Characteristics and deoxy-liquefaction of cellulose extracted from cotton stalk

Fuel 166 (2016) 196–202

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Characteristics and deoxy-liquefaction of cellulose extracted from cotton stalk Jinhua Li a, Shuai Zhang a, Boyang Gao a, Aikai Yang a, Zonghua Wang a,⇑, Yanzhi Xia a, Haichao Liu b a College of Chemical Science and Engineering, Teachers College, Collaborative Innovation Center for Marine Biomass Fiber Materials and Textiles, Qingdao University, Qingdao 266071, China b Beijing National Laboratory for Molecular Sciences, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China

h i g h l i g h t s
Cellulose extracted from cotton stalk showed representative cellulose structure.
The extracted cellulose was converted into liquid oil by deoxy-liquefaction.
The liquid oil was mainly composed of aromatics, phenols, and alkanes.
The liquid oil had low oxygen content and high heating values.

a r t i c l e

i n f o

Article history: Received 10 June 2015 Received in revised form 23 October 2015 Accepted 28 October 2015 Available online 3 November 2015 Keywords: Cellulose Cotton stalk Deoxy-liquefaction Liquid oil

a b s t r a c t Cellulose was extracted from cotton stalk and characterized by using scanning electron microscope (SEM), X-ray diffraction (XRD), nuclear magnetic resonance (NMR), and Fourier transform infrared (FTIR) spectroscopy. The results showed that the extracted cellulose had representative cellulose structure. Furthermore, the extracted cellulose was converted into liquid oil by direct deoxy-liquefaction. The elemental analysis, FTIR spectroscopy and gas chromatography–mass spectrometry (GC–MS) analyses of the liquid oil indicated that the extracted cellulose oil was mainly composed of aromatic hydrocarbons, phenols and alkanes. This oil featured high quality including the low oxygen content of 6.46% and the higher heating value (HHV) of 42.66 MJ/kg. This suggested that deoxy-liquefaction technique may be an effective way to convert cellulose into high-quality liquid fuel and value-added chemicals. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Lignocellulose, as the most abundant and renewable source, has proved to be a potential feedstock to produce fuel and value-added chemicals [1,2], which is mainly composed of three basic structural components: cellulose, hemicellulose and lignin. Generally, dry lignocellulose is roughly comprised of 40–50% cellulose, 20–30% hemicellulose, and 20–35% lignin [3]. Obviously, cellulose is the most abundant component in lignocellulose, which consists of D-glucose units linked by 1,4-b-glycosidic bonds [3]. In practice, cellulose can be converted into useful chemicals such as mannitol [4], furfural [5], and levoglucosenone [6], or fuels including cellulosic ethanol [3] and bio-oil through pyrolysis [1,2]. For example, Wang et al. [7] and Shen et al. [8] similarly investigated the main products from cellulose pyrolysis which were mainly composed

⇑ Corresponding author. Tel.: +86 0532 85950873; fax: +86 0532 85955529. E-mail address: [email protected] (Z. Wang). http://dx.doi.org/10.1016/j.fuel.2015.10.115 0016-2361/Ó 2015 Elsevier Ltd. All rights reserved.

of levoglucosan, furfural, furans, hydroxyacetaldehyde, hydroxyacetone, etc. Obviously, the bio-oils obtained from cellulose by pyrolysis were highly oxygenated complex mixtures, which usually led to lower HHV and poor combustion properties. Furthermore, most of the above researches were performed on the commercially available microcrystalline cellulose, which is different from that extracted from lignocellulosic biomass. In this paper, direct deoxy-liquefaction which has been proved to be effective to produce high-quality liquid oils from biomass [9– 11] was applied to the thermochemical conversion of cellulose extracted from cotton stalk. During the conversion process of biomass, cellulose degradation was accompanied by the degradation of hemicellulose and lignin, which made the process more complicated. So it is necessary to separate the main component-cellulose from lignocellulosic biomass and study its degradation products. Up to date, various kinds of separation methods such as steam explosion, microwave assisted extraction, inorganic acid treatment, alkali treatment and ionic liquid methods have been reported. Recently, for example, Dong et al. [12], obtained cellulose

