Simple chemical transformation of lignocellulosic biomass into furans for fuels and chemicals

Article abstract of DOI:10.1021/ja808537j

Lignocellulosic biomass is a plentiful and renewable resource for fuels and chemicals. Despite this potential, nearly all renewable fuels and chemicals are now produced from edible resources, such as starch, sugars, and oils; the challenges imposed by notoriously recalcitrant and heterogeneous lignocellulosic feedstocks have made their production from nonfood biomass inefficient and uneconomical. Here, we report that N,N-dimethylacetamide (DMA) containing lithium chloride (LiCl) is a privileged solvent that enables the synthesis of the renewable platform chemical 5-hydroxymethylfurfural (HMF) in a single step and unprecedented yield from untreated lignocellulosic biomass, as well as from purified cellulose, glucose, and fructose. The conversion of cellulose into HMF is unabated by the presence of other biomass components, such as lignin and protein. Mechanistic analyses reveal that loosely ion-paired halide ions in DMA-LiCl are critical for the remarkable rapidity (1-5 h) and yield (up to 92%) of this low-temperature (≤140 °C) process. The simplicity of this chemical transformation of lignocellulose contrasts markedly with the complexity of extant bioprocesses and provides a new paradigm for the use of biomass as a raw material for a renewable energy and chemical industries.

Full text of DOI:10.1021/ja808537j

Published on Web 01/21/2009
Simple Chemical Transformation of Lignocellulosic Biomass
into Furans for Fuels and Chemicals
Joseph B. Binder† and Ronald T. Raines*,†,‡
Departments of Chemistry and Biochemistry, UniVersity of WisconsinsMadison,
Madison, Wisconsin 53706
Received October 30, 2008; E-mail: rtraines@wisc.edu
Abstract: Lignocellulosic biomass is a plentiful and renewable resource for fuels and chemicals. Despite
this potential, nearly all renewable fuels and chemicals are now produced from edible resources, such as
starch, sugars, and oils; the challenges imposed by notoriously recalcitrant and heterogeneous lignocellulosic
feedstocks have made their production from nonfood biomass inefficient and uneconomical. Here, we report
that N,N-dimethylacetamide (DMA) containing lithium chloride (LiCl) is a privileged solvent that enables
the synthesis of the renewable platform chemical 5-hydroxymethylfurfural (HMF) in a single step and
unprecedented yield from untreated lignocellulosic biomass, as well as from purified cellulose, glucose,
and fructose. The conversion of cellulose into HMF is unabated by the presence of other biomass
components, such as lignin and protein. Mechanistic analyses reveal that loosely ion-paired halide ions in
DMA-LiCl are critical for the remarkable rapidity (1-5 h) and yield (up to 92%) of this low-temperature
(e140 °C) process. The simplicity of this chemical transformation of lignocellulose contrasts markedly with
the complexity of extant bioprocesses and provides a new paradigm for the use of biomass as a raw material
for a renewable energy and chemical industries.
Introduction
with water. These attributes bode well for the use of DMF as
an alternative liquid fuel for transportation.
Throughout most of human history, renewable biomass
resources have been the primary industrial and consumer
feedstocks. Only in the past 150 years have coal, natural gas,
and petroleum grown into their dominant roles as sources for
energy and chemicals.1 Today these fossil resources supply
approximately 86% of energy and 96% of organic chemicals,
but in as soon as two decades petroleum production is unlikely
to meet the growing needs of humanity and natural gas resources
will be increasingly inaccessible.2 Moreover, consumers and
governments concerned about CO2 emissions and other envi-
ronmental impacts are demanding renewable power and products.
With advances in conversion technology, plentiful biomass
resources have the potential to regain their central position as
feedstocks for civilization, particularly as renewable carbon
sources for transportation fuels and chemicals.3 A hexose
dehydration product, the platform chemical 5-hydroxymethyl-
furfural (HMF), will be a key player in the biobased renaissance.
This six-carbon analogue of commodity chemicals like tereph-
thalic acid and hexamethylenediamine can be converted by
straightforward methods into a variety of useful acids, aldehydes,
alcohols, and amines, as well as the promising fuel 2,5-
dimethylfuran (DMF).4,5 The energy content of DMF (31.5 MJ/
L) is similar to that of gasoline (35 MJ/L) and 40% greater
than that of ethanol (23 MJ/L).6,7 Moreover, DMF (bp 92-94
°C) is less volatile than ethanol (bp 78 °C) and is immiscible
Despite the potential for HMF-based fuels and chemicals,
most efforts toward HMF production have used edible starting
materials, primarily fructose and glucose. In fact, almost all
renewable fuels and chemicals are usually based on food
resources such as starch, sugars, and oils. These simple starting
materials are easy to convert into valuable products, while
inedible lignocellulosic biomass is relatively recalcitrant and
heterogeneous, making its conversion typically inefficient and
uneconomical.8 In the case of HMF, its formation by the
dehydration of fructofuranose is straightforward and has been
demonstrated in water, traditional organic solvents,9 multiphase
(3) (a) Werpy, T.; Petersen, G.; Aden, A.; Bozell, J.; Holladay, J.; White,
J.; Manheim, A. Top value-added chemicals from biomass, volume I:
Results of screening for potential candidates from sugars and synthesis
gas; Report DOE/GO-102004-1992; Office of Scientific and Technical
Information, U.S. Department of Energy, Oak Ridge, TN, 2004;
www1.eere.energy.gov/biomass/pdfs/35523.pdf. (b) Perlack, R. D.;
Wright, L. L.; Turhollow, A. F.; Graham, R. L.; Stokes, B. J.; Erbach,
D. C. Biomass as feedstock for a bioenergy and bioproducts industry:
The technical feasibility of a billion-ton annual supply; Report DOE/
GO-102995-2135; U.S. Department of Energy and U.S. Department
of Agriculture, Oak Ridge, TN, 2005; feedstockreview.ornl.gov/pdf/
billion_ton_vision.pdf. (c) Chheda, J. N.; Huber, G. W.; Dumesic, J. A.
