Dissolution of wood in ionic liquids

Article abstract of DOI:10.1021/jf071692e

The present paper demonstrates that both hardwoods and softwoods are readily soluble in various imidazolium-based ionic liquids (ILs) under gentle conditions. More specifically, a variety of ionic liquids can only partially dissolve wood chips, whereas ionic liquids such as 1-butyl-3-methylimidazolium chloride and 1-allyl-3-methylimidazolium chloride have good solvating power for Norway spruce sawdust and Norway spruce and Southern pine thermomechanical pulp (TMP) fibers. Despite the fact that the obtained solutions were not fully clear, these ionic liquids provided solutions which permitted the complete acetylation of the wood. Alternatively, transparent amber solutions of wood could be obtained when the dissolution of the same lignocellulosic samples was attempted in 1-benzyl-3-methylimidazolium chloride. This realization was based on a designed augmented interaction of the aromatic character of the cation of the ionic liquid with the lignin in the wood. After dissolution, wood can be regenerated as an amorphous mixture of its original components. The cellulose of the regenerated wood can be efficiently digested to glucose by a cellulase enzymatic hydrolysis treatment. Furthermore, completely acetylated wood was found to be readily soluble in chloroform, allowing, for the first time, detailed proton nuclear magnetic resonance (NMR) spectra and NMR diffusion measurements to be made. It was thus demonstrated that the dissolution of wood in ionic liquids now offers a variety of new possibilities for its structural and macromolecular characterization, without the prior isolation of its individual components. Furthermore, considering the relatively wide solubility and compatibility of ionic liquids with many organic or inorganic functional chemicals or polymers, it is envisaged that this research could create a variety of new strategies for converting abundant woody biomass to valuable biofuels, chemicals, and novel functional composite biomaterials.

Full text of DOI:10.1021/jf071692e

9142 J. Agric. Food Chem. 2007, 55, 9142–9148
Dissolution of Wood in Ionic Liquids
ILKKA KILPELÄINEN,*,† HAIBO XIE,§ ALISTAIR KING,† MARI GRANSTROM,†
SAMI HEIKKINEN,† AND DIMITRIS S. ARGYROPOULOS*,†,§
Organic Chemistry of Wood Components Laboratory, Department of Forest Biomaterials Science and
Engineering, North Carolina State University, Raleigh, North Carolina 27695-8005, and Department of
Chemistry, Laboratory of Organic Chemistry, University of Helsinki,
P.O. Box 55, FIN-00014, Finland
The present paper demonstrates that both hardwoods and softwoods are readily soluble in various
imidazolium-based ionic liquids (ILs) under gentle conditions. More specifically, a variety of ionic
liquids can only partially dissolve wood chips, whereas ionic liquids such as 1-butyl-3-methylimida-
zolium chloride and 1-allyl-3-methylimidazolium chloride have good solvating power for Norway spruce
sawdust and Norway spruce and Southern pine thermomechanical pulp (TMP) fibers. Despite the
fact that the obtained solutions were not fully clear, these ionic liquids provided solutions which
permitted the complete acetylation of the wood. Alternatively, transparent amber solutions of wood
could be obtained when the dissolution of the same lignocellulosic samples was attempted in 1-benzyl-
3-methylimidazolium chloride. This realization was based on a designed augmented interaction of
the aromatic character of the cation of the ionic liquid with the lignin in the wood. After dissolution,
wood can be regenerated as an amorphous mixture of its original components. The cellulose of the
regenerated wood can be efficiently digested to glucose by a cellulase enzymatic hydrolysis treatment.
Furthermore, completely acetylated wood was found to be readily soluble in chloroform, allowing, for
the first time, detailed proton nuclear magnetic resonance (NMR) spectra and NMR diffusion
measurements to be made. It was thus demonstrated that the dissolution of wood in ionic liquids
now offers a variety of new possibilities for its structural and macromolecular characterization, without
the prior isolation of its individual components. Furthermore, considering the relatively wide solubility
and compatibility of ionic liquids with many organic or inorganic functional chemicals or polymers, it
is envisaged that this research could create a variety of new strategies for converting abundant woody
biomass to valuable biofuels, chemicals, and novel functional composite biomaterials.
KEYWORDS: Wood; dissolution; derivatization; regeneration; characterization; cellulose; lignin; ionic
liquids; cellulase digestion; NMR; diffusion studies
INTRODUCTION
The main components of wood are cellulose, lignin, hemi-
cellulose, and extractives. From these, lignin binds the plant
cells together, providing wood with mechanical strength and
rigidity to resist external forces, such as wind, and forming a
barrier against microbial attack. According to the current view,
lignin is an irregular, cross-linked polymer network, which is
composed of randomly cross-linked phenylpropanoid units (1–3).
Therefore, in general, it is thought that it is practically impossible
to dissolve wood in its native form, because the three-
dimensional lignin network binds the whole wood architecture
together. In papermaking, the lignin network is fragmented under
alkaline conditions, and cellulose is harvested as cellulose fibers.
The insolubility of wood in common solvents has severely
hampered the development of new methods for the efficient
utilization of wood and its components.
The efficient utilization of biomass is becoming increas-
ingly important due to diminishing resources of fossil fuels
as well as global heating warnings caused by greenhouse gas
emissions. To date, wood is mainly utilized for the production
of paper from cellulose, and most of the other components
of wood are simply incinerated to energy. Therefore, there
is an increasing demand for new processes that could provide
new means to use this resource in a more efficient manner,
not only as a fuel but also as a starting material for our
chemical industry with the aim of producing commodity and
fine chemicals.
* Authors to whom correspondence should be addressed [(D.S.A.)
telephone (919) 515-7708, fax (919) 515-6302, e-mail dsargyro@
NCSU.edu; (I.K.) telephone +358 50 5181148, fax +358 9 19150366,
e-mail ilkka.kilpelainen@helsinki.fi].
