Industrial Crops and Products 23 (2006) 180–193 Comparative study of organosolv lignins from wheat straw Feng Xu a , Jin-Xia Sun b,c , RunCang Sun c,a,∗ , Paul Fowler c , Mark S. Baird d b a State Key Laboratory of Pulp & Paper Engineering, South China University of Technology, Guangzhou 510641, China College of Forestry, The North-Western University of Agricultural and Forest Sciences and Technology, Yangling 712100, China c The BioComposites Centre, University of Wales, Bangor LL57 2UW, UK d Department of Chemistry, University of Wales, Bangor LL57 2UW, UK Received 10 March 2004; accepted 19 May 2005 Abstract Dewaxed wheat straw was treated with acetic acid–H2 O (65/35, v/v), acetic acid–H2 O (80/20, v/v), acetic acid–H2 O (90/10, v/v), formic acid–acetic acid–H2 O (20/60/20, v/v/v), formic acid–acetic acid–H2 O (30/60/10, v/v/v), methanol–H2 O (60/40, v/v) and ethanol–H2 O (60/40, v/v) using 0.1% HCl as a catalyst at 85 ◦ C for 4 h, in which 78.2, 80.0, 88.2, 89.4, 94.1, 23.5 and 37.4% of the original lignin, and 42.4, 58.7, 70.0, 65.1, 76.5, 14.2 and 22.2% of the original hemicelluloses was released, respectively. Lignins obtained were characterized by their content of hemicelluloses, composition of phenolic acids and aldehydes, molecular weight, thermal stability and by UV, Fourier transform infrared (FT-IR), 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy. The results showed that aqueous organic acid was more effective than aqueous organic alcohol for extensive delignification and selective fractionation of cellulose, lignin and hemicelluloses from the straw. In particular, the addition of formic acid gave a significant effect on the dissolution of lignin. All the acid-insoluble lignin fractions contained small amounts of contaminated hemicelluloses as shown by their content of neutral sugars, 0.9–4.3%, and had weight-average molecular weight between 3960 and 4340 g mol−1 . An increase in concentration of acetic acid or formic acid in organosolv resulted in an increment in release of guaiacyl units and in lignin condensation. However, the lignin preparations released during the treatment with aqueous organic alcohol without organic acid contained almost equal amounts of non-condensed guaiacyl and syringyl units with fewer p-hydroxyphenyl units. The ␤-O-4 ether bonds together with ␤-␤, ␤-5 and 5-5′ carbon–carbon linkages were identified to be present in lignin substructures. © 2005 Elsevier B.V. All rights reserved. Keywords: Wheat straw; Lignin; Organic acids; Phenolics; FT-IR; 13 C NMR 1. Introduction ∗ Corresponding author. Tel.: +44 1248 370588; fax: +44 1248 370594. E-mail address: bcs00a@bangor.ac.uk (R. Sun). 0926-6690/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.indcrop.2005.05.008 Large amounts of lignin are produced every year by the pulping processes as a by-product of wood and non-wood delignification. However, the availability of F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 lignin is not as high as could be expected from this fact (Oliet et al., 2001). The commercial pulping processes (kraft and sulphite technologies) lead to high quality pulps, but fractions such as lignin and hemicelluloses (accounting for 50–55% of the dry weight of wood and straw) are employed in low added value applications such as producing process energy (Vila et al., 2003). Organosolv pulping, based on the utilization of organic solvents as delignification agents, provides an interesting alternative to the current commercial technologies, since they lead to a solid phase enriched in cellulose and to liquors containing hemicellulose-degradation products and lignin-degradation products free from sulphur. In organosolv pulping, a mixture of organic solvent and water is used as cooking liquor. The solvent primarily acts on the promotion of vegetal tissue impregnation and the solubilization of the lignin fragments so produced (Balogh et al., 1992; Gilarranz et al., 2000). In non-catalyzed pulping (autocatalyzed), the cooking liquor becomes acidified due to the acetic acid released from the wood. However, in catalyzed pulping the liquor can be acidic, neutral or alkaline depending on the nature of the additives employed (Aziz and Sarkanen, 1989). During organosolv acid delignification, the Acetosolv process (based on the utilization of HCl-catalyzed acetic acid media) and Formacell process (formic acid-catalyzed media) have proved to be promising process to achieve complete utilization of lignocellulosics without impact to environment. Both processes have ability to cause extensive removal of both lignin and hemicelluloses under mild conditions, with no significant cellulose degradation. By either of these processes (Nimz and Casten, 1986; Sano et al., 1990; Pan and Sano, 2000), wood and non-wood can be simply fractionated to pulp, lignin and monosaccharides or hemicellulosic-degradation products, which makes it easy to utilize them for more valuable products. The pulp can be used for either paper or cellulose derivatives, such as carboxymethylcellulose, cellophane, viscose or cellulose acetate. The lignin can be converted to valuable products, such as carbon fibre (Uraki et al., 1995), activated carbon fibres (Uraki et al., 1997) and adhesives (Pan and Sano, 1998). From the monosaccharides or hemicellulosic-degradation products, sugars or other chemicals, sweetening materials, food additives, fuel and polymers can be obtained (Fengel and Wegener, 1984). 