Angew. Chem., Int. Ed. 2007, 46, 7164–7183.
(4) Lewkowski, J. ARKIVOC 2001, 17–54.
(5) Lichtenthaler, F. W. Acc. Chem. Res. 2002, 35, 728–737.
(6) Nisbet, H. B. J. Inst. Petrol. 1946, 32, 162–166.
(7) Roman-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Nature
2007, 447, 982–986.
† Department of Chemistry.
‡ Department of Biochemistry.
(8) Zhang, Y.-H. P.; Ding, S.-Y.; Mielenz, J. R.; Cui, J.-B.; Elander, R. T.;
Lasser, M.; Himmel, M. E.; McMillan, J. R.; Lynd, L. R. Biotechnol.
Bioeng. 2007, 97, 214–223.
(1) Kamm, B., Gruber, P. R., Kamm, M., Eds. BiorefineriessIndustrial
Processes and Products; Wiley-VCH: Weinheim, Germany, 2006.
(2) Facing the Hard Truths about Energy; U.S. National Petroleum
Council: Washington, DC, 2007.
(9) Halliday, G. A.; Young, R. J., Jr.; Grushin, V. V. Org. Lett. 2003, 5,
2003–2005.
9
10.1021/ja808537j CCC: $40.75
2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 1979–1985 1979

A R T I C L E S
Binder and Raines
Table 1. Synthesis of HMF from Fructosea
systems,10,11 and ionic liquids.12-14 Zhao et al.13 reported that
chromium catalysts in alkylimidazolium chloride ionic liquids,
such as 1-ethyl-3-methylimidazolium chloride ([EMIM]Cl),
enable synthesis of HMF from glucose, a less expensive
feedstock, in good yield. While promising, this method depends
on expensive ionic liquid solvents and still uses starch-derived
glucose. In our research, we sought to address these concerns
by both minimizing the use of ionic liquids and utilizing
authentic lignocellulosic biomass as the starting material for
HMF production.
catalyst,
mol %
T
(°C)
time molar
(h) yield (%)
solvent
additives, wt %
DMA-LiCl (10%)
120
80
80
100
100
120
120
140 0.5
140 0.5
80
80
80
120
120 1.5
100
100
100
100
100
100
80
2
65
66
62
63
62
68
71
66
58
70
72
78
67
83
84
71
59
71
48
81
56
63
DMA-LiCl (10%) H2SO4, 6
DMA-LiCl (10%) CuCl, 6
DMA-LiCl (10%) H2SO4, 6
DMA-LiCl (10%) CuCl, 6
DMA-LiCl (10%) H2SO4, 6
DMA-LiCl (10%) CuCl, 6
DMA-LiCl (10%) H2SO4, 6
DMA-LiCl (10%) CuCl, 6
4
5
5
5
1
3
DMA-LiCl (10%) H2SO4, 6 [EMIM]Cl, 5
DMA-LiCl (10%) H2SO4, 6 [EMIM]Cl, 10
DMA-LiCl (10%) H2SO4, 6 [EMIM]Cl, 20
4
4
4
1
DMA-LiCl (10%) CuCl, 6
DMA-LiCl (10%) CuCl, 6
DMA
DMA-LiCl (10%) H2SO4, 6 [EMIM]BF4, 20
DMA H2SO4, 6 [EMIM]BF4, 20
DMA-LiCl (10%) H2SO4, 6 [EMIM]OTf, 20
[EMIM]Cl, 20
[EMIM]Cl, 40
H2SO4, 6 [EMIM]Cl, 20
2
2
4
1
2
2
2
2
DMA
DMA
DMA
DMA
H2SO4, 6 [EMIM]OTf, 20
H2SO4, 6 [EtPy]Cl, 20
H2SO4, 6 KCl, 1.5
H2SO4, 6 KCl, 1.5;
18-crown-6, 5.6
H2SO4, 6 LiF, 10
H2SO4, 6 LiBr, 10
H2SO4, 6 NaBr, 10
H2SO4, 6 KBr, 10
H2SO4, 6 LiI, 10
Noting that the chloride counterions in [EMIM]Cl form
only weak ion pairs,15 we reasoned that other nonaqueous
solvents containing a high concentration of chloride ions
could be as effective as [EMIM]Cl for HMF synthesis. We
were aware that DMA containing LiCl is one of few solvents
that can dissolve purified cellulose without modifying its
chemical structure and can do so to a concentration of up to
15 wt %, and that DMA-LiCl can also dissolve simple
sugars.16,17 These useful properties likely stem from the
association of lithium ions with DMA to form DMA · Li+
macrocations, resulting in a high concentration of weakly
ion-paired chloride ions (2.0 M in saturated anhydrous
DMA-LiCl versus 6.8 M in [EMIM]Cl). These chloride ions
can form hydrogen bonds with the hydroxyl groups of
cellulose, disrupting its otherwise extensive network of intra-
and interchain hydrogen bonds. Here, we report that using
DMA-LiCl as a solvent enables the efficient synthesis of
HMF in a single step from cellulose and even lignocellulosic
biomass, as well as glucose and fructose. This discovery
facilitates access to fuels and chemicals derived from HMF.
