Softwood kraft lignin for value-added applications: Fractionation and structural characterization

Industrial Crops and Products 66 (2015) 220–228

Contents lists available at ScienceDirect

Industrial Crops and Products journal homepage: www.elsevier.com/locate/indcrop

Softwood kraft lignin for value-added applications: Fractionation and structural characterization Marina Alekhina a,∗ , Olga Ershova a , Andreas Ebert b , Sami Heikkinen c , Herbert Sixta a a b c

Department of Forest Product Technology, Aalto University, School of Chemical Technology, P.O. Box 16300, Vuorimiehentie 1, Espoo FIN-00076, Finland Research Division of Biopolymers, Fraunhofer IAP, Geiselbergstraße 69, 14476 Potsdam-Golm, Germany Laboratory of Organic Chemistry, Department of Chemistry, University of Helsinki, A.I Virtasenaukio 1, P.O. Box 55, Helsinki FIN-00014, Finland

a r t i c l e

i n f o

Article history: Received 4 August 2014 Received in revised form 30 October 2014 Accepted 13 December 2014 Keywords: Black liquor Softwood Kraft lignin Selective fractionation Structural features Chemical properties

a b s t r a c t This study focuses on a systematic, structural characterization of lignin samples fractionated from softwood industrial black liquor (BL). In addition to the isolation efficiency, the impact of fractionation and pulping severity on lignin structure was studied. BL samples from different chemical pulping stages were collected and the lignin was fractionated by sequential precipitation. While the alteration in structure and properties of the samples recovered from different delignification stages was only marginal, the spectroscopic characterization of the isolated lignin revealed significant alteration in its structure and functionalities as a function of the pH. Lignin samples precipitated at pH 10.5 exhibited the highest purity, as indicated by them having the lowest content of polysaccharides. In contrast, the samples precipitated at a pH of 2.5 revealed the highest carbohydrate content, rising from 9.7% at the beginning of cooking, to 36.8% at the end. At the same time, these lignin samples had the lowest Mw and the highest number of phenolic hydroxyl groups. Based on spectroscopic analysis, low Mw kraft lignin displayed an unusually high carboxyl content and a low methoxyl group content. All of the kraft lignin samples showed a reduction in the number of primary and secondary OH groups, which continuously decreased over the course of pulping. An opposite trend was observed with decreasing precipitation pH. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The use of renewable raw materials as a source of chemicals, materials, and energy is expected to increase in the foreseeable future due to the gradual depletion of fossil fuels. In this scenario, the development of new products and materials from lignocellulosic biomass in connection with emerging biorefinery strategies will be promoted. The concept of wood biorefinery comprises the selective separation of the three main polymeric wood components cellulose, hemicellulose, and lignin, and their subsequent utilization for the production of fuels, high value-added chemicals, and other related products. Cellulose and hemicellulose fractions can be readily used for the production of pulp, as a source of sugars for fermentation to produce pure chemicals, or as biofuels after further conversion processes. However, the efficient utilization of lignin presents an ongoing challenge (Ragauskas et al., 2014). Lignin constitutes between 15% and 40% of the dry matter of woody plants. It is primarily a material that adds strength and

∗ Corresponding author. Tel.: +358 505672661. E-mail address: marina.alekhina@aalto.fi (M. Alekhina). http://dx.doi.org/10.1016/j.indcrop.2014.12.021 0926-6690/© 2014 Elsevier B.V. All rights reserved.

structure to the cell walls, controlling fluid flow, and protecting against the enzymatic degradation of other components. Lignin is a complex phenolic polymer built up through the oxidative coupling of C6 C3 (phenylpropane) units; namely, guaiacyl alcohol (G), syringyl alcohol (S), and 4-hydroxy phenyl alcohol (H), which form an irregular structure in a three-dimensional network inside the plant cell wall. The structural building blocks of lignin are joined together by ether linkages (C O C) or carbon carbon bonds (C C), while the major inter-unit linkage is of the ␤-O-4 type. In addition to the 20 different types of bonds present within lignin itself, lignin is also associated with the hemicelluloses, forming the so-called lignin–carbohydrate complexes (LCCs) (Garcia et al., 2009). Technical or industrial lignin is generated in large quantities as a by-product of the chemical pulping of wood. During kraft pulping, about 90–95% of the lignin is chemically degraded to fractions that are soluble in aqueous alkali and thus these fractions form the major constituent of the black liquor (BL). Kraft lignin differs significantly from native lignin in its structure and chemical composition. Structural changes such as cleavage of ␣-O-4 and ␤-O-4 linkages lead to formation of lower molecular weight lignin and its dissolution in spent liquors. Additionally, undesirable impurities such as sulfurous compounds or carbohydrates are present in lignin derived

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

from kraft fractionation process. (Marton, 1971; Rudholm, 1965; Santos et al., 2013; Sixta et al., 2006) Currently, the pulp and paper industry produces an estimated 50 million tons of lignin annually (Higson and Smith, 2011). However, only about 2% of it (1.1 million tons) is commercially utilized for low-value products such as dispersing or binding agents and the rest is used for production of steam, heat and power (Higson and Smith, 2011). Although the energy aspect of lignin utilization is very important for the overall economic balance of the process, the potential of the valuable chemical properties and functionality of lignin has not yet been fully realized. Due to development of emerging lignocellulose biorefinery industry for production of biofuels and chemicals, even larger amounts of lignin will be generated in near future. For example, the U.S. Energy Security and Independence Act of 2007 mandates the development of 79 billion liters of second generation biofuels annually by 2022 (Ragauskas et al., 2014). Therefore, about 62 million tons of lignin will be produced annually in the Unites States alone. Without new product applications, the lignin produced worldwide would far exceed the current market (Ragauskas et al., 2014). Additionally, it should be mentioned that the total lignin availability in the biosphere exceeds 300 billion tons and increases by around 20 billion tons every year (Smolarski, 2012). The structure and properties of lignin differ depending on its origin, pulping conditions, and isolation methods. The severity of the kraft pulping process can influence the structure of dissolved lignin. Baptista et al. (2008) reported that the structure of both the residual and dissolved lignin were influenced by the extent of delignification. In contrast, Robert et al. (1984) did not identify any major changes in the chemical structure of lignin as a function of cooking time, which may be attributed to the difference in the delignification rates. The results of their study tend to indicate that the majority of the chemical reactions during kraft cooking take place in the solid phase, thus resulting in comprehensive changes in the lignin structure. Different methods have been proposed for the fractionation of technical lignin. The three main approaches include fractionation based on solubility in organic solvents (Saito et al., 2014), selective acid precipitation at reduced pH values (Mussatto et al., 2007), and membrane ultrafiltration (Toledano et al., 2010). In addition, the recent implementation of commercial processes—such as LignoBoost (Öhman et al., 2006), or the process for the production of Indulin AT—has demonstrated the successful manufacture of kraft lignin from BL with relatively high purity. With these processes, it will be easier for kraft pulp mills to incorporate the isolation of lignin from the excess thick liquor. Due to its aromatic structure, lignin can be an attractive source of aromatic compounds and can be used as starting material for a series of useful products. Several types of lignin and its derivatives are successfully used in the production of vanillin, phenols, benzene, adhesives, dispersants, emulsifiers, chelants, antioxidants, pesticides, fertilizers, vegetal charcoal, concrete additives, and as a component for resins and thermoplastics (Gargulak and Lebo, 1999). In addition to the competing energy value of lignin, other challenges, including heterogeneity, especially the broad distribution of the molecular weight, and undesirable organic and inorganic impurities (sulfurous compounds and carbohydrates, respectively) associated with technical lignin isolated from BL, limit its utilization in high value-added applications. To further advance the utilization of lignin in new applications or products, better knowledge is required on the sourcing, the separation process, and their impact on the structural and chemical characteristics of the obtained materials. Several studies in the field of lignin separation by acid precipitation have been performed earlier. However, the majority of the studies deal with fractionation of BL derived from pulping of hardwood species and annual plants.