Some time ago Lu and Ralph demonstrated that after
extensive ball-milling, wood becomes soluble in dimethyl
sulfoxide–tetrabutyl ammonium fluoride and dimethyl sulfox-
† University of Helsinki.
§ North Carolina State University.
10.1021/jf071692e CCC: $37.00
2007 American Chemical Society
Published on Web 10/02/2007

Dissolution of Wood in Ionic Liquids
J. Agric. Food Chem., Vol. 55, No. 22, 2007 9143
s), 5.48 (2H, s), 6.81–6.82 (1H, d), 6.95–6.99 (2H, m), 7.22–7.24 (1H,
m), 7.32 (1H, d), 7.45 (1H, s), 10.78 (1H, s).
Scheme 1. Structures and Abbreviations of the Examined Ionic Liquids
1-Methyl-3-methylbenzylimidazolium Chloride (5). The ionic liquid
was synthesized using the same procedure as per ionic liquid (3): yield,
96%; δH (300 MHz, CDCl3) δ 2.26 (3H, s), 4.01 (3H, s), 5.44 (2H, s),
7.10 (1H, d), 7.17–7.19 (3H, m), 7.25 (1H, d), 7.51 (1H, d), 10.72
(1H, s).
1-Methyl-3-benzyl-imidazolium Dicyanamide (6). The ionic liquid
was prepared by the anion exchange reaction between 1-methyl-3-
benzylimidazolium chloride (0.20 mol) and NaN(CN)2 (0.21 mol) using
H2O as solvent. The homogenous mixture was stirred at room
temperature for 12 h. After evaporation of the water, 50 mL of CH2Cl2
was added into the residua. The formed NaCl was filtered, and the
organic solvent phase was dried with anhydrous MgSO4. After filtration
of the MgSO4 and evaporation of the solvent, a yellow liquid ionic
liquid was obtained: yield, 93%; δH (300 MHz, CDCl3) δ 3.98 (3H,
s), 5.38 (2H, s), 7.25 (1H, d), 7.32 (1H, d), 7.38–7.43 (5H, m), 9.21
(1H, s).
ide–imidazole binary solvent systems (4). However, it is well-
known that the required extensive ball-milling causes degrada-
tion of both cellulose and lignin (4). Consequently, from both
economic and environmental points of view, the magnitude of
the wood conversion industries to paper and allied products
imposes serious challenges in the development of green process-
ing technologies for wood and other lignocellulosic materials.
It has also been shown that some ionic liquids (ILs) (such as
1-butyl-3-methyl- and 1-allyl-3-methylimidazolium chloride,
[bmim]Cl and [amim]Cl, respectively) can effectively dissolve
biopolymers (5, 6). In our continuing work to characterize the
structure of lignin, we have observed that both [bmim]Cl, and
[amim]Cl were also able to dissolve different types of lignin
samples. Similar results were very recently reported by Pu et
al. (7). These results have prompted us to examine the
dissolution of wood and lignin in these solvents. This paper
reports our initial findings that pertain to the various issues that
determine the complete solubility of hardwoods and softwoods
in ionic liquids.
All IL samples were dried using an oil pump vacuum at 80 °C for
several days to remove all residual water and solvent traces. The water
1
content of the used ILs was below 1% (according to H NMR).
NMR Spin Diffusion Measurements. Such NMR spectra were
acquired with a Varian Inova 500 MHz NMR spectrometer in CDCl3
at 27 °C. The parameters for the diffusion-ordered spectra were as
follows: gradient parameters selected for each gradient in bipolar
gradient pair (BPPSTE sequence); gradient duration, 1.8 ms; gradient
recovery delay, 200 µs; diffusion delay, 200 ms; gradient strengths
(G/cm), 4.6, 6.5, 7.9, 9.2, 10.2, 11.2, 12.1, 13.0, 13.7, 15.2, 16.5, 18.9,
20.0. The number of transients was 256 for each spectrum acquired
with a recycling time of 2.1 s.
Dissolution of Wood in Ionic Liquid. The ionic liquid was charged
into a 50 mL dried flask equipped with a mechanical stirrer, under an
inert atmosphere of argon. The temperature of the dissolution process
was controlled by an oil bath at the specified temperature (Table 1).
The wood sample (particle size ) 0.1–2 mm) was then added into the
ionic liquid quickly, and the dissolution proceeded at specified time
intervals.
Wood Regeneration, Enzymatic Hydrolysis, Glucose Determi-
nation. The wood solution created by the treatment described previously
was gradually added into an excess of rapidly stirred distilled water.
The precipitated bulky material was then filtered using a Buchner funnel,
washed thoroughly with distilled water. Finally, a small sample of it
was taken to determine its solids content by oven-drying it at 110 °C
overnight. The remaining material was then treated with cellulase
(Iogen, Canada; filter paper activity ) 130 FPU mL-1) using a
previously optimized (8) ratio of 40 FPU/g of wood. The enzymatic
hydrolyses were carried out at 40 °C for 48 h using 50 mM citrate
buffer (pH 4.5) at 5% consistency in an orbital water bath shaker. The
filtrate from the enzymatic hydrolysis was then diluted to a volume of
250 mL using a volumetric flask. Using 100 mL of this solution
subsequent dilutions were thus prepared containing 100–1000 µL (in
100 µL intervals) of this filtrate in distilled water. Four milliliters of
each of these dilutions was then added into a test tube. A series of
glucose standards with the following concentrations were also prepared:
1.28 × 10–6, 2.55 × 10–6, 3.83 × 10–6, 5.11 × 10–6, and 6.39 × 10–6
g/mL. In addition, the coloring reagent was prepared by mixing 1 part
of reagent B with 50 parts of reagent A of BCA test kit (protein assay
kit, Sigma). Reagent B is a copper solution, and reagent A is a BCA
solution. The resulting solution is green in color, which should be
freshly prepared for every analysis. The color was finally developed
by adding to each of the 4 mL glucose standard solutions and diluted
filtrates 1 mL of coloring reagent. In addition, a blank solution was
also prepared by mixing 4 mL of distilled water and 1 mL of coloring
reagent. The samples were then mixed using a Vortex mixer for few
seconds. The samples were then allowed to react at 60 °C (protected
from light by covering each tube with aluminum foil) and allowed to
incubate for 2 h (stirring is not necessary). Samples containing glucose
turned purple. The amount of glucose in the various samples was finally
determined spectrophotometrically at 562 nm against a blank.