80
DMA
DMA
DMA
DMA
DMA
DMA
DMA
80
100
100
100
100
100
100
2
4
2
2
6
5
5
0
92
93
92
89
91
92
H2SO4, 6 NaI, 10
H2SO4, 6 KI, 10
a Fructose was reacted at a concentration of 10 wt % relative to the
total mass of the reaction mixture. The solvent composition is indicated
by the weight percent of LiCl relative to DMA with additive
concentrations relative to the total mass of the reaction mixture. Catalyst
loading is relative to fructose. Yields are based on HPLC analysis.
through catalysis by Brønsted acids (e.g., H2SO4) and metal
chlorides (e.g., CuCl, CuCl2, and PdCl2). Nonetheless, these
yields of HMF are moderate when compared to the yields of
up to 85% achieved in solvents such as DMSO and [EMIM]Cl.9,13
In an effort to favor the formation of HMF, we added increasing
amounts of [EMIM]Cl to the DMA-LiCl medium (Table 1).
In general, the addition of ionic liquid improved yields. For
example, [EMIM]Cl in combination with CuCl bolstered the
yield of HMF up to 83%.
Intrigued by the effect of [EMIM]Cl, we investigated the
influence of other additives on the dehydration of fructose in
DMA as catalyzed by H2SO4 (Table 1). The trifluoromethane-
sulfonate and tetrafluoroborate salts of EMIM delivered modest
HMF yields, which were increased by the addition of LiCl.
Adding ionic liquids with a variety of cationic counterions for
chloride delivered HMF in yields around 80%. These results
suggested that chloride ion mediated the crucial advantage of
[EMIM]Cl as an additive. Furthermore, the reaction appeared
to proceed better with the loosely ion-paired chloride afforded
by ionic liquids than with lithium chloride. We also observed
that the potassium-complexing agent 18-crown-6 increased the
yield of HMF in reactions utilizing potassium chloride. These
results indicate that weakly ion-paired halide ions18 favor the
reaction.
Results and Discussion
Synthesis of HMF from Fructose in DMA with Halide
Additives. We began by exploring the reactivity of fructose in
DMA-LiCl, both alone and with added catalysts. When mixed
with DMA-LiCl and heated to sufficiently high temperatures
(80-140 °C), fructose is converted to HMF in moderate yields
(55-65%; Table 1). Increased yields up to 71% can be obtained
(10) Roman-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Science 2006, 312,
1933–1937.
(11) Chheda, J. N.; Roman-Leshkov, Y.; Dumesic, J. A. Green Chem. 2007,
9, 342–350.
(12) Moreau, C.; Finiels, A.; Vanoye, L. J. Mol. Catal. A: Chem. 2006,
253, 165–169.
(13) Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007,
316, 1597–1600.
(14) Zhao, H.; Holladay, J. E.; Zhang, Z. C. U.S. Pat. Appl. 20080033187,
2008.
(15) (a) Chiappe, C.; Pieraccini, D. J. Phys. Org. Chem. 2005, 18, 275–
297. (b) Chowdhury, S.; Mohan, R. S.; Scott, J. L. Tetrahedron 2007,
63, 2363–2389. (c) El Seoud, O. A.; Koschella, A.; Fidale, L. C.;
Dorn, S.; Heinse, T. Biomacromolecules 2007, 8, 2629–2647.
(16) McCormick, C. L.; Callais, P. A.; Hutchinson, B. H., Jr. Macromol-
ecules 1985, 18, 2394–2401.
We found distinct differences in the ability of halide ions to
mediate the formation of HMF from fructose in DMA containing
(17) Potthast, A.; Rosenau, T.; Buchner, R.; Roder, T.; Ebner, G.;
Bruglachner, H.; Sixta, H.; Kosma, P. Cellulose 2002, 9, 41–53.
(18) Weaver, W. M.; Hutchison, J. D. J. Am. Chem. Soc. 1964, 86, 261–
265.
9
1980 J. AM. CHEM. SOC. VOL. 131, NO. 5, 2009

Transforming Lignocellulosic Biomass into Furans
A R T I C L E S
Figure 1. Halide salts in DMA enable previously elusive yields of bio-based chemicals from a variety of carbohydrates. Conditions are optimized for the
conversion of carbohydrates into HMF (1 step) and DMF (2 steps). (Photograph courtesy of Department of Energy/National Renewable Energy Laboratory.)
H2SO4. As expected from their low nucleophilicity and high
basicity, fluoride ions were completely ineffective. On the other
hand, bromide and iodide ions, which tend to be less ion-paired
than fluoride or chloride, enabled exceptionally high HMF yields
in DMA. For example, adding 10 wt % lithium bromide or
potassium iodide enabled the conversion of 92% of fructose to
HMF in 4-5 h at 100 °C (Figure 1). These reactions were highly
selective, resulting in only low levels of the colored byproducts
and insoluble polymeric products (i.e., humins) often formed
concurrently with HMF.4 Kinetic analyses indicate that the rate
of HMF formation has a first-order dependence on halide
concentration (Figure 2).
Influence of Halides on the Formation of HMF from Fruc-
tose. Our results provide mechanistic insights regarding the
conversion of fructose into HMF. In most depictions of this
process,19 a fructofuranosyl oxocarbenium ion forms first and
then deprotonates spontaneously at C-1 to form an enol. To
account for the dramatic influence of halide on yield and rate,
Figure 2. Log-log plot of initial rates of HMF formation from fructose
we propose two variations on this mechanism (Figure 3A). In
vs [LiI]. Linear regression analysis of the data (shown) gives a slope of
one, a halide ion (X-) attacks the oxocarbenium ion to form a
1.1, which is consistent with a first-order dependence of the rate of HMF
formation on [I-].
2-deoxy-2-halo intermediate that is less prone to side reactions
as well as reversion to fructose. This intermediate then loses
is reasonable to expect its being attacked by chloride, bromide,
HX to form the enol (nucleophile pathway). Alternatively, a
and iodide. Bromide and iodide, which are better nucleophiles
halide ion could form the enol merely by acting as a base that
and leaving groups than chloride, are also more effective as
deprotonates C-1 (base pathway).
ionic additives. A mechanism that required the halide ion to
The known reactivity of sugars as well as our observations
act only as a base would invert the order of halide reactivity.
of fructose reactivity support the nucleophile pathway. The
Synthesis of HMF from Glucose in DMA. Next, we sought
to utilize the DMA-halide system to convert glucose into HMF.