181 Lignin is an extremely complex three-dimensional polymer (typically found in vascular plants) formed by dehydrogenative polymerization of phydroxycinnamyl, coniferyl and sinapyl alcohols. These three lignin precursors (‘monolignols’) give rise to the so-called p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S) phenylpropanoid units, which show different abundances in lignins from different groups of vascular plants, as well as in different plant tissues and cell-wall layers. During polymerization of the above p-hydroxycinnamyl alcohols involved, the formation of aryl ether (involving C4 ) interunit linkages is strongly favoured. In addition, a small proportion of lignin units remains as phenolic, being linked only by C–C bonds, such as ␤-5, ␤-1, ␤-5, ␤-␤ and ␣-␤ linkages. Although this phenolic moiety represents a low (and variable) fraction of the total lignin, it can strongly affect the reactivity of the polymer (Camarero et al., 1999). During organosolv acid delignification, lignin is dissolved essentially by acid-catalyzed cleavage of such bonds as ␣-aryl ether and arylglycerol-␤-aryl ether in the lignin macromolecule (Sarkanen, 1990). However, the cleavage of ␤-aryl ether bonds occurs at lower extent (Goyal et al., 1992). The cleavage of ether bonds gives rise to new phenolic hydroxyl groups in lignin, which affect some industrial uses of lignins and lignocellulosic materials, since they increase lignin solubility (favouring its alkaline extraction during paper or pulp manufacture), and modifies the reactivity of technical lignins to be used as raw material for manufacture of lignin-based adhesives and other applications (Camarero et al., 1999). The aim of this work was to study the influence of various organosolvs on physico-chemical properties of lignins dissolved from wheat straw at atmospheric pressure. Two favourable organic acids were chosen to be used, formic acid and acetic acid. In comparison, aqueous methanol and ethanol were also used as cooking solvents. In order to get more information on chemical structures and relationship between physical properties and chemical structures, the lignins were chemically characterized in this paper by their fractional yield, sugar composition, content of phenolic acids and aldehydes, and molecular weights. Other techniques such as Fourier transform infrared (FT-IR) and 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy were also used to investigate the changes occurring 182 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 in lignin structure during the organosolv treatment processes. Wheat straw (Variety Riband) was kindly supplied by B Lloyd Co., Llangefni. The composition (%, w/w) of the straw is cellulose 39.0%, hemicelluloses 38.7%, lignin 17.0%, ash 1.8% and wax 1.9% on a dry weight basis. The deviations of these contents from their respective means were all less than 8%. After being dried at 60 ◦ C in an oven for 16 h, the straw was ground to pass through a 0.7 mm screen and then extracted with toluene–ethanol (2/1, v/v) in a Soxhlet extractor for 6 h. The dewaxed straw was further dried at 60 ◦ C for 12 h. All chemicals used were of analytical or reagent grade. The hemicelluloses (hemicellulosic degradation products) were isolated from the concentrated hydrolysates by precipitation with three volumes of 95% ethanol (22 ◦ C, 12 h). The solubilized lignins were obtained from the corresponding supernatants by reprecipitation after evaporation of all the organic solvents. The acid-insoluble lignin fractions were then washed with acidified water and freeze-dried. Note that the treatments with acetic acid–H2 O (65/35, v/v) was for lignin preparation L1 , acetic acid–H2 O (80/20, v/v) for L2 , acetic acid–H2 O (90/10, v/v) for L3 , formic acid–acetic acid–H2 O (20/60/20, v/v/v) for L4 , formic acid–acetic acid–H2 O (30/60/10, v/v/v) for L5 , methanol–H2 O (60/40, v/v) for L6 and ethanol–H2 O (60/40, v/v) for L7 , respectively. The scheme for organosolv treatment and the isolation of lignin is illustrated in Fig. 1. All experiments were performed at least in duplicate. Yield of lignin is given on a dry weight basis related to the starting material. 2.2. Organosolv treatment 2.3. Characterization of the lignin preparations Organosolv treatment was carried out in a 500 ml glass reactor at atmospheric pressure. The extractive free powder (10.0 g) was treated under seven different organosolv systems (Table 1) using 0.1% HCl as a catalyst at 85 ◦ C for 4 h with a liquor to solid ratio of 20:1 (ml/g) under stirring, respectively. After the reactor was loaded with wheat straw and the cooking liquor, it was heated to the operating temperature, which was then maintained throughout the experiment. After treatment, the residue enriching in cellulose was filtrated on a nylon cloth, then washed with hot acid–water mixture and finally, washed with hot distilled water. The neutral sugar composition of the contaminated hemicelluloses in lignins was determined as their alditol-acetate derivatives by gas chromatography (GC) after hydrolysis with 2 M trifluoroacetic acid for 2 h at 120 ◦ C (Blakeney et al., 1983). The chemical composition of phenolics liberated from alkaline nitrobenzene oxidation of the lignins (175 ◦ C, 2.5 h) was determined on a Hichrom H5ODS HPLC column of dimensions 250 mm × 4.6 mm (Phenomenex Co., Beijing). The identification of the individual compounds was detected at 280 nm by computer comparison of the retention times and peak areas with the authentic 2. Materials and methods 2.1. Materials Table 1 Extraction conditions of organosolv-soluble lignins Experiment no. Organosolv system Catalyst Temperature (◦ C) Liquor to solid ratio (ml/g) Reaction time (h) W1 W2 W3 W4 Acetic acid–H2 O (65/35, v/v) Acetic acid–H2 O (80/20, v/v) Acetic acid–H2 O (90/10, v/v) Formic acid–acetic acid–H2 O (20/60/20, v/v/v) Formic acid–acetic acid–H2 O (30/60/10, v/v/v) Methanol–H2 O (60/40, v/v) Ethanol–H2 O (60/40, v/v) 0.1% HCl 0.1% HCl 0.1% HCl 0.1% HCl 85 85 85 85 20:1 20:1 20:1 20:1 4 4 4 4 0.1% HCl 85 20:1 4 0.1% HCl 0.1% HCl 85 85 20:1 20:1 4 4 W5 W6 W7 183 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 300 spectrometer at 300 and 74.5 MHz, respectively. 