Bioresource Technology 187 (2015) 149–160

Contents lists available at ScienceDirect

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

A biorefining process: Sequential, combinational lignocellulose pretreatment procedure for improving biobutanol production from sugarcane bagasse Haifeng Su a,1, Gang Liu b,1, Mingxiong He a, Furong Tan a,⇑ a b

Biogas Institute of Ministry of Agriculture, Chengdu 610041, Sichuan, PR China Sichuan Academy of Grassland Science, Xipu Chengdu 611731, Sichuan, PR China

h i g h l i g h t s  A biorefining process: SCLPP was designed to pretreat lignocellulose.  Improving biobutanol production from sugarcane bagasse.  Enzymatic hydrolysis with simultaneous saccharification fermentation was effective.  Developing SCLPP that increase biofuel from lignocellulose is feasible.

a r t i c l e

i n f o

Article history: Received 4 February 2015 Received in revised form 22 March 2015 Accepted 23 March 2015 Available online 28 March 2015 Keywords: Pretreatment Sugarcane bagasse Fermentation Biobutanol

a b s t r a c t Here, for the first time, we designed a sequential, combinatorial lignocellulose pretreatment procedure (SCLPP) for microbial biofuel fermentation to reduce generation of microbial growth inhibitors and furthermore increase sugar yields. We tested this pretreatment process using sugarcane bagasse as substrate and assessed the effectiveness by analysis of biobutanol production through microbial clostridium beijerinckii NCIMB 8052 conversion. Our results showed that there were no inhibitory effects when using the hydrolysates as fermentation substrate. Under the SSF scheme, we observed the highest concentrations of butanol (6.4 g/L) and total ABE (11.9 g/L), resulting in a higher ABE productivity, compared with the SHF method. These findings suggest that the SCLPP is a feasible method for improving ABE production, lowering microbial inhibitor generation, and ensuring success in the subsequent fermentation process. Therefore, our work demonstrated developing a tractable integrated process that facilitates to increase biofuel production from agricultural residues rich in lignocellulose is feasible. Ó 2015 Elsevier Ltd. All rights reserved.

1. Background Several well-established pretreatment methods are available to convert cellulosic and hemicellulosic materials into sugars via hydrolysis. Microbial fermentation processes can convert these Abbreviations: SCLPP, sequential, combinational lignocellulose pretreatment procedure; LHW, liquid hot water pretreatment; HMF, hydroxymethylfurfural; HAN, HTR and HPJ, indicate the total hydrolysates were used as fermentation substrates of SHF; HANNY, HTRNY and HPJNY, indicate the total hydrolysates were used as fermentation substrates of SSF; AN ATCC 1015, A. niger ATCC 1015; TR ATCC 26921, T. reesei ATCC 26921; PJ ATCC 44750, P. janthinellum ATCC 44750; AR2, solid residue AN1 + solid residue AN2; TR2, solid residue TR1 + solid residue TR2; PR2, solid residue PJ1 + solid residue PJ2; AAR3, solid residue AAN; ATR3, solid residue ATR; APR3, solid residue APJ; EAR4, solid residue EAN; ETR4, solid residue ETR; EPR4, solid residue EPJ. ⇑ Corresponding author. E-mail addresses: [email protected] (G. Liu), [email protected] (F. Tan). 1 Equally contributed to this work. http://dx.doi.org/10.1016/j.biortech.2015.03.107 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.

sugars into biofuels (Brown and Brown, 2013; Kumar et al., 2009; Yang and Wyman, 2008). Recently, advances in industrial biotechnology have provided additional opportunities for economical utilization of agro-industrial residues, such as sugarcane bagasse, a complex material and a major by-product of the sugarcane industry. Sugarcane is one of the most prevalent forms of agricultural waste and is very abundant each year in southern China. The composition of sugarcane bagasse is approximately 50% cellulose, 25% hemicellulose and 25% lignin. Therefore, using sugarcane bagasse to produce biobutanol is a promising approach for biofuel production. Butanol has a higher energy value and lower hygroscopicity than ethanol. However, the feasibility of producing biobutanol from sugarcane bagasse also depends on other factors such as pretreatment and detoxification (Jönsson et al., 2013). These processes are very important because conventional pretreatment approaches like acid hydrolysis often generate toxic

150

H. Su et al. / Bioresource Technology 187 (2015) 149–160

substances (e.g., furfural and 5-hydroxymethylfurfural) that greatly inhibit the growth of bacteria. Diverse approaches are available to pretreat crop residues rich in cellulose and hemicellulose including steam explosion (Li et al., 2013b) and pretreatments with dilute acids such as phosphoric acid (de Vasconcelos et al., 2013) and sulfuric acid. Acid hydrolysis, especially when using dilute sulfuric acid, is a common and effective pretreatment method because it is simple and inexpensive (Jung et al., 2013). Generally, the hemicellulose fraction of sugarcane bagasse can be hydrolyzed to monomeric sugars by dilute sulfuric acid. However, the microbial inhibitors produced during acid hydrolysis will inhibit microbial growth during fermentation, thereby suppress the fermentation efficiency (Bamufleh et al., 2013). For example, such inhibitory compounds had been found to significantly suppress cell growth and biobutanol production in Clostridium beijerinckii (Guo et al., 2013; Zhang and Ezeji, 2013). Furfural, hydroxymethyl furfural and phenolic compounds are the major inhibitors of biofuel fermentation, (Hayashi et al., 2003; Shimizu et al., 2005). Due to their toxicities, it is essential to remove inhibitory compounds from hydrolysates prior to biobutanol fermentation. Unfortunately, detoxifying pretreated substrates is a complicated process. Although several detoxification approaches have been investigated previously, including lime treatment, evaporation, adsorption using ionic-exchange column chromatography or activated charcoal, and biological treatment (Jennings and Schell, 2011; Shen et al., 2013), most of which are either ineffective at toxin removal or cost prohibitive for industrial application due to high cost for waste disposal. At present, chemically degrading inhibitors is the most common detoxification technique because of its simplicity and low cost. However, the efficiency of toxin removal by chemical methods is dependent on the chemical structure similarity of the inhibitors, resulting in incomplete removal of toxic compounds. In addition, chemical detoxification leads to high salt ion concentrations in the fermentation liquids which can significantly inhibit microbial growth (Chen et al., 2008; Kim et al., 2009a) and result in a reduced biobutanol production. Therefore, to reduce or avoid production of inhibitor compounds in the pretreatment stage, it is crucial to select appropriate pretreatment methods. To be effective, a pretreatment process must break down fiber with a high efficiency, produce a sufficient quantity of sugar, prevent the dissipation of sugar out of the desired fraction (i.e. pentosan fraction), and limit the extent to which the pretreated material inhibits microbial growth during fermentation. Some pretreatment methods can avoid or reduce the production of inhibitors such as microwave pretreatment (Tyagi and Lo, 2013), enzymatic hydrolysis (Ju et al., 2013), and liquid hot water pretreatment (Li et al., 2013a). In particular, biological pretreatment using microbiological degradation has shown excellent success in removing inhibitors from laccase-treated hydrolysates, enabling the substrates to be used for fermentation without generating inhibitory phenolic compounds (Krastanov et al., 2013). Finally, to be economically feasible, the pretreatment process should also aim to minimize energy demands and costs associated with construction materials, treatment of process residues, and reduction in feedstock size. However, if only a single pretreatment method is used to help limit the production of inhibitors, the substrate will not be fully decomposed and the sugar production will not be very high. Therefore, to limit the generation of toxic substances as far as possible and completely break down substrate materials simultaneously, we believe combining various pretreatment methods is the optimal approach, instead of using just one or two. However, there are no reports integrating multiple (i.e., three or more) approaches to pretreat feedstock rich in lignocellulose.