J. Li et al. / Fuel 166 (2016) 196–202

fibers from cotton stalk bark by a combination of steam explosion, potassium hydroxide and peroxide treatments. Differently, ElSakhawy et al. [13] used hydrochloric or sulfuric acid to prepare microcrystalline cellulose from bagasse, rice straw and cotton stalk. While Song et al. [14] applied H2SO4, NaOH and H3PO4 to decompose lignocellulose samples (giant reeds, pennisetum and cotton stalks) and investigated the correlation between cellulose allomorphs. In this study, cellulose was extracted from cotton stalk by using hydrochloric acid and sodium hydroxide to remove hemicellulose and lignin, which was the most simple and economic method. Furthermore, this cellulose was characterized and converted into liquid oil by direct deoxy-liquefaction. Then, the liquid oils obtained from extracted cellulose and cotton stalk feedstock were analyzed using elemental analysis, FTIR and GC–MS analyses. 2. Materials and methods 2.1. Materials

197

500 °C to obtain the liquid oil. Water was also distilled out and the oil floated on the water. When distillation finished, water was separated by an injector. Finally, the solid char was removed and weighed at room temperature. At least three duplicate runs were conducted, and the maximum error under the same conditions was ensured within 5%. 2.4. Analysis methods The heating value was obtained by calculation according to Dulong’s formula [18]. HHV ðMJ=kgÞ ¼ ½338:2  C wt% þ 1442:8  ðH wt%  O wt%=8Þ  0:001

Elemental analyses (C, H, and N) of the raw material and oil were performed on an elemental analyzer (Elemental Varian EL, Germany). The oxygen content was calculated by difference as follows:

O ðwt%Þ ¼ 100  ðC þ H þ NÞ ðwt%Þ

The sample of cotton stalk investigated in this study was collected from Shandong Province, China. Before the experiments, the sample was air-dried, grounded and then sieved to give fractions with particle size of 60 meshes. The determination of extractives, cellulose, hemicellulose, lignin and ash was based on Van Soest method [15]. And the characteristics of the samples on dry basis were shown in Table 1. 2.2. Extraction of cellulose The cotton stalk powder was treated by refluxing in a Soxhlet apparatus with toluene-ethanol (2:1 v/v) for 6 h to remove fats and waxes and then dried in an oven under 110 °C for 8 h. The dried cotton stalk powder was added to 2 N HCl in a threenecked flask with magnetic stirring at 40–60 °C for 3–4 h. The reaction mixture was then filtered and the filter residue was washed repeatedly with distilled water to neutral. Finally, the above washed filter residue was delignified with 3% aqueous sodium hydroxide at 120 °C in autoclave for 2 h [16]. 2.3. Experimental procedure The direct liquefaction experiments were performed in a stainless steel tubular reactor with 15% water as medium, which is similar to the methods described in the literatures [10,11,17]. The reactor was heated at the heating rate of 80 °C/min and maintained for 30 min at final temperature of 400 °C. After heated, the reactor was cooled to room temperature and the volatile products were collected by gas collecting bags. Then, the residue was further distilled with the temperature rising from room temperature to

Table 1 Main characteristics of cotton stalk and the extracted cellulose oil. Ultimate analysis

Cotton stalk

Liquid oil

C (wt%) H (wt%) O (wt%) N (wt%) Empirical formula H/C molar ratio O/C molar ratio HHV (MJ kg1) Extractives (wt%) Cellulose (wt%) Hemicellulose (wt%) Lignin (wt%) Ash (wt%)