As in [EMIM]Cl, CrCl2 and CrCl3 in DMA enable the
fructofuranosyl oxocarbenium ion is known to be attacked
readily by alcohols,20 borohydrides,21 and even fluoride,22 so it
conversion of glucose into HMF in yields up to 69% (Table 2).
We found negligible HMF yields (<1%) from glucose in the
absence of chromium salts. With the addition of CrCl2 to DMA
or DMA-LiCl, we obtained substantial yields of HMF
(47-60%), which were improved further by adding [EMIM]Cl
(19) Antal, M. J., Jr.; Mok, W. S. L.; Richards, G. N. Carbohydr. Res.
1990, 199, 91–109.
(20) Poncini, L.; Richards, G. N. Carbohydr. Res. 1980, 87, 209–217.
(21) Garegg, P. J.; Lindberg, B. Carbohydr. Res. 1988, 176, 145–148.
(22) Defaye, J.; Gadelle, A. Carbohydr. Res. 1985, 136, 53–65.
9
J. AM. CHEM. SOC. VOL. 131, NO. 5, 2009 1981

A R T I C L E S
Binder and Raines
Figure 3. Putative mechanisms for the dehydration of fructose and isomerization of glucose. (A) Putative nucleophilic and basic mechanisms for halide
participation in the conversion of fructose into HMF. X- represents a halide ion. (B) Putative mechanism for the chromium-catalyzed transformation of
glucose into HMF. Chromium species, depicted here as a hexacoordinate chromium(II) complex, catalyze the isomerization of glucose into fructose through
a putative enediolate intermediate.13 Fructose is then converted rapidly into HMF.
Table 2. Synthesis of HMF from Glucosea
fructose, and do not require ionic liquid solvents. Moreover,
glucose syrups like those readily available from corn are also
excellent feedstocks for HMF synthesis by our methods.
time
molar
solvent
catalyst, mol %
additive, wt %
T (°C) (h) yield (%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA
100
6
6
4
3
5
3
2
6
6
6
6
6
6
6
6
6
4
4
4
6
6
6
6
6
3
3
<1
<1
60
47
60
53
55
64
67
67
58
61
62
59
63
67
79
56
54
79
80
76
66
62
28
43
Further investigation of the generality of chromium-catalyzed
HMF synthesis revealed that a wide range of polar aprotic
solvents afforded HMF from glucose in yields higher than those
commonly achieved (Table S4 in Supporting Information).11
On the other hand, some coordinating classes of solvents, such
as amines and alcohols, prevent formation of HMF, perhaps
because of their interactions with the chromium salts. No HMF
was formed upon reaction of glucose in pyridine, and adding
several equivalents (based on chromium) of pyridine or N,N,N′,N′-
tetramethylethylenediamine markedly decreased HMF formation
in DMA-LiCl (Table 2). The basicity of these amines might
also disfavor HMF formation. Although Zhao et al.13 reported
that CrCl2 was markedly less effective than CrCl3, we found
that chromium in either oxidation state gave a similar yield.
[EMIM]Cl, 20 100
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl3, 6
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl3, 6
CrCl3, 6
CrCl3, 6
CrCl2, 6
CrCl2, 6
CrCl2, 6
CrCl3, 6
CrBr3, 6
CrBr3, 2
CrBr3, 1
CrCl2, 6
CrCl2, 6
CrCl2, 6
100
120
100
120
120
DMA
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA
DMA
DMA
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA
DMA
DMA
[EMIM]Cl, 5
100
[EMIM]Cl, 10 100
[EMIM]Cl, 20 100
[EMIM]Cl, 5
100
[EMIM]Cl, 10 100
[EMIM]Cl, 20 100
[EMIM]Cl, 5
100
[EMIM]Cl, 10 100
[EMIM]Cl, 20 100
DMA
DMA-LiCl (5%)
DMA
LiBr, 10
LiBr, 5
LiI, 10
LiBr, 10
LiBr, 10
LiBr, 10
LiBr, 10
100
100
100
100
100
100
100
Chromium likely enables conversion of glucose to HMF by
catalyzing the isomerization of glucose into fructose (Figure
3B).13 The fructose is then converted to HMF. Our observations
suggest that the yield of HMF from glucose in reactions utilizing
chromium correlates with metal coordination. Highly coordinat-
ing ligands such as amines decrease the yield of HMF. On
the other hand, halide ligands enhance HMF yields, with
bromide being the most effective. Our data suggest that the
halide additives must balance two roles in the conversion of
glucose into HMF: serving as ligands for chromium and
facilitating the selective conversion of fructose. Although iodide
excels in the latter role, its large size or low electronegativity
could compromise its ability as a ligand. In contrast, bromide
potentially offers the optimal balance of nucleophilicity and
coordinating ability, enabling unparalleled transformation of
glucose into HMF.
