Comparing the influence of acetate and chloride anions on thestructure of ionic liquid pretreated lignocellulosic biomass
Hyungsup Kim a, Yongjun Ahn b, *, Seung-Yeop Kwak b, **
a Department of Organic and Nano System Engineering, Konkuk University, Seoul, Republic of Koreab Department of Materials Science and Engineering, Seoul National University, Seoul, Republic of Korea
a r t i c l e i n f o
Article history:Received 22 July 2015Received in revised form11 July 2016Accepted 14 July 2016
The effect of the anion type, in an ionic liquid, on the transition of the crystalline structure whenlignocellulose was pretreated in 1-ethyl-3-methylimidazolium acetate (EmimOAC) or 1-ethyl-3-methylimidazolium chloride (EmimCl) was studied. The influence of the pretreatment on the compo-sition, the molecular structure, and the crystalline structure was observed using Fourier-transforminfrared (FT-IR) compositional analysis, thermogravimetric analysis (TGA), rheological behavior, and X-ray diffraction (XRD). Compared to EmimCl, EmimOAC pretreatment substantially decreased the ligninand hemicellulose contents. The pretreatment also significantly changed the entanglement or cross-linking state of polymer chains in the lignocellulose solution. The changes in lignin content and thetransformation from cellulose I to II were dependent on the anion type of the ionic liquid. The pretreatedsamples were recrystallized to cellulose II only in EmimOAC, whereas the samples pretreated withEmimCl had both cellulose I and II structures present at the same time.
The need for alternative resources for fossil-based fuel hasaccelerated investigations on sustainable energy and bio-basedmaterials. Lignocellulosic biomass, the most abundant materialon the earth, is considered to be a promising candidate with greatpotential to serve as a substitute for fossil fuels. Lignocellulosicbiomass (lignocellulose) is a complex composite of cellulose, ligninand hemicellulose as well as other materials. Among these com-ponents, cellulose is the major carbon-source for bio-fuel and sugarproducts. In particular, sugars obtained from cellulose are attractivestarting materials for large-scale production in a bio-refinery [1e6],although the potential of lignocellulosic biomass has not been fullyexploited due to poor accessibility of enzymes. Native lignocellu-losic biomass is inherently recalcitrant to biodegradation, owing tothe presence of lignin, high crystallinity of cellulose, and complexbonding between them. Lignin in lignocellulose acts as a physical orchemical barrier, protecting cellulose and hemicellulose fromenzymatic degradation, and decreasing the efficiency of enzymatichydrolysis [7,8]. The high crystallinity of the native cellulose
structure (cellulose I) also hinders the accessibility of enzymes intopolysaccharides. Accordingly, the conversion of the crystallinestructure of cellulose I to other crystal forms such as cellulose II,cellulose IIIII and cellulose IIII or to an amorphous structure, cansignificantly improve the susceptibility to hydrolysis [9]. Pretreat-ment of the biomass involving physicochemical, structural, andcompositional changes is critical to overcome its recalcitrance tohydrolysis [10e12].
Ionic liquid-pretreatment for lignocellulosic biomass hasemerged as a novel technology. Since Swatloski et al. [13] reportedcellulose dissolution using ionic liquid, many ionic liquids havebeen synthesized and tested for lignocellulose dissolution. Repre-sentatively, superior solubility of those based on acetate andchloride anion combined with dialkylimidazolium cations was re-ported to dissolve high amounts of each components, such as cel-lulose, lignin and hemicellulose, in lignocellulose under milecondition [9]. Pretreatment of the lignocellulose by ionic liquidsremoves lignin as well as decreasing the level of crystallinity byweakening intra-/inter-molecular hydrogen bonds. In early studies,Dadi et al. [14] reported that the enzymatic hydrolysis yield ofAvicel was dramatically improved after pretreatment with 1-buthyl-3-methylimidazolium chloride (BmimCl) since the result-ing decrease in crystallinity improved enzyme accessibility to thesubstrate. Other researchers [15,16] studied the changes of
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253244
crystalline structure and crystallinity using 1-ethyl-3-methylimidazolium acetate to pretreat various biomasses under arange of conditions. Although previous researches demonstratedthe potential of ionic liquids for pretreatment in a bio-refinery, asystematic study on the influence of anion type remains to be done.In point of structural change of lignocellulose, the anion type inionic liquids should be also considered due to its significance fordisturbing the microstructure. Therefore, it is still required for anunderstanding on the physical state of the lignocellulose moleculeassociated differentlywith anion types of the ionic liquid in order tocontrol the crystalline conversion.
In this study, we purposefully focused our investigation usingtypically used ionic liquids, 1-ethyl-3-methylimidazolium acetateand 1-ethyl-3-methylimidazolium chloride, for the lignocellulosepretreatment. The selected ionic liquids are expected to representproperties of other ionic liquids consisting of similar anion, such asformate, iodide and bromide anions, due to their similar properties.Two types of ionic liquid with different anions were compared forpretreatment efficiency of lignocellulose: the influence of the aniontype on the composition and the crystalline structure was investi-gated. The molecular interactions and the mobility were observedby considering the rheological behavior, which was used todemonstrate differences in crystalline change. This study provides afundamental understanding of the action of ionic liquid and pro-vides recommendations for which materials and conditions shouldbe selected for pretreatment of the lignocellulosic biomass. Inaddition, the study is also expected to provide information con-cerning the mechanism of crystalline transition of lignocellulose byinvestigating the relationship between the dynamics of lignocel-lulose molecules and its composition.