42.31 6.19 50.54 0.81 CH1.75O0.90N0.016 1.75 0.90 14.12 5.61 41.83 24.53 21.24 6.79

82.12 11.16 6.46 0.22 CH1.63O0.06N0.002 1.63 0.06 42.66

FTIR spectra of extracted cellulose, cotton stalk and the liquid oil from cellulose were recorded using a Fourier-Transform Infrared Spectrometer (Varian 3100, America) over a range 4000– 400 cm1. X-ray diffraction patterns of extracted cellulose and cotton stalk were carried out on a D8 Advance instrument (Bruker AXS, Germany) with Cu Ka radiation and 2h from 3° to 70°. Scanning electron microscope (SEM) images of extracted cellulose and cotton stalk were performed on a FESEM S-4800 (Japan). Cross polarization/magic angle spinning (CP/MAS) 13C solid-state nuclear magnetic resonance (NMR) experiments were performed for the extracted cellulose and cotton stalk on a Bruker Advance III 400 NMR spectrometer operating at 25 °C. The compositions of the liquid oils were analyzed by a Shimadzu gas chromatography and mass spectrometry (GC–MS) (QP2010S, Japan). The GC was fitted with a 30 m  0.25 mm  0.25 lm fused quartz capillary column and coated with TR-5MS as the stationary phase. Compounds in the oils were identified by comparison with the mass spectra with the NIST (National Institute of Standards and Technology) 08 library, together with the literature data to obtain the highest likelihood of compound identification. 3. Results and discussion 3.1. Characterizations of extracted cellulose 3.1.1. Scanning electron microscopy (SEM) The cotton stalk and extracted cellulose was characterized using SEM to understand its surface morphology. From Fig. 1a and b, it can be seen that the grounded cotton stalk had a rough and non-uniform outer surface. The outer layer of the stalks was mostly composed of lignin, ash and hemicellulose that enclosed the interior cellulose. Most of these surface substances were removed during cellulose extraction resulting in relatively smooth cellulose as shown in Fig. 1c [19,20]. The extracted cellulose displayed a rod-shaped morphology with some of the crystalline regions broken in Fig. 1c, which was different from the images of the cellulose fibers in the literatures [14,21]. Fig. 1d shows irregular aggregated fiber fragments, a network structure and a rough surface morphology which may be attributed to the residual non-cellulose components such as hemicellulose and lignin [22,23]. 3.1.2. X-ray diffraction (XRD) As shown in Fig. 2, the XRD patterns of the extracted cellulose had similar diffraction peaks to those of cotton stalk. Compared

198

J. Li et al. / Fuel 166 (2016) 196–202

Fig. 1. Scanning electron micrographs. (a and b) Cotton stalk. (c and d) Extracted cellulose.

Fig. 2. X-ray diffraction patterns. (a) Extracted cellulose. (b) Cotton stalk.

with the cotton stalk, the relative intensity of diffraction peaks of extracted cellulose increased, which can be attributed to the removal of amorphous, non-cellulose substances in cotton stalk [20]. The diffraction peaks of cellulose at the 2h values of 22° and 35° were corresponding to 2 0 0 and 0 0 4 lattice planes for cellulose I crystallites, respectively [12,22]. However, the characteristic  1 0 lattice planes for cellulose I crystallites were peaks of 1 1 0 and 1 not distinct, but combined into one single broad peak at 2h value of 16°. This may be due to the presence of noncellulosic substances such as hemicellulose and lignin in the sample [16]. This result was similar with that in the previous studies [12,22,23], but different with 17°, 26° and 17°, 27° and 40° in the cellulose reported by Ibrahim and El-Sakhawy [13,19], which may be ascribed to the different treatment methods. 3.1.3. CP/MAS 13C solid-state NMR spectra The CP/MAS 13C solid-state NMR spectra of extracted cellulose and cotton stalk were shown in Fig. 3. Most of the signals were

Fig. 3. CP/MAS 13C solid-state NMR spectra. (a) Extracted cellulose. (b) Cotton stalk.

similar on both spectra. For example, the chemical shifts at 61 and 65 ppm were attributed to C6 of the primary alcohol group in cellulose. The resonances at 72 and 74 ppm were ascribed to C2, C3 and C5 in cellulose. The chemical shifts at 83 and 88 ppm were associated with C4 and the chemical shift at 104 ppm was assigned to C1 in cellulose [16]. In addition, there were obvious differences in the signals at 172, 151, 55, and 29 ppm which may be attributed to carbonyl and methyl resonances. These very weak signals in cellulose spectra indicated trace amount of the residual of lignin or hemicellulose, which was in accordance with the results of SEM and XRD.

