Bioresource Technology 182 (2015) 120–127

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Pyrolysis behaviors of four lignin polymers isolated from the same pine wood Shurong Wang ⇑, Bin Ru, Haizhou Lin, Wuxing Sun, Zhongyang Luo State Key Laboratory of Clean Energy Utilization, Zhejiang University, Hangzhou 310027, China

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Isolation process greatly influences

lignin structure and pyrolysis behavior.  AL and MWL have poor thermal stability due to well-preserved weak ether linkages.  Weights of two reactions in DGDAEM vary among the four lignins pyrolysis.  Higher phenols yield in the pyrolysis of AL and MWL via breaking of weak ether bonds.

a r t i c l e

i n f o

Article history: Received 15 December 2014 Received in revised form 27 January 2015 Accepted 29 January 2015 Available online 7 February 2015 Keywords: Lignin Pyrolysis Isolation Kinetics Mechanism

a b s t r a c t Four lignin polymers, alkali lignin (AL), klason lignin (KL), organosolv lignin (OL), and milled wood lignin (MWL), were isolated from the same pine wood. Structural characterization by FTIR and 13C NMR indicated that the four lignins have different structural features. Their pyrolysis behaviors were analyzed by TG-FTIR and Py-GC/MS. Thermally unstable ether bonds and side branches were well-preserved in AL and MWL, but were broken in OL and KL. Pyrolysis of AL and KL produce more phenols at low temperature by the breakage of ether bonds. AL and KL show lower activation energies in the main degradation stage, quantified by a distribution activation energy model with two linearly combined Gaussian functions. The evolution behaviors of typical gaseous products, CH4, CO2, and CO, were analyzed, and insights about the correlation between chemical structure and pyrolysis behavior were obtained. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Fast pyrolysis is a promising thermo-chemical method for directly converting biomass into liquid biofuels (Mohan et al., 2006). As one of the three main components in lignocellulosic biomass, lignin has a significant influence on the pyrolysis behavior of biomass. Lignin is a three-dimensional, complex polymer comprised of p-hydroxyphenyl, guaiacyl, and syringyl, all of which

⇑ Corresponding author at: State Key Laboratory of Clean Energy Utilization, Zhejiang University, Zheda Road 38, Hangzhou 310027, China. Tel.: +86 571 87952801; fax: +86 571 87951616. E-mail address: [email protected] (S. Wang). http://dx.doi.org/10.1016/j.biortech.2015.01.127 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.

are randomly connected with each other by CAOAC (b-O-4, a-O4, 4-O-5) or CAC (5-5, b-5, b-1) bonds, where b-O-4 is most frequent. Lignin usually has a content of 18–35% in lignocellulosic biomass, and is covalently linked with hemicellulose (Mohan et al., 2006). However, the pyrolysis mechanism of lignin remains unclear. The current popular model compounds used in lignin pyrolysis mechanism studies include monomers (Kotake et al., 2013), dimers (Kawamoto et al., 2007), and oligomers (Chu et al., 2013). They have simplified structures and could be considered as basic fragments in lignin. However, lignin polymer that is directly isolated from biomass has a structure more closely related to its natural form, i.e., high degree of polymerization and more complete branches. Lignin polymers could be isolated from bio-

