Christensen, J. Rass-Hansen, C. C. Marsden, E. Taarning, K. Egeblad,
ChemSusChem 2008, 1, 283–289.
[4] a) M. Bicker, D. Kaiser, L. Ott, H. Vogel, J. Supercrit. Fluids 2005, 36, 118–
126; b) A. Shrotri, L. K. Lambert, A. Tanksale, J. Beltramini, Green Chem.
2013, 15, 2761–2768; c) S. Dutta, S. De, B. Saha, ChemPlusChem 2012,
77, 259–272.
Mass of unreacted solid
Mass of feed solid
Mass Conversion, X ð%Þ ¼ 1À
 100
ð1Þ
Molar yield of 5-HMF was calculated according to the following
equation:
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
646
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[5] a) H. Ait Rass, N. Essayem, M. Besson, Green Chem. 2013, 15, 2240–
2251; b) Y. Romµn-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature
2007, 447, 982–U985; c) Y. Romµn-Leshkov, J. A. Dumesic, Top. Catal.
2009, 52, 297–303.
[6] S. Bhogeswararao, D. Srinivas, J. Catal. 2015, 327, 65–77.
[7] G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Science 2005, 308,
1446–1450.
[18] a) Y. Yang, C. W. Hu, M. M. Abu-Omar, ChemSusChem 2012, 5, 405–410;
b) F. S. Asghari, H. Yoshida, Carbohydr. Res. 2010, 345, 124–131; c) Z. H.
Zhang, Z. B. K. Zhao, Bioresour. Technol. 2010, 101, 1111–1114.
[19] R. Weingarten, A. Rodriguez-Beuerman, F. Cao, J. S. Luterbacher, D. M.
Alonso, J. A. Dumesic, G. W. Huber, ChemCatChem 2014, 6, 2229–2234.
[20] a) C. M. Cai, T. Y. Zhang, R. Kumar, C. E. Wyman, Green Chem. 2013, 15,
3140–3145; b) J. P. Lange, E. van der Heide, J. van Buijtenen, R. Price,
ChemSusChem 2012, 5, 150–166.
[21] X. Hu, L. P. Wu, Y. Wang, Y. Song, D. Mourant, R. Gunawan, M. Gholiza-
deh, C. Z. Li, Bioresour. Technol. 2013, 133, 469–474.
[22] a) R. P. KibbleWhite, M. J. C. Riddell, C. J. A. Shelbourne, New Zealand J.
Forestry Sci. 2000, 30, 458–474; b) R. F. H. Dekker, Biocatal. Biotransform.
1987, 1, 63–75; c) M. Brink, E. G. Achigan-Dako, Fibres, PROTA Founda-
tion, 2012.
[23] a) A. Ferraz, J. Baeza, J. Rodriguez, J. Freer, Bioresour. Technol. 2000, 74,
201–212; b) K. K. Y. Wong, K. F. Deverell, K. L. Mackie, T. A. Clark, L. A.
Donaldson, Biotechnol. Bioeng. 1988, 31, 447–456.
[24] a) G. Koch in Handbook of Pulp, Wiley-VCH, Weinheim, 2008, pp. 21–68;
b) M. G. Grønli, G. Vµrhegyi, C. Di Blasi, Ind. Eng. Chem. Res. 2002, 41,
4201–4208; c) T. E. Timell, Wood Sci. Technol. 1967, 1, 45–70; d) P.
Reyes, R. T. MendonÅa, J. Rodríguez, P. Fardim, B. Vega, J. Chil. Chem.
Soc. 2013, 58, 1614–1618; e) C. Stewart, D. Foster, Aust. J. Chem. 1953,
6, 431–438.
[8] Y. J. Pagµn-Torres, T. F. Wang, J. M. R. Gallo, B. H. Shanks, J. A. Dumesic,
ACS Catal. 2012, 2, 930–934.
[9] a) J. B. Binder, R. T. Raines, J. Am. Chem. Soc. 2009, 131, 1979–1985;
b) R. L. Huang, W. Qi, R. X. Su, Z. M. He, Chem. Commun. 2010, 46, 1115–
1117; c) A. Takagaki, M. Ohara, S. Nishimura, K. Ebitani, Chem. Commun.
2009, 6276–6278; d) M. Von Sivers, G. Zacchi, Bioresour. Technol. 1995,
51, 43–52; e) H. P. Yan, Y. Yang, D. M. Tong, X. Xiang, C. W. Hu, Catal.
Commun. 2009, 10, 1558–1563; f) Y. Yang, C. W. Hu, M. M. Abu-Omar,
Green Chem. 2012, 14, 509–513.
[10] a) J. B. Binder, R. T. Raines, Proc. Natl. Acad. Sci. USA 2010, 107, 4516–
4521; b) K. Shill, S. Padmanabhan, Q. Xin, J. M. Prausnitz, D. S. Clark,
H. W. Blanch, Biotechnol. Bioeng. 2011, 108, 511 –520.
[11] L. C. Peng, L. Lin, J. H. Zhang, J. P. Zhuang, B. X. Zhang, Y. Gong, Mole-
cules 2010, 15, 5258–5272.
[12] B. R. Caes, R. T. Raines, ChemSusChem 2011, 4, 353–356.
[13] Y. Simeó, J. V. Sinisterraa, A. R. Alcantara, Green Chem. 2009, 11, 855–
862.
[25] H. B. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science 2007, 316,
1597–1600.
[14] a) D. F. Aycock, Org. Process Res. Dev. 2007, 11, 156–159; b) S. Shanmu-
ganathan, L. Greiner, P. D. de Maria, Tetrahedron Lett. 2010, 51, 6670–
6672; c) H. Yang, S. Mahapatra, P. H. Y. Cheong, R. G. Carter, J. Org.
Chem. 2010, 75, 7279–7290.
[26] X. Hu, R. J. M. Westerhof, D. H. Dong, L. P. Wu, C. Z. Li, ACS Sustainable
Chem. Eng. 2014, 2, 2562–2575.
[27] J. N. Chheda, Y. Romµn-Leshkov, J. A. Dumesic, Green Chem. 2007, 9,
342–350.
[15] V. Pace, P. Hoyos, L. Castoldi, P. D. de Maria, A. R. Alcantara, ChemSus-
Chem 2012, 5, 1369–1379.
[28] C. M. Cai, N. Nagane, R. Kumar, C. E. Wyman, Green Chem. 2014, 16,
3819–3829.
[16] J. G. Lynam, C. J. Coronella, W. Yan, M. T. Reza, V. R. Vasquez, Bioresour.
Technol. 2011, 102, 6192–6199.
[17] NREL, U.S. Department of Energy; http://www.nrel.gov/biomass/data_
resources.html.
Received: October 8, 2015
Revised: November 10, 2015
Published online on December 8, 2015
ChemCatChem 2016, 8, 640 – 647
www.chemcatchem.org
647
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim