Enhanced yields of furfural and other products by simultaneous solvent extraction during thermochemical treatment of cellulosic biomass

Article abstract of DOI:10.1039/c3ra41857j

Furfural derived from pentose sugars is one of the key reactive intermediates for production of hydrocarbons from cellulosic biomass for use as drop-in fuels. Simultaneous extraction into a solvent immiscible in water can substantially enhance furfural yields by removing it from the aqueous environment where the yield would otherwise be limited by its rapid degradation to chars and other carbon rich compounds that are loosely termed humins. Thus, in this study, the effectiveness of the organic solvent methyl isobutyl ketone (MIBK) in improving furfural yields from maple wood was determined for reactions of 5 wt% solids in 0.1 M sulfuric or hydrochloric acid at 170°C over a range of reaction times. For comparison, pure xylose, glucose, and Avicel cellulose were also reacted under the same conditions. Various process configurations based on simultaneous hydrolysis and dehydration were compared to acid hydrolysis followed by dehydration to furfural with and without simultaneous extraction. Without MIBK extraction, the maximum furfural yields were less than 65% when maple wood was reacted for about 40-45 min with either sulfuric or hydrochloric acid. However, the yield increased significantly to about 85.3% when MIBK extraction was employed in combination with sulfuric acid catalysis for 50 min, while combining MIBK extraction with hydrochloric acid catalysis only increased the yield to ～67.0%. Simultaneous extraction with MIBK also improved yields of other products such as 5-HMF and levulinic acid, compared to results from the acids alone.

Full text of DOI:10.1039/c3ra41857j

View Article Online
View Journal
RSC Advances
Accepted Manuscript
This is an Accepted Manuscript, which has been through the RSC Publishing peer
review process and has been accepted for publication.
Accepted Manuscripts are published online shortly after acceptance, which is prior
to technical editing, formatting and proof reading. This free service from RSC
Publishing allows authors to make their results available to the community, in
citable form, before publication of the edited article. This Accepted Manuscript will
be replaced by the edited and formatted Advance Article as soon as this is available.
To cite this manuscript please use its permanent Digital Object Identifier (DOI®),
which is identical for all formats of publication.
More information about Accepted Manuscripts can be found in the
Information for Authors.
Please note that technical editing may introduce minor changes to the text and/or
graphics contained in the manuscript submitted by the author(s) which may alter
content, and that the standard Terms & Conditions and the ethical guidelines
that apply to the journal are still applicable. In no event shall the RSC be held
responsible for any errors or omissions in these Accepted Manuscript manuscripts or
any consequences arising from the use of any information contained in them.
www.rsc.org/advances
Registered Charity Number 207890

Page 1 of 34
RSC Advances
View Article Online
Enhanced Yields of Furfural and Other Products by Simultaneous Solvent
Extraction during Thermochemical Treatment of Cellulosic Biomass
1
2
3
4
5
Taiying Zhanga, Rajeev Kumara, and Charles E. Wymana,b*
6
7
8
a. Center for Environmental Research and Technology (CE-CERT), Bourns College of
Engineering, University of California, Riverside, 1084 Columbia Avenue, Riverside, California
92507 United States
9
10
11
b. Department of Chemical and Environmental Engineering, Bourns College of Engineering,
University of California, Riverside, 446 Winston Chung Hall, 900 University Ave., Riverside,
CA 92507, United States
12
13
14
15
*Correspondence to: Prof. Charles E. Wyman, cewyman@engr.ucr.edu, Tel: (951) 781-5703, Fax: (951)
781-5790
16
1

RSC Advances
Page 2 of 34
View Article Online
1
2
Abstract
Furfural derived from pentose sugars is one of the key reactive intermediates for
3
4
production of hydrocarbons from cellulosic biomass for use as drop-in fuels. Simultaneous
extraction into a solvent immiscible in water can substantially enhance furfural yields by
removing it from the aqueous environment where its yields would otherwise be limited by its
rapid degradation to chars and other carbon rich compounds that are loosely termed humins.
Thus, in this study, the effectiveness of the organic solvent methyl isobutyl ketone (MIBK) in
improving furfural yields from maple wood was determined for reactions of 5 wt% solids in
0.1M sulfuric or hydrochloric acid at 170oC over a range of reaction times. For comparison,
pure xylose, glucose, and Avicel cellulose were also reacted at the same conditions. Various
process configurations based on simultaneous hydrolysis and dehydration were compared to acid
hydrolysis followed by dehydration to furfural with and without simultaneous extraction.
Without MIBK extraction, the maximum furfural yields were less than 65% when maple wood
was reacted for about 40-45 min with either sulfuric or hydrochloric acids. However, the yield
increased significantly to about 85.3% when MIBK extraction was employed in combination
with sulfuric acid catalysis for 50 min, while combining MIBK extraction with hydrochloric acid
catalysis only increased the yield to ~ 67.0%. Simultaneous extraction with MIBK also improved
yields of other products such as 5-HMF and levulinic acid compared to results from the acids
alone.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Keywords: furfural, levulinic acid, cellulosic biomass, reactive intermediates, biphasic reaction
2

Page 3 of 34
RSC Advances
View Article Online
1
2
1. Introduction
Furfural (furfuraldehyde) has been widely used for decades as a solvent in the petroleum
3
4
industry and as a raw material for production of furfuryl alcohol, maleic anhydride,
tetrahydrofuran, phenolics, and resins 1-4. Beyond the existing furfural market, it has been
promoted as a building block for production of “drop-in” fuels from biomass that are compatible
with the existing petroleum infrastructure by conversion into such compounds as dimethylfuran
or ethyl levulinate with further polymerization to hydrocarbons 5-9. Positive results in engine
performance and reductions in emissions were reported for furan-based blends with diesel10. In
fact, furfural from hemicellulose and 5-HMF from cellulose in biomass have been selected as
two of the top 10 building block chemicals or co-products in a biorefinery along with two
5
6
7
8
9
10
11
furfural derivatives: furan dicarboxylic acid and levulinic acid 3, 5, 11
.
12
13
14
15
16
17
18
19
20
21
22
Furfural is a heterocyclic aldehyde, produced by dehydration of the pentoses xylose and
arabinose which are major components in most forms of biomass hemicellulose. The mechanism
for acid hydrolysis of pentosans and pentose dehydration to furfural have been shown to involve
four steps: 1) protonation of an oxygen link (beta-1, 4 xylosidic bond) resulting in trivalent
oxygen; 2) cleavage of a carbon-oxygen bond resulting in conversion of a hydroxyl group to a
pentose unit on one side and carbocation on the other side of C-O-C bond; 3) water addition to
the carbocation; and 4) liberation of a hydrogen ion from H2O+-C to form a hydroxyl group
attached to carbon(HO-C) 12. Free protons are a prerequisite not only for pentosan hydrolysis to
pentose but also for pentose dehydration to furfural. The dehydration of pentose liberates three
water molecules from one pentose molecule, which includes both 1, 4-elimination and 1,2-
elimination of water through hydrogen ions transforming hydroxyl groups to H2O+13, 14
.
3

RSC Advances
Page 4 of 34
View Article Online
1
2
Among methods for furfural production, dilute acid hydrolysis followed by dehydration
is regarded as the most effective 3, 4, with biomass rich in five carbon sugars such as oat bran,
corn cobs, and sugar cane bagasse used for this service 8, 15-19. The first pilot plant for furfural
production was a batch process that began operation in 1921 at the Quaker Oats Co. in Cedar
Rapids, Iowa and processed 50,000 to 60,000 tons of oat hulls per year 20. The resulting
commercial processes employed sulfuric acid at high temperatures in an iron digester that was
lined with brick to prevent corrosion. Furfural yields from this batch system (1921-1966) were
limited to about 50% of theoretical, and the technology was replaced by a more effective
continuous process in Belle Glade, Florida (1966-1997) 3. Since the 1990’s, several other
approaches have been developed such as those by Biofine®, CIMV, SupraYield®, MTC, and
Westpro®3, 4, 19. For example, Biofine® patented a process in 1990 that was primarily for
levulinic acid production with furfural as a side-product with a yield of approximately 70% of
theoretical 21-23. The Vedernikov furfural process increased yields to 75% with less cellulose
losses 24-26. Current commercial technology uses steam to strip furfural from the aqueous
solution, but even then, the boiling point elevation that results from dissolved sugars in water
3
4
5
6
7
8
9
10
11
12
13
14
15
16
limits maximum yields to about 75%. In addition, use of steam is energy intensive 3, 19
.
17
18
19
20
21
22
23
Acid catalyzed furfural production has employed several different strong mineral acids
such as hydrochloric, sulfuric, phosphoric, and nitric 8, 27, 28, mostly at temperatures of 160-280oC
8, 13, 20, 29. Among these acids, hydrochloric acid realized higher furfural yields from xylose under
otherwise comparable reaction conditions 8, 27, 28. Although furfural can be produced by biomass
pyrolysis, low furfural yields (< 5%) and high temperatures (about 527oC) make this approach
unfavorable 16. Yemis and Mazzaa reported the highest furfural yield of 72.1% for Bronsted
acid-catalyzed conversion of flax shives at 180oC for 20 min, pH 0.13, and a liquid to solid ratio
4

Page 5 of 34
RSC Advances
View Article Online
1
2
3
4
5
of 200:1 28. Acidic cesium salts of 12-tungstophosphoric acid were employed as an active solid
catalyst for cyclodehydration of xylose to give a maximum furfural yield of 33% in a biphasic
system with toluene as a solvent in water at 160oC 29, 30. Other chloride salts (KCl, NaCl, CaCl2,
MgCl2,and FeCl3) were reported to enhance furfural formation from xylose, suggesting that
chloride ions significantly improved xylose dehydration to furfural 31.
6
7
Rapid removal of furfural from the acidic aqueous phase can increase yields by reducing
the opportunity for furfural degradation in low pH water to form polymers and other products
loosely termed humins, and immiscible organic solvents can form a separate phase that extracts
much of the furfural following its formation from xylose 27, 29, 30, 32, 33 . Application of such
solvent systems to cellulosic biomass results in formation of three phases: biomass solids, acidic
aqueous phase, and organic phase. Furfural has been produced from rice straw in an aqueous
system coupled with the solvents2-PrOH, 1-BuOH, tetrahydrofuran (THF), and methyl isobutyl
ketone (MIBK) in a batch reactor without stirring 34. In this case, furfural yields of as high as
118.2 g/kg straw (equal to 78.9% of the theoretical maximum for 20.6% xylan by dry weight) as
well as a 5-HMF yield of 26.3 g/kg straw (equal to 9.0% of the theoretical maximum for the 37.5
% glucan by dry weight) were reported with THF at 150oC for 5 h 35. Furfural in the organic
phase could be simultaneously reacted with ketones by aldol-condensation and oligomerization
to make hydrocarbon fuels 11, 18. Advantage could also be taken of the acid in the water to
convert the cellulose left in the solids following furfural production into levulinic acid.
However, little data is available on how effective biphasic systems are in increasing furfural
yields from actual cellulosic biomass materials along with side-by-side comparisons to results
with pure xylose or xylose extracted from biomass. This is important in that real biomass
streams are expected to behave very differently from the pure xylose solutions obtained from
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
5