In this study, using sugarcane bagasse as example feedstock, we designed a SCLPP involving a series of methods including microwave decomposition, enzyme hydrolysis, ammonia immersion, microbial decomposition, and liquid hot water pretreatment to limit the production of microbial inhibitors and obtain high sugar yields. Finally, we assessed the effectiveness of resulting hydrolysates to drive butanol production using two methods: separate hydrolysis and fermentation (SHF) and simultaneous saccharification fermentation (SSF). 2. Methods 2.1. Sugarcane bagasse Sugarcane bagasse (stumps) for the pretreatment experiments was collected from a farmer in Yulin (Guangxi province, China). The sugarcane juice was removed with a squeezer (ET-ZZJ83, Guangzhao, Guangdong) and the remaining sugarcane bagasse was dried at 65 ± 2 °C for 2 days. 2.2. Experimental design and pretreatment procedure of the SCLPP Experiments were conducted following the methodology illustrated in the pretreatment and fermentation process flow sheet (Fig. 1). The sugarcane bagasse was pretreated with five different methods of decomposition. The composition changes in and out processes were also presented in Fig. 1. Then the hydrolysis products were fermented by the C. beijerinckii strain NCIMB 8052. In addition, pure glucose and mixture sugars were also used to ferment. 2.2.1. Step 1: milling raw bagasse to powder The original dried sugarcane bagasse with lengths of 5–10 cm was partially broken down using a pulverizer (AB03, Weifang city, China). The resulting material was screened to obtain particles with a maximum size of 830 lm (size 20 mesh) prior to pretreatment. 2.2.2. Step 2: liquid hot water pretreatment (LHW) A 500 g sample of sugarcane bagasse powder was pretreated with liquid hot water pretreatment (LHW) as follows. The sugarcane bagasse was positioned at the bottom of a reactor (Autoclaves Sterilizer, SANYO⁄MLS-3750, Japan) and was completely immersed in 2 L of water at room temperature for 12 h to fully saturate the bagasse and facilitate the decomposition. Then, the feedstock was preheated using steam for 45 s before adding liquid water. For the rest of the procedure, the reaction temperature was maintained at 200 ± 3 °C for 1 h. After the pretreatment, the resulting hydrolysates were filtered and divided equally into six portions (Hydrolysate L1 through L6) for the fermentation experiments. The residual solids were dried at 65 °C and the solid loads were determined. Then the solids loads were used in the next pretreatment step. 2.2.3. Step 3: microwave pretreatment (MP) The humidity of the remnants’ solid fraction environment was adjusted to 40% for microwave pyrolysis. Prior to pretreatment, the sugarcane bagasse remnants were positioned into a threenecked, round-bottomed flask reactor. Nitrogen was infused to exclude all oxygen in the reaction flask, which was then kept overnight in a 4 °C refrigerator. Next day, the flask was placed in a microwave pyrolysis furnace (MCR-3, Gongyi Instrument Co. Ltd., Henan, China) at 120 °C for 7 min to allow decomposition. Then heating was stopped; the hot gas was allowed to completely escape the flask; the furnace was cooled to room temperature;

H. Su et al. / Bioresource Technology 187 (2015) 149–160

151

Fig. 1. The flow sheet of experiments following the methodology illustrated for pretreatment and fermentation processes. AN ATCC 1015: A. niger ATCC 1015; TR ATCC 26921: T. reesei ATCC 26921; PJ ATCC 44750: P. janthinellum ATCC 44750. HAN, HTR and HPJ: indicate the total hydrolysates used as fermentation substrates of SHF. HANNY, HTRNY and HPJNY: indicate the total hydrolysates used as fermentation substrates of SSF.

and the flask was removed. After the MP pretreatment step was done, the resulting substance was washed with 90% ethanol solution to remove any possible soluble oily compounds. Then the sample was filtered and separated equally into six portions (Hydrolysate MP1 through MP6) and the filtered hydrolysis solution was retained for fermentation experiments. The resulting residual solids were dried at 65 °C and divided into six equal portions (Solid residue MP1 through MP6) and their solids loads were determined. The residual solid samples were then used in the next step of decomposition. 2.2.4. Step 4: microorganism decomposition (MD) The solid residues remaining after microwaving were dried to remove the ethanol and a microbial growth nutrient solution was added, which contains a trace elements mixture (30 mg H3BO3, 20 mg MnCl2
4H2O, 185 mg ZnSO4
7H2O, 20 mg Na2MoO4
2H2O, 280 mg FeSO4
H2O, 200 mg CuSO4, 5 g NH4NO3, 1 g KH2PO4, 0.15 g MgSO4
H2O, 0.11 g CaCl2 and 1 L ddH2O, at pH 7.0). The six previously divided equal samples (Solid residue MP1 through MP6) were placed in six 10-cm wide mouthed bottles and sterilized at 121 °C in a high-pressure steam sterilization pot. The humidity of the material was adjusted to 30% after sterilization. Then two bottles were inoculated with 20 mL microorganism suspension of either Aspergillus niger ATCC 1015, Trichoderma reesei ATCC 26921, or Penicillium janthinellum ATCC 44750, respectively. The samples containing one of the three strains were cultivated at 30 °C for 10 days. Then the remaining sugarcane residues were removed by washing with ddH2O and prepared for the next step of pretreatment. The filtered hydrolysis solution was also retained for fermentation experiments. The residual solids (Solid residue AN1 and AN2, TR1 and TR2, and PJ1 and PJ2) were dried at 65 °C and their solids loads were determined.

Then the residual solid samples were used in the next step of decomposition. 2.2.5. Step 5: ammonia immersion (AI) The residual solids from Step 4 (Solid residue AN1 and AN2, Solid residue TR1 and TR2, and Solid residue PJ1 and PJ2) were immersed in a beaker containing an ammonia solution, with a mass ratio of 2:1 for ammonia to sugarcane. The beakers were subjected to autoclave at 90 °C for 30 min. Then the pH of the filtered hydrolysis solution was neutralized to 7.0 with 1% H2SO4 and retained for fermentation experiments. The residual solids (Solid residue AAN, ATR, and APJ) were dried at 65 °C and their solid loads were determined. These samples were then used in the next step of decomposition. After Solid residue AN2, TR2, and PJ2 were pretreated by Step 5, the solid–liquid mixtures were neutralized and kept as Solid–liquid mixture AAN, ATR, and APJ, respectively. 2.2.6. Step 6: enzymatic hydrolysis (EH) The remaining sugarcane bagasse residues from Step 5 (Solid residue AAN, ATR, and APJ) were placed into 500 mL ddH2O for 2 h and then concentrated to 300 mL using a rotatory evaporator at 60 °C. Then the pH of the mixture was adjusted to 7.0 with 1% H2SO4 and a 500 mL reaction flask was prepared, which contains 10% of the pretreated sugarcane residue (dry weight/volume), 10 g cellulase (500000 u/g, Thinkly Co., China; enzyme activity (u/g) definition (CMCA) = 1 g enzyme powder decomposes substrate CMC-Na to produce 1 mg glucose at 50 °C, pH 4.8 for 1 h), 100 mL OPTMASH BG (containing 5.4 U b-glucosidase activity and 1.9 U b-xylosidase activity; GENENCOR, International, USA), 10 lg/mL chloramphenicol, and 5 lg/mL kanamycin. The reaction mixture was buffered with 50 mM phosphate buffer (pH 5.0) and

