NC State
Bai, Y., Chen, F., and Liu, M. (2020). "Separation of cellulose and lignin from eucalyptus wood by liquefaction in deep eutectic solvents," BioRes. 15(3), 6886-6901.


Lignin and cellulose were separated from eucalyptus wood powder by partial liquefaction in deep eutectic solvents (DESs). Approximately 97.4% of the original lignin in the eucalyptus wood was separated and recovered with a choline chloride-PEG 200 mixture. Lignin purity that was as high as 99% could be achieved by this separation method. The DES that was comprised of choline chloride, PEG 200, and glycerin could be used to separate the cellulose. The purity and the crystallinity of the isolated cellulose were as high as 92% and 73%, respectively. Experimental data showed that the examined DES systems could effectively separate the constituents of eucalypt wood powder.

Download PDF

Full Article

Separation of Cellulose and Lignin from Eucalyptus Wood by Liquefaction in Deep Eutectic Solvents

Youcan Bai,a Fangeng Chen,a,* and Mingyou Liu b

Lignin and cellulose were separated from eucalyptus wood powder by partial liquefaction in deep eutectic solvents (DESs). Approximately 97.4% of the original lignin in the eucalyptus wood was separated and recovered with a choline chloride-PEG 200 mixture. Lignin purity that was as high as 99% could be achieved by this separation method. The DES that was comprised of choline chloride, PEG 200, and glycerin could be used to separate the cellulose. The purity and the crystallinity of the isolated cellulose were as high as 92% and 73%, respectively. Experimental data showed that the examined DES systems could effectively separate the constituents of eucalypt wood powder.

Keywords: Eucalyptus; Choline chloride; Component separation; Cellulose; Lignin

Contact information: a: State Key Laboratory of Pulp and Paper Engineering, South China University of Technology, Guangzhou 510640, China; b: National Engineering Research Center of Papermaking & Pollution Control, South China University of Technology, Guangzhou 510640, China;

* Corresponding author:


Lignocelluloses are bio-based polymers comprising cellulose, hemicelluloses, and lignin. The fractionation of these components favors the high-valued utilization of biomass. Successive liquefaction, an important separation technique, has been reported in preparing cellulose microfibers from the holocellulose of banana pseudo-stems with a PEG 400-glycerin medium at 150 °C (Li et al. 2015). The majority of the original lignin and hemicelluloses were liquefied, leaving cellulose undissolved. The resulting cellulose microfibers had a crystallinity of 81.3%, whereas that of the starting holocellulose was 52.2%.

Ionic liquids (ILs) can be good solvents for dissolving cellulose, hemicelluloses, and/or lignin (Lan et al. 2011; Wang et al. 2012; Fort et al. 2016). However, ILs have some drawbacks such as high toxicity, nonbiodegradability, complex synthesis requiring purification, and high cost of the starting materials. Additionally, the ability of ILs to liquefy biomass after repeated recovery and reuse has been noted to decrease (Couling et al. 2006).

Deep eutectic solvents (DESs) are widely acknowledged as a new class of ILs, although they are quite different from one another in composition and properties (Kohli 2019). DESs are a eutectic mixture of Lewis or Brønsted acids and bases that can contain a variety of anionic and/or cationic species. In contrast, ILs are composed of one discrete anion and cation pair (Smith et al. 2014). DESs are usually cheaper and less toxic than ILs. In addition, DESs retain most of their original activity after repeated reuse during biomass liquefaction (Procentese et al. 2015).