RSC Advances
Page 6 of 34
View Article Online
1
2
3
4
5
6
7
either pure chemicals or from biomass that are employed in most literature studies in terms of the
reaction system (solid and two different liquid phases) and the reactions (large biopolymers
hydrolysis and xylose dehydration) involved. Furthermore, the low density and hydrophobicity
of actual biomass feedstocks makes it difficult to process them in the aqueous phase to obtain
products at high yields and concentrations. Therefore, studies with real biomass are needed to
determine the value of solvent extraction in enhancing RI production and develop processing
strategies to further enhance yields.
8
9
In this study, hemicellulose (mostly xylan) in maple wood was converted to furfural by
acid catalyzed hydrolysis and dehydration with and without simultaneous extraction by the
organic solvent methyl isobutyl ketone (MIBK) with the objective to improve furfural yields and
reduce steam usage for furfural extraction. Many solvents including ethyl acetate,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
tetrahydrofuran, methyl isobutyl ketone, other organics, and supercritical carbon dioxide have
been applied for such applications. However, methyl isobutyl ketone has higher partition
coefficients (partition coefficients: 18.57 and 8.39 with and without salt) for furfural between the
organic and aqueous phases, low water solubility 0.28 wt% (almost immiscible with water), and
a relatively low boiling point (116.1oC ) compared to other solvents such as 2-butanone, 2-
methylfuran, ethyl acetate, tetrahydrofuran, and toluene 36. This study also employed two
different acids, HCl and H2SO4, at conditions similar to those reported to maximize yields of
furfural from xylose in the literature: 0.1 mol/L concentrations at 170oC for 30 min 32. For
comparison, pure cellulose, glucose, and xylose were also reacted individually with and without
solvent addition to gain insights into how each responded to this environment compared to their
behaviors without solvent added. Changes in furfural yields due to solvent extraction were also
compared for two process configurations: 1) acid hydrolysis of biomass to five carbon sugars
6

Page 7 of 34
RSC Advances
View Article Online
1
2
followed by their subsequent dehydration to furfural and 2) simultaneous hydrolysis and
dehydration.
3
4
5
6
2. Experimental
2.1. Materials
The maple wood (~35% moisture content, particle size ~ 2mm) used in this study was
grown in Auburn, NY and provided by Mascoma Corporation, Lebanon, NH. The raw maple
wood was comprised of 40.9±0.3% glucan, 15.4±0.2% xylan, and 26.1±0.3% Klason-lignin, as
reported in a previous study of different acid pretreatments of maple wood37. All other
chemicals, including MIBK, Avicel® PH101 cellulose, glucose, xylose, furfural, and sulfuric,
hydrochloric, formic, and levulinic acids were purchased from Sigma-Aldrich (St. Louis, MO,
USA).
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
2.2. Experimental methods
All reactions with and without simultaneous MIBK solvent extraction were conducted in
a stirred 1 L Parr reactor (236HC, Parr instrument, Moline, IL) heated in a sand bath (Model
number: SBL-2D, Techne Inc., Burlington, NJ). Maple wood was reacted at various solids
loadings in 0.1M hydrochloric or sulfuric acid at 170oC with stirring at 200 rpm for times of up
to 80 min. Reactions of pure xylose, glucose, and cellulose were conducted over the same range
of conditions. For reactions with simultaneous solvent extraction, an organic to aqueous phase
weight ratio of 1:1 was used. At the end of the reactions, the reactor was cooled down to room
temperature by transferring to a 50 L water bath, and then both the liquid and solid samples were
collected for component analysis.
7

RSC Advances
Page 8 of 34
View Article Online
1
2
3
4
5
6
7
2.3. Analyses
Simultaneous hydrolysis and dehydration of maple wood in the aqueous phase coupled
with simultaneous solvent extraction involved three phases: solid biomass, an aqueous phase,
and an organic phase. Gaseous products formed during the process were not accounted for
because of the difficulty in collecting and quantifying them. However, because the mass balance
could only account for 88% of xylan and 84% of glucan, quantifying gaseous products would be
useful to more completely close material balances.
8
9
The compositions of untreated maple wood and residues left after reactions were
determined by following National Renewable Energy Laboratory Analytical Procedures (LAP,
NREL, Colorado, US) 38. For analysis of compounds dissolved in the aqueous and organic
phases, liquid samples along with an appropriate range of calibration standards were analyzed on
a Waters Alliance HPLC system (Model 2695) equipped with an RI detector (Model 2414,
Waters Co., Milford, MA). Components were separated with a Bio-Rad Aminex® HPX-87H (300
x 7.8 mm; Catalog No.125-0140) column equipped with a micro-guard cation cartridge (Catalog
No.125-0129; 30 x 4.6 mm; Bio-Rad Laboratories, Hercules, CA) with H2SO4 (5 mM) eluent at
a flow rate of 0.6 ml/min.
10
11
12
13
14
15
16
17
18
19
2.4. Yield Calculations
Yields were calculated and reported based on the maximum amount possible from each
component for the measured composition of maple wood as follows:
20
21
Yield (%) = 100*[Product concentration (g/L)* the amount of original liquid added (L)/ Amount
of reactant (g) in biomass* stoichiometric coefficient]
8

Page 9 of 34
RSC Advances
View Article Online
1
2
3
4
The stoichiometric coefficients for xylan to xylose, xylan to furfural, and xylan to formic acid
are 1.136, 0.727, and 0.348, respectively, and the coefficients for glucose, 5-HMF, levulinic acid,
and formic acid from glucan are 1.111, 0.777, 0.716, and 0.284, respectively. The yields in the
organic phase were calculated in a similar way as those in the aqueous phase:
5
6
Yield (%) = 100*[Product concentration (g/L)* the amount of original organic solvent added (L)/
Amount of reactant (g) in biomass* stoichiometric coefficient]
7
8
Conversion of pure sugars and other chemicals that were in solution was calculated to measure
how much of the total available material reacted to desired and undesired products as:
9
Conversion (%) = 100- (Final amount (g)/ Initial amount (g))
10
11
12
3. Results and Discussion
3.1. Acid catalyzed reactions of biomass
As cellulosic biomass is hydrolyzed and dehydrated under acidic conditions at high
temperatures, the reactions involved are as shown in Equations 1 and 2 for xylan and glucan, the
two major carbohydrates in biomass. Thus, dehydration of xylose to furfural and of glucose to 5-
HMF are followed by reaction to organic acids, which then degrade further under acidic
conditions at high temperatures. For these reactions, the xylan products include xylooligomers,
xylose, furfural, formic acid, and unknown degradation products from furfural and sugars. The
glucan products include cellooligomers, glucose, 5-HMF, levulinic acid, formic acid, and
unknown degradation products. The reaction of glucose to 5-HMF involves isomerization
between glucose and fructose and dehydrations of fructose and glucose to 5-HMF with formation
of additional dehydration, fragmentation, and condensation products 6, 27. The direct degradation
13
14
15
16
17
18
19
20
21
22
of sugars and dehydration products to char complicate fully accounting for all products.
9