DMA
DMA
DMA
DMA
DMA-LiCl (10%)b
DMA-LiCl (10%)
DMA-LiCl (10%)
[EMIM]Cl, 20 100
TMEDA, 1
pyridine, 1
120
120
a Glucose was reacted at a concentration of 10 wt % relative to the
total mass of the reaction mixture. The solvent composition is indicated
by the weight percent of LiCl relative to DMA with additive
concentrations relative to the total mass of the reaction mixture. Catalyst
loading is relative to glucose. Yields are based on HPLC analysis. b In
this case, glucose was 45% in water.
to the reaction mixture. Supplementing the solvent with up to
20% [EMIM]Cl allows HMF yields up to 69%, comparable to
those obtained by Zhao et al.13 in [EMIM]Cl alone. As with
fructose, we observed a marked halide effect. Although addition
of iodide salts to the chromium chloride reaction mixture did
not greatly change the HMF yield, using 10 wt % lithium or
sodium bromide increased the HMF yield to 79-81% after 5-6
h at 100 °C. These yields of HMF from glucose exceed those
reported previously, approach typical yields of HMF from
Synthesis of HMF from Cellulose. HMF has been traditionally
obtained from monosaccharides. Cellulosic biomass is, however,
an especially promising source because of its inexpensive
availability from nonfood resources. Unfortunately, the typical
aqueous acid hydrolysis methods for producing HMF from
cellulose rely on high temperatures and pressures (250-400 °C,
9
1982 J. AM. CHEM. SOC. VOL. 131, NO. 5, 2009

Transforming Lignocellulosic Biomass into Furans
A R T I C L E S
Table 3. Synthesis of HMF from Cellulose and Lignocellulosic Biomassa
biomass
solvent
catalyst, mol %
additives, wt %
T (°C)
time (h)
HMF yieldb (%)
furfural yieldc (%)
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
cellulose
corn stover
AFEX corn stover
pine sawdust
corn stover
corn stover
corn stover
corn stover
corn stover
corn stover
corn stover
corn stover
corn stover
corn stover
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (15%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
[EMIM]Cl
[EMIM]Cl, 40
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
140
2
6
6
2
2
4
1
2
2
4
1
3
3
2
6
6
5
3
3
3
3
3
3
1
2
2
1
4
15
17
22
33
33
43
54
47
38
53
<1
<1
37
16
16
19
23
24
36
31
29
26
39
48
47
42
CrCl2, 25
CrCl3, 36
CrCl2, 25; HCl, 10
CrCl3, 25; HCl, 10
CrCl2, 25; HCl, 6
CrCl2, 25; HCl, 6
CrCl2, 25; HCl, 6
CrCl2, 25; HCl, 6
CrCl3, 25; HCl, 6
CrCl2, 25; HCl, 6
CrCl2, 25; HCl, 10
CrCl2, 25; HCl, 10
CrCl3, 25; HCl, 10
CrCl2, 38
[EMIM]Cl, 20
[EMIM]Cl, 40
[EMIM]Cl, 60
[EMIM]Cl, 80
[EMIM]Cl, 40
DMA
DMA
LiI, 10
LiBr, 10
LiBr, 3
[EMIM]Cl, 10
[EMIM]Cl, 10
[EMIM]Cl, 15
[EMIM]Cl, 20
[EMIM]Cl, 40
[EMIM]Cl, 60
[EMIM]Cl, 80
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
[EMIM]Cl
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
DMA-LiCl (10%)
[EMIM]Cl
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
34
CrCl2, 38
CrCl2, 33
CrCl2, 10; HCl, 10
CrCl2, 10; HCl, 10
CrCl2, 10; HCl, 10
CrCl2, 10; HCl, 10
CrCl2, 10; HCl, 10
CrCl3, 10; HCl, 10
CrCl3, 10; HCl, 10
CrCl3, 10; HCl, 10
CrCl3, 10; HCl, 10
CrCl3, 10; HCl, 10
[EMIM]Cl, 20
[EMIM]Cl, 40
[EMIM]Cl, 60
[EMIM]Cl, 80
37
ND
a Cellulose was reacted at a concentration of 4 wt % relative to the total mass of the reaction mixture. Biomass was reacted at a concentration of 10
wt % relative to the total mass of the reaction mixture. Solvent composition is indicated by weight percent of LiCl relative to DMA with additive
concentrations relative to the total mass of the reaction mixture. Catalyst loading and molar yield are relative to moles of glucose monomers contained
in the cellulose in the starting material. Yields are based on HPLC analysis. b Molar yield from pine sawdust assumes a typical cellulose content of
40%; yields from corn stover are based on cellulose analysis of 34.4%. c Molar yields from corn stover are based on xylan analysis of 22.8% for
untreated stover. ND, not determined.
10 MPa) and result in yields of 30%, at most.23 We suspected
that the solubility of cellulose in DMA-LiCl and the efficient
conversion of glucose into HMF with chromium would enable
superior results. Dissolution of purified cellulose in a mixture
of DMA-LiCl and [EMIM]Cl and addition of CrCl2 or CrCl3
produced HMF from cellulose in up to 54% yield within 2 h at
140 °C (Table 3). These yields compare well with reports of
HMF synthesis from cellulose in the patent literature using
aqueous acid and ionic liquids.23,14 Neither lithium iodide nor
lithium bromide alone produced high yields of HMF because
these salts in DMA do not dissolve cellulose.16 Using lithium
bromide along with DMA-LiCl did enable modest improve-
ments in yield. Likewise, using hydrochloric acid as a cocatalyst
boosted yields.
chlorinated organic products,25 which are not observed in our
system. In addition, the weight of their reaction mixture is nearly
150-fold greater than that of the cellulose reactant. Their 1,2-
dichloroethane extractant, which is likewise used in a large
excess relative to cellulose, is a possible carcinogen, and the
concentrated hydrochloric acid presents its own hazards.25 By
comparison, DMA is a common industrial solvent that, along
with LiCl, enables the processing of cellulose at far higher
concentrations (e15 vs e0.7 wt %).
Synthesis of HMF and Furfural from Lignocellulosic Bio-
mass. We discovered another desirable attribute of our process:
the ready conversion of lignocellulosic biomass. Efficient
production of fuels or chemicals from crude biomass often
requires pretreatment processes.26,27 In contrast, HMF can be
produced readily from untreated lignocellulosic biomass such
as corn stover or pine sawdust under conditions similar to those
used for cellulose (Table 3). Yields of HMF from corn stover
subjected to ammonia fiber expansion (AFEX) pretreatment26
were nearly identical to those for untreated stover. Other biomass
components, such as lignin and protein, did not interfere
substantially in the process, as yields of HMF based on the
cellulose content of the biomass were comparable to those from
purified cellulose. To be highly efficient, a process for biomass
Recently, Mascal and Nikitin24 reported an alternative method
for producing furanic products, chiefly 5-chloromethylfurfural
(CMF), from purified cellulose. Heating a solution of highly
purified cellulose in concentrated hydrochloric acid and lithium
chloride and then extracting the products with 1,2-dichloroethane
yielded this chlorinated relative of HMF in 71% isolated yield.