1 H NMR spectrum was recorded at 25 ◦ C from 25 mg of sample dissolved in 1.0 ml DMSO-d6 . For each sample, 16 scans were collected. The 13 C NMR spectrum was recorded at 25 ◦ C from 250 mg of sample dissolved in 1.0 ml DMSO-d6 after 30,000 scans. A 70◦ pulse flipping angle, a 10 ␮s pulse width and a 15 s delay time between scans were used. 3. Results and discussion 3.1. Fractional yield and purity of lignin In comparison to other organosolv processes under neutral and alkaline conditions, the recovery of Acetosolv lignin in this study does not need a previous step of reduction, and consequently the further neutralization of liquors is also avoided. The acid-insoluble lignin fractions were therefore obtained from the supernatants by reprecipitation in water solution after evaporation of all the organic solvents. Table 2 gives the fractional yields of the lignins dissolved in various organosolvs. Obviously, the results indicated that the higher yield of lignin was obtained when the treatment was performed using organic acid as a solvent between lignin preparations L1 and L5 ; in this case, the yield of lignin was 13.3–16.0%, corresponding to 78.2–94.1% on total lignin of straw. Interestingly, an increase in acetic acid concentration from 65 to 80% and to 90% resulted in an increment in lignin yield from 13.3% (in L1 ) to 13.6% (in L2 ) and to 15.0% (in L3 ), respectively. The current results were consistent with the studies on acetic acid Fig. 1. Scheme for isolation of acid-insoluble lignin preparations obtained by treatment of dewaxed wheat straw with organosolv under acidic conditions. phenolics. Methods of UV spectra recording, measurement of the molecular-average weights and thermal analysis of lignin samples have been described in previous papers (Sun et al., 1999, 2003). All nitrobenzene oxidation results represent the mean of at least triplicate samples and each oxidation mixture was chromatographed twice. Other experiments were performed in duplicate. The standard errors or deviations were observed to be lower than 6.5% except for the variation among the triplicate nitrobenzene oxidation (7.0–16.8%). FT-IR spectra were obtained on an FT-IR spectrophotometer (Nicolet 750) using a KBr disc containing 1% finely ground samples. The solution-state 1 H and 13 C NMR spectra were obtained on a Bruker MSLTable 2 The yield (% dry matter) of the lignin preparations Lignin fraction Lignin preparationa L1 L2 L3 L4 L5 L6 L7 Total solubilized lignins Acid-insoluble ligninsb Acid-soluble ligninsc Lignin associated in solubilized hemicelluloses 13.3 9.0 3.8 0.5 13.6 9.1 3.7 0.8 15.0 11.6 2.4 1.0 15.2 11.7 2.5 1.0 16.0 12.5 2.0 1.5 4.0 3.0 0.8 0.2 6.4 4.8 1.4 0.2 a L , L , L , L , L , L and L represent the lignin preparations obtained by treatment of the dewaxed wheat straw with acetic acid–H O 1 2 3 4 5 6 7 2 (65/35, v/v), acetic acid–H2 O (80/20, v/v), acetic acid–H2 O (90/10, v/v), formic acid–acetic acid–H2 O (20/60/20, v/v/v), formic acid–acetic acid–H2 O (30/60/10, v/v/v), methanol–H2 O (60/40, v/v) and ethanol–H2 O (60/40, v/v) and 0.1% HCl as a catalyst at 85 ◦ C for 4 h, respectively. b Represent the lignin fractions obtained by precipitation of the supernatant solution at pH 1.5 after isolation of the solubilized hemicelluloses. c Represent the lignin fractions which are still solubilized in the pH 1.5 supernatant after precipitation of the acid-insoluble lignin fractions and obtained by difference (total solubilized lignin–acid-insoluble lignin–lignin associated in the solubilized hemicelluloses). 184 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 pulping of wheat straw under atmospheric pressure by Pan and Sano (1999). The authors stated that acetic acid had been proved to be an effective organosolv for both extensive delignification of wood or non-wood and selective fractionation of cellulose, lignin and hemicelluloses. Similar results have been reported by Nimz and Casten (1986) and Davis et al. (1986) during the delignification of lignocellulosics using acetic acid–water mixture. More importantly, as the data shown in Table 2, the highest yield of lignin was obtained when formic acid was added into acetic acid. Treatment with formic acid–acetic acid–H2 O (20/60/20 and 30/60/10, v/v/v) led to a release of 88.2% (in L4 ) and 94.1% (in L5 ) of the original lignin from the straw, respectively. This indicated that the addition of formic acid showed a significant effect on the dissolution of lignin. In this case, the yield of lignin increased by 1.9 and 2.7% when formic acid increased from 0 to 20% and to 30%. Based on atmospheric acetic acid pulping of wood, Nimz and Schoene, 1993 concluded that when 5–10% of formic acid was added into acetic acid, the pulps could be obtained with improved qualities and kappa number below 5. In other words, these processes achieved a ‘fractionation’ of the lignocellulosic raw materials into separate streams containing hemicellulose-degradation products, lignin-degradation products and cellulose, all of them being utilizable for different end-product applications. It should be noted that the yield of lignin obtained by using methanol–H2 O (60/40, v/v) and ethanol–H2 O (60/40, v/v) together with 0.1% HCl as a catalyst at 85 ◦ C for 4 h, was less than half the yield of organic acid isolated lignin, 4.0% in L6 and 6.4% in L7 . This confirmed again that organic acid, e.g. formic acid and acetic acid, is more powerful solvent for delignification than aqueous alcohol, such as methanol/ethanol–water mixtures. The fairly low lignin yields of L6 and L7 can partially be explained by the relatively low hydroxonium ion concentration in aqueous alcohol system, since lignin dissolution is expected to be proceeded by the acid-catalyzed cleavage of ␣-aryl and ␤-aryl ether linkages in the lignin macromolecule and the rate of delignification is highly pH-dependent. This effective role of organic acid at high concentration is probably due to rapid hydrolysis of hemicelluloses, resulting in increased porosity and accessibility of the solvent to lignin, since significant amounts of hemi- celluloses (42.4, 58.7, 70.0, 65.1 and 76.5% of the original hemicelluloses obtained by treatment with acetic acid–H2 O (65/35, v/v), acetic acid–H2 O (80/20, v/v), acetic acid–H2 O (90/10, v/v), formic acid–acetic acid–H2 O (20/60/20, v/v/v) and formic acid–acetic acid–H2 O (30/60/10, v/v/v), respectively, data not shown) were released or degraded during the treatments using organic acids, while only small quantities of them (14.2 and 22.2% of the original hemicelluloses, data not shown) dissolved during the treatments with aqueous