S. Wang et al. / Bioresource Technology 182 (2015) 120–127

mass by many different methods. Alkali lignin (AL) and klason lignin (KL) are isolated by removing carbohydrates that are linked with lignin using alkali and acid, respectively. Lignin could also be isolated using neutral solvents. Most popularly, an ethanol– water mixture is used to extract the organosolv lignin (OL) at high temperature. Milled wood lignin (MWL) is another lignin polymer isolated by using neutral organic solvent with the help of ball milling at room temperature. Selecting the lignin isolation method has a significant influence on the chemical structure of the isolated lignin, because lignin’s weak ether linkages and side branches are sensitive to chemical treatments (El Hage et al., 2010; Pan et al., 2006). AL, KL, OL, and MWL have often been used as model compounds in the mechanistic study of lignin pyrolysis. Wang et al. (2009) performed pyrolysis of MWLs isolated from hardwood and softwood, and found that methoxy group content has a significant influence on the resulting pyrolysis behaviors. Yang et al. (2007) found that AL decomposes over a wide temperature range of 160–900 °C, and the final solid char yield was about 40%, much higher than that of cellulose pyrolysis. Bährle et al. (2014) used KL as a model compound for comparing the pyrolysis behaviors of hardwood and softwood lignins. They suggested that radical formation during hardwood KL pyrolysis was more frequent than that of softwood KL pyrolysis, and was mainly due to high methoxy group content in hardwood KL. Patwardhan et al. (2011) studied the effect of pyrolytic temperature on the product distribution of lignin pyrolysis by using a commercial OL, and found that the pyrolytic products from OL were mainly monomeric phenols with alkyl or methoxy groups. Since the chemical structure of lignin changes with the isolation process, it is necessary to compare the structural features and pyrolysis behaviors among lignins isolated by different methods. One of the biggest challenges in using lignin polymer as a model compound for pyrolysis mechanism studies is the limited understanding of its chemical structure. While many previous studies mainly focus on the characterization of pyrolytic products, characterization of the original lignin structure and its influence on pyrolysis behavior are inadequate. In this study, four lignin polymers, AL, KL, OL, and MWL, were isolated from the same pine wood (Pinus bungeana). Their chemical structures were characterized using Fourier transform infrared spectroscopy (FTIR) and 13C nuclear magnetic resonance spectroscopy (13C NMR). Subsequently, the lignin polymers were used to study the mechanistic influence of structural features on pyrolysis behavior.

2. Experimental 2.1. Materials The pine wood (Pinus bungeana) used in this study was obtained from a local timber mill (Hangzhou, China). Before each experiment, the raw material was dried and ground into coarse powder approximately 0.15–0.18 mm in size. The isolation of AL was based on the method proposed by Sun et al. (2000), where the NaOH concentration was 5% and the process was performed at 50 °C for 6 h. KL was obtained from the acid–insoluble fraction by NREL standard procedures (Sluiter et al., 2008). In the method previously described by Pan et al. (2006), OL was extracted by adding a small amount of sulfuric acid (1.2 wt% of raw biomass) to a mixed solution of ethanol and water (65:35, v/v) at 170 °C for 80 min. The solid-to-liquid ratio for this process was 1:8. MWL was isolated by a neutral solvent at room temperature using a method developed by Bjorkman (El Hage et al., 2010). The purity of lignin was calculated by quantitative acid hydrolysis method, and the isolated lignin was treated by 4% H2SO4 for 1 h at 120 °C (Ahmed et al., 2013). The purities of each lignin were 87.3%, 92.2% and 88.7%

121

for AL, OL and MWL, respectively. KL was considered to be pure since it was isolated by high concentration acid treatment. The content of ash in lignin was measured by using thermogravimetric analysis operating under air atmosphere. 2.2. Structural characterization of lignins The ultimate analysis of each lignin polymer was performed on a Vario MICRO Elemental Analyzer (Elenemtar Analysensysteme GmbH, Germany). Information about the functional groups of the four lignin polymers were recorded by a Nicolet 5700 Fourier transform infrared spectroscopy over the wavenumber range of 4000–400 cm1 with a resolution of 4 cm1. Each spectrum was accumulated from 36 scans. Solution-state 13C NMR spectra of MWL, AL, and OL were recorded on an Agilent 600 MHz DD2 spectrometer at 25 °C. For each run, 80 mg sample was dissolved in 0.5 mL dimethyl sulfoxide-d6 (DMSO-d6). Tetramethylsilane was selected as the internal standard. Each spectrum was collected for 21,000 scans. Since KL is difficult to be dissolved, a solid-state 13 C cross-polarization magic angle spinning (13C CP/MAS) NMR was employed to analyze the KL structure. The carbonyl carbon resonance of glycine was used as the external reference. The spectrum was averaged over 1200 scans. 2.3. Pyrolysis characterization of lignins The devolatilization behavior of four lignin polymers were identified by a Netzsch STA 409 thermogravimetric (TG) analyzer with a heating rate of 20 °C/min from 25 to 800 °C. The released volatiles were immediately swept into a coupled FTIR spectrometer for on-line analysis (TG-FTIR). Spectra were recorded between 4000–400 cm1 with a resolution of 4 cm1 and a scan rate of 32 scans per minute. Pure nitrogen was used as the carrier gas and held at a flow rate of 60 mL/min. To minimize the influence of heat transfer, samples were held constant at 5 mg. The product distributions from the pyrolysis of each sample at different temperatures were further analyzed on a pyrolyzer coupled with a gas chromatography–mass spectrometer (Py-GC/MS). Approximately 0.5 mg of sample was loaded in the micro-pyrolyzer (CDS5200, CDS Analytical Inc., USA), and was immediately heated to the required temperature with a residence time of 10 s. The pyrolytic products were analyzed by a GC–MS (Thermo scientific, Trace DSQII, USA) equipped with a DB-WAX capillary column (30 m  0.25 mm  0.25 lm). Pure helium was used as the carrier gas with a constant flow rate of 1 mL/min. The GC oven was heated from 40 °C (held for 1 min) to 240 °C (held for 24 min) at a heating rate of 8 °C/min. The MS was operated in EI mode with an ion-source electron energy of 70 eV. The detected mass-to-charge ratio ranged from 35 to 450. All of the chemicals were identified according to the NIST MS library and previous work (Brebu et al., 2013; Kim et al., 2013). 3. Results and discussion The color of isolated KL was dark brown, while the colors of the other three samples were much lighter. All samples had higher carbon content and lower oxygen content than raw biomass (see Table 1), indicating that lignin is an important carbon source for the resulting pyrolytic products. OL contains the lowest oxygen content, suggesting an intense deoxygenation reaction occurring during its isolation. Elemental sulfur was only found in KL because of the sulfuric acid addition during the isolation process. The ash contents of lignins were 6.22%, 1.66%, 1.73% and 2.37% for AL, KL, OL and MWL, respectively. This indicated that the alkali treatment introduced some extra alkali metal in the AL structure.