Determination of Solubilization Conditions and Solubility of
Wood in Ionic Liquids. Ten grams of ionic liquid was charged into a
EXPERIMENTAL PROCEDURES
Materials. Unbleached Norway spruce thermomechanical pulp
(TMP) and pine TMP were sampled in a Swedish mill, ca. 38% dryness,
85 mL of CSF, standard newspaper quality; the mill has a one-stage
refining and a subsequent reject refining (ca. 20%) stage. This pulp
was taken at a press stage after the refined and refined reject pulps had
been combined. Norway spruce sawdust was prepared in-house. All
wood samples used (spruce TMP, spruce sawdust; particle size ) 0.1–2
mm, pine TMP) were kept in a vacuum oven at 50 °C for 24 h prior
to use. All reactions were carried out under argon atmosphere. Pyridine
and acetic anhydride were purchased from Aldrich, Fluka, and J. T.
Baker and used without further purification.
Synthesis of Ionic Liquids. Structure numbers in bold refer to those
of Scheme 1. [Amim]Cl (2) was synthesized according to the general
procedure provided by Zhang et al., (6) with slight modification; both
allyl chloride (Aldrich) and 3-methylimidazole (Aldrich) were distilled
prior to use. [Amim]Cl was further purified, to remove trace color, by
dissolving the crude [amim]Cl mixture in water and refluxing with
activated charcoal (18 h). The solution was filtered through a silica
plug, and the water was removed by distillation and dried for 2 days
in vacuo to yield [amim]Cl as a pale yellow crystalline solid, with a
melting point of 52 °C: δH (300 MHz, CDCl3) 4.06 (3 H, s), 4.94
(2H, d), 5.40 (1H, d), 5.91–5.97 (1H, m), 7.42 (1H, s), 7.65 (1H, s),
10.44 (1H, s).
1-Methyl-3-benzylimidazolium Chloride (3). The ionic liquid, was
prepared with benzyl chloride (0.25 mol) and 1-methylimidazole (0.23
mol) using CH3CN as solvent in a 250 mL three-neck bottle. The
mixture was refluxed for 48 h under an Ar atmosphere. After
evaporation of the solvent and of the residual benzyl chloride, the pure
ionic liquid was obtained. Drying of the materials took place at 120
°C under vacuum by stirring for 24 h. The product was of a solid
gelatinous nature at room temperature: yield, 95%; δH (300 MHz,
CDCl3) 4.02 (3H, s), 5.54 (2H, s), 7.27–7.34 (4H, m), 7.42–7.45 (2H,
m), 7.47–7.50 (1H, t), 10.56 (1H, s).
1-Methyl-3-m-methoxylbenzylimidazolium Chloride (4). The ionic
liquid was synthesized using the same procedure as per ionic liquic
(3): yield, 95%; Tm ) δH (300 MHz, CDCl3), 3.76 (3H, s), 4.01 (3H,

9144 J. Agric. Food Chem., Vol. 55, No. 22, 2007
Kilpeläinen et al.
Table 1. Dissolution Behavior of Wood-Based Lignocellulosic Materials in Different Imidazolium-Based Ionic Liquids
entry
ionic liquid
wood sample form
wood chips
solubilization conditions
solubility (wt %)
1
2
3
3b
4
5
6
7
8
[bmim]Cl
[amim]Cl
[amim]Cl
[amim]Cl
[bmim]Cl
[amim]Cl
[bmim]Cl
[amim]Cl
130 °C, 15 h
80 °C, 8 h
partially soluble
8
8%
5
8
7
7
2
5
5
ball-milled Southern pine powder
Norway spruce sawdust
Norway spruce sawdust
Norway spruce sawdust
Norway spruce TMP
Norway spruce TMP
Southern pine TMP
Southern pine TMP
Southern pine TMP
Southern pine TMP
110 °C, 8 h
80 °C, 24 h
110 °C, 8 h
130 °C, 8 h
130 °C, 8 h
110 °C, 8 h
130 °C, 8 h
130 °C, 8 h
130 °C, 8 h
130 °C, 8 h
130 °C, 8 h
130 °C, 8 h
130 °C, 8 h
[amim]Cl
[bmim]Cl
9
10
11
12
13
14
[bzmim]Cl
[bzmim]Cl
[bz-ome-mim]Cl
[bz-ome-mim]Cl
BenzylmimDca
5
5
5
2
Norway spruce TMP
Southern pine TMP
Southern pine TMP
Southern pine TMP
2
100 mL dried three-neck flask under an inert atmosphere of argon,
which was heated with a temperature-controlled oil bath at a given
temperature. The wood powder (Norway spruce, particle size ) 0.1–2
mm) or fiber was added in portions of only 1 wt % of ionic liquid
each time with mechanical stirring. Additional sample was introduced
until the solution became homogenous and somewhat clear or until
the disappearance of fibers was evident.
In addition to the particle size, the water content of a wood
sample plays a key role in determining its solubility in ionic
liquids. Overall, water was found to significantly reduce the
solubility of wood in ILs. The mere fact that it is very difficult
to remove water effectively from wood chips provides ample
justification as to why the earlier efforts were unsuccessful in
documenting the complete wood solubilization (9). A hot stage
optical microscopy investigation of a Norway spruce sawdust
sample in [amim]Cl clearly demonstrates the disappearance of
fibrous material (Figure 1) over a period of several hours.