152

H. Su et al. / Bioresource Technology 187 (2015) 149–160

incubated at 55 °C in a rotary shaker (HZQ-X500; Yiheng Co. Ltd., Shanghai, China) at 220 rpm for 60 h. 2.3. Microorganism and inoculum preparation Microorganism propagation: the strains of A. niger, T. reesei, and P. janthinellum mentioned in Step 3 were purchased from the American Type Culture Collection (ATCC, USA). They were cultured in a Potato Dextrose Agar medium at 35 °C for 7 days until the plates were full of spores. The wild-type strain C. beijerinckii NCIMB 8052 was obtained from the National Collections of Industrial Food and Marine Bacteria (Aberdeen, UK) and was stored in sterile distilled water at 4 °C. A TGYM medium for the rejuvenation process was prepared as follows: 5 g meat medium (Sigma–Aldrich, Beijing) was dissolved in 50 mL distilled water; 2 g glucose was added; and the mixture was autoclaved at 121 °C for 20 min and then cooled to 80 °C. Preserved strains were added to the TGYM medium and heat shocked at 80 °C for 2 min. Then the mixture was cooled in ice cold water for 1 min. The heat shocked spores were incubated in an anaerobic jar at 35 °C for 48 h. To prepare the fermentation inoculum for cultivation of the strain, we prepared a 100 mL TGYM medium containing 20 g/L glucose, 5 g/L yeast extract, 3 g/L ammonium acetate, 1 g/L sodium chloride, 1 g/L KH2PO4, 1 g/L K2HPO4, 0.2 g/L MgSO4, 0.02 g/L MnSO4
7H2O, and 0.02 g/L FeSO4
7H2O and autoclaved it at 115 °C for 20 min. Once the medium was cooled to 35 °C, 0.5 g/L cysteine was filtered (0.22 lm Milliporefilter, GEMA Medical SL, Chengdu) and added to the medium. Approximately 5 mL of the prepared rejuvenation bacterial culture was then added to the medium and left at 37 °C until the optical density reached 1.5.

0.001 g/L biotin), and minerals (20 g/L MgSO4 7H2O, 1 g/L MnSO4 H2O, 1 g/L FeSO4 7H2O, and 1 g/L NaCl) were filtered for sterilization (Millipore filter, 0.22 lm) and added to each bottle. The bottles were then inoculated with 3 mL of fermentation inoculum (OD600 = 1.5). 2.4.2. Simultaneous saccharification and fermentation (SSF) The hydrolysates from pretreatment Steps 2 through 4 and the solid–liquid mixture from Step 5 were mixed together and concentrated. Then the pH value of the total hydrolysate mixture was adjusted to 7.0. The fermentation substrate HANNY (Fig. 1, Part B) includes Hydrolysate L4, MP4, AN2, and Solid–liquid mixture AAN from substrates of pretreatment process (Fig. 1, Part A). The fermentation substrate HTRNY (Fig. 1, Part B) includes Hydrolysate L5, MP5, TR2, and Solid–liquid mixture ATR from substrates of pretreatment process (Fig. 1, Part A). The fermentation substrate HPJNY (Fig. 1, Part B) includes Hydrolysate L6, MP6, PJ2, and Solid–liquid mixture APJ from substrates of pretreatment process (Fig. 1, Part A). The sugar composition of the fermentation substrate samples was determined (Table 3). The fermentation procedure was simultaneously conducted using enzymatic hydrolysis. Additive trace elements and fermentation conditions were the same as with SHF. Pure nitrogen was injected into the fermentation bottles to maintain anaerobic conditions. Fermentation was conducted at 37 °C for 96 h, during which the samples were periodically withdrawn with a sterile syringe through the rubber stopper, for analysis of sugars, organic acids, pH and ABE production. The samples were centrifuged at 8000 rpm for 20 min and stored at 4 °C before analysis. 2.5. Fermentation experiments with pure glucose and a sugar mixture

2.4. ABE fermentation experiments To compare the effect of different procedures on fermentation, the experiments were divided into two sets. One set used separate hydrolysis and fermentation (SHF) after all pretreatment steps (fermentation substrate from Step 2 through Step 6) were completed. The second set used simultaneous saccharification and fermentation (SSF) (fermentation substrate from Step 2 through Step 6). The fermentation substrates from the hydrolysates produced Part A: Pretreatment Procedure was showed and Part B: Fermentation Processes were used (Fig. 1). 2.4.1. Separate hydrolysis and fermentation (SHF) The hydrolysates from pretreatment Steps 2 through 6 were collected and concentrated at 80 °C. The fermentation substrates were divided into three parts arised from the pretreatment steps using three fungi. The fermentation substrate HAN (Fig. 1, Part B) includes Hydrolysate L1, MP1, AN1, AAN, and EAN from substrates of pretreatment process (Fig. 1, Part A). The fermentation substrate HTR (Fig. 1, Part B) includes Hydrolysate L2, MP2, TR1, ATR, and ETR from substrates of pretreatment process (Fig. 1, Part A). The fermentation substrate HPJ (Fig. 1, Part B) includes Hydrolysate L3, MP3, PJ1, APJ, and EPJ from substrates of pretreatment process (Fig. 1, Part A). Their sugar compositions were analyzed prior to fermentation (Table 2). 150 mL of the total pretreatment hydrolysates were fermented in 250 mL glass anaerobic bottles (Haimen Huakai experiment glass instrument Co., Ltd, Haimen, China) sealed with butyl rubber. After addition of 0.75 g yeast extract and 1.5 g peptone to each bottle, the pH of the fermentation substrates was adjusted to 6.5 with 1% NaOH. These solutions were sterilized at 115 °C for 20 min and cooled to room temperature. Then, 0.5 mL of a combination of P2 trace elements mixture solution (50 g/L KH2PO4, 50 g/LK2HPO4, and 220 g/L CH3COONH4), vitamins (0.1 g/L para-aminobenzoic acid, 0.1 g/L thiamin, and