DESs prepared from quaternary ammonium salts and different hydrogen bond donors (e.g., carboxylic acids and polyalcohols) can selectively dissolve lignin from plant fibers and leave the majority of the cellulose and hemicelluloses undissolved (Francisco et al. 2012). This procedure can be used as a pretreatment of lignocellulose biomass in its conversion to bio-ethanol (Kumar et al. 2016). DESs could also been used to replace the oxygen delignification after kraft pulping (Majová et al. 2017). Alvarez-Vasco et al. (2016) succeeded in dissolving lignin from wood fibers using DES prepared from choline chloride (as a hydrogen acceptor) combined with various hydrogen-bond donors (i.e., acetic acid, lactic acid, levulinic acid, and glycerol). They reported that 78% of the white poplar wood lignin, as well as 58% of the fir wood lignin, was dissolved. The dissolved lignin (95.4% purity) was precipitated from the DES with deionized water. The recovered lignin had a low molecular weight and a narrow molecular weight distribution. The lignin could not be dissolved completely during the liquefaction, so that the liquefaction residue still contained a considerable portion of lignin.

It is well known that plant fibers can easily be liquefied in hydrogen-donor solvents, such as glycerol and PEG 400, in the presence of a catalytic amount of sulfuric acid. In contrast, the liquefaction of plant fiber in DESs is incomplete. During the liquefaction in DES, lignin can be partially dissolved, leaving cellulose and hemicelluloses in the residue. Increasing the dosage of hydrogen-bond donors in DESs is expected to improve the dissolution of hemicelluloses and cellulose.

In this work, several hydrogen bond donors (i.e., PEG 200, lactic acid, and glycerol) were added to choline chloride respectively to form the liquefying media. PEG 200 was used in place of PEG 400 because it has a lower viscosity and a higher hydroxyl content than PEG 400; this should result in the formation of additional hydrogen bonds in the DES system. Lactic acid and glycerin, when mixed with choline chloride, should improve the dissolution of lignin (Procentese et al. 2015; Kumar et al. 2016). The isolation of cellulose and lignin from eucalypt fibers can be accomplished by successive liquefaction of wood powder with the catalytic addition of sulfuric acid. The influences of solvent type, reaction time, and reaction temperature on the yields and purities of isolation products were investigated.



Eucalyptus wood (Eucalyptus exserta (F. V. Muell)) was obtained from Leizhou (Guangdong Province, China). It was ground to a 50-mesh powder prior to use. The screened wood powder contained 45.3% cellulose, 67.0% holocellulose, and 30.8% lignin. The content of ash and benzene-ethanol extractive were 0.3% and 1.2%, respectively. The data above were determined according to Chinese National Standard GB/T 2677.2~GB/T 2677.10.

All the chemical reagents used in this work were commercial products. Glycerol, PEG 200, and 1,4-dioxane, were of reagent grade and were used without further purification. Lactic acid and choline chloride were produced by Shanghai Macklin Biochemical Co., Ltd. (Shanghai, China).


Preparation and characterization of DESs

Four DESs, designated DES 1, DES2, DES3, and DES4, were prepared by mixing choline chloride with lactic acid, glycerin, PEG 200, and PEG 200-glycerin (4:1, v/v) mixture, with 1:4 weight ratio, respectively. The viscosities of them at 25 °C were 84, 130, 61, and 42 mPa∙s, respectively. The powder was mixed with the selected DES using a 1:4 powder-to-medium ratio (w/w) at room temperature. The mixture was then heated for 0.5 h at 60 °C. The viscosity of the DESs was measured using a NDJ-8S rotary viscometer (Shanghai Fangrui Instrument Co., Ltd.; Shanghai, China) at 251 °C.

Separation of eucalypt wood components

A total amount of 50 mL of DES was added to a 3-neck flask equipped with a thermometer, stirrer, and condenser. To the DES was added 5 g of dried wood powder and 0.2 to 1.5 mL of 98% sulfuric acid. Then, the flask was heated to the desired temperature and kept at that temperature for 3 h. After the liquefaction reaction, the flask was placed into a cool water bath to terminate the reaction.