221

Table 1 Physicochemical properties of industrial black liquors derived from different cooking stages. Properties

Initial

Bulk

Residual

Density, [kg/m3 ] pH TDSa , [g/L] TOCb , [g/L] TCc , [g/L] RAd , [g NaOH/L] Lignin, [g/L] Carbohydrates, [g/L] Arabinose, [mg/L] Xylose, [mg/L] Galactose, [mg/L] Glucose, [mg/L] Mannose, [mg/L]

1.128 13.94 218.0 55.0 58.4 28.8 56.3 3.6 1025 1680 755 65 50

1.130 13.86 249.6 69.9 72.7 26.9 69.6 4.0 1395 1975 605 35 30

1.134 13.82 252.8 77.6 80.6 26.3 71.1 3.8 1400 1545 745 50 45

a b c d

TDS – Total dissolved solids. TOC – Total organic carbon. TC – Total carbon. RA – Residual alkali.

Softwoods represent the principal lignocellulosic in northern hemisphere, and therefore, their utilization is of interest. Additionally, the major part of the work published earlier is focusing on isolation of lignin from BL by gradual acid precipitation. In contrast to the earlier research, this work combines fractionation according to pH with collection of BL at different stages of the pine kraft pulping. The effect of BL fractionation on the lignin yield, structure and purity was studied. The fractionation was coupled with comprehensive characterization of obtained fractions using advanced analytical techniques. Fractionated kraft lignin samples were quantified and comparatively characterized in terms of yield, chemical composition, molecular weight, and functional groups. 2. Material and methods 2.1. Materials Samples of industrial pine kraft BL were supplied by Metsä Fiber, from its Rauma (Finland) pulp mill. This pulp mill operates according to the Superbatch cooking process, in which the temperature is ramped from 90 ◦ C (impregnation time, 58 min) to 162 ◦ C over a period of 52 min and then held at this temperature for 59 min. The kappa number of the produced pulp was 29.1. BL samples were collected at different stages of delignification: initial (H-factor = 185), bulk (H-factor = 550) and residual delignification (H-factor = 670) phases. The original characteristics of BL are shown in Table 1. Density was determined by measuring the weight of the BL in a known volume. Total dissolved solids (TDS) were determined by using a method based on TAPPI T264 cm-97. The lignin concentration was measured using a UV–vis spectrometer (UV-2550, Shimadzu) at 280 nm, an absorptivity constant of 24.6 g/L was used (Fengel et al., 1981). The carbohydrate content was determined according to NREL/TP-510-42618. Total organic carbon (TOC) and total carbon (TC) were measured in the BL using a TOC-VCPH analyzer (Shimadzu) according to SFS-ISO 8245. The residual alkali content of BL was determined with an automatic titrator according to SCAN-N 33:94. All BL characteristics are averages of at least three measurements. Pine wood chips for dissolved wood lignin (DWL) preparation were supplied by Metsä Fiber, Finland. Pine DWL was prepared according to (Fasching et al., 2008) and used as a reference showing properties close to those of native lignin. Deuterated chloroform (CDCl3 ) containing 0.03% tetramethylsilane (TMS) (used as internal standard) and Chromium(III) acetylacetonate (Cr(acac)3 ) (relaxation agent) were purchased from (St. Louis,

222

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

Fig. 1. Scheme for fractionation of softwood kraft lignin.

MO, USA) Sigma–Aldrich. Fine chemicals were purchased from Sigma–Aldrich or Merck. 3. Methods 3.1. Black liquor fractionation Black liqour was fractionated by sequential acid precipitation at pH levels of 10.5, 5 and 2.5, respectively (Fig. 1). The solubilized lignin was precipitated by drop-wise addition of 6 M H2 SO4 under stirring at ambient temperature until target pH. After storage for 30 min, the precipitates were separated by centrifugation and subsequently acidified to pH 2.5. In order to remove the residual acid and the formed salts, the precipitates were washed with deionized water in a centrifuge until the pH of the supernatant was around 5. The fractions precipitated at pH 2.5 were subjected to washing without the acidification step. Subsequently, the precipitates were freeze-dried. The yields of precipitated fractions were determined gravimetrically and expressed as gram of precipitate in liter of black liquor [g/L BL] or as % of total lignin content in black liquor [% on total lignin in BL]. Fractionation of each BL was carried out in duplicate and the yield values are averages of at least two measurements.

and Suprema 100, 20 ␮m, 8 mm I.D.*300 mm) and one pre-column (Suprema 20 ␮m). The columns, injector and UV detector were maintained at 80 ◦ C during the analysis. MWD analysis was carried out at least in duplicate. Nuclear magnetic resonance (NMR) experiments were recorded with Varian Unity Inova 600 NMR-spectrometer (600 MHz 1 Hfrequency) equipped with 5 mm broadband probehead at 27 ◦ C. NMR-samples were prepared in 5 mm NMR-tube using CDCl3 as solvent with 0.03% TMS. All lignin samples (300 mg) were acetylated with 6 mL of pyridine–acetic anhydride (1:1) before analysis (Lin and Dence, 1992). Sample concentrations of 150 mg/mL were used. For quantitative 13 C experiments, inverse gated 1 Hdecoupling and 30 degree excitation pulse flip angle were utilized. Spectral width was 36,182.7 Hz, relaxation delay was 5 s, and acquisition time was 0.2 s. Number of transients varied between 41,215 and 49,683. The free induction decays (FID) were apodized using exponential multiplication with 10 Hz line broadening and zero filled up to 16,384 complex points prior to Fourier transformation. For 13 C experiments, the samples were doped with relaxation agent Cr(acac)3 to concentration of 10 mM. 4. Results and discussion

