Enhancement of methane production from Cotton Stalk using different pretreatment techniques

China produces large amount of cotton stalk (CS) residues as agricultural biomass, which are incinerated on-site, causing air pollution. The high organic content of CS could be utilized for biogas production, but the direct digestion without pretreatment always leads to a low methane yield and biodegradability, due to the complicated structure of lignocellulose. In order to search best fitting pretreatment methods in effective anaerobic digestion (AD) of CS, effects of various pretreatments including KOH, NaOH, Ca(OH)2, alkali hydrogen peroxide (AHP), H2SO4, H3PO4 and steam explosion (SE) were studied. It was seen that all treatments resulted in varying methane yields. Among all the pretreatments, acid pretreatment is not suitable for AD of CS. The results showed that the highest cumulative methane yield (CMY) of 192.4 mL·gVS−1 was obtained after 3% AHP pretreatment of CS, and the methane yield improved by 254.3% than the untreated CS. Therefore, AHP treatment was proven to be an efficient pretreatment technique. XRD and FTIR analyses had shown that pretreated CS had favorable structural changes. This research is beneficial in developing environment friendly and cost-effective pretreatment technologies to utilize CS for methane production in future application.

pretreatment has gained popularity in recent years, because it is one of the most promising chemical method for the removal of lignin 18 . Sun et al. 18 compared five stalks (wheat, rice, maize, rape and cotton stalks) by 2% AHP pretreatment for improving methane yield, and the methane yield of CS improved 55.7% than the untreated CS. As for physical pretreatment, steam explosion (SE) is also an attractive option. Li et al. 12 found that the cumulative methane yield (CMY) of SE pretreated corn stalk at 1.2 MPa for 10 min could improve 63.9% of methane yield compared with untreated corn stalk. Although many researchers have investigated crop straw, the effects of various pretreatments differ considerably, therefore, the most economical and effective pretreatment method has not been fully explored. Additionally, few reports have shown the pretreatment effects on CS for improving methane yield. Such information is important for evaluation of pretreatments and their practical application. Therefore, a systematic study regarding the impacts of different pretreatment methods on CS for methane production is highly demanded.
The objectives of this research were to: (1) carry out KOH, NaOH, Ca(OH) 2 , AHP, H 2 SO 4 , H 3 PO 4 , SE pretreatment to evaluate their varying effects on CS, (2) systematically and comprehensively compare the AD performance of CS after different pretreatments, (3) verify structural changes of different pretreated CS.

Methods
Substrates. CS was obtained from a farm in Henan province, China. The raw CS was cut with a chaff-cutter to less than 5 cm in length and milled to 18-mesh powder by an high-speed grinder (XINGSHILIHE, Beijing, China), and then was put in airtight plastic bag at room temperature for future use. The inoculum sludge was collected from an anaerobic digester at Donghuashan biogas plant in Beijing, China.
1 dry matter weight of CS weight of CS water added 100% (1)

Maximal methane yield (MMY), experimental methane yield (EMY) and biodegradability (B d ).
Maximal methane yield (MMY) of CS was calculated by Eqs (2) and (3) 20 , and B d was analyzed according to Eq. (4) 21 . Methane yield. The daily biogas production was determined by measuring headspace pressure in each digester using 3151 WAL-BMP-Test system pressure gauge (WAL Mess-und Regelsysteme GmbH, Germany). The composition of biogas was determined by using a 7890 B gas chromatograph (Agilent, USA). Biogas production was calculated by Eq. (5) 25 .
where V biogas is daily biogas volume (L), P refers to absolute pressure (kPa), V head represents volume of the head space (L), C is molar volume (22.4 L·mol −1 ), R means universal gas constant (8.314 J·mol −1 ·k −1 ), and T is equal to absolute temperature (K).