2. Experimental section
2.1. Materials
Three different lignocelluloses were investigated in this study:cotton (Gossypium arboretum) stalks, hemp (Cannabis sativa) stalksand acacia (Acacia auriculiformis) pruning. Cotton and hemp werekindly provided by Hemplee korea Co., and acacia was provided byMoorim P&P Co. Microcrystalline cellulose (MCC, from SigmaAldrich) was used as a control. 1-ethyl-3-methylimidazolium ace-tate (EmimOAC, ~95%, water content: 0.27 wt%) and 1-ethyl-3-methylimidazolium chloride (EmimCl, ~95%, water content:0.22 wt%) were purchased from Sigma Aldrich. Anhydrous LithiumChloride (LiCl) and N,N-dimethylacetamide (DMAc) were pur-chased fromDaejung Chemicals&Methals Co. To removemoisture,DMAc was distilled over CaH2 before use.
2.2. Pretreatment and recrystallization
The biomass samples were milled to 20 mesh (particle size:
Table 1Pretreatment conditions and ionic liquids used in this study.
Sample code Abbreviation Structure
EmimOAC 80 �C EmimOAC
EmimOAC 120 �CEmimCl 80 �C EmimClEmimCl 120 �C
200e800 mm) before pretreatment and were then incubated withEmimOAC or EmimCl at 4% (w/w) for 2 h at 80 �C or 120 �C. Thetemperatures were chosen to avoid thermal decomposition of theionic liquid which could affect the efficiency of the pretreatment[17]. After the ionic liquid-pretreatment, the samples were trans-ferred to a beaker, and acetone/water (1:1 v/v) was added in orderto separate and remove the lignin from the cellulose [18]. Themixture was stirred for 1 h at room temperature and then theprecipitated substrates were separated from the coagulant byfiltration through a ceramic funnel with nylon filter paper, undervacuum. The material was washed with water at least four times toremove the ionic liquid and then dried in a vacuum oven for 2 daysat 60 �C. The pretreatment conditions used for the lignocellulosesare summarized in Table 1.
The samples were subjected to FT-IR spectroscopy using a Per-kinElmer spectrum BX spotlight spectrophotometer with diamondattenuated total reflectance attachment. Scanning was conductedfrom 4000 to 700 cm�1 with 64 repetitious scans averaged for eachspectrum. Resolutionwas 4 cm�1 and interval scanningwas 2 cm�1.
2.4. Lignocellulosic biomass composition
The composition of the samples before and after the pretreat-ment was determined according to the analytical procedure of theNational Renewable Energy Laboratory (NREL) No. 002 [19,20].Briefly, samples (150 mg each) were treated with 72% (v/v) sulfuricacid at 30 �C for 3 h, followed by diluted acid (4%) at 121 �C. Hy-drolyzed products were analyzed by HPLC (Young-Lin ModelYL9100, Korea) equipped with an RI detector and a Shodex sugarSP0810 column operated at 85 �C. The mobile phase consisted ofdeionized water with a flow rate of 0.6 ml/min. The amounts ofcellulose and hemicellulose were calculated from the glucose andxylose contents multiplied by a conversion factor of 0.9 and 0.88respectively [21]. The amount of acid-insoluble lignin after acidhydrolysis was measured as the mass of insoluble residueremaining [19]. The amount of acid-soluble ligninwas measured bya UVevis spectrophotometer at 205 nm with an extinction coeffi-cient value of 110 Lg�1 cm�1 (NREL, 1996).
2.5. Thermogravimetric analysis (TGA)
Thermogravimetric analysis of the biomass samples was per-formed using a PerkinElmer instrument, Pyris Diamond TG/DTA.The thermal stability of 0.5 mg of each sample was studied fromroom temperature to 600 �C at a rate of 10 �C/min. The rate of purgegas (Nitrogen) flow was controlled at 70 mL/min.
Name Pretreatment temp. (�C)
1-ethyl-3-methylimidazolium acetate 80
1201-ethyl-3-methylimidazolium chloride 80
120
Table 2Composition of cellulose, hemicellulose and lignin in native and pretreated cotton, hemp and acacia.a
a All measurements are an average of triplicate samples.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253 245
2.6. Rheological properties
In order to avoid cellulose degradation, a DMAc/LiCl solventsystem was employed for measuring steady-state and dynamicviscoelastic behaviors. The lignocellulose was activated using themethod reported by McCormick et al. [22] where the lignocellulosesamples were dissolved in the solvent, 9 wt% DMAc/LiCl, after beingcompletely dried in vacuum below 60 �C.