3.2. Analyses of the black liquor fractions

4.1. Black liquor composition

Acid-insoluble lignin or Klason lignin, acid-soluble lignin (ASL) and carbohydrate contents in the isolated lignin samples were determined after two-stage acid hydrolysis according to NREL/TP-510-42623. Subsequent analysis of the recovered neutral sugar monomers was performed by using a Dionex ICS 3000 high-performance anion exchange chromatograph with pulsed amperometric detection (HPAEC–PAD) equipped with a CarboPacPA20 column (Dionex, Sunnyvale, CA, USA). Water was used as the eluent at a flow rate of 0.4 mL/min at 30◦ C. A minimum of two parallel determinations were performed with an additional determination if a noticeable differences between the two measurements were observed. The infrared (IR) spectra of the lignin samples were recorded by means of the attenuated total reflectance (ATR) technique (Bruker Tensor 27 FTIR-spectrometer, Specac single reflection diamond ATR “Golden gate”). Reported spectra represent baseline corrected average spectra from two measurements, straight line from 2450–1850 cm−1 (absorption range of diamond, necessary for baseline correction). The spectral data was normalized to the band of ca. 1600 cm−1 . The molecular weight distribution (MWD) was determined by gel-permeation chromatography (GPC) equipped with UV detection (UV–vis Detector 2487). The column was eluted with dimethyl sulfoxide (DMSO) with 0.1% LiBr at a flow rate of 1 mL/min. The GPC system consisted of two analytical columns (Suprema 1000

The kinetics of lignin removal during kraft pulping is mainly described as three parallel first-order reactions characterized by different delignification rate constants. About 20% of the lignin in wood is removed during the initial phase, where delignification is controlled by diffusion (Sjöström, 1993). The selectivity of the initial delignification is very poor since up to 40% of hemicelluloses are concomitantly degraded and dissolved during this phase. In contrast, the bulk phase is characterized by a selective delignification where up to about 90% of the original lignin present in wood is removed while only minor amounts of polysaccharides are degraded. In the final residual-delignification phase, the delignification rate is slower and is accompanied by an increased degradation of polysaccharides (Sjöström, 1993). Due to the different dissolution rates during the cooking process, the properties of lignin derived from different delignification stages are expected to vary. The physicochemical properties of the industrial BLs used in this work are summarized in Table 1. As expected, the lignin content increased with an increasing H-factor due to the cleavage of the ␣- and ␤-aryl ether bonds of lignin, followed by its fragmentation and dissolution. In agreement with published data (Sjöström, 1993), over 90% of the total lignin in the BL was already dissolved during the initial and bulk cooking stages, and only a small additional amount was further dissolved during the residual period. The total dissolved polysaccharides in the BL samples varied from

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

3.6 to 4.0 g/L. The level of dissolved carbohydrates reaches a maximum during the bulk delignification stage. The slight decrease in the total carbohydrate content from the bulk to the residual delignification stages is mostly attributed to a decrease in the amount of dissolved xylan (Table 1). The galactose content was relatively high in the BL collected during the initial delignification stage (Table 1), indicating the extensive dissolution and degradation of galactoglucomannan (GGM). GGM is the most abundant hemicellulose in softwoods; the ratio of galactose:glucose:mannose in native softwood GGM is about 1:1:3, and the ␣-d-galactopyranose residue is linked as a single-unit side chain to the glucomannan backbone (Sjöström, 1993). Interestingly, only minor amounts of glucose and mannose were found in the BL (Table 1), indicating the fast degradation of GGM molecules into monomers, and then their further decomposition into degradation products such as hydroxy acids. This could be explained by mannan’s low tendency for the stopping reaction. This is in agreement with data published by (Paananen et al., 2013), who postulated that GGM is mainly subjected to peeling, and that more than 70% of the GGM present in wood already dissolves in the BL during the heating up period. Therefore, it is present in the BL in the form of hydroxyl acids, mainly as isosaccharinic acid. The high galactose content in the BL may result from the possible presence of lignin–carbohydrate covalent associations. The galactose side chain was found to be one of the most commonly proposed linker sugar units of LCCs (Lawoko et al., 2006). It is assumed that a majority of galactose residue is cleaved off from the main chain of GGM at the beginning of the cooking process. Subsequently, as already mentioned above, glucomannan molecules are subjected to degradation. However, galactose residue might be still chemically bonded to the lignin via LCC, thus it is carried on to the BL and further to the precipitates. Additionally, it was shown that pectic galactans are easily dissolved during kraft cooking, and thus the galactan–lignin complexes become enriched in the BL (Tamminen et al., 1995). In contrast, pine xylan is relatively stable toward alkaline peeling when compared to GGM due to the stabilizing effect of the arabinose side groups. Nevertheless, a decrease in xylan content with increasing cooking time was observed (Table 1) as an indication of the intensive degradation (peeling) and fragmentation (retroaldol) reactions leading to the production of various hydroxy carboxylic acids. A similar observation was described by Paananen et al. (2013) where prolonged cooking times for conventional pine kraft pulping resulted in a decrease in the amount of xylan dissolved in the BL, indicating an increase in the competitive degradation pathways of the dissolved xylan. Additionally, the decrease in xylan content in the BL during the final cooking stage may be associated with a partial re-adsorption of xylan onto the pulp fibers due to a decreased alkalinity. The measured data on the lignin and carbohydrates dissolved in the BL are in a good agreement with those previously reported for softwood (Niemelä et al., 2007). 4.2. Yield and composition of lignin fractions The analysis of the precipitated dissolved matter after the acidification showed that the amount of precipitate increased with the time of cooking (Table 2). Overall, approximately 79–93% of the lignin present in the BL could be recovered by sequential precipitation. The present lignin yield results were similar to those obtained previously for straw soda black liquors (Garcia et al., 2009), where as much as 80% of lignin was precipitated when the pH of the BL was decreased to 2. When considering all of the delignification stages, the majority of the lignin was precipitated at pH 5, comprising 60–72% of the total precipitated matter (or 59–69% as pure lignin) of the total lignin in the BL. Together with the lignin precipitated at pH 10.5, 74–89% of the total lignin was recovered. The subsequent