122

S. Wang et al. / Bioresource Technology 182 (2015) 120–127

Table 1 Ultimate analysis of four lignin polymers.

AL KL OL MWL Raw biomass

Cdaf (wt%)

Hdaf (wt%)

Odaf (wt%)

Ndaf (wt%)

Sdaf (wt%)

O/C

H/C

54.91 59.65 65.09 59.25 46.65

5.89 5.80 5.93 5.98 6.70

39.07 33.38 28.89 34.74 46.55

0.13 0.03 0.09 0.03 0.10

– 1.14 – – –

0.53 0.42 0.33 0.44 0.75

1.29 1.17 1.09 1.21 1.72

3.1. Structural features of lignin The functional groups of the four lignin polymers were similar according to the FTIR characterization. All the lignins showed sharp peaks corresponding to the stretching vibrations from phenolic OAH, aliphatic OAH and CAH bonds. Guaiacyl was the major basic unit in the four lignin polymers, and was responsible for the peaks at 1270 cm1 (guaiacyl ring plus CAO stretching), 1137 cm1 (CAH in-plane deformations of guaiacyl units), 858 cm1, and 812 cm1 (both attributed to the CAH out-of-plane vibrations in positions 2, 5 and 6 of guaiacyl units) in the FTIR analysis (Jahan et al., 2007; Tejado et al., 2007; Wang et al., 2009). This analysis result is in agreement with a previous study proposing that softwood lignin usually contains guaiacyl as the primary basic unit, while hardwood lignin was enriched in both guaiacyl and syringyl units (Wang et al., 2009). Conjugated and unconjugated C@O groups (1618 cm1 and 1712 cm1, respectively) were also detected. C@O conjugated with the aromatic ring was the major C@O vibration, which usually had lower reaction activity than unconjugated C@O (Wang et al., 2014b). It was found that lignins contained abundant branches, such as the functional groups detected at 1224 cm1 (phenolic OAH plus ether CAO stretching), 1087 cm1 (CAO deformation at Cb and aliphatic ether), and 1033 cm1 (CAO deformation at Ca and aliphatic ether) (Tejado et al., 2007). The peak at 1328 cm1 is assigned to the condensed guaiacyl ring (Jahan et al., 2007). The relative absorbance intensity (Iw/Iar) was introduced to further compare the relative contents of the functional groups in the four lignin polymers. Iar, the intensity at 1511 cm1 corresponding to the aromatic skeletal vibration, was used as a standard. This method has previously been successfully employed to compare the structural difference between hardwood and softwood lignins (Pandey, 1999). As shown in Fig. 1, the relative intensities for the four samples were significantly different. OL shows a lower relative intensity for five oxygen-containing functional groups (1712 cm1, 1618 cm1, 1224 cm1, 1087 cm1, and 1033 cm1). This is in agreement with the finding that the oxygen content in OL is the lowest of the four lignins (see Table 1), and is caused by intense hydrolysis of esters during organic solvent treatment at 170 °C (El Hage et al., 2010). The sulfuric acid added during the OL isolation process is capable of catalyzing the breakage of Ca- and Cbether linkages, as well as the dehydration of aliphatic hydroxyls (Pan et al., 2006). AL shows a higher relative intensity at 1712 cm1 and 1618 cm1, indicating that each aromatic ring in the AL structure contains more conjugated and unconjugated C@O groups. The higher relative intensity of KL at 1328 cm1 demonstrates that an intense condensation reaction occurs during the Klason process. 13 C NMR was employed to further investigate the structure of the lignin polymers. Chemical shift assignments were made according to previous studies (El Hage et al., 2010; Huang et al., 2011), and the results are listed in Table 2. The integral of the aromatic region (d 160–102 ppm) was set to 6.12 (six from the aromatic ring carbons plus 0.12 carbon contribution from coniferyl