Rheological measurements were carried out on a strain-controlled rheometer (RS-1, Thermo Electron Co., Germany) with40 mm plate-plate geometry. Frequency sweeps within the linearregion controlled strain were performed at various temperaturesranging from 30 to 110 �C. All data were reduced to a referencetemperature of 30 �C by time-temperature superposition (TTS).Steady shear experiments were carried out at 30 �C with a shearrate of 0.01e500 s�1. To prevent moisture uptake by the sampleduring experiments, the edges of the specimen were coated withsilicon oil between the plates.
In order to estimate and quantify the chain mobility, the acti-vation energy (DEa) of the lignocellulose solutions was calculatedusing the Williams-Landel-Ferry equation and the results aresummarized in Table 3.
lnaT ¼ �c1ðT � TrÞc2 þ ðT � TsÞ
where c1 and c2 denote experimental constants and Tr is referencetemperature (here Tr ¼ 303 K). Then, DEa at the reference tem-perature was estimated from the following equation:
DEaðT ¼ TrÞ ¼ RdlnaTdð1=TÞ
��T¼Tr
¼ 2:303RðC1=C2ÞT2r
Table 3Dependence of activation energy on pretreatment using different ionic liquids.a
a Activation energy is the average of triplicates.
2.7. X-ray diffraction (XRD)
The microstructures of the native and the pretreated sampleswere analyzed by wide-angle X-ray diffraction (WAXD, Ultima IV,Rigaku, Japan). The scans were collected from 2q ¼ 8 to 40� with astep size of 0.01� at 1 s per step. The following empirical equationwas adopted to estimate the cellulose crystallinity in the native andpretreated biomass samples [23].
CrI ¼ ðItotal � IamÞ=Itotal � 100
where Itotal is the diffraction intensity at peak position for 22.5� forcellulose I and Iam is the intensity at suitable locations for theamorphous background (18�). For cellulose II, the main peak ap-pears as a doublet at 21.5�, and the amorphous peak appears at 16�.
3. Results and discussion
3.1. Composition of ionic liquid-pretreated lignocellulose
Fig. 1 shows the FT-IR spectra of the native and the pretreatedMCC, cotton, hemp and acacia. The characteristic peaks of ligninand hemicellulose in the lignocelluloses appeared at1732 cm�1 (C]O unconjugated stretching, hemicellulose), 1515-1505 cm�1 (C]C aromatic symmetrical stretching, lignin),1425 cm�1 (C]C stretching in aromatic groups, lignin and hemi-cellulose) and 1240 cm�1 (CeO aryl group, lignin) [24]. With theexception of acacia, the characteristic peaks were decreased withincreasing temperature, regardless of the type of ionic liquid.However, with EmimOAC pretreatment, there was a more rapiddecrease as the temperature increased, than after treatment withEmimCl. When the biomasses were pretreated with EmimOAC at120 �C, the peaks almost disappeared. For acacia, the significantdifference of FT-IR spectra was not observed in any of the experi-mental conditions. When MCC, cotton and hemp were pretreatedwith EmimOAC, a new peak appeared at 3447 cm�1 (OH stretchingintramolecular hydrogen bonds of cellulose II) [25]. This result in-dicates that the crystalline structure is transformed from cellulose Ito II only by EmimOAC.
The composition of each biomass was measured using HPLC(Table 2). The table shows more efficient decomposition andremoval of lignin and hemicellulose pretreated in EmomOAC thanEmimCl as confirmed by the FT-IR results. The removal of lignin andhemicellulose seems to be related to the solubility of the ionicliquids for cellulose, lignin and hemicellulose. Compared to otheranions in ionic liquids, EmimOAC has better solubility for cellulose,
Fig. 1. FT-IR spectra of MCC, cotton, hemp and acacia samples pretreated with EmimOAC or EmimCl at 80 �C and 120 �C.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253246
lignin and hemicellulose [9,26,27], suggesting that non-cellulosiccomponent would have more freedom to separate from thebiomass after dissolutionwhen solvent has more solubility for eachcomponent [20]. As a result, the lignin and hemicellulose in theEmimOAC-pretreated biomass was substantially reduced. However,EmimCl is less effective dissolution at separating each componentsbecause of low solubility compared to EmimOAC [9,26,27]. Theinsufficient dissolution (heterogeneous state) of lignocellulose inEmimCl results in the low efficiency of lignin and hemicellulosereduction. On the basis of this explanation, the effect of increasingpretreatment temperature can be interpreted as the solubilityenhancement for the lignocelluloses. At high temperature, thelignocellulose in ionic liquids became more homogeneous andloosening, leading to the efficient removal of non-cellulosiccomponents.
The removal efficiency of the lignin and hemicellulose was alsodependent on their initial content. In low lignin content, lignocel-lulose samples were more effectively delignified than for sampleswith a higher content. The lignin removal efficiency in cotton wasapproximately 70% when it was pretreated with EmimOAC at120 �C, however, the efficiency for the pretreated acacia under thesame conditions was only 20%. This trend was similar to hemicel-lulose. The difference in the efficiency is attributed to limited sol-ubility of the ionic liquid and the known recalcitrance of lignin for
dissolution. Lignin in lignocellulose surrounds and combines withcellulose, which acts as chemical barrier [28]. There is an increasedlikelihood of forming strong complex between cellulose and non-cellulose, which further impedes their separation when the non-cellulosic content is high. Consequentially, lignocelluloses havinghigh lignin content is not easily dissolved in ionic liquid.