To compare the effect of production of ABE in our pretreatment design, we compared our results with experiments using just glucose or a mixture of sugars mimicking the sugar composition of typical hydrolysates of duckweed (10 g/L glucose, 42 g/L xylose, 3.5 g/L arabinose, 3.5 g/L cellobiose, 3 g/L galactose, and 1.5 g/L mannose) as the fermentation substrate to produce ABE under the same fermentation conditions. 2.6. Analytical procedures The compounds (acetone, butanol, ethanol, ABE) were measured with a gas chromatograph (GC) equipped with a flame ionization detector. The system was a model 6890 GC (Agilent Technologies, Santa Clara, CA, USA) with a model 7673A automatic injector, sampler, and controller (Hewlett–Packard). Alcohol compounds were separated using a ZB-WAX capillary column (30 m, 0.25 mm inside diameter, 0.25 lm film thickness, Phenomenex Inc., PA, USA). The GC oven temperature was held initially at 40 °C for 5 min, and then raised stepwise, by 15 °C/min, until it reached 150 °C. It was then raised by 50 °C/min up to 250 °C, and held for 4 min. Helium was used as the carrier gas, with an inlet pressure of 9.3 lb/in2. The injector and detector were maintained at 220 °C. A 1 lL volume of supernatant from the culture broth was injected in split-injection mode at a 1:30 split ratio. Isobutanol was used as the internal standard. Total residual carbohydrate was determined with the phenol– sulfuric acid method (Liu et al., 1973; Masuko et al., 2005; Saha and Brewer, 1994). The capacity of strain was calculated using the ratio of solvent produced (g)/carbohydrate consumed (g). Cellulose content was measured with spectrophotometry: 50 g sugarcane bagasse powder was placed in 1 L water, 60 mL of 60% H2SO4 were added, and the plants were left to decompose for 30 min. We added 2% anthrone reagent (v/v) to the hydrolyzed

H. Su et al. / Bioresource Technology 187 (2015) 149–160

mixture, left it for 2 min, and then measured the absorbance at 620 nm (Black, 1951; Hansen and Møller, 1975; Viles and Silverman, 1949). We then calculated the cellulose content of sample according to the standard curve. Cellulose content Y (%) of sugarcane bagasse = X (cellulose content of standard sample)  a (diluted multiples)  100/W (total weight of samples). The content of lignin was determined using acetyl bromide according to the standard methods (Iiyama and Wallis, 1990, 1988). The content of hemicellulose was determined using acid methanolysis according to the previously reported methods (Bertaud et al., 2002). The ash content was determined after burned at 500 °C in a muffle furnace. The constituent sugars were detected with DIONEX UltiMate3000 liquid chromatograph in a column packed with Aminex HPX-87H (Hercules, CA, USA: carbohydrate analysis column Aminex HPX-87P Column 300  7.8 mm, catalog 125-0098 serial 426070, 5 mM H2SO4, 0.6 mL/min; column temperature at 65 °C). Running conditions of the RI detector (detecting liquid refractive index): an RI detector temperature of 45 °C. Concentrations of the sugars were determined using extrapolation from standard curves. Butyric and acetic acids glucuronic acid, pcoumaric acid, syringic acid, and ferulic acid were determined with a DIONEX UltiMate 3000 liquid chromatograph in a column packed with Aminex HPX-87H and 0.05 mM H2SO4 on Chromosorb WAW. The chromatography was conducted at an injector temperature of 175 °C, detector temperature of 180 °C, and oven temperature of 125 °C. Determination of furfural and 5-Hydroxymethylfurfural was performed with HPLC according to the previously reported methods (Lee et al., 1986; Theobald et al., 1998). 2.7. Statistical analysis Three replicates of each experiment and assay were carried out unless otherwise indicated. For each treatment, we calculated the mean response variables and their standard deviation (SD). Analysis of variance (ANOVA) was performed and the mean separation was calculated using the Fisher’s Least Significant Difference test (with P 6 0.05) using the SPSS 21.0 program. 3. Results and discussion 3.1. Pretreatment process 3.1.1. Selection of appropriate pretreatment methods of the SCLPP Choosing an appropriate pretreatment procedure is a key factor influencing the yield of target sugars produced from lignocellulose. The pretreatment process determines whether simple sugar production is high enough for microbes to use and whether the fermentation conditions are suitable for microbial growth. Generally, common acid hydrolysis pretreatment with dilute sulphuric acid is a simple and effective method (Cai et al., 2012; Zheng et al., 2012; Zhou et al., 2012). Although sulphuric acid hydrolysis can generate enough sugar effectively, it also produces toxic compounds that inhibit the growth of microorganisms. This means an additional detoxification step with complicated pretreatment technologies is typically needed following acid hydrolysis. Therefore, when selecting a pretreatment process, the ideal approach would aim to maximize monosaccharide production and minimize the generation of inhibitor compounds, thereby avoiding the need for subsequent detoxification processes. The first step in a cellulose hydrolysis pretreatment process typically takes place at a solid–liquid interface. Previous work suggests that the level of decomposition directly correlates with the solid component particle size (Poletto et al., 2012). Therefore, in this study, we first mashed sugarcane bagasse into tiny particles to facilitate pretreatment by

153

increasing substrate accessibility (Step 1). Then, we chose several different individual pretreatment methods and combined them sequentially to design an optimal, multi-step SCLPP. To assess the potential advantages of the SCLPP, we tested it on sugarcane bagasse. Experiments were conducted following the methods illustrated in the pretreatment (Step 2  Step 6) and fermentation process flow sheet (Fig. 1).

3.1.2. Appropriate combination and managing order of the pretreatment methods of the SCLPP In this study, the sugarcane bagasse was comprised of 48.67% cellulose, the major component, 27.13% hemicellulose and 16.35% lignin. LWH, a common used method for lignocellulose pretreatment was selected for our experiment (Van Walsum et al., 1996). Compared to dilute acid pretreatment, LWH offers several potential advantages. LHW does not require acid use or transitioning to special, non-corrosive reactor materials. LHW also benefits from lower production of hydrolysate neutralization residues. Treatment of cellulose with LHW resulted in destruction of cellwall of solid interface (Kim et al., 2009b; Li et al., 2010; Pérez et al., 2008). Therefore in Step 2 of the SCLPP, the sugarcane bagasse was pretreated using LHW to improve decomposition productivity during the SCLPP. Although the bagasse has achieved a certain degree of ‘‘decomposition’’ by Step 2, it still has not been fully degraded. Therefore, a third pretreatment step is necessary. For Step 3, we chose the MP method because the solid surface structure of cellulose has been damaged during Step 2. Previous studies showed that MP can lead to higher yields of reducing sugar, shorter reaction time, and lower energy consumption for pretreating starch-free wheat fibers, switch grass, and rice hulls (Chen et al., 2012b; Janker-Obermeier et al., 2012). These benefits make MP a suitable technique for further decomposition of cellulose. The destruction of lignin structures can also be accelerated via creating nonthermal effects by the electromagnetic field in MP by Step 3. These effects may contribute to facilitating microbial growth and obtaining higher decomposition rates during subsequent pretreatment. Therefore, MD was chosen as our fourth pretreatment step of the SCLPP to further process the residual sugarcane solids. Ultimately, the primary aim for the first four steps (Step 2  Step 5) was to decompose cellulose and hemicellulose. However, some lignin may still not be fully decomposed and could facilitate EH in the final step of the SCLPP, so further processing is required. In Step 5, we chose to treat the residual solids from the three different MD steps (Step 4) with AI. The AI functions to separate lignin from the overall lignocellulosic structure, rather than destroy it. We specifically investigated the pretreatment methods using NH3
H2O (i.e., aqueous ammonia) as a solvent to separate lignin (Masuko et al., 2005; Pandey et al., 2000; Poletto et al., 2012). The EH was chosen as the last step of pretreatment. The goal of the previous pretreatment steps prior to EH is to disrupt the lignocellulosic matrix, thus rendering the substrate more accessible to enzyme action. Typically this involves only a single step of pretreatment, rarely using multiple steps, prior to enzymatic treatment according to literatures. For example, using NH3
H2O treatment before adding enzymes can efficiently hydrolyze some lignocellulosic biomasses like corn stover (Cao and Aita, 2013; Qin et al., 2013). Additionally, enzymatic hydrolysis of solids following microwave heating led to glucose production and decreased fermentation time to some extent (Chen et al., 2012a). However, using a single pretreatment step is usually not sufficient to thoroughly decompose lignocellulose and thus results in lower sugar yields. In this study, each step of the SCLPP has its own advantages, to attempt to break down lignocellulose more efficiently and completely and facilitate enzymolysis in the last step before fermentation.