On a macroscopic scale, the appearance of [bmim]Cl and
[amim]Cl wood solutions was not fully clear. Because the
dissolution of cellulose, as well as various hemicelluloses (6, 12),
has been demonstrated to produce clear viscous solutions in
[bmim]Cl, the hazy appearance of wood solutions could be due
to the presence of lignin. In the literature, the dissolution of
cellulose in various solvent systems has been discussed in
detail (5, 13–18). A common property for these solvents is their
ability to break down the extensive inter- and intramolecular
hydrogen bonding network of cellulose. In ILs, both the cation
and anion of the salt play a crucial role in the dissolution of
cellulose. In addition, the viscosity of the IL plays a role in the
dissolution speed (12). For the efficient dissolution of cellulose
in ILs, the cation is usually an imidazolium salt with a C2–C6
side chain. (5) From these, the butyl and allyl derivatives are
the most efficient, and growing chain length has been found to
decrease the solubility of cellulose (5, 6). From anions, the
chloride is usually the most efficient due to its capability to
participate in breaking down hydrogen bonds (5, 18). Notably,
also various strong Lewis acid anions (such as aluminum
chlorides) have the potential of being efficient solvents for
cellulose. However, these often pose problems associated with
solvent recycling, and their enhanced reactivity precludes them
from a variety of applications.
RESULTS AND DISCUSSION
Dissolution of Wood into Ionic Liquids. Complete wood
dissolution [up to 8% (w/w), Figure 1] can be carried out by
simply mixing a dried wood sawdust sample with the ionic
liquid ([amim]Cl or [bmim]Cl) and stirring the mixture me-
chanically at 80–120 °C. In a previous study, similar solvents
were utilized, but it was claimed that the sample did not dissolve
completely and that dissolution was limited, leading to partial
fractionation of the sample (9). However, in that work, the
researchers used wood chips for dissolution experiments. The
two crucial parameters that define wood solubilzation are as
follows: First, the dissolution rate of wood is highly dependent
on the particle size of the wood sample. This is because the
complex and compact structure of the wood cell wall and
between the lignin, cellulose and hemicellulose would essentially
inhibit the diffusion of the ionic liquid into its interior, resulting
in only a partial dissolution of wood chips.
During our work we discovered that the solubilization
efficiency of the lignocellulosic materials in ILs was found to
be of the order ball-milled wood powder > sawdust g TMP
fibers . wood chips. The dissolution of fine sawdust (Norway
spruce, particle size ) 0.1–2 mm) in ILs takes place within a
few hours, even under ambient conditions. Furthermore, ther-
momechanical pulp (TMP) samples readily dissolved in [bmim]Cl
and/or [amim]Cl. However, we discovered that it takes several
weeks for wood chips (size in excess of 5 mm × 5 mm × 1
mm) to completely dissolve at elevated temperatures. It has also
been previously reported that wood dissolution into ILs is
possible under microwave irradiation, which enhances the rate
of dissolution (10). However, this also leads to delignification
and to partial degradation of the wood components (11).
Beyond the mass transfer and diffusion considerations
outlined above aimed to rationalize the solubilization order
already discussed previously (see Table 1), additional explan-
tions are in order if one is to comprehend the reason for hazy
Figure 1. Optical photomicrographic images (magnification ×10) of Norway spruce fibers from sawdust in AmimCl as a function of time (from left to
right: initial, 0.5 h, 2 h, 4 h at 120 °C).

Dissolution of Wood in Ionic Liquids
J. Agric. Food Chem., Vol. 55, No. 22, 2007 9145
versus transparent solutions. The highly crystalline character
of cellulose in wood is driven by a set of regular intermolecular
and intramolecular hydrogen-bonding interactions that when
coupled with the three-dimensional network character of lignin
and its possible covalent linkages with the carbohydrates are
primarily responsible for the complex and compact structure
of wood. Accounts of π–π interactions among the aromatic
groups in lignin have also appeared in the literature, accounting
for the conformationally stable supermolecular structure of lignin
(19). Because significantly attractive interactions between
π-systems have been known for over half a century (20), in
recent years they have seen wide utility toward understanding
crystal structure, self-assembly in nano-supermolecular structural
chemistry, materials chemistry, etc. Ionic liquids have a more
complex solvent behavior compared with traditional solvents.
The various types of interactions between ILs with many solutes
include dispersion, π–π, n–π, hydrogen bonding, dipolar, and
ionic/charge–charge and have been invesitigated by Anderson
et al. (21). This team has reported that although the Bmim cation
does not have the analogous electron aromatic system, the
chloride anion (with nonbonding electrons), in conbination with
the Bmim cation, forms an IL that exhibits the ability to interact
with π-systems of probe molecules. In our case, conceivably
the active chloride ions in the BmimCl ionic liquid would disrupt
the hydrogen-bonding interactions present in wood, allowing it
to diffuse into the interior of the wood and resulting in a viscous
but hazy solution. Both, however, BmimCl and AmimCl still
cannot effectively and fully interact and solvate the aromatic
character of lignin, imparting a hazy characteristic to such
solutions.
Figure 2. X-ray spectra of (a) spruce sawdust, (b) regenerated spruce
from [amim]Cl using H2O as the nonsolvent, and (c) 8% wt spruce sawdust
in [bmim]Cl solution.
was also discovered that blending [benzylmim]Cl with [amim]Cl
could effectively reduce the viscosity of [benzylmim]Cl, whereas
a transparent solution of wood material could be obtained (entry
15, Table 1).
Furthermore, we have determined that the facility and the
kinetics of the process of wood dissolution in ionic liquids are
dependent on the applied temperature, which is consistent with
the results in a previous publication related to the dissolution
of wool keratin fibers (23). At higher temperatures (>100 °C),
all of the examined wood samples dissolved relatively rapidly.