methanol and ethanol, respectively. Furthermore, as expected, the acid-insoluble lignin fraction in Table 2 was the major fraction, comprising 67.7–78.1% of the total solubilized lignins, while the lignin fraction associated in the solubilized hemicelluloses, was accounted only 3.1–9.4% of the released lignins. This result suggested that treatment with various organosolvs under the acidic conditions given substantially cleaved the ether bonds between lignin and hemicelluloses from the cell walls of wheat straw. In this study, UV–vis absorption measurements of the seven acid-insoluble lignin fractions were performed using a dioxane–water mixture, which verify the purity of lignins at λ = 250–380 nm. Spectra of acid-insoluble lignin fractions L2 (spectrum L2 ), L3 (spectrum L3 ), L4 (spectrum L4 ) and L6 (spectrum L6 ) solubilized during the treatment with acetic acid–H2 O (80/20, v/v), acetic acid–H2 O (90/10, v/v), formic acid–acetic acid–H2 O (20/60/20, v/v/v) and methanol–H2 O (60/40, v/v), respectively, are illustrated in Fig. 2. The maximum absorption at 280 nm originates from non-conjugated phenolic groups in the lignin. The presence of a second characteristic region of lignin absorption around 318 nm can be assigned to the presence of both ferulic and p-coumaric acids (Scalbert et al., 1986). Interestingly, as shown in the spectra, a highest absorption coefficient occurred in L6 fraction, suggesting that the most pure lignin preparation can be obtained when aqueous alcohol was used as an organosolv. On the other hand, the lowest absorption coefficient of L4 fraction, released during the treatment with formic acid–acetic acid–H2 O (20/60/20, v/v/v), was undoubtedly due to the highest amounts of bound hemicelluloses and other non-lignin materials. It can be seen that the purity of the lignin decreased in the order: aqueous alcohol-soluble lignin, acetic acid–watersoluble lignin, formic acid–acetic acid–water-soluble lignin. 185 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 Fig. 2. UV spectra of acid-insoluble lignin preparations isolated with acetic acid–H2 O (80/20, v/v, sample L2 , spectrum L2 ), acetic acid–H2 O (90/10, v/v, sample L3 , spectrum L3 ), formic acid–acetic acid–H2 O (20/60/20, v/v/v, sample L4 , spectrum L4 ) and methanol–H2 O (60/40, v/v, sample L6 , spectrum L6 ) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. Organosolv treatment under the acidic conditions used resulted in intensive hydrolysis of lignin– hemicellulose bonds and, accordingly, in low hemicellulose content in the isolated lignins. Table 3 shows the content of sugars determined after acid hydrolysis of acid-insoluble lignin fractions. Even under mild organic alcohol reaction conditions, the total hemicellulose content never exceeded 5%. The relatively high amount of xylose and arabinose suggested that the main hemicelluloses bonded to lignin were arabinoxylans. Glucose and galactose could also be found in all lignin samples. Trace amount of mannose was present in some cases. No clear effect of treatment conditions on the amount of hemicelluloses in lignin preparations was observed except for the lignin preparation L4 , which had a noticeable amount of hemicelluloses as determined by the sugar content, 4.3%. This high- est content of hemicelluloses in L4 preparation implied that there were more chemical linkages between lignin and polysaccharides in L4 than in other lignin preparations, which were stable to aqueous formic acid–acetic acid hydrolysis. p-Coumaric and ferulic acids present in wheat straw lignin might be favourable to this kind of linkages. In particular, ferulic acid might form crosslinks between lignin by ether bonds through its phenolic oxygen in the cell walls of wheat straw and hemicelluloses by simultaneous esterification of their carboxyl group to the C-5 position of arabinose substituents of arabinoglucuronoxylans (Himmelsbach and Barton, 1980). The presence of linkages between hemicelluloses and lignin makes it difficult to prepare wheat straw lignin free from polysaccharides. Similar results were reported for purification of wheat straw lignins by BenGhedalia and Yosef (1994). Table 3 The content of neutral sugars (% lignin sample, w/w) in the acid-insoluble lignin preparations Neutral sugars/uronic acids Lignin preparationa L1 L2 L3 L4 L5 L6 L7 Arabinose Xylose Mannose Galactose Glucose 0.6 1.0 Trb 0.2 0.2 0.4 0.8 Tr 0.2 0.4 0.4 0.8 Tr 0.3 0.4 0.9 2.1 Tr 0.6 0.7 0.6 1.1 Tr 0.4 0.6 0.3 0.3 Tr 0.1 0.2 0.3 0.4 Tr 0.1 0.2 Total 2.0 1.8 1.9 4.3 2.8 0.9 1.0 a Corresponding to the lignin preparations in Table 2. 186 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 increment in release of guaiacyl units. This means that the fraction with more guaiacyl units were much easier to extract than that with more syringyl units during the preparation of organosolv lignins with relatively high concentration of acetic acid or formic acid, or that guaiacyl units were less condensed or cross-linked than syringyl units in the cell walls of wheat straw. On the other hand, these data also suggested that syringyl units were condensed to a great extent during the treatment with the concentration (90%) of aqueous acetic acid and formic acid–acetic acid–water mixture than other units. The total yield of oxidation products decreased from 38.69% (L2 ) to 24.68% (L3 ) with an increase in acetic acid concentration from 80 to 90% and reduced from 23.89% (L4 ) to 17.97% (L5 ) when the formic acid increased from 20 to 30% in the treating mixture of formic acid–acetic acid–water, implying that condensation degree of the lignins increased as the concentration of acetic acid or formic acid raised. In other words, the lignin preparations extracted with a relatively high concentration of acetic acid or formic acid were highly condensed and difficult to oxidize while the lignin preparation dissolved during the treatment with 60% aqueous methanol (L6 ) was least condensed and easiest to oxidize, due to its highest yield of nitrobenzene oxidation products (50.68%). These different degrees of lignin condensation demonstrated 