RSC Advances
Page 10 of 34
View Article Online
H+
H+
H+
xylan
xylose
furfural
formic
humins
1
(1)
H+
H+
H+
glucan
levulinic formic
glucose
humins
5-HMF
2
3
4
5
(2)
3.2. Comparison of different processes
Furfural was produced from maple wood with HCl catalysis according to three primary
routes, as summarized in Table 1 for preliminary tests with HCl. Method 1 applied two different
solids loadings (20 g and 40 g of maple wood in 0.1 M HCl to total 400 g in the presence of 400
g MIBK for 30 min at 170oC, equivalent to 5 and 10 wt% solids in the aqueous phase) to
determine whether this variable affected performance. Table 1 shows that the monomeric sugars
yields (glucose and xylose) in the aqueous phase were very similar for 5 wt% and 10 wt% solids
with simultaneous MIBK extraction, but the lower solids loadings resulted in much higher
furfural and 5-HMF yields in the organic phase. In addition, overall yields of these products
were higher for reactions for 30 min at 170oC, as were yields of levulinic acid and formic acid.
The difference in product yields might be caused by differences in the ratio of acid to biomass,
mixing, and/or the reaction mechanisms. Based on these results, a 5 wt% solids loading was
employed for the following studies.
6
7
8
9
10
11
12
13
14
15
16
17
18
In contrast to the single step method (Method 1 in Table 1) that combined hydrolysis and
dehydration with simultaneous MIBK extraction, acid hydrolysis was conducted to produce
10

Page 11 of 34
RSC Advances
View Article Online
1
2
sugars before sugar dehydration for Methods 2 and 3. The highest xylose yields of 73.8%
without solvent were achieved in our previous work by reacting maple wood for 10 min in 0.1 M
HCl at 160oC. Furfural could then be produced from the xylose in these solutions with expected
yields of about 90% 32, 39. The residual solids could either be left with the liquid hydrolyzate for
dehydration according to Method 2 or separated from the liquid hydrolyzate according to
Method 3. When hydrolysis was applied before dehydration, furfural yields increased from
50.8% in the single step method (Method 1) to 63.5% and 73.2% for the two step method with
(Method 3) and without (Method 2) solid-liquid separation, respectively. The 73.2% furfural
yield from extracted hemicellulose solution of maple wood by Method 2 was comparable to
literature results from cellulosic biomass with organic solvent extraction 33, 39. Overall, Method 2
achieved a higher furfural yield of 73.2% without solid-liquid separation than the 63.5% yield
from Method 3 with solid-liquid separation under otherwise identical reaction conditions. Thus,
solid-liquid separation was not necessary or even desirable to maximize furfural yields from
cellulosic biomass. Although separate hydrolysis and dehydration enhanced furfural yields and
appeared very promising compared to simultaneous hydrolysis and dehydration, multi-step
methods increase operation complexity and would be more capital intensive for commercial
applications. In addition, because the focus of this study was to enhance furfural yields in a
single step, Method 1 was chosen for optimization of furfural production from maple wood with
both HCl and H2SO4.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
3.3. Product distributions, concentrations, and yields
For simultaneous hydrolysis and dehydration reactions of maple wood in sulfuric acid
(0.1M) with simultaneous solvent extraction, product yields in the organic and aqueous phases at
170oC are shown in Figure 1a and 1b, respectively, versus reaction time. The combined overall
11

RSC Advances
Page 12 of 34
View Article Online
1
2
yields from both phases are shown in Figure 2a, with yields shown for reactions without
simultaneous extraction in Figure 2b as a control. Similarly, the results for reactions in 0.1 M
HCl with and without organic solvent extraction are shown in Figure 3 and 4, respectively.
Reactions with just hot water (without acid) with and without solvent extraction were also run as
controls, with data shown in Figure 5a and 5b, respectively. Yields of xylose, furfural, glucose,
HMF, levulinic acid, and formic acid are summarized in Table 2 for 30 min acid catalyzed
reactions of maple wood, xylose, glucose, and cellulose at 170oC with sulfuric and hydrochloric
acids. Thus, for reactions with simultaneous solvent extraction, sugars and mineral acid were
mostly located in the aqueous phase, furfural was mainly located in the organic phase, and 5-
HMF and its breakdown product, levulinic acid, were found in both phases.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
3.3.1. Furfural and other xylose derivatives
Figure 1a shows that furfural yields in the organic phase for reaction with 0.1M H2SO4 at
170oC increased from 42.1% at 20 min to 81.8% at 50 min and then dropped to 66.1% after 75
min. As shown in Figure 1b, furfural quickly transferred to the organic phase where it was more
soluble soon after it was formed in the aqueous phase, with only small amounts detectable in the
aqueous phase (~6%) 40. The plots of total furfural yields from maple wood for the aqueous and
organic phases combined versus reaction time in Figure 2a show the highest yield to be 85.3%
for a reaction time of 50 min. This yield was about 20% higher than measured for reactions
without solvent extraction, as shown in Figure 2b. Reaction for longer times reduced furfural
yields as furfural degradation became faster than formation.
21
22
23
Since xylose is the intermediate from conversion of xylan in maple wood to furfural, the
profiles of xylose yields for these reactions were also tracked. Because hydrolysis and
dehydration of xylan to xylose and on to furfural occurs primarily in the aqueous phase, xylose
12