CMF can be converted subsequently to potential fuels like DMF
and 5-ethoxymethylfurfural. Although this process avoids the
use of chromium and results in higher yields (84% isolated yield
of furanic products versus 54%), it has notable drawbacks
relative to our DMA-LiCl system. Mascal and Nikitin use
chloride stoichiometrically and produce potentially hazardous
(25) Marshall, K. A. Chlorocarbons and chlorohydrocarbons. In Kirk-
Othmer Encyclopedia of Chemical Technology, 5th ed.; Wiley-
Interscience: New York, 2004; Vol 6, pp 226-253.
(26) Chundawat, S. P. S.; Venkatesh, B.; Dale, B. E. Biotechnol. Bioeng.
2006, 96, 219–231.
(23) Kono, T.; Matsuhisa, H.; Maehara, H.; Horie, H.; Matsuda, K. Jpn.
Pat. Appl. 2005232116, 2005.
(24) Mascal, M.; Nikitin, E. B. Angew. Chem., Int. Ed. 2008, 47, 7924–
7926.
(27) Aden, A.Biochemical Production of Ethanol from Corn Stover: 2007
State of Technology Model; Report NREL/TP-510-43205; National
Renewable Energy Laboratory, Golden, CO, 2008; http://
www.nrel.gov/docs/fy08osti/43205.pdf.
9
J. AM. CHEM. SOC. VOL. 131, NO. 5, 2009 1983

A R T I C L E S
Binder and Raines
conversion should also utilize the pentoses present in hemicel-
lulose. Notably, the industrial chemical furfural is formed from
the hemicellulose component of biomass under our reaction
conditions in yields similar to those obtained in industrial
processes (34-37% vs ∼50%).28
inexpensive catalysts to transform cellulose into a valuable
product in an ample yield. In addition, our privileged solvents
enable rapid biomass conversion at useful solid loadings (10
wt %). Under our best conditions, we transform 42% of the
dry weight of cellulose into HMF and 19% of the dry weight
of corn stover into HMF and furfural in one step. For
comparison, cellulosic ethanol technology, which has been
optimized extensively, enables the conversion of 24% of the
dry weight of corn stover into ethanol in a complex process
involving multiple chemical, biochemical, and microbiological
steps.27
Our process is also competitive on the basis of energy yield.
The HMF and furfural products contain 43% of the combustion
energy available from cellulose and xylan in the corn stover
starting material, whereas ethanol from corn stover preserves
62% of the sugar combustion energy (see Supporting Informa-
tion). Biomass components that cannot be converted into HMF,
such as lignin, could be reformed to produce H2 for HMF
hydrogenolysis (Figure 1) or burned to provide process heat.32
Realizing all of the intrinsic advantages of our process
requires additional improvements. The high loading of the
chromium catalyst and the toxicity of this metal could be barriers
to its large-scale use. To address this issue, we have already
found that decreasing chromium loading by two-thirds decreases
the yield of HMF from glucose only slightly (Table 2).
Additionally, the yields of HMF from cellulose and lignocel-
lulosic biomass are still modest. Finally, methods for recycling
solvents and salts would make the process more economical.
We also anticipate that further mechanistic studies of this
fascinating reaction cascade will enable the design of enhanced,
chromium-free catalysts that can accomplish the transformation
of cellulose into HMF in higher yield. With these types of
improvements, this selective chemistry could become a highly
attractive process for the conversion of lignocellulosic biomass
into an array of fuels and chemicals.
We propose that the formation of HMF from cellulose in
DMA-LiCl occurs via saccharification followed by isomeriza-
tion of the glucose monomers into fructose and dehydration of
fructose to form HMF. The saccharification of cellulose in water
is thought to occur via Brønsted acid-catalyzed hydrolysis of
its glycosidic bonds.29 A similar Lewis acid-catalyzed process
could be responsible for the hydrolysis activity of chromium
halides. Additionally, the improved HMF yield with addition
of hydrochloric acid suggests that Brønsted acid catalysis also
occurs in DMA-LiCl. This area of cellulose chemistry offers
rich opportunities for mechanistic understanding in addition to
enabling HMF synthesis and possibly offering an alternative to
enzymatic or aqueous acid-catalyzed hydrolysis for saccharifi-
cation of lignocellulosic biomass.
Conversion of Lignocellulosic Biomass into 2,5-Dimethyl-
furan. The conversion of lignocellulosic biomass to HMF in a
single step offers straightforward access to a wide variety of
useful HMF derivatives, such as DMF. Dumesic and co-workers
have shown that DMF, a promising HMF-derived fuel,6,7 can
be prepared by hydrogenolysis of fully purified HMF using
copper catalysts.30 We sought instead a process to synthesize
DMF in two chemical reactions from lignocellulosic biomass.
In the first step of our process, we formed HMF from
untreated corn stover in DMA-LiCl. We then removed the
chloride ions from the crude HMF by ion-exclusion chroma-
tography in water.31 This separation step prevented the chloride
from poisoning the copper hydrogenolysis catalyst. Finally, we
subjected the crude HMF from corn stover to hydrogenolysis
in 1-butanol with a carbon-supported copper-ruthenium catalyst
and obtained a 49% molar yield of DMF, similar to that obtained
by Dumesic and co-workers7 using HMF that contained trace
chloride. The overall molar yield of DMF based on the cellulose
content of the stover was 9% (Figure 1). We expect that
optimization of the process could readily improve upon this
result.