223

decrease in the pH to 2.5 resulted in a further precipitation of 4 and 5% of the total lignin. Concurrently with the decreasing pH from 10.5 to 2.5, the color of the processed BL changed from dark brown to light yellow. The dark color of the BL is due to the chromophoric functional groups formed during the lignin degradation, such as quinones, carbonyl groups, phenolic hydroxyl groups, etc., (Fengel and Wegener, 1989). A similar observation has been reported by (Garcia et al., 2009; Mussatto et al., 2007), when the pH of the BLs derived from the alkaline cooking of wheat straw and of brewer’s spent grain was decreased. Table 3 also shows that the proportion of ASL-to-Klason lignin substantially increased along with a decrease in the precipitating pH. Yasuda et al. (2001) studied the structure and formation mechanism of ASL and concluded that ASL is composed of low-molecular-weight degradation products and hydrophilic derivatives of lignin. Therefore, the proportion of low molecular weight lignin continuously increased with decreasing precipitation pH, while at the same time, the hydrophilicity of the lignin increased. The ASL content was especially high for lignin samples precipitated at pH 2.5. Additionally, the ASL content in the recovered lignin fraction slightly increased with rising cooking severity, which may be attributed to a progressive depolymerization of the lignin with increased cooking time. All of the lignin samples were washed with deionized water to remove any undesired contaminants. As can be observed (Table 2), the ash content was relatively low for all samples, from 0.1 to 0.4%, indicating efficient washing of the samples. In contrast, all of the isolated kraft lignins were contaminated with carbohydrates, ranging from 2.1 to 36.8% of the oven-dry amount of precipitate. Amount of hemicellulose contaminates increased as a function of precipitation pH. The lignin fractions obtained at pH 10.5 contained the lowest amounts of polysaccharides, with galactose as the major sugar unit (Table 3). It may be assumed that the carbohydrates detected in the samples precipitated at pH 10.5 are associated with this lignin fraction as LCCs (Lawoko et al., 2006). In contrast, the lignin fractions precipitated at pH 2.5 contained the highest amounts of carbohydrates, which can probably be attributed to both LCCs and non-bounded sugars. The highest contamination rate in the fraction recovered by acidic precipitation is related to the lower solubility of the hemicelluloses in acidic media and their subsequent precipitation with decreasing precipitation pH. These results differ from those reported by Santos et al. (2014), who suggested that an increase precipitation pH increased the proportion of hemicellulose that co-precipitated with the lignin. However, it should be noted that those authors used a gradient acid precipitation, in contrast to the sequential acid fractionation implemented in this study. Therefore, the different isolation methods may explain the differences in lignin purity. Additionally, the black liquor derived from Eucalyptus pulping was used in study performed by (Santos et al., 2014). It is well know that softwoods BL differ from hardwoods BL in terms of carbohydrate composition, hemicelluloses structure, lignin content and structure and Mw of lignin and carbohydrates. This is probably the main reason behind the discrepancy in the findings. Additionally, the carbohydrate content increased significantly with increasing cooking severity. This effect was observed particularly for samples precipitated at pH 2.5. The amount of xylan (xylose) in the pH 2.5-precipitated lignin isolated from the initial, bulk, and residual phases of kraft cooking increased from 8.0% to 13.8%, and to 31.7% (based on the oven-dry precipitates), respectively. In all cases, the xylan (xylose) content comprised more than 80 wt% of the total carbohydrates. This was also reported for kraft Eucalyptus globulus lignin and soda brewer’s spent grain, where xylan was found to be the main polysaccharide linked to the lignin (Fernández-Costas et al., 2014; Sun and Tomkinson, 2001).

224

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

Table 2 Yield and chemical composition of different lignin fractions precipitated from black liquors (BLs). Namea

pHb

Cooking stage

Acid con.c

Yield of precipitated fractions

[g/L BL]

Yield as pure lignin [% on lignin in BL]

[% on lignin in BL]

Chemical composition of precipitated fractions

Klason lignin ASL [% of o.d. sample]

Carb.d

Ash

I 10.5 I5 I 2.5

Initial Initial Initial

10.5 5 2.5

0.46 0.40 0.20

8.9 34.0 3.1

15.8 60.4 5.5

15.6 58.5 4.9

97.37 92.33 80.43

1.36 4.6 8.8

2.12 2.18 9.69

0.24 0.29 0.28

Sum B 10.5 B5 B 2.5

Bulk Bulk Bulk

10.5 5 2.5

1.06 0.46 0.37 0.18

46 13.1 44.7 4.3

81.7 18.8 64.2 6.2

79.1 18.3 60.8 5.0

– 95.31 89.64 69.75

– 2.04 4.99 10.98

– 2.10 3.87 16.49

– 0.19 0.34 0.28

Sum R 10.5 R5 R 2.5

Residual Residual Residual

10.5 5 2.5

1.01 0.44 0.38 0.19

62.1 14.3 51.4 4.4

89.2 20.1 72.3 6.2

84.1 19.6 69.4 3.6

– 94.97 90.36 46.90

– 2.34 5.69 10.78

– 3.33 3.37 36.80

– 0.21 0.37 0.26

1.01

70.1 –

98.6 –

92.6 –

– 85.40

– 6.80

– 5.24

– 0.00

Sum DWL a b c d

Name – sample name. pH – precipitation pH. Acid con. – acid consumption [mol H2 SO4 /L BL]. Carb. – carbohydrates.

Table 3 Content of residual carbohydrates in lignin fractions in % with respect to the quantity of lignins. Sample

Xylose, %

Galactose, %

Arabinose, %

Glucose, %

Mannose, %

Total sugars, %

I 10.5 I5 I 2.5 B 10.5 B5 B 2.5 R 10.5 R5 R 2.5 DWL

0.39 1.45 7.98 0.80 2.87 13.75 0.92 2.39 31.70 0.87

1.42 0.26 0.41 0.87 0.26 0.51 1.79 0.30 0.54 0.85

0.25 0.40 1.21 0.36 0.65 2.11 0.57 0.61 4.43 n.d.