However, it was possible to dissolve up to 5% (w/w) of Norway
spruce sawdust in [amim]Cl at 80 °C within 24 h. In fact,
complete dissolution was also found to take place at ambient
temperature. For example, when wood sawdust is gently
homogenized with [amim]Cl in a mortar and the sample is
subsequently transferred into a test tube (under argon), it slowly
turns liquid over time. Among the examined ILs, the [amim]Cl
was found to possess the lowest viscosity, which permits using
somewhat lower temperatures. Alternatively, ILs with aromatic
side chains (Scheme 1) required somewhat higher temperatures
due to their higher melting points and viscosities (Table 1).
Properties of Regenerated Wood. After dissolution of the
lignocellulosic material into ILs, it is possible to recover the
sample by simply pouring the solution into an excess of a
nonsolvent, that is, water. If the regeneration is carried out under
rapid mechanical stirring, fully amorphous material is obtained.
This is clearly visible in the X-ray spectra of the regenerated
material (Figure 2), because the X-ray diffraction signals from
the crystalline regions of spruce sawdust have disappeared after
the dissolution–regeneration process. The data of Figure 2 show
that spruce sawdust/[amim)Cl displays a slight broad amorphous
diffraction peak centered around 2θ of 25°, in the absence of
the characteristic diffraction pattern of wood cellulose, which
usually displays a prominent sharp peak near 15° and 23°. In
an earlier work, Wei et al. (24) were investigating the structural
differences between cellulose regenerated from [bmim]Cl and
untreated cellulose, using Fourier transform IR (FT-IR) spec-
troscopy, X-ray diffraction, and thermogravimetric (TG) mea-
surements. The data of this team showed that the crystalline form
of wood pulp cellulose was transformed completely from cellulose
I to cellulose II after regeneration from [bmim]Cl solution.
However, the material regenerated from dissolved wood does not
show any signs for cellulose II. This may be explained by the
presence of lignin and other wood components, as well as by the
regeneration method, which may contribute to the crystallinity of
The Abraham solvation equation is an effective approach used
to characterize the polymer solvation capacity of ILs (21). The
various parameters involved in this equation include the ability
of the ionic liquid to interact with π- and n-electrons of the
solute, a measure of the dipolarity/polarizability of the ionic
liquid, and the ionic liquid’s hydrogen bond acidity and basicity.
Upon close inspection of the Abraham solvation equation we
concluded that for our work a cationic moiety with an electron-
rich aromatic π-system may create stronger interactions (higher
retention factors) for polymers capable of undergoing π–π and
n–π interactions. With this in mind, we designed and synthesized
a series of phenyl-containing ionic liquids and consequently
examined their solubilization capabilites for various wood
samples (entries 10–13, Table 1). The emerging data were found
to be consistent with the above hypothesis as demonstrated by
the fact that 1-benzyl-3-methylimidazolium chloride ([bzmim]Cl)
ionic liquids/wood solutions were completely transparent, amber
colored, and viscous solutions. However, one major drawback
of [bzmim]Cl is its high viscosity. As such, efforts were directed
to reduce it by decreasing the symmetry of the cation by
synthesizing ILs 4 and 5 (Scheme 1) as well as altering the
nature of the anion (IL 6, Scheme 1) within the structure of the
ionic liquid. Although the obtained solutions were moderately
clear (Table 1), their wood-dissolving abilities were not
adequate as evidenced by the amount of residual fibrous material
remaining when these structures were used. It is likely that any
increase in the mass and bulk of the cation within a given ionic
liquid may decrease the concentration of active chloride ion
(5), thus reducing its solvating capacity for both cellulose and
lignin. In addition, although –N(CN)2 anion based ionic liquids
have shown excellent solvating abilities for carbdrohydrates (22),
IL 6 (Scheme 1), which was of significantly lower viscosity,
even at room temperature, was only partially capable of
dissolving Norway spruce TMP fibers (entry 14, Table 1). It

9146 J. Agric. Food Chem., Vol. 55, No. 22, 2007
Kilpeläinen et al.
Figure 4. IR spectra (ATR) of vacuum-dried (50 °C) spruce wood sawdust
(top spectrum) and acetylated sample regenerated from [amim]Cl (bottom
spectrum; entry 3b, Table 1). The signals for the -OH and -CdO
stretching bands are indicated.
Figure 3. Amount of glucose released during enzymatic hydrolysis with
the IL pretreatment (in [amim]Cl or [bmim]Cl), in millimoles per gram of
wood and in percentage terms compared with the theoretical maximum
and with a control experiment (spruce wood sawdust without pretreatment).
The effect of galactoglucomannan.hemicellulose is not taken into account.
the sample. In the present study, the material was regenerated with
vigorous stirring, which may partially inhibit the induction of
crystalline ordering within the regenerated sample.
As demonstrated in Figure 2, the crystallinity of the cellulose
in the wood was eliminated with the IL dissolution–regeneration
treatment. Such a transformation is anticipated to allow a greater
accessibility for the hydrolytic enzymes to rapidly penetrate and
hydrolyze the wood (25). In an effort to further investigate the
reactivity of regenerated materials toward cellulolytic enzymes,
the pretreated lignocellulosic material in the IL was then
submitted to an enzymatic hydrolysis (26). Our initial, and
completely nonoptimized experiments, showed that about 60%
of the theoretical amount of glucose was enzymatically released
from the wood when predissolved in [amim]Cl and regenerated
by precipitation in water. This compared to only 12% of glucose
units being released from the untreated control wood sample
(Figure 3). Similar pretreatments in [bmim]Cl were also found
to improve the release of glucose unit but to a significantly lower
degree compared to [amim]Cl. The use of ILs on straw as the raw
material prior to cellulolytic hydrolysis has also been documented
recently with somewhat less impressive results (27).