Experimental Methods
General. Commercial chemicals were of reagent grade or better
and were used without further purification. With the exception of
hydrogenolysis, reactions were performed in glass vessels heated
in a temperature-controlled oil bath with magnetic stirring. The term
“concentrated under high vacuum” refers to the removal of solvents
and other volatile materials using a rotary evaporator at vacuum
attained by a mechanical belt-drive oil pump while the water-bath
temperature was maintained below 30 °C. Conductivity was mea-
sured with an Extech Instruments ExStik II conductivity meter.
1-Ethyl-3-methylimidazolium chloride (99.5%, [EMIM]Cl) was
from Solvent-Innovation (Cologne, Germany). 1-Ethyl-3-meth-
ylimidazolium tetrafluoroborate (97%, [EMIM]BF4), 5-hydroxy-
methylfurfural, and 2,5-dimethylfuran were from Aldrich (Milwau-
kee, WI). 1-Ethyl-3-methylimidazolium triflate (98.5%, [EMIM]OTf),
1-butyl-3-methylpyridinium chloride (97%, [BMPy]Cl), 1-ethyl-
3-methylimidazolium bromide (97%, [EMIM]Br), and 1-propyl-
3-methylimidazolium iodide (97%, [PMIM]I) were from Fluka
(Geel, Belgium). 1-Ethylpyridinium chloride (98%, [EtPy]Cl),
1-ethyl-2,3-dimethylimidazolium chloride (98%, [MMEIM]Cl), and
furfural were from Acros (Buchs, Switzerland). Cu-Ru/carbon
catalyst (3:2 mol ratio Cu:Ru) was prepared by the method of
Dumesic and co-workers7 using 5% Ru/carbon from Aldrich
(Milwaukee, WI). Cellulose (medium cotton linters, C6288) was
from Sigma (St. Louis, MO). Milled and sieved corn stover and
AFEX-treated corn stover were generously provided by B. E. Dale
Conclusions
Our two-step process represents a low-temperature (<250 °C),
nonenzymic route from lignocellulosic biomass to fuels. Most
other chemical methods for the conversion of lignocellulosic
biomass to fuels use extreme temperatures to produce pyrolysis
oil or synthesis gas, incurring substantial energy costs. Our low-
temperature chemical conversion also has inherent advantages
over bioprocessing for cellulosic fuels and chemicals. Fermenta-
tion of lignocellulosic feedstocks requires saccharification
through extensive pretreatment, fragile enzymes, and engineered
organisms. In contrast, our chemical process uses simple,
(28) Zeitsch, K. J. The Chemistry and Technology of Furfural and Its Many
By-Products; Elsevier: Amsterdam, 2000.
(29) Blazej, A.; Kosik, M. Degradation reactions of cellulose and ligno-
cellulose. In Cellulose and its DeriVatiVes: Chemistry, Biochemistry,
and Applications; Kennedy, J. F., Phillips, G. O., Wedlock, D. J.,
Williams, P. A., Eds.; Ellis Horwood Ltd.: Chichester, England, 1985;
pp 97-117.
(30) Adkins, H. B. Reactions of Hydrogen with Organic Compounds oVer
Copper-Chromium Oxide and Nickel Catalysts; University of Wis-
consin Press: Madison, WI, 1937.
(31) (a) Rapp, K. M. U.S. Patent 4740605, 1988. (b) Fritz, J. S.
J. Chromatogr. 1991, 546, 111–118.
(32) Navarro, R. M.; Pen˜a, M. A.; Fierro, J. L. Chem. ReV. 2007, 107,
3952–3991.
9
1984 J. AM. CHEM. SOC. VOL. 131, NO. 5, 2009

Transforming Lignocellulosic Biomass into Furans
A R T I C L E S
and co-workers (Michigan State University).26 Cellulose and corn
stover were dried to constant weight at 120 °C prior to use.
Analytical Methods. All reaction products were analyzed by
HPLC and quantified with calibration curves generated from
commercially available standards. Following a typical reaction, the
product mixture was diluted with a known mass of deionized water,
centrifuged to sediment insoluble products, and analyzed. The
concentrations of products were calculated from HPLC-peak
integrations and used to calculate molar yields. HPLC was
performed with either a Waters system equipped with two 515
pumps, a 717 Plus autosampler, and a 996 photodiode array detector
or an Agilent 1200 system equipped with refractive index and
photodiode array detectors. HMF and furfural were analyzed either
by reversed-phase chromatography on a Varian Microsorb-MV
100-5 C18 column (250 × 4.6 mm; 93:7 water/acetonitrile, 1 mL/
min, 35 °C) or by ion-exclusion chromatography on a Bio-Rad
Aminex HPX-87H column (300 × 7.8 mm; 5 mM H2SO4, 0.6 or
0.9 mL/min, 65 °C). DMF was analyzed by reversed-phase
chromatography on a Varian Microsorb-MV 100-5 C18 column
(250 × 4.6 mm; 55:45 water/acetonitrile, 1 mL/min, 35 °C).
Representative ion-exclusion HPLC traces for HMF and furfural
analysis are shown in Figures S1 and S2 in Supporting Information.
Representative Procedure for Synthesis of HMF from Sugars.
Fructose (27.1 mg, 150 µmol) and LiCl (24 mg) were mixed in
DMA (203 mg). Following addition of concentrated H2SO4 (0.5
µL, 9 µmol), the reaction mixture was stirred at 80 °C for 5 h. At
1-h intervals, aliquots of the reaction mixture were removed for
HPLC analysis. For reactions of glucose that involved[EMIM]Cl,
chromium salts were mixed with a portion of the ionic liquid (25
mg) before addition to the reaction mixture. HMF yields for
optimized reactions of sugars were reproduced to within 2-3%.