3.2. Thermal stability of ionic liquids and pretreated lignocellulose
The thermal decomposition of ionic liquids during pretreatmentis a critical issue because the decomposed ionic liquids alter theefficiency of lignocellulose pretreatment [17]. To confirm thechemical integrity of ionic liquids, the observation for theirdecomposition behavior was conducted using TG analysis. Fig. 2(a)shows that both EmimOAC and EmimCl maintained their weight inthe temperature range used in this study (80 and 120 �C). Inaddition, Fig. 2(b) demonstrates that the weight loss of EmimOACand EmimCl for 2 h under isothermal condition was only 0.6% and0.4%, respectively. This behaviors indicate that both ionic liquidsnearly maintained their chemical structure, suggesting the thermaldecomposition of ionic liquids can be neglected in this study.
Molecular weight is well known to plays an important role inexplaining crystallization behavior. However, the molecular weightof lignocellulose is not easy to be measured directly using gel
Fig. 2. (a) TG curves and (b) isothermal TG analysis of EmimOAC and EmimCl at 120 �C.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253 247
permeation chromatography due to its complex composition.Therefore, the relationship between the thermal stability and themolecular weight as alternative method was investigated toobserve the cellulose molecular weight. Fig. 3 shows the derivativethermogravimetric (DTG) curves of the native and the pretreatedsamples. The broad signal below 100 �C indicates the removal ofwater in the samples. The DTG curves of the native samples show asharp and narrow peak at 360 �C, which is typical thermaldecomposition temperature of cellulose. When the samples werepretreated, the thermal degradation started at a lower temperature.The pretreatment temperature differently affected the thermalbehavior according to anion type. For the EmimOAC-pretreatmentat 80 �C, MCC, cotton and hemp show two major decompositionsat 280 and 330 �C. However, pretreatment at 120 �C resulted in asingle decomposition below 300 �C. In particular, the acacia sam-ples treated at 120 �C initiated degradation below 200 �C andshows a broad range of degradation temperature. This indicatesthat the cellulose was decomposed during the pretreatment at hightemperature. The change of the decomposition temperature can beexplained in terms of the decrease in molecular weight [29,30]. It isgenerally known that the decomposition temperature is decreasedwith a lower thermal resistance of polymer chain relating with themolecular weight. Therefore, the cellulose in lignocelluloses is ex-pected to have lower molecular weight with ionic liquid-pretreatment at high temperature. This speculation may beconsidered to be valid, regarding the result the similar behaviorwas observed for DTG curves of cellulose samples having differentmolecular weight [31].
The degradation temperature of the samples treated withEmimCl was lower and split when the treatment was carried out athigh temperature, which is similar to the results with EmimOAC.However, each decomposition peak for the sample treated withEmimCl appeared at a higher temperature than for the samplestreated with EmimOAC. As explained above, the thermal stability ofcellulose could be depended on the molecular weight, suggestingthat EmimOAC-pretreated lignocellulose hydrolyzed more exten-sively than the EmimCl-pretreated sample. In a previous study oncation-anion interactions in ionic liquid, it was reported that theacetate anion has a stronger ability to form ionic bonding withcations thanwith the chloride anion [26]. This ability means that anacetate anion and a proton from a cation can combine easily to formcarbene and acid, which in turn results in the possibility of acid-catalyzed hydrolysis occurring during the pretreatment [29]. Inaddition, the generated carbene leads to ring-opening reaction atend group of glucose. Based on these points, the sample pretreatedwith EmimOAC showed reduced thermal stability compared to theone with EmimCl.
3.3. Rheological behavior of ionic liquid-pretreated lignocellulose
Fig. 4 shows the viscosity changes of 4% solutions for the nativeand the pretreated samples. The viscosity represents the molecularmobility, such as chain rigidity, chain to chain interaction andmolecular weight. The change of the quantitative value in therheological behavior demonstrated the transition of physicalinteraction between lignocellulose molecules by ionic liquid-pretreatment. For the native lignocellulose solutions, a conven-tional polymeric fluid behavior with high viscosity was seen,exhibiting a Newtonian plateau at low shear rate and a shearthinning at high shear rate. This indicates that the lignocellulosesolutions have chain entanglements or junction points between thelignocellulose molecules, meaning that the cellulose is constrainedby the lignin and hemicellulose. For the pretreated lignocellulosesolutions, the viscosity was rapidly decreased while the Newtonianplateau evidently became broad and shear-thinning shifted to ahigher shear rate. The EmimOAC pretreated sample solution, inparticular, exhibited a significant viscosity drop with Newtonianflow behavior. The shear-thinning shifting and the viscosity dropcan be attributed primarily to molecular weight and a restrictedchain effect. It resulted from the decreased molecular weight andthe removal of lignin and hemicellulose as explained above [32].