Acetylation of Wood Dissolved in Ionic Liquids. One of
the major advantages of ILs is that they allow the use of a wide
range of chemical reactions to be performed in them with
significant alterations in the reaction rates and the stabilization
of the various transitions state complexes (28, 29). In the
literature, the acetylation of cellulose has been studied for many
ILs (6, 30, 31). By subjecting fully dissolved wood in [amim]Cl
(entry 3b, Table 1) to an incremental addition of a 1:1 mixture
of acetic anhydride/pyridine it became possible to essentially
completely acetylate wood. This is evident from the complete
disappearance of the hydroxyl IR stretch band located at 3500
cm-1 (Figure 4) as well as from the appearance of strong –CdO
stretching band at 1750 cm-1. The rigid and compact nature of
wood is known to be attributed to an intricate hydrogen-bonded
network that precludes its solubility in common molecular
solvents. In this light, the demonstration of complete acetylation
depicted in Figure 4 is rather compelling. The complete
derivatization of all of the hydroxyl functionalities also em-
phasizes that the obtained wood solutions are true solutions,
and they are not simply gels or larger aggregates.
Figure 5. 1H NMR spectrum (500 MHz) of acetylated Norway spruce
sawdust sample dissolved in CDCl3. The assignments for acetylated
cellulose signals are shown according to ref 35. In addition, the HR signal
due to the lignin ꢀ-O-4 strucures is also shown.
of intact acetylated wood (Figure 5). However, the amount of
material that could actually be incorporated in chloroform
solutions (as well as DMSO, DMF, or acetone) was rather
limited, and this limited our attempts to collect more informative
two-dimensional correlation data (such as 1H–13C HSQC) from
the whole sample. Our continuing efforts, however, in the
direction of synthesis new wood derivatives and ionic liquids
may eventually allow us to overcome these limitations.
1
At first sight, the H spectrum of acetylated wood seems
somewhat surprising, because the signals from lignin are of
rather low intensity as compared to the signals from cellulose,
which are seen to dominate the spectrum. There are, however,
two parameters that need to be carefully considered at this point.
Because cellulose is a highly homogenous polymer, all of its
signals are anticipated to show a similar proton chemical shift
(i.e., the H-1 resonating at about δ 5.1). Lignin, on the other
hand, is a highly irregular polymer. The lignin polymer is
composed of different types of subunits, and even the signals
from certain types of subunits will have a larger chemical shift
dispersion caused by the neighboring structures (3). From the
subunits of lignin, the signal from the R proton in ꢀ-O-4
substructures is known to possess a unique chemical shift at
6.0 ppm (3). Furthermore, during our work we determined the
amounts of cellulose and lignin in the examined Norway spruce.
This was found to be approximately 50% cellulose and 26%
lignin. On the basis of the 26% lignin content, only about 20%
of these are known to belong to ꢀ-O-4 subunits (1, 2). Therefore,
Another significant manifestation of our work with fully
acetylated wood is its solubility in chloroform. This allowed
the recording, for the first time, of solution state 1H NMR spectra

Dissolution of Wood in Ionic Liquids
J. Agric. Food Chem., Vol. 55, No. 22, 2007 9147
cellulose and lignin signals (as well as chloroform) follows
essentially single exponential behavior. The implications of these
measurements suggest that there are no covalent linkages
between cellulose and lignin in native wood. On the basis of
the present data it seems probable that if there are covalent
lignin–carbohydrate linkages in wood, they are more likely to
exist between hemicelluloses and lignin rather than cellulose
and lignin. Alternatively, the situation may be different between
different species of wood, and this remains to be explored in
detail. The possible linkage between lignin and carbohydrates
is a question that has been puzzling researchers for a long time.
It is practically impossible to isolate completely carbohydrate-
free lignin samples from wood, but it has been very difficult to
obtain precise structural information from intact wood.
In the present article, we have demonstrated that wood is
fully soluble into certain ionic liquids. More specifically, we
have discovered that most ionic liquids can only partially
dissolve wood chips, whereas ionic liquids such as [Bmim]Cl
and [Amim]Cl have good solvating power for Norway spruce
sawdust, Norway spruce, and southern pine TMP fibers.
Transparent amber solutions of wood could be obtained when
the dissolution of the same lignocellulosic materials was
attempted in [benzylmim]Cl. The fact that it became possible
to fully acetylate wood is testimony to the complete accessibility
of the various wood components to the chemical reagents,
providing additional evidence supporting its total solubilization.
During the dissolution, the components of wood remained intact,
which opens a whole range of possibilities to utilize wood in
an efficient manner. From these solutions wood can be readily
regenerated as an amorphous material. After this dissolution–
regeneration treatment, wood cellulose can be efficiently
digested to glucose, which could be processed to various
commodity chemicals (and fuels) (25). Considering the relatively
wide solubility characteristics and compatibility of ionic liquids
with many organic or inorganic functional chemicals or
polymers, one may envisage that this research could create a
variety of new strategies for converting our abundant woody
biomass to valuable chemicals and novel functional composite
biomaterials (33, 34). A variety of such promising technology
platforms are currently under investigation in our laboratories
aimed at further augmenting and exploring the full potential of
the described research, details of which will be published in
due course.
Figure 6. Diffusion measurements on acetylated Norway spruce sample
dissolved in CDCl3: diffusion rates of CDCl3 ([), lignin ꢀ-O-4 R proton
(9), and cellulose H-1 (2).
the intensity of the ꢀ-O-4 signals for lignin should be ap-
proximately 8% of the proton signal intensity of cellulose.
1
Careful integration of the H spectrum depicted in Figure 6
yielded a value of 7.3%, which is in good agreement with the
anticipated value.