The reaction mixture was filtered with a G3 sintered glass funnel. As the DESs were viscous, the mixture was diluted with a 1,4-dioxane-water mixture (4:1 v/v) before filtration. The solid residue was washed with 1,4-dioxane-water mixture until the filtrate was colorless. The filtered and washed solid was then vacuum-dried. The cellulose content in the resulting residual was measured by the nitric acid-ethanol method of Loow et al. (2017). The filtrate was concentrated by rotary evaporation. The resulting concentrate was mixed with 400 mL of deionized water to form a precipitate, which was filtered with a G4 sintered glass funnel. The precipitate (which was predominantly lignin) was washed with deionized water repeatedly until the eluent became colorless. Then, the washed precipitate was vacuum-dried to obtain the crude lignin. The purity of the crude lignin was measured using TAPPI T222 om-11 (2011). The yield and purity of the separated components of the wood powder were calculated according to the following equations,

YC=mC/m0 × 100% (1)

YL=mL/m0 × 100% (2)

PC=mPC/mC × 100% (3)

PL=mPL/mL × 100% (4)

where YC and YL represent the yields of crude cellulose and crude lignin, respectively; PC and PL represent the purity of the crude cellulose and crude lignin, respectively; mC is the mass of crude cellulose; mL is the mass of crude lignin; mPC is the mass of the residual solid obtained by the nitric acid-ethanol method (i.e., pure cellulose); mPL is the total mass of lignin calculated from the sum of acid-insoluble lignin and acid-soluble lignin (determined by TAPPI T222 om-11 (2011)); and m0 is the mass of raw wood powder. All the liquefaction experiments were repeated 3 times, and the results were averaged.

The liquefaction conditions were optimized based on the yield of lignin. In subsequent experiments, the samples used were prepared from the crude cellulose and crude lignin that were obtained under optimal liquefaction conditions. The crude lignin and crude cellulose samples were designated L1 through L4 and C1 through C4, respectively.

Gel permeation chromatography (GPC) analysis

The molecular weight and molecular weight distribution of lignin were measured by GPC on an Agilent 1100 HPLC instrument (Agilent Technologies Inc.; Palo Alto, CA, USA) equipped with PL1110-6300 and PL1110-6530 columns. Tetrahydrofuran (THF) was used as the eluent. The lignin samples were acetylated with acetic anhydride in pyridine prior to GPC analysis. The GPC columns were calibrated using narrow-distributed polystyrene standards. The injection volume was 20 μL for each run. The column temperature was maintained at 40±0.1 °C.

Fourier transform infrared (FT-IR) spectroscopy

The FT-IR spectra of lignin and cellulose samples were recorded using a Vector 33 FT-IR spectrophotometer (Bruker; Ettlingen, Germany) over the wavenumber range of 4000 to 400 cm-1. Approximately 3-mg samples were well mixed into 200 to 250 mg of fine KBr powder, then finely pulverized and placed into a pellet-forming die. A force of approximately 8 tons was applied under a vacuum for 5 min to form transparent pellets.

Thermogravimetric analysis (TGA)

TGA was performed using a Q500 Thermogravimetric Analyzer (TA Instruments; New Castle, DE, USA). The samples were heated from 40 to 700 °C at a heating rate of 10 °C /min under a nitrogen blanket at a flow rate of 25 mL/min.

Determination of degree of polymerization of cellulose

The average degree of polymerization (DP) of cellulose was calculated from the intrinsic viscosity of cellulose dissolved in cupriethylenediamine (CED) solution (in accordance with ISO 5351 (2010). The viscosity was measured using an Ubbelhode viscometer with a capillary diameter of 0.8 mm at 251 °C. The average DP of cellulose was calculated based on the Mark-Houwink-Sakurada equation,

[η= K × DPα (5)

where [η] is the intrinsic viscosity of the cellulose solution. The values of K and α are 1.33 g/mL and 0.905, respectively, for cellulose in a CED solution, according to ISO 5351 (2010).