In order to fundamentally understand the flow behavior of thelignocellulose, the changes of storage (G0) and loss (G00) moduli as afunction of angular frequency for the lignocellulose solutions at30 �C were measured by time-temperature superposition princi-ples and the results are shown in Fig. 5. The data was shifted alongthe vertical axis by factor B to prevent data overlap, with the cor-responding B value of �3, 0 and 3. For the native lignocellulosesolutions, a plateau region and cross-over point between G0 and G00
where the relations of G0 f u2 and G00 f u1 hold, were observed.According to the rubber elasticity theory, this suggests the presenceof an extensive quantity of entangled networks or crosslinking. Thecross-over point at low frequency indicates a slow relaxation of thecellulose chain constricted by entanglement and crosslinking fromneighboring chains and lignin. The Newtonian flow was observedin the EmimOAC pretreated lignocellulose solution. The curve of G00
was larger than that of G0 in the whole u-region where the mea-surements were carried out, indicating that the solutions weremore like viscous liquids and suggesting that the intermolecularinteraction was weakened by the pretreatment.
For all solutions, the DEa value of the EmimOAC-pretreatedlignocellulose solutions was lower than that of EmimCl-pretreated lignocellulose solutions as shown in Table 3. This
Fig. 3. TG-DTG curves of MCC, cotton, hemp and acacia samples pretreated with EmimOAC or EmimCl at 80 �C and 120 �C.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253248
indicates that EmimOAC pretreatment improved the chain mobilitymore efficiently than EmimCl, which may also be explained interms of changes of molecular weight and chemical bonding. Asdescribed above, the cellulose decomposition and the removal ofthe lignin and hemicellulose occurred simultaneously during theionic liquid pretreatment. The hydrolysis of the lignocellulose re-sults in weakening of the entanglement and chemical bondingbetween cellulose and other components. The reduction of theintermolecular interaction by the pretreatment enhances the chainmobility, which may lead to improve the rearrangement andrecrystallization of cellulose during precipitation.
3.4. Microstructure of ionic liquid-pretreated lignocellulose
Fig. 6 shows the X-ray diffraction patterns of the native and thepretreated biomasses. The main peak of four native samples waslocated around 22.5�, which indicates the distance betweenhydrogen bonds in cellulose I. There was a broad peak at ~16�,which is known to be a composite of two peaks, forms Ia (16.7� and14.9�) and Ib (16.8� and 14.3�) [16]. Results after pretreatment withthe ionic liquid showed a significant influence on the crystallinestructure and a strong dependence on the anion types present.With the exception of the acacia samples, the main peak at 22.5�
disappeared and a new asymmetric doublet peak at 20.0� and 21.5�
appeared while the broad peak at ~16� disappeared and a new peakwas detected at 12� after the pretreatment with EmimOAC. Thisindicates that the samples were recrystallized as cellulose II, whichis in accordance with the FT-IR results. For the EmimCl pretreat-ment, all biomasses substrates were recrystallized to the cellulose Ior remained as an amorphous structure. The dependency of thetransition on anion type can be explained by the lignin and hemi-cellulose content during the pretreatment. Pretreatment withEmimCl did not significantly affect the lignin and hemicellulosecontent due to the low solubility [9] as discussed in results of FT-IRand composition analysis. This indicates that the lignocellulosemaintained a microfibril structure with a strong attachment be-tween cellulose and other components (lignin and hemicellulose)described in previous studies [33e35], suggesting the microfibrilformation of a network structure consisting of cellulose, lignin andhemicellulose, which reduced cellulose mobility and hinderedrecrystallization. The interruption of recrystallization by lignin canbe also understood from a consideration of neat cellulose recrys-tallization under mechanical tension. Samayam et al. [16] reportedthat mechanical tension during recrystallization of cellulose hin-dered the transformation to cellulose II although cellulose I couldreadily be converted to cellulose II under low/no tension. In asimilar way, the strong combination of cellulose, lignin and hemi-cellulose retains the crystalline structure of cellulose I despite the
Fig. 4. Steady shear viscosity change of MCC, cotton, hemp and acacia samples pretreated with EmimOAC or EmimCl at 120 �C.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253 249
pretreatment, as shown in Fig. 7. The conversion from cellulose I toII, therefore, requires sufficiently swollen or dissolved state torearrange cellulose molecules. This speculation can be demon-strated additionally in terms of homogenously swollen structurederived from high temperature. The loosen structure by increasingtemperature should allow that the cellulose molecules are easilydiffused and moved in ionic liquid, indicating fast rearrangementfor crystalline transition. The pretreatment at high temperaturehelp the transition from cellulose I to cellulose II.
The XRD pattern responded to the increase of the pretreatmenttemperature by displaying different trends according to the ionicliquid. The cellulose II structure was still observed in the XRDcurves of the cotton and the hemp treated with EmimOAC at highertemperature. However, the EmimCl-pretreated samples showedsmaller and broader peak intensity, indicating that the samplesconsisted almost of amorphous structure. This results from thedifferences in chain mobility as explained in the section dealingwith rheological behavior. Compared to the chloride anion, theacetate anion effectively weakens intermolecular interaction dur-ing the pretreatment and improves the chain mobility, which inturns improves the chance for recrystallization. In addition to this,
the sample pretreated with EmimCl still contained more lignincompared to the sample after treatment with EmimOAC (Table 2).Because the presence of bound lignin restricts range of motion ofthe cellulose chains, lignocellulose with more lignin has less like-lihood of being recrystallized.