0.05 0.06 0.10 0.07 0.09 0.12 0.05 0.07 0.13 1.79

n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 1.73

2.12 2.18 9.69 2.10 3.87 16.49 3.33 3.37 36.80 5.24

As already mentioned above, softwood xylan is relatively stable under alkaline pulping conditions; however, its structure undergoes a substantial alteration, such as through the extensive cleavage of 4-O-MeGlcA and the arabinose side groups, thus, yielding a linear xylan polysaccharide that is essentially free from side groups (Alekhina et al., 2014). Even though all of the lignin samples were washed extensively, the polymeric form of xylan and its linear structure probably caused its insolubility in water. It is possible that the relatively high molecular weight fraction of xylan re-absorbed onto the fiber surface in the final cooking stage, while the fraction of low molecular weight xylan remained dissolved in the liquors. Therefore, lignin samples precipitated at a higher pH were nearly free of xylan contaminants, while lignin precipitated at pH 2.5 contained up to 32% of xylan.

of the lignin fractions, the Mw , number-average molecular weight (Mn ), and polydispersity (PD) were calculated (Table 4). The Mw of lignin might be closely related to the number of ␣and ␤-aryl ether bonds and C C bonds between the structural units present in native lignin (Sjöström, 1993). A significant number of lignin fragments dissolved in the BL during the initial and bulk phases of delignification due to the degradation of lignin’s ␣-O-4 and ␤-O-4 bonds, as demonstrated by the relatively low molecular weight of these fractions. The significantly higher Mw of the lignin obtained from residual delignification could be explained by the formation of stable covalent C C bonds, or so-called condensed structures. Additionally, selective dissolution of lignin molecules

Table 4 Weight-average (Mw ) and number-average (Mn ) molecular weights and polydispersity (PD) of the lignin fractions.

4.3. Molecular weight distribution (MWD) GPC was carried out to obtain the MWD of the lignin fractions. The results clearly indicated that efficient fractionation of the lignin can be achieved by sequential acidification. There was a clear correlation between the molecular weight and the pH of precipitation; the weight-average molecular weight (Mw ) decreased proportionally to the decreasing pH levels that were adjusted for precipitation (Fig. 2). Similarly, the increase in the Mw of lignin with increasing precipitation pH was reported by Santos et al. (2014) and Wang and Chen (2013) for hardwood kraft lignin. The dissimilarity between the samples that originated from the different delignification stages was only marginal. In order to obtain a quantitative comparison

Sample name

Mn a [g/mol]

Mw b [g/mol]

PDc

I 10.5 I5 I 2.5 B 10.5 B5 B 2.5 R 10.5 R5 R 2.5 DWL Pine

1508 1072 788 1347 1008 787 1471 1087 798 2106

6111 3099 2349 5542 3048 2663 8092 3474 3537 6098

4.1 2.9 3.0 4.1 3.0 3.4 5.5 3.2 4.4 2.9

a b c

Mw molecular weight average. Mn number average. PD = Mw/Mn, polydispersity.

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

225

Fig. 2. Molecular weight distribution of isolated lignin fractions.

that occurs during the delignification process can lead to a fractionation process, where more accessible low molecular weight lignin fragments are dissolved in the first stages of cooking. However, it should be noted that the relatively high Mw of I-10.5 and I-5 could be related to recirculation of the BL in industry (the BL from a previous cook used for chip impregnation may contain high molecular weight lignin). The highest Mw was found for lignin precipitated at pH 10.5 from the sample collected from the residual stage of the cook, which resulted in a Mw = 8100 g/mol with a PD of 5.5. Both the molecular weight and the PD of the lignin decreased substantially when precipitated at pH 5. Although the lignin samples precipitated at pH 2.5 had a similar Mw to those of samples precipitated at pH 5, the somewhat higher PD of the samples precipitated at pH 2.5 indicated that both the higher Mw (condensed lignin) and lower Mw (ASL) fractions were present. This was particularly pronounced for the lignin isolated from the residual phase of delignification, where, possibly due to the high amount of condensed lignin, the Mw of R-2.5 was even higher than that from the pH 5-precipitated lignin. 4.4. ATR-IR spectroscopy For further investigation of the heterogeneity of the kraft lignin structure derived from different cooking stages, the obtained lignin samples were analyzed by ATR-IR spectroscopy, and compared with the DWL that was representative of native lignin (Fig. 3). The absorption bands in the ATR-IR spectra can be readily assigned to both lignin and polysaccharides, providing information on the structure of lignin—on its purity—including the composition of the polysaccharide components. All assignments are based on previously published data (Faix, 1992). A comparison between the kraft lignin spectra and the spectrum of DWL revealed differences that would indicate an alteration of the lignin structure occurring during the cooking processes. Additionally, the ATR-IR spectra of the lignin fractions obtained from different cooking stages, but precipitated at a similar pH, were quite comparable (Fig. 3b and d), suggesting that the general structure of lignin dissolved in the BL was not affected by the duration of the pulping. All of the studied samples showed a strong and broad absorption band centered at 3400 cm−1 that originated from the O H stretching vibration in the phenolic and aliphatic hydroxyl groups (Fig. 3a and b). The intensity of this band increased with decreasing precipitation pH, demonstrating a higher content of hydroxyl groups in the lignin that was precipitated at a lower pH. A higher hydroxyl content is an indication of a fragmented lignin structure and is a result of ␣- and ␤-O-4 linkage cleavages during kraft pulping. These results were in agreement with MWD data showing the lowest molecular weight of lignin being precipitated at pH 2.5 (Table 4).