Pretreatment solution and biogas slurry.
For pretreatment solution, the soluble chemical oxygen demand (COD) was determined by using HACH meter (DR2800 spectrophotometer). The soluble saccharides were measured by using phenol-sulfuric acid method with a spectrophotometer (UV-1800PC, Mapada, China) at 490 nm. Total alkalinity (TA), total ammonia nitrogen (TAN), total volatile fatty acid (TVFA), pH value of the biogas slurry after digestion were analyzed. The pH value was measured with a le438 pH electrode (Mettler Toledo, USA). TAN was determined using HACH meter (DR2800 spectrophotometer). TA was observed by titrimetric method with 1.6 mol·L −1 of sulfuric acid. TVFA was measured by gas chromatography (GC) system (7890A, Agilent, USA).
Kinetics model. In this study, the CMY has been modeled by employing modified Gompertz equation (6), which used to analyze the results of AD 26 0 m 0 where B refers to the simulated cumulative methane production (mL·gVS −1 ), B 0 represents the simulated highest methane yield (mL·gVS −1 ), μ m stands for the maximum methane production rate (mL·gVS −1 ·day −1 ), e is equal to 2.72, λ is the lag phase time (day), and t means the digestion time (day).
Micro-structure observation. XRD analysis. The crystallinity index (CrI) of the cellulose component, which was obtained under different pretreatment, was measured by X-ray diffraction (XRD). The Bruker D8-Advance (Germany) at 40 KV and 40 mA with Cu Kα radiation was used to examine the XRD patterns. Samples were scanned from 5 to 60° at a rate of 5°/min. CrI of these pretreated CS was calculated by Eq. (7)  where I 002 is the maximal diffraction intensity of the 002 lattice plane, I amorphous refers to the intensity of the background peak at 2θ = 18°. Statistical analysis. All the data and graphs were processed using Origin Pro 8.0 (Origin Lab, USA).
Analysis of variance (ANOVA) was used to statistically analyze the data.