3.2. Composition of phenolic acids and aldehydes The results of alkaline nitrobenzene oxidation of seven organosolv lignin preparations are listed in Table 4. As can be seen, all the lignin preparations gave vanillin and syringaldehyde as main products accompanied with their corresponding aromatic acids in minor quantity, suggesting that the organosolv lignins mainly resulted from guaiacyl–syringyl units. The presence of fewer p-hydroxybenaldehyde and p-hydroxybenzoic acid was considered most probably to the indicative of non-condensed p-hydroxyphenyl units, indicating the incorporation of p-hydroxycinnamoyl alcohol in wheat straw organosolv lignin. The occurrence of significant amounts of non-condensed guaiacyl and syringyl units with relatively fewer p-hydroxyphenyl units revealed that the seven organosolv lignins could be justified as GSH-lignin such as grass type lignin. The relative molar ratios of G (the relative total moles of vanillin, vanillic acid and acetovanillin) to S (the relative total moles of syringaldehyde, syringic acid and acetosyringone) and to H (the relative total moles of p-hydroxybenzaldehyde and p-hydroxybenzoic acid) were found to be 4:3:1 in L1 and L2 , 7:4:1 in L3 , 4:4:1 in L4 , 5:4:1 in L5 , 4:4:1 in L6 and 6:6:1 in L7 . Apparently, an increase in acetic acid or formic acid concentration in a mixture of organosolv resulted in an Table 4 The composition (% lignin sample, w/w) of phenolic acids and aldehydes from nitrobenzene oxidation of the acid-insoluble lignin preparations Phenolic acids and aldehydes Lignin preparationa L1 L2 L3 L4 L5 L6 L7 p-Hydroxybenzoic acid p-Hydroxybenzaldehyde Vanillic acid Vanillin Syringic acid Syringaldehyde Acetovanillin Acetosyringone p-Coumaric acid Ferulic acid Cinnamic acid 1.49 1.68 1.36 12.86 0.63 11.38 0.53 1.00 0.38 0.47 0.030 1.50 2.36 2.81 15.89 1.27 11.37 0.54 1.88 0.40 0.64 0.026 0.25 1.03 0.68 9.68 0.32 9.65 0.38 0.69 0.35 0.33 0.041 0.77 1.34 1.02 9.20 0.89 8.93 0.36 0.72 0.24 0.30 0.12 0.57 0.78 0.72 7.22 0.46 7.17 0.23 0.40 0.18 0.16 0.078 1.34 2.72 2.52 17.18 0.82 21.46 1.65 1.60 0.72 0.58 0.092 0.40 1.58 2.02 11.78 1.22 14.26 0.80 0.76 0.56 0.28 0.080 Total 31.81 38.69 24.68 23.89 17.97 50.68 33.74 4:3:1 4:3:1 7:4:1 4:4:1 5:4:1 4:4:1 6:6:1 Molar ratio a (G:S:H)b Corresponding to the lignin fractions in Table 2. G represents the sum of total moles of vanillin, vanillic acid and acetovanillin; S represents the sum of total moles of syringaldehyde, syringic acid and acetosyringone; H represents the sum of total moles of p-hydroxybenzaldehyde and p-hydroxybenzoic acid. b 187 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 the different behaviour of lignins during the various organosolv treatments. In addition, the low yields of L3 , L4 and L5 could also be explained either as the removal of p-coumaric and ferulic acids ester and/or ether-linked to protolignin, which exist in a remarkable amount in gramineae lignin (Sun et al., 2000), or as the occurrence of condensation between the units of the lignin during the concentration (90%) of aqueous acetic acid and formic acid–acetic acid–water mixture in the atmospheric organic acid treatment with 0.1% HCl as a catalyst. 3.3. Molecular weight Weight-average (M̄w ) and number-average (M̄n ) molecular weights and polydispersity (M̄w /M̄n ) of the seven acid-insoluble lignin preparations from wheat straw are summarized in Table 5. Clearly, the seven lignin fractions showed no significant difference in their molecular-average weights, which ranged M̄w from 3960 to 4340 g mol−1 . These data indicated that the organosolv under the conditions given had no substantial influences on the M̄w of the lignins from wheat straw. In other words, aqueous organic acids and aqueous alcohols appeared to have equal effect on the degradation of lignin into small segments by cleavage of ␣-aryl and ␤-aryl ether linkages when 0.1% HCl was used as a catalyst. In addition, the seven lignin fractions also gave a fairly analogous polydispersity, ranging between 1.56 and 1.71. 3.4. FT-IR spectra Among the analysis techniques described in the literature, FT-IR spectroscopy shows interesting characteristics such as high sensitivity and selectivity, high signal-to-noise ratio, accuracy, data handling facility, mechanical simplicity and short time and small Fig. 3. FT-IR spectra of acid-insoluble lignin preparations isolated with acetic acid–H2 O (80/20, v/v, sample L2 , spectrum 1), acetic acid–H2 O (90/10, v/v, sample L3 , spectrum 2) and formic acid–acetic acid–H2 O (20/60/20, v/v/v, sample L4 , spectrum 3) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. amount of sample required for the analysis (Hortling et al., 1997). In addition, the spectrum of a lignin sample gives an overall view of its chemical structure (Gilarranz et al., 2001). Fig. 3 shows FT-IR spectra of acid-insoluble lignin preparations isolated with acetic acid–H2 O (80/20, v/v, sample L2 , spectrum 1), acetic acid–H2 O (90/10, v/v, sample L3 , spectrum 2) and formic acid–acetic acid–H2 O (20/60/20, v/v/v, sample L4 , spectrum 3) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. The spectral profiles and the relative intensities of the bands were rather similar in three spectra, which confirmed that the ‘core’ of lignin structure did not change significantly during the aqueous organic acid treatment. Table 6 gives the assignments of FT-IR absorption bands of the lignins, aromatic skeleton vibrations occur at 1606, 1507 and 1434 cm−1 , in which the aromatic semicircle vibration (a vibration involving Table 5 Weight-average (M̄w ) and number-average (M̄n ) molecular weights and polydispersity (M̄w /M̄n ) of the acid-insoluble lignin preparations Lignin preparationa M̄w M̄n M̄w /M̄n a L1 L2 L3 L4 L5 L6 L7 4130 2530 1.63 4330 2760 1.57 3960 2330 1.70 4140 2650 1.56 4170 2660 1.57 4340 2630 1.65 4280 2500 1.71 Corresponding to the lignin fractions in Table 2. 