The bands at 2935 cm−1 and 2850 cm−1 represent the symmetrical and asymmetrical CH stretching in the methyl and methylene groups, respectively. The band representing the asymmetric deformation of C H stretching also appears at 1458 cm−1 . These bands were less intensive for kraft lignin samples than that of DWL, suggesting a decrease in the number of aliphatic methylene and methyl groups during kraft pulping. Additionally, a reduction in the methyl and methylene group contents was observed with decreasing precipitation pH, probably due to aliphatic chain shortening as a result of C and Cˇ elimination. As indicated in the magnified region of the spectra (Fig. 3c and d), the phenylpropane units (lignin skeleton) were identified in all of the isolated samples, showing the bands that are typical for aromatic skeletal vibrations, assigned at 1510 and 1422 cm−1 . A decrease in the intensity of these bands was observed with decreasing precipitation pH and this was probably related to the presence of high amounts of carbohydrates in samples precipitated at lower pH. The aromatic C H deformation at 1030 cm−1 appears as a complex vibration associated with the C O, C C stretching and C OH bending in carbohydrates (Boeriu et al., 2004). Polysaccharide originating vibrations are associated also with other vibrations in the spectral region 1000–1300 cm−1 . For instance, B-2.5 showed a lower intensity of bands typically associated with lignin and a high intensity of carbohydrate signals when compared to the spectra of other samples. These results were in agreement with and were supported by those obtained through composition analyses, which also indicated the lower purity of B-2.5 (shown in Table 3). A discernible difference was found in the absorption intensity at 1710 cm−1 , corresponding to the C O in the unconjugated ketones, carbonyl, and ester groups stretching. A significant increase in the intensity of this band was observed with decreasing precipitation pH, suggesting an increasing tendency for carbonyl/carboxyl groups with decreasing molecular weight, most likely caused by oxidation during alkaline pulping. However, the increased absorption intensity at 1710 cm−1 for samples precipitated at a lower pH could also be partially related to the carboxyl groups in hemicellulose that are present in lignin as an impurity. In contrast, the DWL spectrum was characterized by a strong absorption centered at 1725 cm−1 , indicating partial acetylation of DWL during the purification step involving 75% acetic acid. The region below 1300 cm−1 is more complex to analyze due to a signal overlap between lignin and the carbohydrates; however, the typical bands associated with the lignin structure could still be properly examined. Signals assigned to guaiacyl (G) units in the lignin were detected, specifically the bands at 1265 cm−1 and 1220 cm−1 , indicating G ring and C O stretch, and vibration at 1153 cm−1 associated with aromatic C H in-plane deformation. The intensity of the bands at 853 and 817 cm−1 that originated from C H out-of-plane vibrations in positions 2, 5, and 6 of the G units were quite similar for kraft lignin samples from different

226

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

Fig. 3. (a and c) ATR-IR spectra of isolated lignin fractions obtained from the pine kraft black liquor derived from bulk delignification stage and precipitated at various pH values. (a) Wave number from 4000 to 600 cm−1 and (c) magnified region of IR spectra (700–1900 cm−1 ). (b and d) ATR-IR spectra of isolated lignin fractions obtained from different delignification stages of kraft pine pulping and precipitated at pH 10.5. (b) Wave number from 4000 to 600 cm−1 and (d) magnified region of IR spectra (700–1900 cm−1 ).

delignification stages. The band at 1030 cm−1 appeared as a complex vibration associated with the C H aromatic deformation in the guaiacyl structure, and the C O bending in the aliphatic alcohols was significantly more intense in DWL than in kraft lignin. Additionally, the intensity of this peak increased dramatically with decreasing precipitation pH. This may be due to a higher amount of aromatic C H in lignin precipitated at a lower pH, with a higher number of un-substituted positions in the ring, and consequently, a lower degree of condensation (DC). In addition, the somewhat high intensity of this band for B-2.5 might be due to overlapping signals that originated from the C O in polysaccharides. 4.5. NMR spectroscopy The chemical structure of selected lignin fractions was investigated using 13 C NMR spectroscopy (Fig. 4). The assignment of the resonances is based on work reported by (Capanema et al., 2004; Lin and Dence, 1992), and our own HSQC spectra (not shown). All of the integrations were performed by setting the aromatic region (106–155.5 ppm) to 6.12, assuming that it includes six aromatic carbons and 0.12 vinylic units per C9 unit (Robert and Chen, 1989). The integral values for other structural moieties were calculated in relation to the aromatic ring (Ar). The results are shown in Table 5, including the corresponding values for a spruce DWL. All of the studied lignin samples showed resonances between 90

and 102 ppm attributed to carbohydrates (Fig. 4). The bands grow proportionally with decreasing precipitation pH, confirming the highest carbohydrate content in the lignin sample being precipitated at pH 2.5. The results of the 13 C NMR indicated a significant decrease in the methoxyl group content of all kraft lignin samples when compared to DWL as a result of demethylation/demethoxylation caused by the nucleophilic attack by hydrosulfide ions on the methyl aryl ether bonds. The demethylation/demethoxylation of kraft lignin was found to be particularly high at the beginning of the cook, as indicated by 0.68 OMe per Ar. As delignification progressed, the extent of demethylation/demethoxylation decreased, as demonstrated by an increase in the methoxyl group content. Additionally, it seems that kraft lignin precipitated at a lower pH, and therefore, lignin with a low molecular weight had a significantly lower methoxyl group content than that of high molecular weight lignin. The aromatic region of the 13 C spectrum can be divided into three main regions representing the oxygenated aromatic (155.5–140.5 ppm), the condensed aromatic (140.5–124.3 ppm), and the protonated aromatic (124.3–106 ppm) regions. In theory, the oxygenated aromatic region consists primarily of etherified aromatic carbons involved in ␣- and ␤-O-4 linkages and methoxylated carbons, and should contain around 2 atoms per Ar (atom C3 and C4 ). A rapid reduction in CAr-O was observed with decreasing precipitation pH, indicating that the aryl ethers present in the

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

227

Fig. 4. 13 C NMR spectra of dissolved wood lignin (DWL) and lignin precipitated at different pH (left, all lignin samples isolated from bulk delignification stage at pH 10.5 (B 10.5), 5 (B 5) and 2.5 (B 2.5)) and isolated from different cooking stages (right, all lignin samples precipitated at pH 10.5 from initial (I 10.5), bulk (B 10.5) and residual (R 10.5) delignification stages).

lignin structure are highly degraded. Further, the results reveal that demethoxylation is likely occurring via a loss of formaldehyde rather than through demethylation, as a reduction in the OMe content was observed in connection with a decrease in the amount of CAr-O . In general, there was no considerable difference in the number of aromatic Ar-C bonds between the studied kraft lignins, showing only a marginal increase in the CAr-C with an increasing degree of delignification, indicating a higher DC. This is in agreement with the results obtained from ATR-IR and MWD. A somewhat higher Ar-C content was found for samples precipitated at lower pH. The relatively high content of Ar-C bonds in B-5 and B-2.5 may originate from hydroxymethyl groups in position C5 formed during kraft pulping as a result of a reaction with formaldehyde (Marton, 1971). The significantly lower molecular weight of the B-2.5 sample originates from more fragmented lignin molecules due to the cleavage of aryl ether bonds, as indicated by the increased Ph-OH content of this sample. Moreover, the condensation reaction involving the aliphatic side chain, and the consequent formation of new linkages such as ␤-5, ␣-5, ␣-1, and ˛–˛ of lignin samples precipitated at a higher pH may be responsible for the high Mw of these samples. As discussed previously (Holtman et al., 2006), the oxygenated and condensed aromatic regions may have some degree of overlap, leading to a less precise calculation of the DC. In contrast, the protonated region does not suffer from any overlap and the DC can be estimated more precisely. A significant increase in the number of protonated aromatic carbons per aromatic ring has been noted to occur with decreasing precipitation pH. A higher amount of aromatic C H is evidence of a higher number of free positions in the ring, and, therefore, a lower DC. The DC was calculated by subtracting the integration of CAr-H from 3, based on the assumption that uncondensed softwood lignin would have 3 protons at C2 , C5 , and C6 . Thus, the DC of kraft lignin G rings decreased with decreasing