J. Li et al. / Fuel 166 (2016) 196–202

3.1.4. FTIR spectra As shown on FTIR spectra in Fig. 4, the spectra of extracted cellulose was similar with that of cotton stalk except for the bands at 1730 cm1 and 1505 cm1 which were associated with the absorption of carbonyl stretching of ester or carboxyl groups in hemicellulose and the aromatic skeletal vibration in lignin, respectively [21]. The strong absorption at 3425 cm1 and 2905 cm1 were assigned to OAH and CAH stretching vibration in cellulose, respectively. The absorption near 1635 cm1 was indicative of the bending mode stretch of the absorbed water, since the pure cellulose has strong affinity for water [16]. The absorption at 1060 cm1 was attributed to the CAOAC stretching vibration [24]. The small sharp band at 898 cm1 was characteristic absorption of b-glycosidic linkages between monosaccharides of cellulose, which suggested that the glucose that formed the backbone of the cellulose was linked by b-form bonds [16]. 3.2. Analysis of the liquid oils obtained from extracted cellulose and cotton stalk 3.2.1. Elemental analysis The main characteristics of the liquid oil obtained from extracted cellulose were shown in Table 1. The H/C and O/C molar ratios were 1.63 and 0.06, respectively, which were lower than that of the original feedstock (1.75 and 0.90). This can be attributed to the departure of oxygen atoms during the deoxy-liquefaction process. Obviously, the present oxygen content (6.46%) was much lower than those reported by many literatures [25,26], which indicated that the deoxy-liquefaction technique was more effective to remove the oxygen from the feedstock. Accordingly, the HHV of the extracted cellulose oil was 42.66 MJ/kg, which was higher than those mentioned in the above Refs. [25,26]. So, it can be concluded that cellulose, as the most abundant component in lignocellulosic biomass, was favorable for production of high quality liquid oil. 3.2.2. GC–MS analysis The characterizations of the components in the liquid oils from extracted cellulose and cotton stalk were carried out by GC–MS. About 1 1 0 peaks were observed in each spectrum. MS spectra of these peaks were identified by the MS data available as standards in the Nist 08 database of the GC–MS. The composition of the major compounds is expressed as percentage peak area (%) based on the total area of the selected peaks in the chromatograms. The molecular formulae, relative peak area of selected compounds

Fig. 4. FTIR spectra. (a) Extracted cellulose. (b) Cotton stalk.

199

from both liquid oils were listed in Table 2, except for those compounds with relative peak area less than 0.2%. Similarly, the compounds contained in both liquid oils were classified as aromatic hydrocarbons, phenols, alkanes and some other oxygen containing compounds such as alcohols, ketones, and esters, among which aromatic hydrocarbons, alkanes and phenols were the three major compounds. Interestingly, compounds such as levoglucosan, furfural, and aldehydes, which abounded in the bio-oils obtained from cellulose by pyrolysis [2,7,8], were not found in the present cellulose oil. This may be the remarkable difference between the technique of deoxy-liquefaction and several kinds of pyrolysis methods. Furthermore, either the compounds or their relative contents varied between both oils, which can be attributed to the differences in components between the two kinds of feedstock. The relative content of aromatic hydrocarbons in extracted cellulose oil was 14.12%, which was lower than 20.10% in cotton stalk oil. These aromatic hydrocarbons can be classified into monocyclic aromatic hydrocarbons and polycyclic aromatic hydrocarbons (PAHs) with two rings. In the cotton stalk oil, there were 18 kinds of monocyclic aromatics, while only 10 kinds were found in the cellulose oil. Besides the similar kinds such as benzene, toluene, xylene, ethyl-benzene and trimethyl-benzene, the cotton stalk oil also included some complicated aromatics such as 4-ethyl-1,2dimethyl-benzene, 1-ethenyl-4ethyl- benzene, 1-methyl-4-(1-me thyl-2-propenyl)-benzene, 3-ethyl-1,2,4,5-tetramethyl-benzene, indane, and indene. Moreover, the PAHs were similarly characterized as methyl-naphthalene, dimethyl-naphthalene, trimethylnaohthalene, tetrahydro-methyl-naphthalene and tetrahydro-dim ethyl-naphthalene, etc. Generally speaking, the aromatics mainly derived from the decomposition of lignin. For example, some aromatic hydrocarbons can be obtained during the process of deoxygenation of oxygenated aromatic compounds like catechol and pyrogallol produced from lignin decomposition. However, the aromatic hydrocarbons can also form in the gas phase by secondary polymerization of unsaturated light compounds during the cellulose degradation [27]. This is the reason why the extracted cellulose oil still contained some aromatic hydrocarbons. More remarkably, long-chain alkanes with carbon distribution range of C14AC21 existed in both liquid oils. Furthermore, most of the alkanes were straight chain alkanes, which were found to be similar to that in diesel oil. The relative content of alkanes was 12.43% in cotton stalk oil and 10.72% in extracted cellulose oil, respectively. As discussed in the previous research [11], alkanes in the oils obtained by deoxy-liquefaction may derive from the decomposition of extractives in the biomass. Therefore, it is reasonable that the relative content of alkanes in cotton stalk oil was higher than that in extracted cellulose oil. However, it was difficult to explain the formation of long-chain alkanes from extracted cellulose after separation of the extractives from cotton stalk. But one point was certain that removal of great amount of oxygen in feedstock could be one of the most important factors for the formation of long-chain alkanes. This indicated that deoxy-liquefaction technique which could produce the liquid oil with low oxygen content was quite different from the conventional pyrolysis [2,7]. The oxygen content of the oil was mostly ascribed to the phenol derivatives, which were 41.12% and 33.96% in the oils from cotton stalk and extracted cellulose, respectively. It was proposed that decomposition of lignin led to the production of numerous phenolic compounds [9]. So the higher content of phenol derivatives in cotton stalk oil was consistent with the high content of lignin in the feedstock. However, Wang et al. [10] postulated that the formation of phenolic compounds could also be formed from cellulose by dehydration, deoxygenation, condensation or cyclization. This can explain why there were still relatively high content of phenolic compounds in the extracted cellulose oil after removal of lignin