The spectroscopic NMR characterization of acetylated intact
wood in deuterated chloroform offers additional opportunities
for addressing remaining questions related to the macromo-
lecular and connectivity issues among the various wood
components. To demonstrate this, we subjected the acetylated
sample to NMR diffusion measurements (30). Such diffusion
data provide ideal mobility information of the various molecules
present in solution. Small molecules usually show characteristic
large diffusion rates (such as the solvent present in an NMR
sample), whereas larger molecules and macromolecules show
smaller diffusion rates. Furthermore, it needs to be emphasized
that the diffusion rates of all molecules are dependent on their
hydrodynamic volume, which is related to the molecular weight,
providing direct information about chemical bonding between
molecules. In a diffusion experiment, the sample magnetization
is first excited by a normal RF pulse, which is followed by a
gradient pulse. Then follows a diffusion period (τ), which is a
period when molecules are allowed to diffuse in the NMR tube.
After τ, another gradient pulse with the opposite sign to the
first one is applied to refocus the residual magnetization, and
1
the spectrum (otherwise normal H spectrum, but attenuated
LITERATURE CITED
proportionally to the diffusion speed of the molecule in question)
is collected. If a molecule has high diffussion speed (small
molecules), its signal decays more quickly than those of
molecules with small diffusion speed (large molecules). There-
fore, the decay speed of the signals provides direct information
of the mobility of the molecules in solution; covalently attached
molecules should display exactly similar diffusion speeds. In
practice, the diffusion measurements are usually carried out by
varying the strength of the gradient pulses instead of varying
the diffusion delay τ. The outcome of such experiments is
similar to varying the diffusion delay but avoids the influence
of relaxation during the experiment (32). Such diffusion
measurements carried out on a sample of completely acetylated
Norway spruce sawdust are shown in Figure 6. Notably, the
diffusion rates of lignin (R signal of the ꢀ-O-4 subunits) and
cellulose (H-1 protons) are clearly smaller than that of the
solvent (i.e., the decay rate of the signal). Most significant,
however, is the fact that cellulose and lignin have clearly
different diffusion rates. Cellulose has a much lower overall
diffusion constant as compared to lignin. The decay of the
(1) Adler, E. Lignin chemistry–past, present and future. Wood Sci.
Technol. 1977, 11, 169–218.
(2) Ammalahti, E.; Robert, D.; Bardet, M.; Brunow, G.; Kilpelainen,
I. Identification of side-chain structures in a poplar lignin using
three-dimensional HMQC-HOHAHA NMR spectroscopy. J. Ag-
ric. Food Chem. 1998, 46, 5113–5117.
(3) Ralph, J.; Marita, J. M.; Ralph, S. A.; Hatfield, R. D.; Lu, F.;
Ede, R. M.; Peng, J.; Quideau, S.; Helm, R. F.; Grabber, J. H.;
Kim, H.; Jimenez-Monteon, G.; Zhang, Y.; Jung, H.-J. G.;
Landucci, L. L.; MacKay, J. J.; Sederoff, R. R.; Chapple, C.;
Boudet, A. M.; Solution-state NMR of lignins. In AdVances in
Lignocellulosics Characterization; Argyropoulos, D. S., Ed.; Tappi
Press: Atlanta, GA, 1999; pp 55–108.
(4) Lu, F.; Ralph, J. Non-degradative dissolution and acetylation of
ball-milled plant cell walls: high-resolution solution-state NMR.
Plant J. 2003, 35 (4), 535–544. Guerra, A.; Filpponen, I.; Lucia,
L.; Saquing, C.; Baumberger, S; Argyropoulos, D. S. Toward a
better understanding of the lignin isolation process from wood.
J. Agric. Food Chem. 2006, 54, 5939–5947.
(5) Swatloski, R. P.; Spear, S. K. H.; Rogers, R. D. Dissolution of
cellulose with ionic liquids. J. Am. Chem. Soc. 2002, 124, 4974–

9148 J. Agric. Food Chem., Vol. 55, No. 22, 2007
Kilpeläinen et al.
4975. Phillips, D. M.; Drummy, L. F.; Conrady, D. G.; Fox, D. M.;
Naik, R. R.; Stone, M. O.; Trulove, P. C.; De Long, H. C.; Mantz,
R. A. Dissolution and regeneration of Bombyx mori silk fibroin
using ionic liquids. J. Am. Chem. Soc. 2004, 126, 14350–14351.
Xie, H. B.; Zhang, S. B.; Li, S. H. Chitin and chitosan dissolved
in ionic liquids as reversible sorbants of CO2. Green Chem. 2006,
8, 630–633.
a mechanistic working hypothesis. Phytochem. ReV. 2003, 2, 235–
255.
(20) Hunter, C. A.; Sanders, J. K. M. The nature of PI–PI interactions.
J. Am. Chem. Soc. 1990, 112, 5525–5534.
(21) Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W.
Characterizing ionic liquids on the basis of multiple solvation
interactions. J. Am. Chem. Soc. 2002, 124, 14253–14254.
(22) Forsyth, S. A.; MacFarlane, D. R.; Thomson, R. J.; von Itzstein,
M. Rapid, clean, and mild O-acetylation of alcohols and carbo-
hydrates in ionic liquid. Chem. Commun. 2002, 714–715.
(23) Xie, H. B.; Li, S. H.; Zhang, S. B. Ionic liquids as novel solvents
for the dissolution and blending of wool keratin fibers. Green
Chem. 2005, 7, 606–608.
(24) Zhai, W.; Chen, H. Z.; Ma, R. Y. Structural characteristics of
cellulose after dissolution and regeneration from the ionic liquid
[bmim]Cl. Beijing Huagong Daxue Xuebao, Ziran Kexueban
2007, 34 (2), 138–141.
(25) Argyropoulos, D. S. Use of lignocellulosics solvated in ionic
liquids for production of biofuels. U.S. Patent Appl. 60,888,438,
Feb 6, 2007.