precipitation pH. In addition, the DC was rather constant throughout the cook, as indicated by only a slight decrease in the number of aromatic protonated carbons with increasing cooking severity. Concurrently, as kraft cooking progresses, the number of primary OH groups decreased substantially for all kraft lignin samples as compared to that of pine DWL. The primary hydroxyl groups are assumed to correspond to the number of ␥-hydroxymethyl groups. The decrease in primary and secondary hydroxyl groups showed that the lignin that was already dissolved in the early stage of the kraft cook underwent an extensive elimination of the terminal hydroxymethyl groups in the form of formaldehyde, followed by the fragmentation and progressive dissolution of the propyl side chains. The number of primary and secondary aliphatic hydroxyl groups decreased further with an increase in the degree of delignification. This is also supported by the decrease in aliphatic carbon content obtained by integrating the aliphatic region and subtracting the CDCl3 signal. This is in agreement with the work of Rutkowska et al. (2009), where a continuous decrease in the primary and secondary hydroxyl groups in the course of pulping of Eucalyptus globulus was reported. Additionally, the results demonstrated an increased primary aliphatic OH content for the lignin precipitated at a lower pH. This can be explained by the formation of alkali-stable enol aryl ether structures as a result of the cleavage of ␥-C. All of the kraft lignin samples showed a distinct reduction in the number of secondary OH groups, and they continuously decreased over the course of pulping, thus indicating dehydration at the ␣␤ carbon atoms, and the subsequent formation of unsaturated and/or condensed structures. However, the secondary hydroxyl group content was found to be significantly higher for samples precipitated at low pH, probably due to the introduction of new OH groups at the C˛ and Cˇ atoms as a result of the formation of arylglycerol-, diol-, or thioglycol-type structures.

Table 5 Relative abundance of main functional groups and linkages present in the examined lignin samples obtained by quantitative 13 C NMR measurements. Sample name

Aromatic methoxyl

Oxygenated aromatic C (C O)

Condensed aromatic C (C C)

Protonated aromatic C (C H)

Region ı (ppm) I 10.5 B 10.5 R 10.5 B 10.5 B5 B 2.5b DWL

58.0 − 53.7 0.68 0.73 0.79 0.73 0.72 0.51 0.94

155.5 − 140.5 1.49 1.50 1.47 1.50 1.41 1.27 1.69

140.5 − 124.3 2.41 2.45 2.49 2.45 2.50 2.48 1.9

124.3 − 106.0 2.20 2.17 2.15 2.17 2.21 2.37 2.53

a b

DCa

0.80 0.83 0.85 0.83 0.79 0.63 0.47

Degree of condensation. Value might be overestimated because carbohydrates signals may interfere with the lignin moieties

Primary aliphatic OH

Secondary aliphatic OH

Phenolic OH

171.7 − 170.2 0.37 0.34 0.31 0.34 0.35 0.48 0.71

170.2 − 169.1 0.23 0.19 0.17 0.19 0.22 0.83 0.42

169.1 − 167.0 0.42 0.48 0.60 0.48 0.64 0.71 0.21

228

M. Alekhina et al. / Industrial Crops and Products 66 (2015) 220–228

Only 0.21 per Ar-free Ph-OH was found in DWL, since most of PhOH are occupied in linkages to neighboring phenylpropane units. In contrast, the frequency of free phenolic hydroxyl groups was rather high in all kraft lignin samples. A substantial increase in phenolic hydroxyl group content was observed with decreasing precipitation pH and increasing cooking severity due to the formation of phenolic OH groups as a result of the extensive cleavage of the ␣and ␤-ether bonds. 5. Conclusions Selective acid precipitation was shown to be an efficient, simple and inexpensive method to separate and fractionate lignin from softwood black liquor. Our results indicated that up to 85% of the lignin solubilized in softwood black liquor can be recovered in high purity by lowering the pH to 5. A further pH reduction leads to increased precipitation of carbohydrates. Extended cooking resulted in increased molecular weight of the precipitated lignins; a consequence of progressive condensation reactions. The increase in Mw was especially pronounced for the pH 10.5 and pH 2.5 precipitated lignins, respectively. The demethylation of kraft lignin was found to be particularly high in the beginning of cook. Additionally, it seems that kraft lignin with a low molecular weight had significantly high phenolic and aliphatic hydroxyl group and low methoxyl group contents. Based on the results, the pulping stage and precipitation acidity should be carefully selected to tailor the lignin properties to those required for the final application. Acknowledgements The financial support for this work by TEKES and the Forest Cluster as part of the Wood Wisdom project is gratefully appreciated. The authors acknowledge Dr Kari Kovasin and Metsä Fiber for supplying the BL samples. We thank Hannes Moosbauer (Lenzing AG) for providing assistance with ATR-IR analysis and Dr Hendrik Wetzel (Fraunhofer IAP) for his help with the GPC analysis. We also thank Dr Miro Suchy, Dr Marc Borrega and Dr Agnes Stepan for helpful discussions. References Alekhina, M., Mikkonen, K.S., Alén, R., Tenkanen, M., Sixta, H., 2014. Carboxymethylation of alkali extracted xylan for preparation of bio-based packaging films. Carbohydr. Polym. 100, 89–96. Baptista, C., Robert, D., Duarte, A.P., 2008. Relationship between lignin structure and delignification degree in Pinus pinaster kraft pulps. Bioresour. Technol. 99, 2349–2356. Boeriu, C.G., Bravo, D., Gosselink, R.J.A., van Dam, J.E.G., 2004. Characterisation of structure-dependent functional properties of lignin with infrared spectroscopy. Ind. Crops Prod. 20, 205–218. Capanema, E.A., Balakshin, M.Y., Kadla, J.F., 2004. A comprehensive approach for quantitative lignin characterization by NMR spectroscopy. J. Agric. Food Chem. 52, 1850–1860. Faix, O., 1992. Fourier transform infrared spectroscopy. In: Lin, S., Dence, C. (Eds.), Methods in Lignin Chemistry. Springer, Berlin Heidelberg, pp. 233–241. Fasching, M., Schröder, P., Wollboldt, R.P., Weber Hedda, K., Sixta, H., 2008. A New and Facile Method for Isolation of Lignin from Wood Based on Complete Wood Dissolution. Holzforschung, pp. 15. Fengel, D., Wegener, G., 1989. Wood: Chemistry, Ultrastructure, Reactions. Walter De Gruyter. Fengel, D., Wegener, G., Feckl, J., 1981. Characterization of Analytical and Technical Lignins 2. Physicochemical and Electron Microscopical Studies, 35. Holzforschung, pp. 111–118.