Fig. 1. Relative intensities of typical functional groups for the four lignin polymers.

Table 2 Quantitative comparison of Chemical shift (ppm)

195–190 172 168 160–140 140–123 123–102 89–57 89–78 73–71 64–61 57–54 31–29 15 a b

13

C NMR signals for four lignin polymers.

Assignment

Carbonyl C@O Ester COO-R Carboxyl C@O Aromatic CAO Aromatic CAC Aromatic CAH Aliphatic CAO Cb in b-O-4, Ca in b-5, b-b Ca in b-O-4 Cc-OR AOCH3 Ca and Cb in-CH2A Cc in ACH3

Functional group distributions (mol/100 g) ALa

KLb

OLa

MWLa

0.003 0.011 0.079 0.844 1.086 0.784 0.841 0.281

– – – 1.178 1.023 0.899 0.960 –

0.026 0.000 0.105 0.753 1.277 1.245 0.573 0.153

0.017 0.003 0.077 0.770 1.055 1.027 0.858 0.271

0.112 0.205 0.597 0.069 0.011

– – 0.407 – –

0.046 0.190 0.622 0.075 0.088

0.106 0.273 0.742 0.043 0.023

Data obtained by solution-state 13C NMR spectrum. Data obtained by solid-state 13C CP/MAS NMR spectrum.

and coniferaldehyde) (El Hage et al., 2010). The area integration of each functional group could be adjusted based on the structural moieties of each aromatic ring. Therefore, the C9 formulae for the four lignin polymers could be calculated: C9H9.28O4.14(OCH3)1.35 for AL, C9H9.03O3.31(OCH3)0.80 for KL, C9H7.62O2.19(OCH3)1.16 for OL, and C9H8.05O3.07(OCH3)1.59 for MWL. The OL molecular weight is the lowest (186.9 g/mol), this was because that ester groups underwent severe cleavage or breakdown during isolation, and thus the molecular weight was decreased. The OL ester signal at 172 ppm nearly disappears, indicating the primary reaction was the hydrolysis of esters. This agrees well with previous studies where only trace ester groups could be detected by 13C NMR for ethanol-extracted OL (El Hage et al., 2010). Instead, OL exhibits a strong signal between 195 and 190 ppm caused by the formation of Hibbert ketone from the hydrolysis of esters (El Hage et al., 2010). For the same reason, OL also contains higher carboxyl content (the signal at 168 ppm) than AL and MWL. The integrated values of the aromatic region (160–102 ppm) per 100 g of KL and OL were higher than those of AL and MWL. This confirms that the organosolv extraction and Klason treatment were the two most intense processes that greatly changed the original lignin structure. Many of the side branches in the OL structure were eliminated, while the KL structure was condensed. The signals at 89–57 ppm are ascribed to aliphatic