A difference of crystalline structure according to lignocellulosetypes was also observed. The crystalline structure of acacia (with alignin content above 20%) was different from the other substrates(with a lignin content below 10%). Although the peak of acaciabecame broader, the peak location still remained at 16� and 22.5�,regardless of the ionic liquid type. This means that the pretreatedacacia still had cellulose I structure rather than cellulose II. Asmentioned in the composition results, lignin interrupts swellingand dissolution of cellulose and hinders the conversion of thecrystalline structure from cellulose I to II.
The crystallinity index of each sample was calculated from theXRD curves (Table 4) [36]. Regardless of the ionic liquid type, thecrystallinity index of the all pretreated samples rapidly decreased,corresponding to the results of the crystalline structure. As thepretreatment temperature increased, the crystallinity displayed atendency that was dependent on the ionic liquid. The crystallinities
Fig. 5. Dependence of dynamic storage modulus, G0 (filled symbols), and loss modulus, G00 (open symbols), on reduced angular frequency for MCC, cotton, hemp and acacia samplespretreated with EmimOAC or EmimCl at 120��C. These curves are shifted vertically to avoid overlap except for samples pretreated with EmimCl.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253250
of MCC, cotton and hemp treated by EmimOAC were slightlyincreased with the higher pretreatment temperature. In contrast tothe results of the EmimOAC pretreatment, the crystalline structureof the cotton and the hemp pretreated with EmimCl almost dis-appeared completely at high temperature. As explained above, theswollen structure after treatment with EmimOAC was sufficient toresult in an improvement of chainmobility which in turn, improvedthe crystallinity.
In the point of lignocellulose type, the different trends of thecrystallinity change were observed according to the pretreatmenttemperature. The crystallinity of the lignocellulose containing highlignin content above 10% was decreased with increasing tempera-ture, whereas the crystalline having small amount of lignin waswell developed. Inwell swollen state of the cellulose, the crystallineconversion readily occurred after structural collapse by the pre-treatment as described above. The transition can be facilitated with
Fig. 6. XRD patterns of MCC, cotton, hemp and acacia samples pretreated with EmimOAC or EmimCl at 80 �C and 120 �C.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253 251
well development of the crystalline at high temperature. In the caseof structure constrained strongly by lignin, the cellulose could bemore collapsed at high temperature, but it is insufficient to convertcrystalline form because external network of lignin and cellulose isstill sustained. This results suggest that the effect of temperature isdifferently applied to the lignocellulose consisting of thick cell wallsuch as acacia.
The difference in the crystallinity behavior could also beexplained in terms of molecular weight. The cellulose pretreatedwith EmimOAC at high temperature underwent more extensivehydrolysis (Fig. 3). The reduction in the molecular weight of cel-lulose decreased the inter chain entanglements (Fig. 5). Thisimproved the chainmobility during recrystallization; therefore, thecellulose had a higher probability of forming cellulose II with highcrystallinity. As a result, the sample from the EmimOAc pretreat-ment contains cellulose II with high crystallinity. Previous studiesshowed that the crystalline structure of regenerated cellulose wasdependent on hydrolysis conditions. Some researchers [37e40]have noted that while cellulose hydrolyzed under harsh conditionhad the possibility to achieve the cellulose II crystalline form, ifmild conditions were used, recrystallization to cellulose II wasmore difficult. For the acacia sample pretreated with both Emi-mOAC and EmimCl, the crystallinity decreased as the temperature
increased. Despite effective dissolution at high temperature, thecellulose chain was still strongly constrained by the remaininglignin and hemicellulose. The combination of cellulose and othercomponents does not permit the cellulose to be rearranged to thecrystalline form, which is in accordance with the results of therheological behavior and activation energy (Table 3 and Fig. 5). Thisresulted in a decrease in crystallinity after pretreatment at hightemperature.
4. Conclusions
The compositional and structural transitions of three lignocel-luloses pretreated with two different ionic liquid were investigated.The type of anion in the ionic liquid significantly influenced on theefficiency of the removal of lignin and hemicellulose. Compared tothe chloride anion, the acetate anion was more efficient in theremoval of lignin and hemicellulose as well as in the extraction ofcellulose from lignocellulose. Concurrently, the thermal stability ofthe lignocelluloses pretreated with EmimOAC was significantlydecreased, indicating the rapid reduction of the molecular weight.Due to the hydrolysis and separation of cellulose, lignin andhemicellulose, the intermolecular interaction of the lignocellulosebecame remarkably faster, accompanying the increase in chain
Fig. 7. Schematic representation of pretreatment and recrystallization of lignocellulosic biomass showing the penetration of ionic liquid into the biomass and dissolution of cel-lulose. Depending on lignin content and ionic liquid type, the cellulose recrystallizes to either uniform cellulose II or a mixed crystalline form including cellulose I and II.