200

J. Li et al. / Fuel 166 (2016) 196–202

Table 2 GC–MS characterization of typical compounds of the liquid oils from extracted cellulose and cotton stalk. Compound name

Formula

Benzene Toluene o-Xylene p-Xylene Ethylbenzene Indane Benzene, 1-ethyl-3-methylBenzene, 1-ethyl-2-methylBenzene, 1,2,3-trimethylBenzene, 4-ethyl-1,2-dimethylBenzene, 1-methyl-4-(2-propenyl)Benzene, 1-ethenyl-4-ethyl1H-Indene, 2,3-dihydro-4,7-dimethylBenzene, 1-methyl-4-(1-methyl-2-propenyl)Naphthalene, 1,2,3,4-tetrahydro-5-methylNaphthalene, 1-methylNaphthalene, 2-methylNaphthalene, 1,4-dimethylNaphthalene, 2,3-dimethylNaphthalene, 1,2,3,4-tetrahydro-1,8-dimethylNaphthalene, 1,2,3,4-tetrahydro-6,7-dimethylNaphthalene, 1,4,6-trimethylBenzene, hexamethylNaphthalene, 1,4,5-trimethylNaphthalene, 1,2-dihydro-2,5,8-trimethylBenzene, 2-(2-butenyl)-1,3,5-trimethylNaphthalene, 1,2,3,4-tetrahydro-6-propylNaphthalene, 1,2,3,4-tetrahydro-1,5,7-trimethylNaphthalene, 1,2,3,4-tetrahydro-2,5,8-trimethylPhenol Phenol, 2-methylPhenol, 3-methylPhenol, 3-ethylPhenol, 2-ethylPhenol, 3,5-dimethylPhenol, 4-ethylPhenol, 2,4-dimethylPhenol, 2,6-dimethylPhenol, 3,4-dimethylPhenol, 2,3,5-trimethylPhenol, 2-ethyl-5-methylPhenol, 3-ethyl-5-methylPhenol, 3-(1-methylethyl)Phenol, 3-ethyl-5-methylPhenol, 3,4,5-trimethylPhenol, 2,3,6-trimethyl2,5-Diethylphenol 3,4-Diethylphenol Phenol, 2-methyl-5-(1-methylethyl)Phenol, 4-(1-methylpropyl)Phenol, 2,3,5,6-tetramethylPhenol, 2-(1,1-dimethylethyl)-5-methylUndecane Tetradecane Pentadecane Hexadecane Heptadecane Octadecane Nonadecane Eicosane Heneicosane 2-Pentanone 2-Hexanone Ethanone, 1-(1-cyclohexen-1-yl)2-Cyclopenten-1-one, 2,3,4,5-tetramethyl2-Decanone Cyclohexanone, 2-(2-methylpropylidene)1-Methylindan-2-one Benzofuran, 4,7-dimethylEthanone, 1-(4-ethylphenyl)1H-Inden-1-ol, 2,3-dihydro-2-methylEugenol 1-Naphthalenol