Page 13 of 34
RSC Advances
View Article Online
1
2
3
4
5
6
7
8
yields for reactions with and without organic solvent could be easily compared based on
compositions in the aqueous phase, as shown in Figures 1b and 2b. As shown in Figure 2b
without solvent extraction, xylose yields rapidly increased from 11.2% at 30 min to 36.5% at 40
min and then dropped to 26.6% at 50 min and further to 16.9% at 60 min. In contrast, Figure 2a
shows that the presence of organic solvent reduced xylose yields from maple wood as the
reaction time was extended, as xylose reaction to furfural and other products was faster than its
formation from xylan. Overall, it is clear that reaction with 0.1M acid with or without solvent
extraction at 170oC was unable to obtain high xylose yields.
9
10
11
12
13
14
15
Similar results as presented in Figures 1 and 2 for sulfuric acid are shown in Figures 3
and 4 but with 0.1 M hydrochloric acid as the catalyst. Furfural yields in the organic phase for
reaction of maple wood in 0.1 M HCl at 170oC with organic solvent extraction shown in Figure
3a increased from 40.0% at 20 min to 65.0% at 60 min before dropping to 43.1% at 75 min,
while xylose yields in the aqueous phase peaked at 45.0% in 20 min and then dropped to 11.8%
for 75 min, as shown in Figure 3b. Overall, xylose yields were higher and furfural yields were
lower for reaction of maple wood in 0.1 M HCl than with 0.1M H2SO4.
16
17
18
19
20
21
22
23
Plots of overall furfural yields in the aqueous and organic phases are shown in Figure 4a
against time for reaction of maple wood in 0.1M HCl at 170oC. The highest yield of 72.8% was
realized in 60 min, a difference of 9.5% compared to a maximum of 63.3% without solvent
extraction for the same reaction time, as shown in Figure 4b. Without acid, furfural yields were
only 9.6% and 7.0% for reaction with and without organic solvent, as shown in Figure 5a and
Figure 5b, respectively. However, xylose yields without acid but with simultaneous solvent
extraction in control experiments were slightly higher compared to those with no solvent
extraction or acid.
13

RSC Advances
Page 14 of 34
View Article Online
1
2
For experiments with pure xylose at 170oC, the maximum furfural yield was only 4.0%
for the control reaction without acid or solvent. However, it is interesting to note that the
furfural yield with simultaneous extraction and no acid increased to 44.5% and was higher than
that with 0.1 mol/L HCl without simultaneous extraction (36.5%). Besides a significantly higher
furfural yield, xylose accountability (in the forms of residual xylose, furfural, and formic acid)
for reactions in water with simultaneous extraction was close to 100% of the initial xylose
amount, compared to 92.9% for reactions in acid without extraction. The furfural yield was only
46.5% for reaction of pure xylose in 0.1 M H2SO4 without solvent extraction, lower than the
results with simultaneous extraction of 63.9% and 76.1% with 0.1 M H2SO4 and 0.1 M HCl,
respectively. The furfural yield of 76.1% for reaction of pure xylose catalyzed by 0.1 M HCl
with simulations MIBK extraction was comparable to that reported elsewhere for pure xylose
with MIBK and other solvents extraction 27, 32. However, furfural yields from pure xylose did
not reach the 85% of the theoretical as predicted in the literature for kinetic modeling of pure
xylose with MIBK32, suggesting that the reaction system affects the applicability of the kinetic
models. Therefore, to accurately predict furfural production from xylose, more refined models
are needed. In addition to improving furfural yields, MIBK enhanced overall recovery of
unreacted xylose, furfural, and formic acid together in xylose reactions catalyzed by acid, likely
by reducing degradation to side products from xylose and furfural degradation. Overall, this
study showed for the first time that simultaneous organic solvent extraction can increase furfural
yields to greater than 80% in a batch process for real biomass, a somewhat better result than the
75% yields realized in commercial processes by stripping furfural from a continuous process
with steam19.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
14

Page 15 of 34
RSC Advances
View Article Online
1
2
3.3.2. 5-HMF and other glucose derivatives
During reaction of xylan in maple wood to xylose and on to furfural, some of the glucan
was hydrolyzed to glucose that further dehydrated to 5-HMF before being rehydrated to levulinic
acid (LA). Similar to furfural, most of the 5-HMF was in the organic phase for the solvent
system. As shown in Figure 1a for reaction in 0.1 M H2SO4 at 170oC with simultaneous solvent
extraction, 5-HMF yields in the organic phase increased from only 8.1% at 20 min to peak at
33.0% at 60 min and then dropped to 28.6% at 75 min. The small amount of levulinic acid also
found in the organic phase increased with reaction time, as shown in Figure 1a. However,
glucose yields in the aqueous phase increased slightly from 15.2% at 20 min to 16.1% at 30 min
before dropping to 2.9% at 60 min, as shown in Figure 1b.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Figure 2a shows that reaction with 0.1 M H2SO4 at 170oC increased the overall 5-HMF
yields from 10.3% at 20 min to a peak value of 35.8% at 60 min. Then the overall 5-HMF yield
dropped slightly to 31.6% at 75 min. However, yields of 5-HMF and levulinic acid in the
aqueous phase were low. The product with the highest yields from glucan in maple wood during
furfural production with solvent extraction in 0.1 M H2SO4 at 170oC was 5-HMF in both aqueous
and organic phases. Without solvent extraction, the most abundant product derived from glucan
was glucose with a yield of 28.7%, as shown in Figure 2b. Similar to 5-HMF and levulinic acid
production with solvent extraction in Figure 1b, yields of 5-HMF and LA in the aqueous phase
for reaction with 0.1 M H2SO4 at 170oC without solvent extraction were very small as well, as
shown in Figure 2b.
21
22
23
Total 5-HMF yields with 0.1M HCl at 170oC with simultaneous organic solvent
extraction increased from 5.8% to 30.4% as the reaction time progressed from 20 min to 30 min
before peaking at 32.2% for a 45 min reaction time, as shown in Figure 4a. Although the highest
15