X-ray powder diffraction (XRD) analysis

The crystallinity of crude cellulose was quantified by XRD using a Rigaku SmartLab SE polycrystalline X-ray diffractometer (Rigaku; Tokyo, Japan). The testing conditions were as follows: Cu K radiation of wavelength 0.15418 nm, scanning step length 0.04°, and scanning speed 0.2 s/step. The crystallinity index (CrI) was calculated as follows,

CrI = [(I2-I1)/I2× 100% (6)

where I1 and I2 represent the peek intensities recorded at 18.5° and 22.6°, respectively, for crystalline and amorphous cellulose.

13C-NMR spectra characterization

The 13C NMR spectra of the samples were recorded using an Avance-3-HD 400 spectrometer (Bruker; Ettlingen, Germany) following Robert (1992). Two hundred milligrams of each sample were dissolved in 1 mL of deuterated dimethylsulfoxide (DMSO-d6) at room temperature. The main parameters were as follows: NMR tube diameter 5 mm, scan width 200 ppm, NS = 15,000, TD = 32,768, LB = 10 Hz, pulse angle 90°. No internal standards were used. The 13C-NMR scanning time was 8 h for each sample.


Selection of Liquefaction Agents

Theoretically, polyethylene glycols (PEGs) of any molecular weight can be used in the partial liquefaction of woody biomasses because all PEGs are hydrogen bond donors. However, the preliminary experiments indicated that the lignin yield was low when PEG 400 or PEG 400-glycerin was mixed with choline chloride as the liquefaction medium. In this study, PEG 200 was used instead of PEG 400 because the former contains more hydroxyl groups and is less viscous, which helps to improve the ability to penetrate the fibers to rapidly dissolve them. The composition of the PEG-glycerin mixture was set with reference to the authors’ previous work (Liu et al. 2009).

Unlike the liquefaction in DESs, the liquefaction of lignocellulose in polyhydric alcohols should be catalyzed by sulfuric acid or other strong acids. The liquefaction procedure in this work combined two types of liquefaction to improve the dissolution of lignin and hemicellulose. The influences of sulfuric acid concentration on the yield of liquefaction residue and crude lignin were studied. The reaction lasted 120 min at 140 °C.

Fig. 1. Yield of residue (a) and crude lignin (b) at various sulfuric acid concentrations (wt%)

Figure 1 shows the yields of residue and crude lignin by addition of different dosage of sulfuric acid. It was found that increasing the concentration of sulfuric acid significantly reduced the residue yield in the liquefaction media containing PEG or glycerin, while the effect was not so obvious in the choline chloride-lactate liquefaction medium. As a matter of fact, sulfuric acid is seldom used in the choline-lactic acid system. In this study, the dosage of sulfuric acid was chosen at 1 mL to ensure that the reaction proceeded at a relatively high rate.

Influence of Reaction Conditions on the Yield of Separated Products

Figure 2 shows the impact of temperature on the yield of liquefaction residue and crude lignin. Reduced residue and increased crude lignin yields could be achieved at enhanced temperature. The reason for these observations was that increasing the temperature provided more energy to cleave the covalent bonds connecting the various constituents, which promoted wood dissolution. The yields of both residue and lignin remained relatively unchanged when the DES temperature was increased above 140 °C, which indicated that most of the lignin and hemicelluloses, as well as amorphous cellulose, was transferred into the liquid phase. Based on the lignin yield levels, 140 °C was recognized as the optimal DES temperature for the liquefaction.