C6H6 C7H8 C8H10 C8H10 C8H10 C9H10 C9H12 C9H12 C9H12 C10H14 C10H12 C10H12 C11H14 C11H14 C11H14 C11H10 C11H10 C12H12 C12H12 C12H16 C12H16 C13H14 C12H18 C13H14 C13H16 C13H18 C13H18 C13H18 C13H18 C6H6O C7H8O C7H8O C8H10O C8H10O C8H10O C8H10O C8H10O C8H10O C8H10O C9H12O C9H12O C9H12O C9H12O C9H12O C9H12O C9H12O C10H14O C10H14O C10H14O C10H14O C10H14O C11H16O C11H24 C14H30 C15H32 C16H34 C17H36 C18H38 C19H40 C20H42 C21H44 C5H10O C6H12O C8H12O C9H14O C10H20O C10H16O C10H10O C10H10O C10H12O C10H12O C10H12O2 C10H8O

Relative peak area ECO

CSO

2.19 1.50 1.52 – – – – – 0.97 – 0.95 0.89 0.54 – 0.37 – – – 0.36 0.51 3.07 – – – – 0.21 0.21 0.47 0.36 3.02 3.56 4.53 0.52 1.89 3.34 1.87 2.54 1.69 1.23 0.83 0.74 1.59 0.72 1.23 0.76 0.60 0.58 – – – 1.56 – – 1.87 1.47 2.01 1.41 1.32 1.49 0.81 1.98 – – 2.32 1.49 2.07 0.44 1.21 – 0.58 1.44 – –

0.43 1.11 0.61 1.17 0.77 0.90 0.48 0.34 0.85 0.42 0.47 1.28 0.83 0.37 – 1.16 2.37 0.74 – – – 0.48 1.10 0.72 0.46 – – – – 2.40 3.22 3.17 1.46 2.03 2.67 2.43 3.02 1.24 2.16 1.45 1.70 1.58 – 1.61 1.58 1.82 2.32 0.43 0.56 1.78 2.02 0.47 0.34 1.93 – 1.85 1.76 1.97 1.95 – 0.92 0.35 0.44 1.76 – – – – 2.57 – 1.99 0.99

201

J. Li et al. / Fuel 166 (2016) 196–202 Table 2 (continued) Compound name

1-Naphthalenol, 2-methyl1-Dodecanol, 3,7,11-trimethyl2-Pentadecanone, 6,10,14-trimethyl2-Nonadecanone n-Nonadecanol-1 Heptadecanoic acid, methyl ester Nonadecanoic acid, methyl ester Behenic alcohol Eicosyl acetate