RSC Advances
Page 16 of 34
View Article Online
1
2
3
4
5
6
5-HMF yields from maple wood with 0.1 M H2SO4 or HCl at 170oC were very close with most
of the 5-HMF in the organic phase, the reaction time needed to reach the maximum yield was
shorter for 0.1 M HCl, suggesting that 5-HMF benefits from the greater acid strength of HCl.
However, as shown in Figure 4b, reaction with HCl without simultaneous solvent extraction
resulted in 5-HMF yields below 4.5% and glucose yields below 26.7%, slightly lower than the
results shown in Figure 2b for catalysis with H2SO4.
7
8
As shown in Figure 5, control experiments without acid (just hot water) with or without
solvent extraction gave much lower product yields from glucan. At reaction conditions that gave
the highest furfural yields (0.1 M sulfuric acid at 170oC for 50 min), yields were only 7.1% for
glucose, 8.0% for levulinic acid, and 27.3% for 5-HMF, together accounting for about 43% of
the cellulose available in the untreated biomass.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
In order to further investigate breakdown of glucan in maple wood to glucose, 5-HMF,
and levulinic acid, sulfuric and hydrochloric acids at 0.1M concentrations were applied to pure
glucose and pure cellulose at 170oC for 30 min with and without simultaneous solvent extraction,
with results summarized in Table 2. Solvent extraction almost doubled the 5-HMF yields from
glucose from a maximum of 23.8% and 19.6% with H2SO4 and HCl, respectively, without
organic solvent to about 45.0% for both acids in the presence of solvent. These results indicated
that water contributed to glucose dehydration to 5-HMF as a reactant and solvent, but the
difference between the effects of the two acids was insignificant. However, acid was critical to
conversion of pure glucose to 5-HMF and 5-HMF to levulinic acid, since without acid, very little
levulinic acid was detected, and the highest 5-HMF yields from glucose were only 12.6% and
8.7% with and without organic solvent extraction, respectively.
16

Page 17 of 34
RSC Advances
View Article Online
1
2
3
4
5
6
7
8
9
Table 2 shows that acid was particularly essential to obtain reasonable product yields
from cellulose, as yields without acid were negligible. However, for reaction of cellulose in
either 0.1 M HCl or 0.1 M H2SO4 at 170oC for 30 min, product yields were much higher for
systems employing simultaneous extraction than those without, suggesting a high fraction of 5-
HMF was extracted in the organic phase. On the other hand, solvent extraction did not have
much impact on glucose yields, undoubtedly because glucose is far more soluble in water than it
is in MIBK 36. Product yields from cellulose with HCl were significantly higher than those with
H2SO4, both with and without solvent extraction, showing that high acid strength is more
important to obtaining high yields from cellulose than from hemicellulose.
10
11
12
13
14
15
16
17
18
19
20
21
These results with cellulose suggest that the sulfuric and hydrochloric acids have
significantly different effects on cellulose hydrolysis to glucose and the further reactions of
glucose and its degradation products. Due to the similarity in glucose dehydration with the two
acids with and without solvent extraction, the major difference appears to be in cellulose
hydrolysis to glucose. In addition to differences in 5-HMF production from glucose and from
cellulose, differences in results with the homogenous solution of glucose in water and the
heterogeneous mixture of cellulose with water could also contribute to lower 5-HMF yields
without acid. Compared to the 8.7% 5-HMF yield from glucose without acid and solvent,
products from cellulose hydrolysis and subsequent dehydration were negligible (0.1% and 0.2%)
under the same conditions. This result is consistent with what one would expect in light of the
activation energy for cellulose hydrolysis being ~180 kJ/mol and that for glucose decomposition
being ~140 (kJ/mol), as determined a few decades ago 41, 42
.
22
23
Reactions of Avicel cellulose at 1 wt% solids loading in 0.1 M HCl with MIBK
extraction gave similar yields (glucose yield of 28.5% and 5-HMF yield of 35.6%) as for maple
17

RSC Advances
Page 18 of 34
View Article Online
1
2
3
4
5
6
7
8
9
wood at a 5wt% solids loading ( glucose yield of 30.3% and 5-HMF yield of 32.4%). However,
similar reactions of Avicel in 0.1 M H2SO4 with MIBK extraction gave different yields (glucose
yield of 15.1% and 5-HMF yield of 16.2%) than maple wood (7.1% glucose and 35.8% 5-HMF)
at a solid loading of 5wt%. Therefore, H2SO4 appeared to be facilitate conversion of cellulose in
maple wood but not pure Avicel® PH101 microcrystalline cellulose to 5-HMF. Levulinic acid
yields from Avicel cellulose were slightly lower than those from maple wood, but glucose yields
were much higher, possibly due to structural differences between Avicel® cellulose and the
cellulose in maple wood. Thus, Avicel® cellulose was more easily hydrolyzed to glucose than
the cellulose in maple wood, but this result deserves further study to understand why.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
3.4. Solids residue
The solids left after furfural production were analyzed to track how the mass of key
components and compostion changed. Figure 6 compares the mass of each major component left
in the solid residues following maple wood acid hydrolysis reactions with and without
simultaneous organic solvent extraction adjusted to a basis of 100 g of original maple wood. As
shown, about 48% of the initial weight was removed after 40 min of hydrolysis in 0.1M sulfuric
acid without solvent extraction and about 50% (~ 26.4 g) of the remaining mass was Klason-
lignin/chars. In other words, the amount of Klason-lignin/chars increased from 20 g in the
untreated biomass to 26.4 g due to degradation of sugars, furfural, and other compounds to
insoluble chars. For reaction times greater than 40 min, however, the mass of solid residue
dropped, the fraction of humins/lignin increased, and the total amount of humans/lignin stayed
almost constant at about 26 g. On the other hand, as shown in Figure 6b, with simultaneous
solvent extraction, the amount of solids residue was ~15% lower after 60 min of reaction than for
reaction for the same time without organic solvent. The amount of lignin/chars measured after
18