(26) Mabee, W.; Gregg, D. J.; Arato, C.; Berlin, A.; Bura., R.; Gilkes,
N.; Mirochink, O.; Pan, X.; Pye, K.; Saddler, J. Updates on
softwood to ethanol process development. Appl. Biochem. Bio-
technol. 2006, 129–132, 55–70.
(27) Liu, L. Y.; Chen, H. Z. Enzymatic hydrolysis of cellulose materials
treated with ionic liquid [BMIM] Cl. Chinese Sci. Bull. 2006, 51
(20), 2432–2436.
(28) Wasserscheid, P.; Welton, T. Ionic Liquids in Synthesis; Wiley-
VCH: Weinheim, Germany, 2003.
(29) Rogers, R. D.; Seddon, K. Ionic liquids–solvents for future.
Science 2003, 302 (31), 792–793.
(30) Schlufter, K.; Schmauder, H.-P.; Dorn, S.; Heinze, T. Efficient
homogeneous chemical modification of bacterial cellulose in the
ionic liquid 1-N-butyl-3-methylimidazolium chloride. Macromol.
Rapid Commun. 2006, 27 (19), 1670–1676.
(31) Barthel, S.; Heinze, T. Acylation and carbanilation of cellulose
in ionic liquids. Green Chem. 2006, 8 (3), 301–306.
(32) Wu, D.; Chen, A.; Johnson, C. S., Jr. An improved diffusion-
ordered spectroscopy experiment incorporating bipolar-gradient
pulses. J. Magn. Reson. A 1995, 115 (2), 260–264.
(33) Argyropoulos, D. S.; Xie, H. B. Polymer derivatives and
composites from the dissolution of lignocellulosics in ionic liquids.
U.S. Patent Appl. 60,888,447, Feb 6, 2007.
(34) Argyropoulos, D. S.; Xie, H. B.; Filpponen, I. Product preparation
and recovery from thermolysis of lignocellulosics in ionic liquids.
U.S. Patent Appl. 60,888,453, Feb 6, 2007.
(35) Buchanan, C. M.; Hyatt, J. A.; Lowman, D. W. Two-dimensional
NMR of polysaccharides: spectral assignments of cellulose
triesters. Macromolecules 1987, 20, 2750–2754.
(6) Wu, J.; Zhang, J.; Zhang, H.; He, J.; Ren, Q.; Guo, M.
Homogenous acetylation of cellulose in a new ionic liquid.
Biomacromolecules 2004, 5, 266–268.
(7) Pu, Y.; Jiang, N.; Ragauskas, A. J. Ionic liquid as a green solvent
for lignin. J. Wood Chem. Technol. 2007, 27 (1), 23–33.
(8) Argyropoulos, D. S.; Sun, Y.; Palus, E. Isolation of residual Kraft
lignin in high yield and purity. J. Pulp Paper Sci. 2002, 28 (2),
50–53.
(9) Fort, D. A.; Remsing, R. C.; Swatloski, R. P.; Moyna, P.; Moyna,
G.; Rogers, R. D. Can ionic liquids dissolve wood? Processing
and analysis of lignocellulosic materials with 1-n-butyl-3-meth-
ylimidazolium chloride. Green Chem. 2007, 9 (1), 63–69.
(10) Myllymaki, V.; Aksela, R. Dissolution and delignification of
lignocellulosic materials with ionic liquid solvent under microwave
irradiation. FI 1156 (2003) and PCT Int. Appl., 2005.
(11) Li, X.; Simonsen, J.; Li, K. C. Wood dissolution and the
regeneration of its components using ionic liquids; Abstracts of
Papers, 227th ACS National Meeting, Anaheim, CA, March 28–
April 1; American Chemical Society: Washington, DC, 2004.
(12) Fukaya, Y.; Sugimoto, A.; Hiroyuki, O. Superior solubility of
polysaccharides in low viscosity, polar, and halogen-free 1,3-
dialkylimidazolium formats. Biomacromolecules 2006, 7, 3295–
3297.
(13) Zhu, S.; Wu, Y.; Chen, Q.; Yu, Z.; Wang, C.; Jin, S.; Ding, Y.;
Wu, G. Dissolution of cellulose with ionic liquids and its
application: a mini-review. Green Chem. 2006, 8, 325–327.
(14) Unger, E. W.; Fink, H. P.; Philipp, B. Morphometric investigation
of the swelling dissolution process of cellulose fibers in FeTNa
and LiCl/dimethylacetamide. Papier 1995, 49, 297–300.
(15) Wu, J.; Zhang, J.; Zhang, H.; He, J.; Ren, Q.; Guo, M. L.
Homogeneous acetylation of cellulose in a new ionic liquid.
Biomacromolecules 2004, 5, 266–268.
(16) Zhang, H.; Wu, J.; Zhang, J.; He, J. 1-Allyl-3-methylimidazolium
chloride room temperature ionic liquid: A new and powerful
nonderivatizing solvent for cellulose. Macromolecules 2005, 38,
8272–8277.
(17) Striegel, A. M. Advances in the understanding of the dissolution
mechanism of cellulose in DMAc/LiCl. J. Chilean Chem. Soc.
2003, 48, 73–77.
(18) Remsing, R. C.; Swatloski, R. P.; Rogers, R. D.; Moyna, G.
Mechanism of cellulose dissolution in the ionic liquid 1-n-butyl-
3-methylimidazolium chloride: a 13C and 35/37Cl NMR relaxation
study on model systems. Chem. Commun. 2006, 12, 1271–1273.
(19) Guerra, A.; Gaspar, A.; Contreras, I. S.; Lucia, L.; Crestini, C.;
Argyropoulos, D. S. On the propensity of lignin to associate, part
I. A size exclusion chromatographic study with native lignins
isolated from different wood species. Phytochemistry 2007, 68
(19). Chen, Y. R.; Sarkanen, S. Macromolecular lignin replication:
Received for review June 9, 2007. Revised manuscript received August
13, 2007. Accepted September 1, 2007.
JF071692E