Formula

Relative peak area

C11H10O C15H32O C18H36O C19H38O C19H40O C18H36O2 C20H40O2 C22H46O C22H44O2

ECO

CSO

– 1.39 0.61 – 0.72 0.61 1.09 1.47 0.45

0.92 – – 0.86 – 0.99 – – –

ECO = extracted cellulose oil. CSO = cotton stalk oil. ‘‘–”: Not detected

from cotton stalk. More interestingly, almost all the phenolic compounds in both oils contained methyl-, dimethyl-, trimethyl-, tetramethyl-, ethyl-, and diethyl-groups. This is quite different from those found in the oils produced by pyrolysis reactions [28,29], in which the phenolic products usually featured methoxy groups, such as 2-methoxy-phenol, 4-ethyl-2-methoxy-phenol and 4-methoxy-phenol. This indicated that deoxy-liquefaction reaction was favorable for the demethoxylation of aromatic rings. The energy provided by deoxy-liquefaction was high enough to break the aromatic CAO of methoxyl group (356 kJ/mol) [30]. While, the CAO bonds in phenolic hydroxyl groups could not be broken completely owning to the high bond energy (414 kJ/mol), resulting in the presence of certain amount of phenolic compounds, which also confirmed the suitability of the liquid oil to be considered for value-added chemicals. In addition, some other oxygen-containing compounds such as alcohols, ketones, esters and acids were also found in both oils. For example, 1-(1-cyclohexen-1-yl)-ethanone, 2-decanone, 1-methylindan-2-one, 1-(4-ethylphenyl)-ethanone, n-nonadeca nol-1, heptadecanoic acid-methyl-ester, nonadecanoic acidmethyl-ester, behenic alcohol, etc. may derive from the decomposition of cellulose [2]. While, eugenol, 1-naphthalenol and 2-methyl-1-naphthalenol, etc. which probably came from lignin were found only in the cotton stalk oil. 3.2.3. FTIR analysis The FTIR spectra of the liquid oil from extracted cellulose in Fig. 5 showed typical peaks. The stretching vibrations of 3430 cm1 were characteristic of OAH group in phenols and alcohols. The spectrum peaks of 2924, and 1460, 1380 cm1 were characteristic of the stretching and bending vibrations of CAH, respectively. Additionally, the absorptions at 1160 and 1110 cm1 can be attributed to the CAO stretching existing in the primary, secondary and tertiary alcohols and phenols. The absorptions of C@O stretching vibrations at 1700 cm1 indicated the presence of ketones in the liquid oil. The absorbance peak at 1652 cm1 were due to the stretching vibrations of C@C groups in aromatics and the absorptions of substituted aromatic groups were also observed at 750 and 620 cm1. So it can be concluded that aromatics, phenols, alkanes, alchhols, and ketones coexisted in the liquid oils, which was in good agreement with the result of GC–MS analysis. 4. Conclusions Cellulose extracted from cotton stalk was characterized and converted into the liquid oil by direct deoxy-liquefaction. This liquid oil was further compared with that obtained from cotton stalk feedstock through GC–MS analyses. The most remarkable difference was that the relative content of the aromatic derivatives including aromatic hydrocarbons and phenolic compounds in

Fig. 5. FTIR spectra of the liquid oil from extracted cellulose.

extracted cellulose oil was lower than that in cotton stalk oil. This can be attributed to the removal of the lignin from cotton stalk. The results showed that the types and contents of the compounds in the oils were closely connected with the components of feedstock. Attractively, the extracted cellulose oil featured lower oxygen content (6.46%) and higher heating value (42.66 MJ/kg), which indicated that cellulose can produce high-quality oil via deoxyliquefaction. And this oil can be upgraded to become a potentially renewable fuel.

Acknowledgments This study was financially supported by National Natural Science Foundation of China (Nos. 21475071 and 21275082) and Beijing National Laboratory for Molecular Sciences (BNLMS).

References [1] Cao XF, Zhong LX, Peng XW, Sun SN, Li SM, Liu SJ, et al. Comparative study of the pyrolysis of lignocellulose and its major components: characterization and overall distribution of their biochars and volatiles. Bioresour Technol 2014;155:21–7. [2] Shra’ah AA, Helleur R. Microwave pyrolysis of cellulose at low temperature. J Anal Appl Pyrol 2014;105:91–9. [3] Lin T, Goos E, Riedel U. A sectional approach for biomass: modeling the pyrolysis of cellulose. Fuel Process Technol 2013;115:246–53. [4] Fukuoka A, Dhepe PL. Catalytic conversion of cellulose into sugar alcohols. Angew Chem Int Ed 2006;45:5161–3. [5] Amarasekara AS, Ebede CC. Zinc chloride mediated degradation of cellulose at 200 °C and identification of the products. Bioresour Technol 2009;100:5301–4.