S. Wang et al. / Bioresource Technology 182 (2015) 120–127

CAO, in which 89–78 ppm and 73–71 ppm correspond to Cb and Ca, respectively. The integrated values of these two signals are very similar for AL and MWL, while the integrated value of OL was much lower due to the breakage of ether bonds (Pan et al., 2006). The OL signal from the CcAO group (64–61 ppm) was also lower than that of AL and MWL due to the dissociation of hydroxyls and esters linked with Cc (Sun et al., 2000). The signals at 57–54 ppm are ascribed to methoxy groups, and play an important role in the lignin pyrolysis (Bährle et al., 2014; Wang et al., 2009). KL contains the lowest content of methoxyl, which was eliminated during the Klason treatment, as reported in previous studies (Akiyama et al., 2003; Kim et al., 2013). The Ca/Cb methylene and Cc methyl groups in the propyl side chains are responsible for the signals at 31–29 ppm and 15 ppm, respectively. OL shows a higher signal than AL and MWL in these two bands due to the hydroxyl dehydration and ether cleavage accompanied with Hibbert ketones formation (El Hage et al., 2010). Thus, it could be concluded that, unlike OL and KL, the branched nature of lignin is well-preserved in AL and MWL. 3.2. Pyrolysis kinetics The devolatilization of all four lignin polymers occurs over a wide temperature range and achieves a maximum weight loss rate at 300–450 °C. The ether linkages have poor thermal stability during pyrolysis. b-O-4, the most frequent and weakest ether bond in the lignin structure, is easily broken in the range of 250–350 °C (Chu et al., 2013). According to structural characterization, weak ether bonds in the original lignin were well-preserved in AL and MWL, but not in OL and KL. Thus, AL and MWL have weaker thermal stability and decompose more easily at low temperatures. The temperature at maximum weight loss rate for AL (346 °C) and MWL (359 °C) are lower than those for OL (396 °C) and KL (405 °C). The final residue weight percentages of the four samples were MWL (41.86%) < OL (44.69%) < AL (49.43%) < KL (49.86%). Combined with the analytical results listed in Table 2, the final residue weights decrease with increasing methoxyl content. The structure with lower methoxyl content has better thermal stability and more condensed structural units (Jakab et al., 1997), leading to more char production. Furthermore, methoxyl groups may decompose into small molecular radicals, which would further stabilize large molecule fragments produced during lignin pyrolysis, and would prevent their polymerization leading to char. Pyrolysis kinetic study could in depth investigate the devolatilization behavior of biomass during pyrolysis (Chen et al., 2013). In this study, a modified distributed activation energy model (DAEM) is introduced to simulate the thermal degradation behavior of lignin. DAEM assumes that a series of independent first order reactions occur during devolatilization, and the activation energies for all parallel reactions satisfy a probability-distribution function f(E), as shown in Eq. (1).

1  aðTÞ ¼

Z 0

1

  Z k0 T E=RT exp  e dT f ðEÞdE b 0

Eq. (2). This method has previously been successfully employed in the kinetic study of MWL pyrolysis (Wang et al., 2014b). The two functions correspond to two reactions: first, the decomposition and release of volatiles (reaction I), and second, polymerization to form char (reaction II).

  1 ðE  E01 Þ pffiffiffiffiffiffiffi exp þ ð1  wÞ 2r2E1 rE1 2p   1 ðE  E02 Þ pffiffiffiffiffiffiffi exp  2r2E2 rE2 2p

f ðEÞ ¼ w

ð2Þ

E01 and E02 are the mean activation energies for the first and second Gaussian distributions, and rE1 and rE2 are the corresponding standard deviations. A weight factor, w, scales the devolatilization contributions for the two reactions. A pattern search algorithm is adopted to perform the iterative operation and obtain kinetic parameters. DG-DAEM fits to the pyrolysis data of the four lignin polymers are excellent (see Fig. 2). Their activation energy distributions are similar (see Fig. 3). The first sharp peak with low E01 and narrow rE1 corresponds to reaction I, while the shoulder peak with high E02 and wide rE2 corresponds to reaction II. Reaction I primarily includes the breakage of ether linkages and side branches, and thus, the pyrolysis of AL (139.7 kJ/mol) and MWL (145.6 kJ/mol) exhibit lower activation energies than OL (151.0 kJ/mol) and KL (166.2 kJ/mol). This observation is in agreement with the temperatures at maximum weight loss rate in the TG analysis. The structures of AL and MWL contain more weak linkages than OL and KL. The dissociation of side branches and the breakage of ether bonds usually have low energy barriers, resulting in low activation energies during pyrolysis. During reaction II, the polymerization and cross-linking reaction to form char, AL pyrolysis shows the lowest activation energy for char formation, and might be due to the catalytic effect of alkali metal in AL (Jakab et al., 1997). During the alkali treatment, sodium could be preserved in AL in two forms, mostly in the form of organic sodium, such as carboxylate sodium and phenolic sodium that are linked with AL structure; besides, inorganic sodium, such as NaOH and Na2CO3, might also exist (Guo et al., 2012). KL pyrolysis shows the lowest value of w,

Pyrolysis behaviors of four lignin polymers isolated from the same pine wood.

Four lignin polymers, alkali lignin (AL), klason lignin (KL), organosolv lignin (OL), and milled wood lignin (MWL), were isolated from the same pine w...
2MB Sizes 0 Downloads 11 Views