154

H. Su et al. / Bioresource Technology 187 (2015) 149–160

Table 1 Solid recovery yields after each pretreatment step was completed. Pretreatments

Pretreated substrates

SRRY (w/w%)

SR (g)

TSYa (%)

Weight loss (g)

TSYb (%)

Achieved TS (g/L)

The Step 1: smash The Step 2: LHW The Step 3: MP

Sugarcane bagasse Bagasse powder Residue from step 1

71.31 66.29

356.55 237.14

22.28 13.63

143.45 119.12

77.67 40.80

111.42 48.61

The Step 4: MD A. niger ATCC 1015 T. reesei ATCC 26921 P. janthinellum ATCC 44750

Residue from step 2(AR2) Residue from step 2(TR2) Residue from step 2(PR2)

68.59 61.17 72.29

54.72 47.57 56.66

3.82 9.03 6.02

12.11 15.35 10.95

24.93 46.51 43.47

3.02 7.14 4.76

Residue from step 3(AAR3) Residue from step 3(ATR3) Residue from step 3(APR3)

88.38 85.67 89.80

24.23 20.71 25.66

2.62 3.76 2.17

3.18 3.46 2.91

22.64 24.30 21.30

0.72 0.91 0.62

Residue from step 4(EAR4) Residue from step 4(ETR4) Residue from step 4(EPR4)

31.11 28.39 25.59

7.54 5.88 6.57

56.21 70.11 57.67

16.69 14.83 19.09

81.61 97.90 77.31

13.62 14.52 14.76

The Step 5: AI

The Step 6: EH

AR2 = solid residue AN1 + solid residue AN2 (Fig. 1); TR2 = solid residue TR1 + solid residue TR2 (Fig. 1); PR2 = solid residue PJ1 + solid residue PJ2 (Fig. 1); AAR3 = solid residue AAN (Fig. 1); ATR3 = solid residue ATR (Fig. 1); APR3 = solid residue APJ (Fig. 1); EAR4 = solid residue EAN (Fig. 1); ETR4 = solid residue ETR (Fig. 1); EPR4 = solid residue EPJ (Fig. 1). SRRY: solid residue recovery yields (SRRY) is equal to reclaimed solid residue divided by total pretreated substrate; TS: total sugars; SR: solid residue. a TSY: total sugars yield is equal to achieved total sugars divided by total pretreated substrate. b TSY: total sugars yield is equal to achieved total sugars divided by weight loss.

3.1.3. Effect of LHW (Step 2) and MP (Step 3) of the SCLPP The analysis results of the pretreated materials are summarized in Table 1. When sugarcane bagasse was crushed into tiny particles and treated with LHW, we achieved a solid residue recovery yield (SRRY) of 71.31%. 28.69% of sugarcane bagasse was broken down by LHW, leading to a weight loss of 143.45 g and a total sugars (TS) production of 111.42 g/L. The total sugar yield from substrate and weight loss after pretreatment reached 22.28% (TSYa) and 77.67% (TSYb), respectively (Table 1). The solid residue was further pretreated with MP (Step 3). After processing through Step 3, the SRRY of sugarcane bagasse reduced to 66.29%, in other words, the decomposition rate reached 33.71%. The weight loss reached 119.12 g and the total sugar (TS) concentration was 48.61 g/L based on SRRY value. The total sugar yield (TSY) from total pretreated substrate and weight loss was 13.63% (TSYa) and 40.80% (TSYb), respectively (Table 1). We found that MP lowered the SRRY from 71.31% to 66.29%, resulting in an additional sugarcane breakdown of 5.02% (Table 1), suggesting that MP is an effective pretreatment method for sugarcane bagasse. The LHW pretreatment step significantly facilitated decomposition in the MP step when the sugarcane was crushed into particles because sugarcane bagasse residues became fully saturated. In this experiment, we used nitrogen-rich, oxygen-limited reaction flasks to help limit the production of unwanted microbial growth inhibitors because the existence of oxygen may produce more kinds of chemical

Table 2 Compositions from hydrolysate of sugarcane bagasse when all pretreatment steps were completed (fermentation substrates of SHF at 0 h). Sugars components in different pretreatment substrates Sugars compound

HAN

HTR

HPJ

Xylose (g/L) Glucose (g/L) Arabinose (g/L) Cellobiose (g/L) Galactose (g/L) Mannose (g/L) Total sugars (g/L)

41.67 ± 2.89 9.53 ± 1.29 3.16 ± 0.89 2.96 ± 0.41 3.19 ± 0.72 1.12 ± 0.31 61.63

43.52 ± 3.23 15.21 ± 1.65 2.14 ± 0.36 2.98 ± 0.58 2.17 ± 0.74 0.85 ± 0.19 66.87

41.47 ± 1.92 13.49 ± 1.28 2.76 ± 0.61 2.87 ± 0.79 2.39 ± 0.41 1.38 ± 0.22 63.36

HAN, HTR and HPJ: indicate the total hydrolysates were used as fermentation substrates of SHF.

Table 3 Compositions from hydrolysate of sugarcane bagasse before the fifth step of enzymatic hydrolysis (fermentation substrates of SSF at 0 h). Pretreatment substrates produced sugars components Sugars compound

HANNY

HTRNY

HPJNY

Xylose (g/L) Glucose (g/L) Arabinose (g/L) Cellobiose (g/L) Galactose (g/L) Mannose (g/L) Total sugars (g/L)

27.15 ± 2.31 5.56 ± 1.67 2.42 ± 0.73 2.31 ± 0.93 2.51 ± 0.83 0.71 ± 0.075 40.66

28.34 ± 2.21 8.13 ± 1.37 1.14 ± 0.33 2.17 ± 0.96 2.09 ± 0.89 0.79 ± 0.085 43.01

27.21 ± 2.54 9.24 ± 1.67 1.54 ± 0.23 1.91 ± 0.26 1.72 ± 0.39 1.13 ± 0.095 42.23

HANNY, HTRNY and HPJNY: indicate the total hydrolysates used as fermentation substrates of SSF.

reactions. The product was also washed with 90% ethanol to remove any possible soluble oily compounds. 3.1.4. Effect of MD (Step 4) using different fungi of the SCLPP Many microorganisms including bacteria, yeasts, and fungi have been used in sugarcane cultivation. Filamentous fungi especially basidiomycetes are the preferred, most widely used choice for the decomposition of lignocellulose (Pandey et al., 2000). The remnant sugarcane bagasse was divided into six portions and then further biodegraded by microbial action (Fig. 1). The three different species of fungi we used (A. niger ATCC 1015, T. reesei ATCC 26921, and P. janthinellum ATCC 44750) were all able to degrade sugarcane bagasse to some extent in Step 4. After processing through Step 4, the SRRY of sugarcane bagasse was reduced to 68.59%, 61.17%, and 72.29% with weight loss of 12.11 g, 15.35 g, and 10.95 g, and small quantitative TS of 3.02 g/L, 7.14 g/L, and 4.76 g/L by using A. niger ATCC 1015, T. reesei ATCC 26921, and P. janthinellum ATCC 44750, respectively. The TSYa from total pretreated substrate was 3.82%, 9.03%, and 6.02%, and TSYb from weight loss was 24.93%, 46.51%, and 43.47%, by using A. niger, T. reesei, and P. janthinellum, respectively (Table 1). The decomposition yield of sugarcane bagasse was 31.41% for AR2, 38.83% for TR2, and 27.71% for PR2 (Table 1), resulting in a mean degradation rate of 32.65%. T. reesei ATCC 26921 showed the best decomposition ability, which might be because the cellobiohydrolase I (CBHI) of this species has a stronger activity to break down the structure of cellulose.