188 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 Table 6 Assignments of FT-IR absorption bands (cm−1 ) Absorption bands Assignment 3429 2945 OH stretching CH stretching of methyl, methylene or methane group C O stretch in unconjugated ketone and carboxyl group C O stretch in conjugated ketone Aromatic skeletal vibrations Aromatic skeletal vibrations Aromatic methyl group vibrations Aromatic skeletal vibrations Aliphatic C–H stretch in CH3 Syringyl ring breathing with C–O stretching Aromatic C–O stretching C–O stretch in ester groups Aromatic C–H in-plane deformation for syringyl type Aromatic C–H in-plane deformation for guaiacyl type Aromatic C–H out of plane bending 1732, 1726 1660, 1653 1606 1507 1460 1434 1374 1328 1242 1165 1135 1043 855, 844 Fig. 4. FT-IR spectra of acid-insoluble lignin preparations isolated with formic acid–acetic acid–H2 O (30/60/10, v/v/v, sample L5 , spectrum 1), methanol–H2 O (60/40, v/v, sample L6 , spectrum 2) and ethanol–H2 O (60/40, v/v, sample L7 , spectrum 3) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. both carbon–carbon stretching and a change of the H–C–C bond angle) is assigned at 1507 cm−1 (Colthup et al., 1990). The band at 1732 cm−1 is originated from the carbonyl and unconjugated ketone and carboxyl Fig. 5. 1 H NMR spectrum of lignin preparation L4 obtained by treatment with formic acid–acetic acid–H2 O (20/60/20, v/v/v) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 group stretching, while the small band at 1653 cm−1 is attributed to conjugated carbonyl stretching in organosolv lignins. Moreover, aliphatic C–H stretch in CH3 and syringyl ring breathing with C–O stretching are clearly seen at 1374 and 1328 cm−1 , respectively. The bands at 1135 and 1043 cm−1 are indicative of the aromatic C–H in-plan deformation for syringyl type and guaiacyl type, respectively. Aromatic C–H out of bending exhibits at 844 cm−1 (Collier et al., 1997; Faix, 1991). Fig. 4 illustrates FT-IR spectra of acid-insoluble lignin preparations isolated with formic acid–acetic acid–H2 O (30/60/10, v/v/v, sample L5 , spectrum 1), methanol–H2 O (60/40, v/v, sample L6 , spectrum 2) and ethanol–H2 O (60/40, v/v, sample L7 , spectrum 3) combining with 0.1% HCl as a catalyst. As can be seen, the changes of the carbonyl absorption region 189 at 1726 and 1660 cm−1 might enable the evaluation of the effects of organic acid and organic alcohol treatment. Obviously, a remarkable increase of carboxyl absorption observed in spectrum 1 at 1726 cm−1 revealed that a noticeable oxidation of the lignin structure did occur during the organic acid treatment. On the other hand, a much stronger peak at 1660 cm−1 in spectra 2 and 3 demonstrated that the lignin preparations obtained by treatment with organic alcohols contained higher amounts of conjugated carbonyl groups than those of the lignins isolated with organic acids. Furthermore, the intensity of band at 1374 cm−1 for aliphatic C–H stretch in CH3 in spectrum 1 was much lower than that of the lignins in spectra 2 and 3, indicating that the oxidation occurred mainly at ␥position of lignin side chains during the organic acid treatment. Fig. 6. 13 C NMR spectrum of lignin preparation L4 obtained by treatment with formic acid–acetic acid–H2 O (20/60/20, v/v/v) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. 190 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 3.5. 1 H and 13 C NMR spectra Analysis of the 1 H NMR signal intensity in the range of 8.0–6.2 ppm provides an indirect method of monitoring the level of substitution on the aromatic ring of lignin. The 1 H NMR spectrum of lignin preparation L4 obtained by treatment with formic acid–acetic acid–H2 O (20/60/20, v/v/v) is shown in Fig. 5. The integrals of signals between 6.2 and 6.8 ppm are attributed to aromatic protons in syringylpropane and guaiacylpropane structures (Faix et al., 1992), indicating the presence of similar relative contents of syringyl and guaiacyl units in the lignin. The small signals around 7.3–7.4 ppm are assigned to the aromatic protons in positions 2 and 6, in structures containing a C␣ O group, to aromatic protons in positions 2 and 6 of phydroxyphenyl units conjugated with a double bond, to the proton in C␣ C␤ structures and to aromatic protons in p-coumaric and ferulic acids, confirming the presence of p-hydroxyphenyl units, C␣ O groups, and p-coumaric and ferulic acids in the lignin fraction (Seca et al., 2000). The H␤ in ␤-O-4 structures exhibits two weak signals between 5.0 and 4.8 (data not shown). Methoxyl protons (–OCH3 ) give a sharp signal at 3.7 ppm. The signal at 3.3 ppm is due to protons in water in DMSO. Two intense signals around 2.5 ppm are indicative of protons in DMSO. Protons in aliphatic groups such as in the side chains of lignin or in acetyl groups in xylans produce signals between 2.0 and 0.8 ppm. In order to gain a more complete understanding of the structures in the isolated lignins, a qualitative 13 C NMR spectrum of lignin preparation L , isolated 4 with formic acid–acetic acid–H2 O (20/60/20, v/v/v), was investigated (Fig. 6), and their chemical shifts (δ, ppm), intensity and assignment list in Table 7. Most of the observed signals have been previously assigned in straw and wood lignin spectra (Nimz et al., 1981; Scalbert et al., 1986). It should be noted that the spectra of the lignin sample do show some signals at 174.5, 102.0, 97.6, 92.4, 82.6, 76.6, 74.6, 73.1, 70.1 and 69.7, 65.6 and 63.0 ppm attributed to hemicellulose carbons. This suggested the partial cleavage of linkages between hemicelluloses and lignin during the treatment with formic acid–acetic acid–water mixture under the condition used. Table 7 Carbon chemical shifts (δ, ppm), intensity and assignment of the lignin preparation L4 in 13 C NMR spectrum δ (ppm) Intensitya Assignmentsb δ (ppm) Intensity Assignments 181.8 174.5 171.9 170.3 166.5 159.8 157.3 152.5 149.3 147.8 145.2 134.9 130.2 127.9 125.5 123.1 119.2 115.7 111.1 103.9 w w s m w w w s m s m m s w w w m s m s C O in dienone or quinone C-6 in 4-O-MeGlcA –COOH, aliphate –C O in acetyl group C-␥, PC ester C-4, PC ester C-4, H C-3/C-5, S C-3, G etherified C-4, G etherified C-4, G non-etherified C-1, S etherified; C-1, G etherified C-2/C-6, PC ester C-2/C-6, H C-1, PC ester C-6, FE ester