Results and Discussion
Characterization of CS and inoculum. Table 1 shows the basis characteristics of raw CS and inoculum sludge. The VS/TS ratio of CS reached 96.75%, indicating a high organic matter, which was preferred for AD. Moreover, the content of cellulose and lignin was up to 50.42% and 16.32%, higher than some other widely used crop straw and biomass 21 . According to the data of the elements analysis in Table 1, the MMY of CS was calculated using Eqs (2) and (3) to be 362.9 mL·gVS −1 .
Effects of different pretreatment methods on cumulative methane yield. Figure 1a and b show the CMY of H 2 SO 4 -treated CS and H 3 PO 4 -treated CS. In these groups, two different peaks of daily methane yield occurred during the digestion process of H 2 SO 4 -treated CS and H 3 PO 4 -treated CS, similar results were reported in the pretreatment of corn stover in anaerobic digestion process 9,28,29 . At the beginning, the daily methane yield of H 2 SO 4 -treated CS and H 3 PO 4 -treated CS quickly reached the first peak on 3 rd or 4 th day, which present in Figure S1 (supplementary data  Fig. 1c. The two highest cumulative methane production at 1.2 MPa for 15 min and 1.5 MPa for 5 min were 177.3 mL·gVS −1 and 171.8 mL·gVS −1 respectively, which showed no significant difference (P > 0.05), and improved 226.6%, 216.4% than the untreated CS. The maximum daily methane yield was 25.6 mL·gVS −1 on 2 nd day at 1.2 MPa for 15 min which was 7.5 times higher than the untreated CS (see Supplementary Fig. S1). Nevertheless, considering the costs and energy consumption, the pressure of 1.5 MPa for 5 min was the best choice in SE pretreatment of CS. When the pretreated pressure was 0.9 MPa and 1.2 MPa, the CMY increased with the rising of pretreatment time which was completely contrary to the data of 1.5 MPa. The difference indicated that some inhibitors might produce at the severe conditions. Moreover, SE pretreatment at 1.5 MPa for 15 min showed the peak of maximum daily methane yield on 9 th day which was quite different from other groups (see Supplementary Fig. S1). It also implied that the pretreatment process had some inhibitory effects on AD.
The CMY were 120.3 mL·gVS −1 , 117.7 mL·gVS −1 , 118.2 mL·gVS −1 from 1%, 2%, and 3% Ca(OH) 2 -treated of CS (Fig. 1d), which had no significant difference (P > 0.05) in CS pretreated by the concentration of 1%, 2%, 3%. The methane yield of KOH-treated CS at the concentration of 3% reached maximum at 177.1 mL·gVS −1 , and improved 226.14% than the untreated CS (Fig. 1e). When the KOH concentration increased to 6%, the CMY reduced to 59.3 mL·gVS −1 , which had no significant difference (P > 0.05) as compared to the untreated CS. This might be because many of the degradable organic matters in CS were dissolved by the pretreatment of 6% KOH. The maximal methane yield was found from 1.5% NaOH-treated CS (178.6 mL·gVS −1 ), which showed a highly significant (P < 0.01) reduction with increase of concentration of NaOH (Fig. 1f). It can be seen that the highest methane yield of CS pretreated by KOH and NaOH was superior to Ca(OH) 2 -treated CS. Considering the efficiency and cost of the pretreatment process in AD, 3%, 1.5%, 1% were an optimal concentration for KOH, NaOH, Ca(OH) 2 pretreatment, respectively.
Comparison of different pretreatments effect. The optimal results and kinetics of methane production after different pretreatments. The optimum results for each of the seven pretreatment processes are shown in Fig. 2a. The results showed that the highest CMY obtained from 3% AHP-pretreated CS, and improved 254.3% than the untreated CS. The results had no significant difference (P > 0.05) compared to SE pretreated CS at 1.5 MPa for 5 min, 3% KOH-treated CS, and 1.5% NaOH-treated CS. The CMY of 1% Ca(OH) 2 -treated CS was 47.3% and 48.5% lower than 3% KOH-treated CS and 1.5% NaOH-treated CS because Ca(OH) 2 was less soluble in water. No significant difference (P > 0.05) within the CMY was observed between 1% H 2 SO 4 -treated CS and 4% H 3 PO 4 -treated CS. Obviously, the CMY of H 2 SO 4 and H 3 PO 4 treated CS was inferior to KOH, NaOH, Ca(OH) 2 , AHP, and SE pretreated CS. Moreover, the CMY of 2% H 3 PO 4 -treated CS reduced 20.1% than the untreated CS due to the long lag period. Therefore, the results implied that H 2 SO 4 and H 3 PO 4 pretreatment was not suitable for CS in AD process.
The selected optimum results were fitted by the Modified Gompertz model for statistical analysis (Fig. 2b). The kinetic parameters of the model are shown in Table 2. The correlation coefficients R 2 values fell within the range of 0.967-0.995, respectively. It indicated that the Modified Gompertz model was well fitted to the CMY in all selected experimental groups. Moreover, the B 0 values of the model were very close to the EMY at different pretreated conditions. The parameters of B 0 and μ m showed great enhancements than the untreated CS which illustrated that all pretreatment in this research could enhance the ultimate methane production and the maximum methane production rate. The start-up time of AD process is shown by the lag phase time (λ). Obviously, a lower λ means a short time of start-up. The λ of 1% H 2 SO 4 -treated CS and 4% H 3 PO 4 -treated CS was longer than the untreated CS, implying that a long lag period appeared which was consistent with the results of Fig. 1a,b. Compared with the untreated CS, the λ of other pretreated CS significantly reduced, especially CS pretreated at 1.5 MPa for 5 min which reduced from −0.21 to −6.97. The highest B d values after pretreating by 3% AHP was 53.02% which implied that the digestion efficiency still had room to be further improved.  