155

H. Su et al. / Bioresource Technology 187 (2015) 149–160

3.1.5. Effect of AI (Step 5) and EH (Step 6) of the SCLPP After biodegradation, the remnant sugarcane bagasse was soaked in NH3
H2O. After processing through Step 5, the SRRY of sugarcane bagasse was reduced to 88.38%, 85.67%, and 89.80%, with weight loss of 3.18 g, 3.46 g, and 2.91 g, and TS of 0.72 g/L, 0.91 g/L, and 0.62 g/L by using A. niger ATCC 1015, T. reesei ATCC 26921, and P. janthinellum ATCC 44750, respectively. The TSYa from total pretreated substrate was 2.62%, 3.76%, and 2.17%, and TSYb from weight loss was 22.64%, 24.30%, and 21.30%, by using A. niger, T. reesei, and P. janthinellum, respectively (Table 1). The resulting decomposition yield of sugarcane bagasse was decreased to 11.62% for AAR3, 14.33% for ATR3, and 10.20% for APR3. Finally, following enzymatic decomposition, the final recovery yield of sugarcane solids was 68.89% for EAR4, 71.61% for ETR4 and 74.41% for EPR4 with a mean value of 71.64% (Table 1). The amount of solid sugarcane did not decrease significantly after pretreatment by AI. However, AI did facilitate high decomposition rates during the following enzymatic hydrolysis step because NH3
H2O is a good solvent for lignin. NH3
H2O achieves dissolution by cleaving the linkages between lignin fragments as well as between lignin polymers and carbohydrates, which increases contact at the enzymesubstrate surface. After all the pretreatment steps were finished, the amount of recycling for SR was from 39.52 g at Step 4 to 7.54 g, 5.88 g, and 6.57 g at Step 6, namely the decomposition yield was 80.92%, 85.12% and 83.37%, respectively, suggesting the T. reesei ATCC 26921 was the best one at decomposing bagasse (Fig. 1 and Table 1). 3.1.6. Determination of starting sugar concentrations for fermentation after the SCLPP After the completion of all preprocessing steps, all hydrolysates were collected and concentrated (Fig. 1 Part B). The sugar compositions of the hydrolysates were determined before (Table 3, Step 1  Step 5) and after (Table 2, Step 1  Step 6) the EH step. We also estimated the specific sugar yield from the EH step by comparing the sugar content after Step 5 with the sugar content after Step 6 (Table 4). We found that sugarcane bagasse was decomposed completely through our comprehensive pretreatment procedure and that the highest degradation yield (38.83%, g/g) was achieved using T. reesei ATCC 26921 in Step 4. The highest glucose yields were obtained using T. reesei ATCC 26921 and P. janthinellum ATCC 44750, which were 15.21 ± 1.65 g/L and 13.49 ± 1.28 g/L, respectively. 3.1.7. Determination of microbial growth inhibitors in hydrolysates after the SCLPP After the pretreatments, we tested for the presence of toxic compounds such as furfural and hydroxymethylfurfural (HMF) in hydrolysates, which could inhibit microbial growth (Table 5). Unlike other strong chemical hydrolyzation or physical decomposition methods, the methods we used for pretreatment occur under Table 4 The productivity of various sugars from the fifth step: enzymatic hydrolysis method. The productivity of various sugars (%) Sugars Compound

HAN

HTR

HPJ

Xylose Glucose Arabinose Cellobiose Galactose Mannose Total sugars

34.85a 41.66a 25a 21.95a 21.31a 36.61a 34.03a

34.88a 46.54a 46.72b 27.18a 3.69b 7.06b 35.68a

34.39a 31.5b 44.20b 33.45b 28.03a 18.11c 33.35a

HAN, HTR and HPJ: indicate the total hydrolysates used as fermentation substrates of SHF. The different letters indicate significant differences based on multiple comparisons for different alcohol products (P < 0.05).

Table 5 Inhibitor concentrations after all pretreatment steps were completed. Compound

Furfural Hydroxymethylfurfural (HMF) Glucuronic acid p-Coumaric acid Syringic acid Ferulic acid

Inhibitor concentration (mg/L) HAN

HTR

HPJ

14.35 ± 2.11 6.29 ± 1.45 ND ND ND ND

11.12 ± 1.89 4.31 ± 2.69 ND ND ND ND

8.37 ± 3.18 7.39 ± 1.31 ND ND ND ND

HAN, HTR and HPJ: indicate the total hydrolysates used as fermentation substrates of SHF. ND: not detected.

relatively mild conditions that are likely not strong enough to generate fermentation-inhibiting chemicals. Therefore our pretreatment method may ultimately lead to improved butanol yields. This conclusion was confirmed when we tested for the presence of microbial growth inhibitors. In our experiments, the highest yield of furfural and HMF in HAN, HTR and HPJ was only 14.35 ± 2.11 and 7.39 ± 1.31 g/L, 11.12 ± 1.89 and 4.31 ± 2.69 g/L, and 8.37 ± 3.18 and 7.39 ± 1.31 g/L respectively (Table 5), which were below the levels known to disrupt the growth of C. beijerinckii. Other inhibitors such as glucuronic acid, p-coumaric acid, syringic acid, and ferulic acid were not detected likely because (1) the concentrations of these compounds were too low to detect or (2) these compounds were volatilized off in the ethanol fraction. These results indicate that the well-organized SCLPP is an effective method for limiting the generation of toxic organic compounds. Additionally, our method does not produce high concentrations of mineral salts from neutralization of pH in acid hydrolysis pretreatment which are known to seriously impede microorganism fermentation. Due to the limited inhibitor production and not producing high concentrations of mineral salts, higher concentrated total sugars were obtained. Therefore, the SCLPP allows a better microbial growth and helps ensure the subsequent fermentation. 3.2. Effect of fermentation process 3.2.1. ABE fermentation for different processes Butanol is a potential substitute for fossil fuels and is usually obtained via ABE fermentation (Ezeji et al., 2004). In this study, we aimed to verify the feasibility of using the SCLPP to hydrolyze lignocellulose in sugarcane bagasse for butanol fermentation. In biofuel production via fermentation, SHF is the most conventional and widely used method. The SSF was another method used for conversion of sugars to avoid inhibitory effect occurs in SHF method. Specifically, the SHF method suffers from slow sugar uptake rates due to the feedback inhibition caused by an accumulation of sugars after enzymatic hydrolysis. To assess the effectiveness of our proposed pretreatment procedure, we conducted biobutanol fermentation experiments using our hydrolysates as the substrate. We attempted to compare the ABE production from hydrolysate of sugarcane bagasse by using SSF or SHF methods. An additional enzymatic saccharification pretreatment step (Step 6) was optimized and used to maximize reducing sugar yields before SHF was performed on the collected sugarcane bagasse hydrolysates from Step 2 through Step 6. Alternatively, SSF was performed on the collected hydrolysates from Step 2 through Step 4 as well as the solid–liquid mixture from Step 5 (Fig. 2). We compared the ability of C. beijerinckii NCIMB 8052 to produce ABE from our hydrolysates using SHF or SSF method. We also tested ABE production using SHF or SSF method with pure glucose and a defined sugar mixture to mimic our hydrolysates in order to test whether our hydrolysates produce inhibitory action for