Kinetic Analysis of HMF Formation from Fructose. Fructose
(425.6 mg, 2.362 mmol) and concentrated sulfuric acid (7.9 µL,
142 µmol) were dissolved in DMA (3.8185 g). Four known weights
of this stock solution (approximately 0.5 g) were combined with
four different mixtures of LiI and LiBF4 (added to maintain constant
salt concentration of 0.55 mmol/g) with LiI concentrations from
72.7 to 5.1 mg/g. After the salts were dissolved completely, the
solutions were heated at 75 °C for 10 min and cooled rapidly on
ice prior to HPLC analysis.
Representative Procedure for Synthesis of HMF from Cel-
lulose and Lignocellulose. Cellulose (21.2 mg, 131 µmol of glucose
units in cellulose), LiCl (26 mg), [EMIM]Cl (159 mg), and DMA
(252 mg) were mixed at 50 °C for 24 h to form a viscous solution.
To this cellulose solution were added concentratedd HCl (1 µL,
12 µmol) and CrCl2 (4 mg, 30 µmol) in [EMIM]Cl (50 mg), and
the reaction mixure was heated to 140 °C with vigorous stirring
for 2 h. At 1-h intervals aliquots of the reaction mixture were
removed for HPLC analysis. Reactions with lignocellulose were
similar except that the dissolution step was performed at 75 °C
and the biomass was added at a concentration of 10 wt %. In other
cases the cellulose was dissolved according to the methods of
McCormick et al.16 HMF yields for optimized reactions of cellulose
were reproduced to within 2-3%.
Ion-Exclusion Chromatographic Separation of HMF. Fructose
(1.20 g, 6.67 mmol) and LiCl (600 mg) were mixed in DMA (6.0
g) and stirred at 140 °C for 1 h. Deionized water (1.0 g) was mixed
with a portion of this reaction mixture (3.87 g) containing HMF
(127 mg). This solution was loaded onto a column of ion-exclusion
resin (Dowex 50X8-200, Li+ form, 70 × 1.5 cm) and eluted with
deionized water at a rate of 3 cm/min. Fractions (25 mL) were
collected and analyzed by HPLC. HPLC analysis indicated that
>75% of the HMF was recovered in the HMF-containing fractions.
Synthesis of DMF from Fructose. Fructose (1.805 g, 10 mmol)
and LiCl (1.690 g) were dissolved in DMA (14.67 g). Following
addition of concentrated H2SO4 (33 µL, 0.6 mmol), the reaction
mixture was stirred at 120 °C for 1 h. A portion of the reaction
mixture (3.01 g, 16.5%) was then diluted with deionized water (2
g), loaded onto a column of ion-exclusion resin (Dowex 50X8-200,
Li+ form, 70 × 1.5 cm), and eluted with deionized water at a rate
of 2 cm/min. HMF-containing fractions with conductivity <50 µS
were pooled and concentrated under high vacuum. The residue was
taken up in 1-butanol (45 g) and placed in a Parr reactor with
Cu-Ru/carbon catalyst (100 mg). The reactor was purged three
times with H2(g), pressurized with 6.8 bar of H2(g), and heated to
220 °C with stirring. After 10 h, the reactor was cooled and vented.
The contents were analyzed by HPLC for DMF (51.6 mg, 537 µmol,
32.5% based on fructose).
Synthesis of DMF from Corn Stover. Corn stover (1.504 g,
3.19 mmol of glucose units in cellulose), LiCl (0.755 g), [EMIM]Cl
(5.630 g), and DMA (6.756 g) were mixed at 75 °C for 24 h. To
this viscous mixture were added concentrated HCl (26 µL, 310
µmol) and CrCl3 (50.5 mg, 319 µmol) in [EMIM]Cl (0.55 g), and
the reaction mixure was heated with 140 °C with vigorous stirring.
After 2 h, the reaction mixture was cooled to room temperature,
diluted with deionized water (7 g), and subjected to centrifugation
to sediment insoluble material. After removal of the supernatant,
the pellet was resuspended in deionized water (3 g), and the process
was repeated. The combined supernatant solutions were loaded onto
a column of ion-exclusion resin (Dowex 50X8-200, Li+ form, 70
× 1.5 cm), and eluted with deionized water at a rate of 3 cm/min.
HMF-containing fractions with conductivity <50 µS were pooled
and concentrated under high vacuum. After they were concentrated
under high vacuum, the pooled HMF-containing fractions with
conductivity >50 µS were loaded a second time on the ion-
exchange column and eluted with deionized water. HMF-containing
fractions from the second pass with conductivity <50 µS were
pooled, concentrated under high vacuum, and combined with the
earlier HMF concentrate. This residue was taken up in 1-butanol
(45 g) and placed in a Parr reactor with Cu-Ru/carbon catalyst
(79 mg). The reactor was purged three times with H2(g), pressurized
with 6.8 bar of H2(g), and heated to 220 °C with stirring. After
10 h, the reactor was cooled and vented. The contents were analyzed
by HPLC for DMF (27.7 mg, 288 µmol, 9.0% based on cellulose).
Acknowledgment. This work was supported by the Great Lakes
Bioenergy Research Center, a DOE Bioenergy Research Center.
J.B.B. was supported by an NSF Graduate Research Fellowship.
We are grateful to H. Brown, J. E. Holladay, and Z. Zhang for
helpful conversations on biomass chemistry in ionic liquids; to B. R.
Caes, L. Summers, and R. Landick for contributive discussions; to
J. J. Blank and A. V. Cefali for experimental assistance; to S. Stahl
for access to a Parr reactor; to B. E. Dale and B. Venkatesh for
corn stover samples; and to K. Bersch for pine sawdust samples.
Supporting Information Available: Calculations of energy
yields, two figures, and six tables. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA808537J
9
J. AM. CHEM. SOC. VOL. 131, NO. 5, 2009 1985