Table 4Crystallinity index (CrI) and crystalline structure of MCC, cotton, hemp and acacia samples pretreated with EmimOAC or EmimCl at 80 �C and 120 �C.a
MCC 0.82 ± 0.01 Cell Ib 0.72 ± 0.01 Cell IIc 0.78 ± 0.02 Cell II 0.67 ± 0.02 Cell II 0.71 ± 0.03 Cell IICotton 0.79 ± 0.01 Cell I 0.59 ± 0.03 Cell II 0.64 ± 0.03 Cell II 0.55 ± 0.04 Cell I ed e
Hemp 0.75 ± 0.02 Cell I 0.62 ± 0.03 Cell II 0.64 ± 0.04 Cell II 0.54 ± 0.05 Cell I e e
Acacia 0.62 ± 0.03 Cell I 0.60 ± 0.05 Cell I 0.35 ± 0.07 Cell I 0.61 ± 0.05 Cell I 0.57 ± 0.05 Cell I
a All measurements are an average of triplicate samples.b Cell I: cellulose I.c Cell II: cellulose II.d The calculation of the crystallinity index in cotton and hemp pretreated with EmimCl at 120 �C was not available due to unclear diffraction patterns.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253252
mobility. The high pretreatment temperature enhanced theremoval efficiency and cellulose hydrolysis, in agreement with thereduction of inter-chain constraints. We propose that cellulose IIforms from fully swelled lignocellulose with EmimOAC whereascellulose I or mixture of cellulose I and II forms from the materialwith the lignocellulose pretreated with EmimCl. The relaxation oflignocellulose microstructure accelerated chain rearrangement,which resulted in an improvement of structural regularity.
Acknowledgements
This work was supported by the National Research Foundationof Korea (NRF) grant funded by the Korean Government (MSIP) (No.2015R1A2A2A01007933).
References
[1] Y. Ahn, D.-H. Hu, J.H. Hong, S.H. Lee, H.J. Kim, H. Kim, Effect of co-solvent onthe spinnability and properties of electrospun cellulose nanofiber, Carbohydr.Polym. 89 (2) (2012) 340e345.
[2] K. David, A.J. Ragauskas, Switchgrass as an energy crop for biofuel production:a review of its ligno-cellulosic chemical properties, Energy Environ. Sci. 3 (9)(2010) 1182e1190.
[3] A. Hendriks, G. Zeeman, Pretreatments to enhance the digestibility of ligno-cellulosic biomass, Bioresour. Technol. 100 (1) (2009) 10e18.
[4] G.W. Huber, S. Iborra, A. Corma, Synthesis of transportation fuels from
biomass: chemistry, catalysts, and engineering, Chem. Rev. 106 (9) (2006)4044e4098.
[5] P. Kumar, D.M. Barrett, M.J. Delwiche, P. Stroeve, Methods for pretreatment oflignocellulosic biomass for efficient hydrolysis and biofuel production, Ind.Eng. Chem. Res. 48 (8) (2009) 3713e3729.
[6] N. Mosier, C. Wyman, B. Dale, R. Elander, Y. Lee, M. Holtzapple, et al., Featuresof promising technologies for pretreatment of lignocellulosic biomass, Bio-resour. Technol. 96 (6) (2005) 673e686.
[7] A.R. Esteghlalian, S.D. Mansfield, J.N. Saddler, Cellulases: agents for FiberModification or Bioconversion? the effect of substrate accessibility on cellu-lose enzymatic hydrolyzability, Prog. Biotechnol. 21 (2002) 21e36.
[8] P. Sannigrahi, D.H. Kim, S. Jung, A. Ragauskas, Pseudo-lignin and pretreatmentchemistry, Energy Environ. Sci. 4 (4) (2011) 1306e1310.
[9] A. Brandt, J. Gr€asvik, J.P. Hallett, T. Welton, Deconstruction of lignocellulosicbiomass with ionic liquids, Green Chem. 15 (3) (2013) 550e583.
[10] L. Laureano-Perez, F. Teymouri, H. Alizadeh, B.E. Dale, Understanding factorsthat limit enzymatic hydrolysis of biomass, Appl. Biochem. Biotechnol. 124(1e3) (2005) 1081e1099.
[11] C. Li, B. Knierim, C. Manisseri, R. Arora, H.V. Scheller, M. Auer, et al., Com-parison of dilute acid and ionic liquid pretreatment of switchgrass: biomassrecalcitrance, delignification and enzymatic saccharification, Bioresour.Technol. 101 (13) (2010) 4900e4906.
[12] Y.H.P. Zhang, L.R. Lynd, Toward an aggregated understanding of enzymatichydrolysis of cellulose: noncomplexed cellulase systems, Biotechnol. Bioeng.88 (7) (2004) 797e824.
H. Kim et al. / Biomass and Bioenergy 93 (2016) 243e253 253
Transition of cellulose crystalline structure and surface morphology ofbiomass as a function of ionic liquid pretreatment and its relation to enzy-matic hydrolysis, Biomacromolecules 12 (4) (2011) 933e941.