156

H. Su et al. / Bioresource Technology 187 (2015) 149–160

fermentation. After Step 5 of pretreatment, sugarcane bagasse hydrolysates were divided into two portions, one for SHF and the other for SSF (Fig. 1, Part B). After a continuous fermentation for 96 h, the changes in various parameters were measured and the target products of fermentation are shown in Fig. 2. 3.2.2. Final yields of ABE fermentation productions The butanol production during SSF peaked at 80 h, which was 6.86 g/L with HANNY as substrate, followed by 6.11 g/L with HTRNY as substrate and 6.21 g/L with HPJNY as substrate (Fig. 2 and Table 6). However, the butanol yields during SHF were lower than those during SSF. During SHF, the highest butanol concentration was 5.57 g/L, which was obtained with HPJ as substrate, followed by 5.21 g/L with HTR as substrate and 5.09 g/L with HAN as substrate. The highest concentrations of acetone during SSF was 3.66 g/L with HTRNY as substrates, followed by 3.48 g/L with HANNY and

HPJNY as substrates. In contrast, the concentrations of acetone were lower during SHF, which were 2.66 g/L with HAN as substrate, 2.77 g/L with HTR as substrate, and 2.89 g/L with HPJ as substrate (Fig. 2 and Table 6). For ethanol, the highest concentration was 1.83 g/L with HTRNY as substrate during SSF, while the use of HANNY and HPJNY as substrates produced 1.52 g/L and 1.611 g/L of ethanol, respectively. The ethanol productions during SSF were not always higher than those observed during SHF (Fig. 2 and Table 6). 3.2.3. Fermentation productions using pure glucose and a defined mixture of sugars To validate availability and efficiency of the SCLPP, we also used pure glucose and a defined mixture of sugars as fermentation substrates to compare the butanol yields with those achieved using sugarcane bagasse hydrolysates as substrates. The highest butanol production of 5.52 g/L was obtained using pure glucose. There was

Fig. 2. Changes of fermentation products over time using the hydrolysate of sugarcane bagasse as fermentation substrate. HAN, HTR and HPJ hydrolysates were used as fermentation substrates with SHF method. HANNY, HTRNY and HPJNY: indicate the total hydrolysates used as fermentation substrates of SSF. Error bars indicate SD (n = 3).

157

H. Su et al. / Bioresource Technology 187 (2015) 149–160 Table 6 Performance of ABE fermentation using hydrolysates of sugarcane bagasse, pure glucose and mixture sugars as substrates by SHF and SSF. Parameters

Butanol (g/L) Acetone (g/L) Ethanol (g/L) Total ABE (g/L) Total acids (g/L) Total solvents (g/L) Residual carbohydrate (g/L) ABE yield (g/g) Productivity (g/L h) Butanol/total solvents (%) Fermentation time (h)

SHF

SSF

Comparison

HAN

HTR

HPJ

HANNY

HTRNY

HPJNY

Pure glucose

Mix sugar

5.09 ± 0.67a 2.66 ± 0.21a 0.88 ± 0.08a 7.63 ± 0.52a 2.59 ± 0.35a 10.22 ± 0.42 13.19 ± 0.73a 0.12 0.08 49.8 96

5.21 ± 0.45a 2.77 ± 0.31a 1.03 ± 0.11b 9.01 ± 0.42b 2.96 ± 0.57a 11.97 ± 0.38 9.42 ± 0.69b 0.13 0.094 43.4 96

5.57 ± 0.64a 2.89 ± 0.34a 1.13 ± 0.13b 9.59 ± 0.51b 2.31 ± 0.25a 11.9 ± 0.42 12.16 ± 0.77a 0.15 0.1 46.8 96

6.86 ± 0.73b 3.48 ± 0.41b 1.52 ± 0.21b 11.86 ± 0.61c 1.77 ± 0.13b 13.63 ± 0.56 5.91 ± 0.43c 0.20 0.123 50.55 96

6.11 ± 0.52b 3.66 ± 0.45b 1.83 ± 0.12c 11.6 ± 0.47c 1.85 ± 0.21b 13.45 ± 0.61 5.447 ± 0.51c 0.17 0.123 46.4 96

6.21 ± 0.71b 3.48 ± 0.32b 1.611 ± 0.17b 12.11 ± 0.45c 1.86 ± 0.16b 13.97 ± 0.49 6.63 ± 0.46c 0.20 0.13 45.67 96

5.52 ± 0.35a 2.96 ± 0.26a 1.08 ± 0.24a 9.56 ± 0.31b 2.61 ± 0.24a 12.17 ± 0.53 3.72 ± 0.6d 0.16 0.1 45.35 96

6.13 ± 0.47b 2.43 ± 0.21a 1.23 ± 0.11b 9.79 ± 0.42b 2.76 ± 0.32a 12.55 ± 0.47 8.96 ± 0.63b 0.16 0.13 0.48 96

HAN, HTR and HPJ: indicate the total hydrolysates used as fermentation substrates of SHF. HANNY, HTRNY and HPJNY: indicate the total hydrolysates used as fermentation substrates of SSF. ABE yield (g/g): total ABE production divided by gross weight of sugarcane bagasse. Productivity (g/L h): total solvents divided by fermentation time. a,b,c Different letters in table indicate significant differences based on multiple comparisons (P < 0.05).

no significant difference in the yield from these two kinds of substrates using SHF, but a significant difference was observed when using SSF. The highest yield of total ABE was 9.56 g/L from pure glucose, which has a significant difference from that obtained using SSF (Fig. 4 and Table 6). The highest butanol yield was 6.13 g/L from the mixture sugar, which has no significant difference from that obtained using SSF, but the highest total ABE yield was 9.79 g/L, which was lower than the mean yield of 11.9 g/L obtained using SSF (Fig. 5 and Table 6). Overall, these results demonstrated that the SCLPP followed by the SSF method is more effective than pure glucose fermentation for producing butanol. However the SHF method showed a similar performance when compared to pure glucose fermentation. Therefore, the SCLPP provides additional benefit and meets our desired objectives. 3.2.4. Changes of acetic acid and butyric acid productions during fermentation Conversion of butyric acid to butanol stopped after 72 h of fermentation and butanol reached the highest level with either glucose or mixed sugars as carbon source, regardless of the hydrolysate used (Figs. 2, 4 and 5). Production of butyric and acetic acids, and products of the acidogenesis phase of fermentation changed during the course of fermentation (Fig. 2) and the amount of total acids produced from the SHF method was higher than that from the SSF method. The highest concentrations of acetic acid and butyric acid were 1.45 g/L and 1.51 g/L, respectively, which were produced using HTR as substrate. The final total acids from the SHF method were higher than those from the SSF method (Table 6). More acids were produced using SHF method, meaning less butyric acid was converted into butanol. Moreover, the pH values after 72 h in the SHF method were higher than those in the SSF method. The results showed SSF was a more efficient method than SHF after conducting the SCLPP. 3.2.5. Productivity of fermentation with different substrates Hydrolysates obtained after all pretreatment steps (SHF) and after the fourth step (SSF) were used for ABE fermentation. The HAN, HTR, and HPJ substrates fermented using the SHF method produced an ABE yield (g/g) of 0.12, 0.13, and 0.15 from each kg of initial sugarcane, respectively. On the other hand, the ABE yield (g/g) from the HANNY, HTRNY, and HPJNY substrates fermented using the SSF method was 0.20, 0.17, and 0.20 respectively per kg sugarcane bagasse. Therefore, the SSF method resulted in an increase of nearly 34% in butanol yield and a higher overall ABE yield (Table 6). A higher total ABE production (11.7 g/L) was obtained using the SSF method for all three substrates, compared to that (8.5 g/L)