C-6, G C-3/C-5, PC ester C-2, G C-2/C-6, S 102.1 97.6 92.4 84.8 82.6 79.5 76.6 74.6 73.1 72.2 70.1 69.7 65.6 63.0 59.7 59.3 56.0 33–24 22.0 13.9 s m w w w w m w w w m m w w m m vs w s m C-1, Xyl internal unit C-1, Xyl reducing end unit, ␤-anomer C-1, Xyl reducing end unit, ␣-anomer C-␤ in ␤-O-4 C-4, 4-O-MeGlcA C-3 in arabinfuranose C-3, Xyl non-reducing end unit C-3, Xyl internal unit C-2, Xyl internal unit C-␣ in ␤-O-4 C-4, Xyl non-reducing end unit C-4, Xyl non-reducing end unit C-5, Xyl non-reducing end unit C-5, Xyl internal unit, C-␥ in ␤-5 C-␥ in ␤-O-4 C-␥ in ␤-O-4 OCH3 in S and G units ␣-, ␤-Methylene groups CH3 in acetyl group ␥-Methyl in n-propyl side chains a Abbreviations: s, strong; m, mean, w, weak; vs, very strong. Abbreviations: G, guaiacyl unit; S, syringyl unit; H, p-hydroxylphenyl unit; PC, p-coumaric acid; FE, ferulic acid; Xyl, xylose; 4-O-MeGlcA, 4-O-methylglucuronic acid. b F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 Signal at 181.8 ppm could be assigned to carbonyl groups in dienone or quinone substructures (Zhang and Gellerstedt, 1999). The carbonyl resonances from uronic acids and esters may contribute to signal at 174.5 ppm, which corresponds to C-6 in methyl urinates (Himmelsbach and Bar- 191 ton, 1980). A strong signal at 171.9 ppm arises from –COOH groups in aliphatic acids. The acetyl groups in xylans give a peak at 170.3 ppm. This phenomenon confirmed again that hemicelluloses are strongly associated with lignin in the cell walls of wheat straw. Fig. 7. The thermograms of the acid-insoluble lignin preparations: (a) L2 , isolated with acetic acid–H2 O (80/20, v/v) and (b) L7 , isolated with ethanol–H2 O (60/40, v/v) combining with 0.1% HCl as a catalyst at 85 ◦ C for 4 h from dewaxed wheat straw. 192 F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 One of most important features of lignin can be seen in the aromatic region (103.9–168.0 ppm). The syringyl (S) residues were identified by signals at 152.5 and 152.2, 147.8 and 147.4, 134.9 and 134.7, and 103.9 and 103.5 ppm. Guaiacyl (G) residues were verified by signals at 149.3, 147.8 and 147.4, 145.2, 134.9 and 134.7, 119.2 and 119.0, 114.9, and 111.1 ppm. The p-hydroxyphenyl (H) residues give a small signal at 127.9 ppm. These signals confirmed that the lignin preparation could be justified as GSH grass lignin. The signals at 166.5 and 166.3, 159.8, 130.2 and 130.0, 125.5 and 124.9, and 115.7 ppm are attributed to esterified p-coumaric acid. Etherified ferulic acids exhibit signals at 168.0 ppm (data not shown). Esterified ferulic acid was detected with a signal at 123.1 ppm. These observations indicated that p-coumaric acid is linked to lignin by ester bonds, while the ferulic acid is linked to lignin by both ether and ester bonds. The most abundant substructure in lignin is an ␤O-4 structure. Side chain carbons: C-␣, C-␤ and C-␥ in ␤-O-4 can be seen at 72.2, 84.8 and 59.7–59.3 ppm, respectively. In addition, other common carbon–carbon linkages such as ␤-␤, ␤-5 and 5-5′ type structures, are also important substructures in lignin. The presence of ␤-␤ substructures can be seen from the C-␥ signal at 70.1 ppm, although C-␣ and C-␤ signals are overlapped with other signals. Side chain carbons of C-␥ in ␤-5 substructures can be observed at 63.0 ppm. Signal at 125.5 ppm could also relate to 5-5′ substructures. These observations indicated that the lignin preparation is mainly composed of ␤-O-4 ether bonds together with small amounts of ␤-␤, ␤-5 and 5-5′ carbon–carbon linkages. The strong signal at 56.0 ppm corresponds to the OCH3 group in syringyl and guaiacyl units. The signals representing the ␥-methyl, ␣ and ␤-methylene groups in n-propyl side chains occur at 13.9 and 18.4–33.6 ppm, respectively. In addition, CH3 in acetyl group give a strong signal at 22.0 ppm, indicating that the lignin fraction contained noticeable amounts of acetyl groups, coming from the residual solvent. lignin samples began to decompose at 190 (L2 ) 186 ◦ C (L7 ). When weight loss reached 50%, the temperature increased to 467 ◦ C (L2 ) and 486 ◦ C (L7 ). It is clear that after about 200 ◦ C the thermal degradation of the lignin took place rapidly. This almost same thermal stability of lignins corresponded to their equal values of molecular weights. The DSC curve of the lignin preparations showed a wide exothermic peak ranged between 200 and 600 ◦ C and centred at 476 ◦ C (L2 ) and 380 ◦ C (L3 ), due to exothermic reactions of the lignin. 4. Conclusion The treatment of wheat straw with aqueous organic acids is very well suitable for separation of the principal chemical components. This is practically true for isolation of lignins. It was found that aqueous organic acid was much effective for delignification of wheat straw than aqueous organic alcohol under the conditions used. FT-IR spectra indicated that the lignins dissolved during the treatment with aqueous organic acid, had more unconjugated and fewer conjugated carbonyl groups than those of the lignins isolated with aqueous organic alcohol. In addition, a remarkable increase of carboxyl absorption observed in the spectra of organic acid lignins suggested that a noticeable oxidation of the lignin structure did occur during the organic acid treatment. Furthermore, the treatment with organic acid, particularly acetic acid, also resulted in noticeable amounts of acetyl groups in the lignin preparations. 