Page 19 of 34
RSC Advances
View Article Online
1
2
3
4
5
6
7
8
9
60 min of acid hydrolysis with simultaneous organic solvent extraction dropped from 20 g in raw
maple wood to 13.9 g, as shown in Figure 6b, possibly due to reduced formation of chars from
sugars and other compounds and lignin dissolution in the aqueous and organic phases. At this
condition, all of the xylan in maple wood was removed whether solvent extraction was applied or
not. Thus, the significant reduction in loss of sugars to chars resulting from reactions between
sugars/intermediates and furfural was another important advantage of simultaneous organic
solvent extraction. Because about 7 to 16% of the lignin was measured to dissolve during
solvent extraction, the drop in lignin/chars indicated that simultaneous organic solvent extraction
reduced loss of sugars to chars, thereby reducing carbon losses.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
3.5. The effect of different acids
Because HCl, H2SO4, and other mineral acids are immiscible with the organic phase but
hydrolysis, dehydration, rehydration, and decarboxylation are catalyzed by hydrogen ions, most
of the reactions occurred only in the aqueous phase. Therefore, both furfural and 5-HMF were
protected from further degradation as they were partitioned into the organic phase. Although
both HCl and H2SO4 were able to improve hydrolysis and dehydration product yields compared
to reaction without added acid, the dissociation constants of pKa = -7 for hydrochloric acid and
pKa1 = -3 and pKa2 =1.99 for sulfuric acid having different effects on product yields and
distributions 36. Hydrochloric acid is stronger than sulfuric acid and stands out for its ability to
catalyze xylose decomposition into furfural and its breakdown products. In the case of glucan
and glucose reactions, hydrochloric acid hydrolyzed cellulose to 5-HMF, which broke down
further if there was no organic solvent extraction. Although H2SO4 is a weaker acid catalyst than
HCl at the same concentration, it gave higher furfural and similar 5-HMF yields from maple
wood with organic solvent extraction at 170oC. When coupled with the fact that H2SO4 is less
19

RSC Advances
Page 20 of 34
View Article Online
1
2
3
4
corrosive, toxic, and expensive than HCl, it would likely be favored for furfural and 5-HMF
production from cellulosic biomass. This finding is consistent with our previous study of
hemicellulose release with different acid pretreatment technologies, which showed that sulfuric
acid was a weak oxidizer and promoted less xylose dehydration than hydrochloric acid37.
5
6
4. Conclusions
Simultaneous organic solvent extraction was successfully employed in a single step batch
system to increase furfural yields for maple wood from about 64% without extraction to about
85% for otherwise identical reaction conditions of 170oC in 0.1 M sulfuric acid. These results
also showed that solvent extraction could reduce xylan degradation, lower carbon losses to chars,
and improve overall accountability of xylan and its breakdown products. Solvent extraction also
improved recovery of other important products such as 5-HMF and levulinic acid, suggesting
this approach would be able to substantially improve conversion of cellulosic biomass to these
valuable chemicals.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
5. Acknowledgements
This work was supported by the Defense Advanced Research Projects Agency (DARPA)
and Army Research Laboratory (ARL) through the Defense Science Office Cooperative
Agreement W911NF-09-2-0010/09-005334 C 00 (Surf-Cat: Catalysts for production of JP-8
range molecules from lignocellulosic biomass). The views, opinions, and/or findings contained
in this article are those of the authors and should not be interpreted as representing the official
views or policies, either expressed or implied, of the Defense Advanced Research Projects
Agency or the Department of Defense. We also acknowledge Mascoma Corporation for
providing the red maple. We thank the Center for Environmental Research and Technology
20

Page 21 of 34
RSC Advances
View Article Online
1
2
3
4
(CE-CERT) at the University of California, Riverside, for providing facilities used for this
research. The corresponding author is also grateful to the Ford Motor Company for funding the
Chair in Environmental Engineering at CE-CERT in the UCR Bourns College of Engineering
that greatly facilitates many projects such as this one.
5
6. References
6
7
8
1.
2.
3.
J. Binder, J. Blank, A. Cefali and R. Raines, ChemSusChem, 2010, 1268-1272.
G. Dautzenberg, M. Gerhardt and B. Kamm, Holzforschung, 2011, 65, 439-451.
W. De Jong and G. Marcotullio, International Journal of Chemical Reactor Engineering, 2010, 8,
Ariticle A69.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
4.
A. Mamman, J. Lee, Y. Kim, I. Hwang, N. Park, Y. Hwang, J. Chang and J. Hwang, Biofuels
Bioproducts & Biorefining-Biofpr, 2008, 438-454.
J. J. Bozell and G. R. Petersen, Green Chemistry, 2010, 12, 539-554.
J. Chheda and J. Dumesic, Catalysis Today, 2007, 59-70.
G. Huber and J. Dumesic, Catalysis Today, 2006, 119-132.
G. Marcotullio, M. A. T. Cardoso, W. De Jong and A. H. M. Verkooijen, International Journal of
Chemical Reactor Engineering, 2009, 7, 1-16.
Y. Roman-Leshkov, C. J. Barrett, Z. Y. Liu and J. A. Dumesic, Nature, 2007, 447, 982-U985.
A. Brassat, Thewes, M., Müther, M., Pischinger, S. et al., Analysis of the Effects of Certain Alcohol
and Furan-Based Biofuels on Controlled Auto Ignition, 2012.
A. Aden, J. Bozell, J. Holladay, J. White, A. Manheim, D. Elliot, L. Lasure, S. Jones, M. Gerber, K.
Ibsen, L. Lumberg and S. Kelley, Top Value Added Chemicals from Biomass Volume I—Results of
Screening for Potential Candidates from Sugars and Synthesis Gas, Pacific Northwest National
Laboratory (PNNL), National Renewable Energy Laboratory (NREL), U.S. Department of Energy,
2004.
J. Z. Karl, in Sugar Series, ed. J. Z. Karl, Elsevier, 2000, vol. Volume 13, p. 14.
M. J. Antal Jr, T. Leesomboon, W. S. Mok and G. N. Richards, Carbohydr. Res., 1991, 217, 71-85.
J. Z. Karl, in Sugar Series, ed. J. Z. Karl, Elsevier, 2000, vol. Volume 13, pp. 15-18.
M. Demirbas, Energy Sources Part a-Recovery Utilization and Environmental Effects, 2006, 1181-
1188.
C. Di Blasi, C. Branca and A. Galgano, Industrial & Engineering Chemistry Research, 2010, 2658-
2671.
S. Makishima, M. Mizuno, N. Sato, K. Shinji, M. Suzuki, K. Nozaki, F. Takahashi, T. Kanda and Y.
Amano, Bioresource Technology, 2009, 2842-2848.
R. Karinen, K. Vilonen and M. Niemela, ChemSusChem, 2011, 4, 1002-1016.
D. T. Win, AU J.T, 2005, 8, 185-190.
C. S. M. H. J. Brownlee, Ind. Eng. Chem. , 1948, 40, 201– 204.
S. Fitzpatrick, Feedstocks For the Future: Renewables For the Production of Chemicals and
Materials, 2006, 271-287.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
S. Ritter, Chem Eng News, 2006, 84, 47-47.
21