[17] M.T. Clough, K. Geyer, P.A. Hunt, J. Mertes, T. Welton, Thermal decompositionof carboxylate ionic liquids: trends and mechanisms, Phys. Chem. Chem. Phys.15 (47) (2013) 20480e20495.
[18] N. Sun, M. Rahman, Y. Qin, M.L. Maxim, H. Rodríguez, R.D. Rogers, Completedissolution and partial delignification of wood in the ionic liquid 1-ethyl-3-methylimidazolium acetate, Green Chem. 11 (5) (2009) 646.
[19] S.H. Lee, T.V. Doherty, R.J. Linhardt, J.S. Dordick, Ionic liquid-mediated selec-tive extraction of lignin from wood leading to enhanced enzymatic cellulosehydrolysis, Biotechnol. Bioeng. 102 (5) (2009) 1368e1376.
[20] A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton and D. Crocker,Determination of Structural Carbohydrates and Lignin in Biomass, NationalRenewable Energy Laboratory, Golden, CO, NREL Technical Report No. NREL/TP-510-42618, 1-13
[22] C.L. McCormick, P.A. Callais, B.H. Hutchinson Jr., Solution studies of cellulose inlithium chloride and N, N-dimethylacetamide, Macromolecules 18 (12) (1985)2394e2401.
[23] M.L. Nelson, R.T. O'Connor, Relation of certain infrared bands to cellulosecrystallinity and crystal lattice type. Part II. A new infrared ratio for estimationof crystallinity in celluloses I and II, J. Appl. Polym. Sci. 8 (3) (1964)1325e1341.
[24] S. Naik, V.V. Goud, P.K. Rout, K. Jacobson, A.K. Dalai, Characterization of Ca-nadian biomass for alternative renewable biofuel, Renew. Energy 35 (8)(2010) 1624e1631.
[25] F. Carrillo, X. Colom, J. Sunol, J. Saurina, Structural FTIR analysis and thermalcharacterisation of lyocell and viscose-type fibres, Eur. Polym. J. 40 (9) (2004)2229e2234.
[26] H. Du, X. Qian, The effects of acetate anion on cellulose dissolution and re-action in imidazolium ionic liquids, Carbohydr. Res. 346 (13) (2011)
1985e1990.[27] H. Wang, G. Gurau, R.D. Rogers, Ionic liquid processing of cellulose, Chem. Soc.
Rev. 41 (4) (2012) 1519e1537.[28] D. Fengel, Ideas on ultrastructural organization of cell-wall components,
J. Polym. Sci. Part C-Polym. Symp. 36 (1971) 383.[29] O.M. Gazit, A. Katz, Dialkylimidazolium ionic liquids hydrolyze cellulose under
mild conditions, ChemSusChem. 5 (8) (2012) 1542e1548.[30] Y. Kang, Y. Ahn, S.H. Lee, J.H. Hong, M.K. Ku, H. Kim, Lignocellulosic nanofiber
prepared by alkali treatment and electrospinning using ionic liquid, FibersPolym. 14 (4) (2013) 530e536.
[31] M. Calahorra, M. Cortazar, J. Eguiazabal, G. Guzm�an, Thermogravimetricanalysis of cellulose: effect of the molecular weight on thermal decomposi-tion, J. Appl. Polym. Sci. 37 (12) (1989) 3305e3314.
[32] N. Tamai, D. Tatsumi, T. Matsumoto, Rheological properties and molecularstructure of tunicate cellulose in LiCl/1, 3-dimethyl-2-imidazolidinone, Bio-macromolecules 5 (2) (2004) 422e432.
[33] M. Lewin, I.S. Goldstein, Wood structure and Composition, M. Dekker, 1991.[34] A. Plaza, P. Montoro, A. Benavides, C. Pizza, S. Piacente, Phenylpropanoid
glycosides from Tynanthus panurensis: characterization and LC-MS quanti-tative analysis, J. Agric. Food Chem. 53 (8) (2005) 2853e2858.
[36] A.D. French, M.S. Cintr�on, Cellulose polymorphy, crystallite size, and the Segalcrystallinity index, Cellulose 20 (2013) 583e588.
[37] Y. Cao, H. Li, Y. Zhang, J. Zhang, J. He, Structure and properties of novel re-generated cellulose films prepared from cornhusk cellulose in room tem-perature ionic liquids, J. Appl. Polym. Sci. 116 (1) (2010) 547e554.
[38] Z.-D. Ding, Z. Chi, W.-X. Gu, S.-M. Gu, J.-H. Liu, H.-J. Wang, Theoretical andexperimental investigation on dissolution and regeneration of cellulose inionic liquid, Carbohydr. Polym. 89 (1) (2012) 7e16.
[39] M. Iguchi, T.M. Aida, M. Watanabe, R.L. Smith Jr., Dissolution and recovery ofcellulose from 1-butyl-3-methylimidazolium chloride in presence of water,Carbohydr. Polym. 92 (1) (2013) 651e658.
[40] H. Zhang, J. Wu, J. Zhang, J. He, 1-Allyl-3-methylimidazolium chloride roomtemperature ionic liquid: a new and powerful nonderivatizing solvent forcellulose, Macromolecules 38 (20) (2005) 8272e8277.