1 H and 13 C NMR spectra revealed that the lignin preparation comprised guaiacyl, syringyl and a small amount of p-hydroxyphenyl units. p-Coumaric acid is linked to lignin by ester bonds, while the ferulic acid is linked to lignin by both ether and ester bonds. The results also showed that the lignin preparation is mainly composed of ␤-O-4 ether bonds together with small amounts of ␤-␤, ␤-5 and 5-5′ carbon–carbon linkages. Acknowledgements 3.6. Thermal stability Fig. 7 illustrates the thermograms of the acidinsoluble lignin preparations L2 (Fig. 7a) isolated with acetic acid–H2 O (80/20, v/v) and L7 (Fig. 7b) isolated with ethanol–H2 O (60/40, v/v). As can be seen, the two The authors are grateful for the financial support of this research from National Natural Science Foundation of China (Nos. 30271061 and 30430550), Guangdong Natural Science Foundation (No. 013034) and Ministry of Education China for a major project. F. Xu et al. / Industrial Crops and Products 23 (2006) 180–193 References Aziz, S., Sarkanen, K., 1989. Organosolv pulping: a review. Tappi J. 72, 169–175. Balogh, D.T., Curvelo, A.A.S., De Groote, R.A.M.C., 1992. Solvent effects on organosolv lignin from Pinus caribea bondurensis. Holzforschung 46, 343–348. Ben-Ghedalia, D., Yosef, E., 1994. Effect of isolation procedure on molecular weight distribution of wheat straw lignins. J. Agric. Food Chem. 42, 649–652. Blakeney, A.B., Harris, P.J., Henry, R.J., Stone, B.A., 1983. A simple and rapid preparation of alditol actates for monosaccharide analysis. Carbohydr. Res. 113, 291–299. Camarero, S., Bocchini, P., Galletti, G.C., Martinez, A.T., 1999. Pyrolysis–gas chromatography/mass spectrometry analysis of phenolic and etherified units in natural and industrial lignins. Rapid Commun. Mass Spectrom. 13, 630–636. Collier, W., Kalasinsky, V.F., Schulz, T.P., 1997. Infrared study of lignin: reexamination of ary–alkyl ether C–O stretching peak assigment. Holzforschung 46, 523–528. Colthup, N.B., Daly, L.H., Wiberley, S.E., 1990. Introduction to Infrared and Raman Spectroscopy. Academic Press Limited, London. Davis, L.L., Young, R.A., Deodhar, S.S., 1986. Organic acid pulping of wood (3). Acetic acid pulping of spruce. Mokuzai Gakkaishi 32, 905–914. Faix, O., 1991. Classification of lignins from different botanical origins by FT-IR spectroscopy. Holzforschung 45, 21–27. Faix, O., Grunwald, C., Beinhoff, O., 1992. Determination of phenolic hydroxyl group content of milled wood lignins (MWL’s) from different botanical origins using selective aminolysis, FTIR, 1 H NMR and UV spectroscopy. Holzforschung 46, 425–532. Fengel, D., Wegener, G., 1984. Wood: Chemistry, Ultrastructure, Reactions. Walter de Gruyter, Berlin, pp. 539–543. Gilarranz, M.A., Rodriguez, F., Oliet, M., 2000. Lignin behavior during the autocatalyzed methanol pulping of Eucalyptus globules—changes in molecular weight and functionality. Holzforschung 54, 373–380. Gilarranz, M.A., Rodriguez, F., Oliet, M., Garcia, J., Alonso, V., 2001. Phenolic OH group estimation by FTIR and UV spectroscopy. Application to organosolv lignins. J. Wood Chem. Technol. 21, 387–395. Goyal, G.C., Lora, J.H., Pye, K.E., 1992. Autocatalyzed organosolv pulping of hardwoods: effect of pulping conditions on pulp properties and characteristics of soluble and residual lignin. Tappi J. 75, 110–116. Himmelsbach, D.S., Barton, F.E., 1980. 13 C nuclear magnetic resonance of grass lignin lignins. J. Agric. Food Chem. 28, 1203–1208. Hortling, B., Tamminen, T., Kentta, E., 1997. Determination of carboxyl and non-conjugated carbonyl groups in dissolved and residual lignins by IR spectroscopy. Holzforschung 51, 405–410. Nimz, H.H., Casten, R., 1986. Chemical processing of lignocellulosics. Holz Roh-Werkstoff 44, 207–212. Nimz, H.H., Robert, D., Faix, O., Nemr, M., 1981. Carbon-13 NMR spectra of lignins, 8: structural differences between lignins of 193 hardwood, softwoods, grasses and compression wood. Holzforschung 35, 16–26. Nimz, H.H., Schoene, M., 1993. Nonwaste pulping and bleaching with acetic acid. In: Proceedings of the 7th International Symposium on Wood and Pulping Chemistry, vol. 1, Beijing, pp. 258–265. Oliet, M., Rodriguez, F., Garcia, J., Gilarranz, M.A., 2001. The effect of autocatalyzed ethanol pulping on lignin characteristics. J. Wood Chem. Technol. 21, 81–95. Pan, X.J., Sano, Y., 1998. Characterization and utilization of acetic acid lignins: methylolation of lignins and preparation of ligninbased adhesives. In: Proceedings of the 43rd Lignin Symposium, Fuchu, Tokyo, Japan, October 26–27, pp. 5–8. Pan, X.J., Sano, Y., 1999. Atmospheric acetic acid pulping of rice straw II: behavior of ash and silica in rice straw during atmospheric acetic acid pulping and bleaching. Holzforscung 53, 49–55. Pan, X.J., Sano, Y., 2000. Comparison of acetic acid lignin with milled wood and alkaline lignins from wheat straw. Holzforschung 54, 61–65. Sano, Y., Nakamura, M., Shimamoto, S., 1990. Pulping of wood at atmospheric pressure II: pulping of hardwoods with aqueous acetic acid containing s small amount of sulfuric acid. Mokuzai Gakkaishi 36, 207–211. Sarkanen, K.V., 1990. Chemistry of solvent pulping. Tappi J. 73, 215–219. Scalbert, A., Monties, B., Guittet, E., Lallemand, J.Y., 1986. Comparison of wheat straw lignin preparations. I. Chemical and spectroscopic characterizations. Holzforschung 40, 119– 129. Seca, A.M.L., Cavaleiro, J.A.S., Domingues, F.M.J., Silvestre, A.J.D., Evtuguin, D., Neto, C.P., 2000. Structural characterization of the lignin from the nodes and internodes of Arundo donax reed. J. Agric. Food Chem. 48, 817–824. Sun, J.X., Sun, X.F., Sun, R.C., Fowler, P., Baird, M.S., 2003. Inhomogeneities in the chemical structure of sugarcane bagasse lignin. J. Agric. Food Chem. 51, 6719–6725. Sun, R.C., Tomkinson, J., Bolton, J., 1999. Effect of precipitation pH on the physico-chemical properties of the lignins isolated from the black liquor of oil palm empty fruit bunch fibre pulping. Polym. Degrad. Stab. 63, 195–200. Sun, R.C., Tomkinson, J., Mao, F.C., Sun, X.F., 2000. Physicochemical characterization of lignins from rice straw by hydrogen peroxide treatment. J. Appl. Polym. Sci. 79, 719–732. Uraki, Y., Kubo, S., Nigo, N., Sano, Y., Sasaya, T., 1995. Preparation of carbon fibres from organosolv lignin obtained by aqueous acetic acid pulping. Holzforschung 49, 343–350. Uraki, Y., Kubo, S., Kurakami, H., Sano, Y., 1997. Activated carbon fibres from acetic acid lignin. Holzforschung 51, 188– 192. Vila, C., Santos, V., Parajo, J.C., 2003. Recovery of lignin and furfural from acetic acid–water–HCl pulping liquors. Bioresour. Technol. 90, 339–344. Zhang, L., Gellerstedt, G., 1999. Detection and determination of carbonyls and quinones by modern NMR techniques. In: 10th ISWPC, vol. 2, Yokohama, Japan, pp. 164–170.