RSC Advances
Page 22 of 34
View Article Online
1
2
3
23.
D. Hayes, J.; Fitzpatrick, Steve, Hayes, M.H.B.; Ross, J.R.H.;, in Ullmann's Encyclopedia of
Industrial Chemistry, eds. B. Kamm, P. R. Gruber and M. Kamm, Wiley-VCH Verlag GmbH & Co.
KGaA, 2005, ch. Biorefineries – Industrial Processes and Products
4
5
6
7
8
9
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Letvia Pat., Process for producing of furfural and acetic acid (LV11032), 1996.
Letvia Pat., Process for producing of furfural and acetic acid (LV11950), 1998.
Letvia Pat., Process for producing of furfural and acetic acid (LV1131), 1998.
J. Chheda, Y. Roman-Leshkov and J. Dumesic, Green Chemistry, 2007, 342-350.
O. Yemis and G. Mazza, Bioresource Technology, 2011, 102, 7371-7378.
A. S. Dias, M. Pillinger and A. A. Valente, Journal of Catalysis, 2005, 229, 414-423.
A. S. Dias, S. Lima, M. Pillinger and A. A. Valente, Carbohydrate Research, 2006, 341, 2946-2953.
G. Marcotullio and W. De Jong, Green Chemistry, 2010, 12, 1739-1746.
R. Weingarten, J. Cho, W. C. Conner and G. W. Huber, Green Chemistry, 2010, 12, 1423-1429.
E. I. Gurbuz, S. G. Wettstein and J. A. Dumesic, ChemSusChem, 2012, 5, 383-387.
O. Yemis and G. Mazza, Bioresource Technology, 2012, 109, 215-223.
H. Amiri, K. Karimi and S. Roodpeyma, Carbohydrate Research, 2010, 345, 2133-2138.
D. R. Lide, CRC Handbook Of Chemistry And Physics, 88th Edition CRC Press Taylor & Francis,
2007-2008.
T. Zhang, R. Kumar and C. E. Wyman, Carbohydrate Polymers, 2013, 92 334– 344.
A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton and D. Crocker, Determination
of Structural Carbohydrates and Lignin in Biomass Laboratory Analytical Procedure (LAP), 2008.
R. Xing, W. Qi and G. Huber, Energy & Environmental Science, 2011, 4, 2193-2205.
J. R. Croker and R. G. Bowrey, Ind Eng Chem Fund, 1984, 23, 480-484.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
37.
38.
39.
40.
41.
42.
K. R. Heimlich and A. N. Martin, J Am Pharm Assoc, 1960, 49, 592-597.
H. E. Grethlein, J Appl Chem Biotechn, 1978, 28, 296-308.
25
26
27
22

Page 23 of 34
RSC Advances
View Article Online
1
2
3
Table Captions
4
5
Table 1. Conditions applied to reaction of maple wood with HCl and simultaneous solvent
extraction and resulting product yields for each.
6
7
8
9
Table 2. Yields of xylose (Xyl), furfural (FF), glucose (glu), hydroxymethyl furfural (HMF),
levulinic acid (LA), and formic acid (FA) from xylose, glucose, cellulose, and maple wood for
30 min reactions catalyzed by dilute sulfuric or hydrochloric acid at 170oC with and without
simultaneous organic solvent extraction.
10 Figure Captions
11
12
13
14
15
Figure 1. Effect of reaction time on product yields in the (a) organic phase and (b) aqueous
phase for reaction of maple wood at 170oC in 0.1 M H2SO4 with simultaneous MIBK extraction.
The total reaction mass was 800 g, and the mass ratio of organic (MIBK) to aqueous phase was
1:1. The solids loading of maple wood was 5wt% in the aqueous phase before solvent addition.
16
17
18
19
20
Figure 2. Overall product yields from the overall process for (a) organic solvent extraction and
(b) no organic solvent extraction for reaction of maple wood at 170oC in 0.1 M H2SO4. The total
reaction mass was 800 g, and the mass ratio of organic (MIBK) to aqueous phase for organic
solvent extraction was 1:1. The solids loading of maple wood was 5wt% in the aqueous phase
before solvent addition.
21
22
23
24
Figure 3. Effect of reaction time on product yields in the (a) organic phase and (b) aqueous
phase for reaction of maple wood at 170oC in 0.1M HCl with simultaneous MIBK extraction.
The total reaction mass was 800 g, and the mass ratio of organic (MIBK) to aqueous phase was
1:1. The solids loading of maple wood was 5wt% in the aqueous phase before solvent addition.
23

RSC Advances
Page 24 of 34
View Article Online
1
2
3
4
Figure 4. Overall product yields for (a) organic solvent extraction and (b) no organic solvent
extraction for reaction of maple wood at 170oC in 0.1 M HCl. The total reaction mass was 800
g, and the mass ratio of organic (MIBK) to aqueous phase for organic solvent extraction was 1:1.
The solid loading of maple wood was 5wt% in the aqueous phase before solvent addition.
5
6
7
8
9
Figure 5. Overall product yields from maple wood reacted at 170oC (a) without adding acid but
with organic solvent extraction and (b) without adding acid and without organic solvent
extraction. The total reaction mass was 800 g, and the mass ratio of organic (MIBK) to aqueous
phase for organic solvent extraction was 1:1. The solids loading of maple wood was 5wt% in the
aqueous phase before solvent addition.
10
11
12
Figure 6. Yields of xylan, glucan, lignin, and unidentified other compounds left in the solids for
(a) no organic solvent extraction and (b) organic solvent extraction for reaction of maple wood at
170oC in 0.1 M H2SO4.
13
14
24

Page 25 of 34
RSC Advances
Figure 1

RSC Advances
Page 26 of 34
Figure 2

Page 27 of 34
RSC Advances
Figure 3

RSC Advances
Page 28 of 34
Figure 4

Page 29 of 34
RSC Advances
Figure 5