The COD and soluble saccharides were the indicators of degradable and solubilized lignin and carbohydrate in pretreatment solution 30 . From Table 3, the concentration of COD and saccharides followed the same trends. However, the value of COD and soluble saccharides showed significant difference, which indicated that they contained some other organic compounds such as aromatic acids, ethanol, furfural derivatives, and uronic acids in pretreatment solution 18 . The two highest COD concentrations appeared in 3% KOH and 1.5% NaOH loading in the black liquor were 13880 mg·L −1 and 14220 mg·L −1 , respectively, which indicated that KOH and NaOH could remove a large quantity of lignin and carbohydrate. The COD and soluble saccharides of 3% AHP pretreatment were 5760 mg·L −1 and 580 mg·L −1 in the liquid, less than KOH, NaOH, Ca(OH) 2 pretreatment, implying that AHP pretreatment could decrease the loss of degradable organic matters. Obviously, the pretreatment solution of KOH, NaOH, Ca(OH) 2 and AHP treated CS had much more COD and soluble saccharides than H 2 SO 4 and H 3 PO 4 treated CS. The results indicated that more lignin and carbohydrate could be deconstructed by alkali pretreatment compared with acid pretreatment.
The digestion process stability was shown by the characteristics of pH values, TA, TAN, and TVFA in biogas slurry after AD ( Table 3). The pH values of all pretreatment were in range of 7.36-7.67, which were suitable for the growth of bacteria. TA indicates the buffering capacity to neutralize acids, which could protect against the low pH of AD system. Some researches held that the TAN was one of the most important indicators in the process of    AD because the ammonia could cause the accumulation of VFA, which resulted in a sharp drop of pH 31 , and then affected the activity of microbes. However, the pH value did not drop and accumulation of TVFA and inhibition of microbial activity were not observed in this study. Besides, the concentration of TAN in blank which contained the same amount of sludge and water was about 1500 mg·L −1 . Therefore, effluent parameters collectively suggested that the AD system after different pretreatments were stable, and pretreatment didn't influence the anaerobic digestion system stablity.
Analysis of modifications in the chemical structure followed by different pretreatments.
(a) XRD Analysis XRD was employed to determine and calculate the crystallinity index of cellulose after pretreatment (Fig. 3). There existed similar peaks of raw CS and pretreated CS at 18°, 22° and 35°, which were obtained from cellulose I 32 , illustrated that cellulose crystalline allomorph did not change after the pretreatment. The CrI of untreated CS was 34.1% and it increased to 50.4%, 51.3%, 51.2%, and 47.0% after 3% KOH, 1.5% NaOH, 3% AHP, and SE pretreatment at 1.5 MPa for 5 min, respectively. This was probably because a great proportion of non-crystalline components, including lignin and hemicellulose, were removed as a result of these pretreatments. An increase in the portion of exposed crystalline structure was noticed in the pretreated samples, hence, the CrI of pretreated CS improved compared with the untreated CS. These results were in conformity with the previous studies 7, 12,16 .
The FTIR spectra of raw CS and pretreated CS was shown in Fig. 4. The wide band at around 3418.5 cm −1 is linked with the level of intermolecular and intramolecular hydrogen bonding 33 , whereas, the peak at 2920.0 cm −1 with shoulder bands is due to aliphatic CH 2 and symmetric and asymmetric CH 3 , respectively 14,33 . The absorbance at 1739.2 cm −1 was owing to the ester groups, indicating the sign existed in lignin 12  SE pretreatment has lower pollution, however it is not highly recommended, due to higher energy consumption. Besides, many inhibitory compounds such as furfural and HMF might be generated from SE pretreatment process 34 . The appropriate conditions such as pressure and time should be chosen, otherwise the process might influence the subsequent digestion reactions. The compared results showed that KOH and NaOH were both suitable for pretreating CS in AD process. But Na + ion from the pretreatment of NaOH might influence the digestion process, causing the inhibition of methanogenesis 10 . Moreover, the liquid after NaOH pretreatment could result in serious environment problems such as soil salinization and water pollution 10 . On contrary, the waste liquid containing KOH might have environmental benefits, including the production of fertilizer 10,30 . KOH is much more expensive than Ca(OH) 2 and NaOH. Therefore, it is possible to consider that using combined KOH and Ca(OH) 2 alkaline pretreatment to reduce the cost by decreasing the KOH dosage and maintaining the efficiency of anaerobic digestion 16 . Compared with NaOH pretreatment, AHP pretreatment could considerably decrease the dosage of NaOH, and reduce Na + ion pollution. Furthermore, hydrogen peroxide in the pretreatment solution could be easily decomposed into water and oxygen to avoid secondary pollution. Therefore, AHP pretreatment was considered to be the promising method for treating CS.

Conclusion
This study compared the performance of various pretreated CS in anaerobic digestion. Acid treated CS could not achieve high methane yield due to the long lag period. Therefore, H 2 SO 4 and H 3 PO 4 are not suitable to pretreat CS in AD process. Although SE pretreatment has given high methane yield, but the treatment is not suitable due to the higher energy consumption and lack of availability of the industrial scale equipment, because it is hard to build big continuous machine with low cost. The efficiency of AD process after KOH and NaOH pretreatment was better than Ca(OH) 2 pretreatment. Results indicated that CS could efficiently be converted to methane after AHP pretreatment at concentration of 3%. In summary, AD after AHP pretreatment could be an effective method to improve the CS utilization. Besides, this study offers a systemic insight into the biogas productions from pretreated CS and provides useful information on the pretreatment of CS in a cost effective and environment friendly way.