Fig. 2. Yield of residue (a) and crude lignin (b) at various DES temperatures, with a reaction time of 120 min

It was found that the dominant component of the solid residue was cellulose. For example, more than 90 wt.% of the residue was cellulose for the DES4 liquefaction medium. The residue yield in this work was higher than that of wood liquefaction in ethylene glycol with sulfuric acid (Rezzoug and Capart 2003), in which most of the lignocellulose was converted into liquid products. PEG with or without glycerin, when being used in the presence of sulfuric acid, is a strong liquefaction medium to break the linkage between the components of lignocellulose. The reaction medium used in this research, however, was a mixture of a hydrogen bond donor and choline chloride. The DES can selectively cleave the ether linkage between lignin units and then produce hydroxyls (Alvarez-Vasco et al. 2016; Loow et al. 2017). Lignin and hemicelluloses are easily dissolved by hydrogen-donor solvents because these two constituents are not crystalline. The ether linkages connecting the hemicelluloses to the lignin are thus easily cleaved. However, the highly ordered structure with crystalline cellulose restricts dissolution in hydrogen-donor media. The recovery of relatively pure crude lignin and cellulose fractions demonstrates that the liquefying agent used in this work is selective for separating these constituents.

Fig. 3. Yield of crude cellulose (a) and crude lignin (b) at various reaction times in various DESs

The yields of crude cellulose and crude lignin versus time are shown in Fig. 3. The experiments were conducted in various DESs at 140 °C. At the beginning of the liquefaction, the residue yield could be thought of as 100% because no wood components were dissolved. The yield of residue (could be recognized as “crude cellulose”) was reduced at time 20 min, indicating that the wood components were being separated. Afterward, the yield of crude cellulose decreased slowly, while that of crude lignin increased slowly. Among the wood components, hemicelluloses and lignin reacted with the liquefying agent more easily than cellulose did. This rapidly reduced the amount of solid residue within a short time period. The liquefaction of amorphous cellulose was easier than that of crystalline cellulose. After 120 min, the yields of lignin and cellulose leveled out; hence, it was determined that a reaction time of 120 min was optimal.

In this work, the changes of hemicelluloses were not investigated because the recovery of hemicellulose was not included.

Influence of Reaction Conditions on the Purity of Cellulose and Lignin

Figure 4 shows that increasing the DES treatment temperature increased the purity of the crude cellulose. The results indicate that increasing the temperature increases the difference in the dissolution rate of wood components. The use of DES1 and DES2 resulted in lower cellulose purity in the crude solids than did the use of DES3 and DES4. The former DES media were not as effective at dissolving lignin and hemicelluloses in the eucalypt wood powder as the latter DES media. Undissolved lignin and hemicelluloses remained in the crude cellulose as impurities. Generally, the off-color of the crude cellulose could be attributed to the presence of dark-colored lignin. As demonstrated in Fig. 4, the cellulose separated by DES4 had the lightest color, demonstrating that it had the least amount of residual lignin remaining. Based on the experimental results, PEG 200 or PEG 200/glycerol dissolved more lignin than did lactic acid or glycerol. Part of the discoloration of the crude cellulose could be due to carbonization of the wood fibers in the eucalypt powder. From TGA, the char residue contents of crude cellulose obtained from DES1 and DES2 were relatively high in elemental carbon. This was attributed to the wood fibers being partially carbonized in the presence of sulfuric acid. Highly pure crude cellulose was achieved when using DES3 or DES4. For instance, a maximum cellulose purity of 92% was obtained using the DES4 treatment.

Fig. 4. Purity of crude cellulose (a) and crude lignin (b) obtained at different reaction temperatures with various DESs

Figure 4 indicates that all the crude lignin obtained from various DESs, except for that obtained from DES4, were high in purity. The purity could be up to 99%. It appears that the use of DES4 resulted in the extraction of some water-insoluble impurities from the hemicelluloses and/or cellulose, so that the purity of crude lignin obtained from DES4 was low. Some of the DESs used in this study contained PEG 200 and glycerin, which are widely used liquefying media for wood. Almost all wood constituents, including cellulose, can be easily dissolved in these media (Yao et al. 1993). Therefore, during the separation of lignin with DES4, cellulose and hemicelluloses, as well as their degradation products, may have been extracted along with lignin; this could then be co-precipitated in subsequent steps, reducing the purity of the fractionated lignin. The color of crude cellulose samples (Fig. 5) was in good agreement with the purity of crude cellulose.