202

J. Li et al. / Fuel 166 (2016) 196–202

[6] Wei XL, Wang Z, Wu Y, Yu ZM, Jin J, Wu K. Fast pyrolysis of cellulose with solid acid catalysts for levoglucosenone. J Anal Appl Pyrol 2014;107:150–4. [7] Wang SR, Guo XJ, Liang T, Zhou Y, Luo ZY. Mechanism research on cellulose pyrolysis by Py-GC/MS and subsequent density functional theory studies. Bioresour Technol 2012;104:722–8. [8] Shen DK, Gu S. The mechanism for thermal decomposition of cellulose and its main products. Bioresour Technol 2009;100:6496–504. [9] Li JH, Wang C, Yang ZY. Production and separation of phenols from biomassderived bio-petroleum. J Anal Appl Pyrol 2010;89:218–22. [10] Wang C, Pan JX, Li JH, Yang ZY. Comparative studies of products produced from four different biomass samples via deoxy-liquefaction. Bioresour Technol 2008;99:1778–86. [11] Wu LB, Guo SP, Wang C, Yang ZY. Production of alkanes (C7–C29) from different part of poplar tree via direct deoxy-liquefaction. Bioresour Technol 2009;100:2069–76. [12] Dong Z, Hou XL, Sun FF, Zhang L, Yang YQ. Textile grade long natural cellulose fibers from bark of cotton stalks using steam explosion as a pretreatment. Cellulose 2014;21:3851–60. [13] El-Sakhawy M, Hassan ML. Physical and mechanical properties of microcrystalline cellulose prepared from agricultural residues. Carbohydr Polym 2007;67:1–10. [14] Song YL, Zhang JZ, Zhang X, Tan TW. The correlation between cellulose allomorphs (I and II) and conversion after removal of hemicellulose and lignin of lignocellulose. Bioresour Technol 2015;193:164–70. [15] Van Soest PJ. Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fiber and lignin. J Assoc Off Agric Chem 1963;46:829–35. [16] Jiang M, Zhao MM, Zhou ZW, Huang T, Chen XL, Wang Y. Isolation of cellulose with ionic liquid from steam exploded rice straw. Ind Crops Prod 2011;33:734–8. [17] Chen YG, Guo SP, Wang C, Yang F, Yang ZY. Production and evaluation of hydrocarbon oil from co-deoxy-liquefaction of waste lard and locust leaves. Fuel 2012;93:528–32. [18] Scholze B, Meier D. Characterization of the water-insoluble fraction from pyrolysis oil (pyrolytic lignin). Part I: PY-GC/MS, FTIR, and functional groups. J Anal Appl Pyrol 2001;60:41–54.

[19] Reddy N, Yang YQ. Properties and potential applications of natural cellulose fibers from the bark of cotton stalks. Bioresour Technol 2009;100:3563–9. [20] Du SK, Zhu XN, Wang H, Zhou DY, Yang WH, Xu HX. High pressure assist-alkali pretreatment of cotton stalk and physiochemical characterization of biomass. Bioresour Technol 2013;148:494–500. [21] Ibrahim MM, Agblevor FA, El-Zawawy WK. Isolation and characterization of cellulose and lignin from steam-exploded lignocellulosic biomass. Bioresources 2010;5(1):397–418. [22] Wang D, Shang SB, Song ZQ, Lee MK. Evaluation of microcrystalline cellulose prepared from kenaf fibers. J Ind Eng Chem 2010;16:152–6. [23] Adel AM, Abd EI-wahab ZH, Ibrahim AA, Al-Shemy MT. Characterization of microcrystalline cellulose prepared from lignocellulosic materials. Part III: physicochemical properties. Carbohydr Polym 2011;83:676–87. [24] Su YP, Du RY, Guo H, Cao M, Wu QF, Su RX, et al. Fractional pretreatment of lignocellulose by alkaline hydrogen peroxide: characterization of its major components. Food Bioprod Process 2015;94:322–30. [25] Czernik S, Bridgwater AV. Overview of applications of biomass fast pyrolysis oil. Energy Fuels 2004;18:590–8. [26] Mullen CA, Boateng A, Hicks KB, Goldberg NM, Moreau RA. Analysis and comparison of bio-oil produced by fast pyrolysis from three barley biomass/ byproduct streams. Energy Fuels 2010;24:699–706. [27] Evans RJ, Milne TA. Molecular characterization of the pyrolysis of biomass. Energy Fuels 1987;1:123–37. [28] He R, Ye XP, Engligh BC, Satrio JA. Influence of pyrolysis condition on switchgrass bio-oil yield and physicochemical properties. Bioresour Technol 2009;100:5305–11. [29] Butt DAE. Formation of phenols from the low-temperature fast pyrolysis of Radiata pine (Pinus radiate): Part I. Influence of molecular oxygen. J Anal Appl Pyrol 2006;76:38–47. [30] Lu WP, Ma M, Zhang XD, Zhang B, Guo YC, Wang C. Comparison of high calorific fuels obtained from fresh and dried Hydrilla verticillata: effects of temperature, residence time, catalyst content and material state. J Anal Appl Pyrol 2015;111:76–87.