Fernández-Costas, C., Gouveia, S., Sanromán, M.A., Moldes, D., 2014. Structural characterization of kraft lignins from different spent cooking liquors by 1D and 2D nuclear magnetic resonance spectroscopy. Biomass Bioenergy 63, 156–166. Garcia, A., Toledano, A., Serrano, L., Egüés, I., González, M., Marín, F., Labidi, J., 2009. Characterization of lignins obtained by selective precipitation. Sep. Purif. Technol. 68, 193–198. Gargulak, J.D., Lebo, S.E., 1999. Commercial Use of Lignin-Based Materials, Lignin: Historical, Biological, and Materials Perspectives. American Chemical Society, pp. 304–320. Higson, A., Smith, C., 2011. Renewable Chemicals Factsheet: Lignin. NNFCC http://www.nnfcc.co.uk/publications/nnfcc-renewable-chemicals-factsheet-lignin Holtman, K.M., Chang m, H., Jameel, H., Kadla, J.F., 2006. Quantitative 13C NMR characterization of milled wood lignins isolated by different milling techniques. J. Wood Chem. Technol. 26, 21–34. Lawoko, M., Henriksson, G., Gellerstedt, G., 2006. Characterisation of Lignin-carbohydrate Complexes (lccs) of Spruce Wood (Picea abies L.) Isolated with Two Methods, 60. Holzforschung, pp. 156. Lin, S.Y., Dence, C.W., 1992. Methods in Lignin Chemistry. Springer-Verlag. Marton, J., 1971. Reactions in alkaline pulping. In: Sarkanen, K.V., Ludwig, C.H. (Eds.), Lignin: Occurrence, Formation, Strcuture and Recations. John Wiley & Sons, USA, pp. 639–694. Mussatto, S.I., Fernandes, M., Roberto, I.C., 2007. Lignin recovery from brewer’s spent grain black liquor. Carbohydr. Polym. 70, 218–223. Niemelä, K., Tamminen, T., Ohra-aho, T., 2007. Black Liquor Components as Potential Raw Materials. 14th ISWFPC, Durban, South Africa. Paananen, M., Liitiä, T., Sixta, H., 2013. Further insight into carbohydrate degradation and dissolution behavior during kraft cooking under elevated alkalinity without and in the presence of anthraquinone. Ind. Eng. Chem. Res. 52, 12777–12784. Ragauskas, A.J., Beckham, G.T., Biddy, M.J., Chandra, R., Chen, F., Davis, M.F., Davison, B.H., Dixon, R.A., Gilna, P., Keller, M., Langan, P., Naskar, A.K., Saddler, J.N., Tschaplinski, T.J., Tuskan, G.A., Wyman, C.E., 2014. Lignin valorization: improving lignin processing in the biorefinery. Science 344 (6185), http://dx.doi.org/10.1126/science.1246843, Art. no. 1246843. Robert, D., Chen, C.-L., 1989. Biodegradation of Lignin in Spruce Wood by Phanerochaete chrysosporium: Quantitative Analysis of Biodegraded Spruce Lignins by 13C NMR Spectroscopy., 43. Holzforschun, pp. 323. Robert, D.R., Bardet, M., Gellerstedt, G., Lindfors, E.L., 1984. Structural changes in lignin during kraft cooking part 3. On the structure of dissolved lignins. J. Wood Chem. Technol. 4, 239–263. Rudholm, S.A., 1965. Chemical Pulping, Pulping Processes. John Wiley & Sons, New York, pp. 439–714. Rutkowska, E.W., Wollboldt, R.P., Zuckerstätter, G., Weber Hedda, K., Sixta, H., 2009. Characterisation of structural changes in lignin during continuous batch kraft cooking of Eucalyptus Globulus. Bioresources 4, 172–193. Saito, T., Perkins, J.H., Vautard, F., Meyer, H.M., Messman, J.M., Tolnai, B., Naskar, A.K., 2014. Methanol fractionation of softwood kraft lignin: impact on the lignin properties. ChemSusChem 7, 221–228. Santos, P.S.B.d., Erdocia, X., Gatto, D.A., Labidi, J., 2014. Characterisation of kraft lignin separated by gradient acid precipitation. Ind. Crops Prod. 55, 149–154. Santos, R.B., Jameel, H., Chang, H.-m., Hart, P.W., 2013. Impact of lignin and carbohydrate chemical structures on degradation reactions during hardwood kraft pulping process. Bioresources 8, 158–171. Sixta, H., Potthast, A., Krotschek, A.W., 2006. Chemical pulping processes. In: Sixta, H. (Ed.), Handbook of Pulp. WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, pp. 109–510. Sjöström, E., 1993. Wood Chemistry: Fundamentals and Applications. Academic Press. Smolarski, N., 2012. High-value Opportunities for Lignin: Unlocking its Potential. Frost & Sullivan, Paris, pp. 15. Sun, R., Tomkinson, J., 2001. Fractional separation and physico-chemical analysis of lignins from the black liquor of oil palm trunk fibre pulping. Sep. Purif. Technol. 24, 529–539. Tamminen, T., Vuorinen, T., Tenkanen, M., Hausalo, T., Hortling, B., 1995. Analysis of lignin and lignin-carbohydrate complexes isolated from black liquor. In: The 8th International Symposium on Wood and Pulping Chemistry, Helsinki, Finland, pp. 297–302. Toledano, A., Serrano, L., Garcia, A., Mondragon, I., Labidi, J., 2010. Comparative study of lignin fractionation by ultrafiltration and selective precipitation. Chem. Eng. J. 157, 93–99. Wang, G., Chen, H., 2013. Fractionation and characterization of lignin from steam-exploded corn stalk by sequential dissolution in ethanol–water solvent. Sep. Purif. Technol. 120, 402–409. Yasuda, S., Fukushima, K., Kakehi, A., 2001. Formation and chemical structures of acid-soluble lignin I: sulfuric acid treatment time and acid-soluble lignin content of hardwood. J. Wood Sci. 47, 69–72. Öhman, F., Theliander, H., Tomani, P., Axegard, P., 2006 Method for separating lignin from black liquor.