using the SHF method (Table 6). Additionally, the SSF method produced lower total acids (mean 1.8 g/L), compared to the SHF method (mean 2.6 g/L). While SSF method produced a higher yield of ABE (0.19 g/g) compared to the SHF method (0.14 g/g). We achieved higher productivity using the SSF method (0.123 g/L h) from total sugarcane bagasse substrate compared to the SHF method (0.09 g/L h) (Table 6). The final EH pretreatment in the SSF method was found to be a key step. Interestingly, in terms of the fermentation time for ABE production, the yields of sugar consumption and ABE production in the SHF process (13.19 g/L, 9.428 g/L, and 12.16 g/L of residual carbohydrate concentration and 0.08 g/L h, 0.094 g/L h, and 0.1 g/ L h of ABE productivity for the HAN, HTR, and HPJ substrates respectively) were comparable to the results observed using the SSF process (5.917 g/L, 5.447 g/L, and 6.63 g/L of residual carbohydrate concentration and 0.123 g/L h, 0.123 g/L h, and 0.13 g/L h of ABE productivity for the HANNY, HTRNY, and HPJNY substrates respectively). In addition, the use of pure glucose (3.72 g/L of residual carbohydrate concentration and 0.1 g/L h of ABE productivity) and the defined sugar mixture (8.96 g/L of residual carbohydrate concentration and 0.13 g/L h of ABE productivity) seems to mimic the results observed when using sugarcane bagasse hydrolysates as the fermentation substrate (Table 6). This suggests that our method is a suitable pretreatment protocol to increase ABE production from sugarcane bagasse owing to its mild conditions. A higher percentage of butanol/total solvents production rate was observed in the SSF process (46.6%) compared to the SHF process (47.54%) over the same fermentation time. This is possibly a result of low sustained sugar levels during fermentation. The higher ABE production may also be attributed to the low inhibitor concentrations produced in the hydrolysates. Overall, these investigations suggest our pretreatment approach has the potential to be exploited for efficient ABE production from lignocellulose. The results showed SSF have higher productivity leading to producing more butanol than SHF after the SCLPP. 3.2.6. Changes of utilized sugar compositions during fermentation More residual sugars (unused sugars) remained in the SHF method than the SSF method (Fig. 3 and Supplementary file, Fig. S2). The residual xylose concentration following the fermentation of the HAN, HTR and HPJ substrates using SHF was 8 g/L, 3.01 g/L, and 2.3 g/L respectively. However, the residual xylose concentrations following the fermentation of the HANNY, HTRNY, and HPJNY substrates using SSF was 0.99 g/L, 1.56 g/L and 1.11 g/L, respectively. So the sugar consumption levels were higher in SSF methods than in SHF method.

158

H. Su et al. / Bioresource Technology 187 (2015) 149–160

Fig. 3. Changes of the utilization of various sugar components during SSF and SHF fermentation using the hydrolysates of sugarcane bagasse as fermentation substrates. (a, c and e) Changes of the utilization of various sugar components during SHF. (b, d and f) Changes of the utilization of various sugar components during SSF. HAN, HTR and HPJ: indicate the total hydrolysates used as fermentation substrates of SHF. HANNY, HTRNY and HPJNY: indicate the total hydrolysates used as fermentation substrates of SSF. Error bars indicate SD (n = 3).

The residual arabinose concentration was about 0.3 g/L across all fermentation experiments, while the residual concentration of cellobiose, a disaccharide, was about 0.5 g/L because it is more difficult for the microbes to use. Galactose and mannose were used more thoroughly, in SSF method than in SHF method. The average concentration of galactose was 0.6 g/L in SHF method while 0.35 g/ L in SSF method. A similar pattern was observed for mannose. After fermentation, the total residual carbohydrate concentration in SSF method (6.0 g/L) was lower than that in SHF method (11.01 g/L, Fig. 3 and Table 4). Higher consumption levels for different sugars compositions by SSF explained why using SSF can lead to obtaining more butanol than SHF after the SCLPP. 3.2.7. Changes of pH during fermentation After fermentation using the SSF method, the final pH value was 4.5, 4.16 and 4.09 when using HANNY, HTRNY and HPJNY as substrates, respectively. Higher pH values (5.1, 4.97 and 5.32) were observed in SHF method with the same substrates (Supplementary

files, Fig. S1). The conversion of butyric acid to butanol is catalyzed by butyraldehyde dehydrogenase and butanol dehydrogenase, which require two molecules of NADH as electron source. Therefore, we believe the electron source in the SHF medium was less than that in the SSF medium, perhaps due to a lower NADH concentration in the SHF medium. This lack of electron donors may lead to less butanol production at the end of the fermentation when using hydrolysates as the carbon source in the SHF method. In addition, the residual carbohydrate concentrations were higher when using the SHF method, illustrating the advantage of conducting EH and fermentation at the same time. The higher final ABE concentrations and faster ABE productions observed using the SSF method are possibly due to more efficient removal of sugars during the saccharification process, therefore eliminating the product feedback inhibition. 3.2.8. Effect of inhibitors on microbial growth during fermentation Comparing the ability of C. beijerinckii NCIMB 8052 to produce ABE from decomposed sugarcane bagasse, the pure glucose, and

H. Su et al. / Bioresource Technology 187 (2015) 149–160

159

Fig. 4. Changes of different parameters during fermentation using pure glucose as fermentation substrates. Error bars indicate SD (n = 3).

a defined sugar mixture lead to an interesting observation: the butanol yields from SHF and pure glucose were similar, while the butanol yields from SSF and mixed sugars were similar. The similar performance of this bacterium to ferment pure glucose and a defined sugar mixture compared with our pretreated sugarcane bagasse hydrolysates highlights the feasibility of the SCLPP. In our approach, a detoxification step, such as overliming for reportorial straw and maize stover, is unnecessary prior to fermentation. The nearly equal ABE yields from the hydrolysates fermented using the SSF method and from the defined sugar mixture substrate imply that the low levels of inhibitors (furfural and HMF) detected in the pretreated substrate did not significantly inhibit bacterial growth or production. Thus, the relatively mild conditions used in the SCLPP may be advantageous for fermentation and therefore yielded more butanol. In addition, the ABE yields from the substrates fermented using the SSF method were higher than those using the SHF method, perhaps partly because butanol can also be produced from lignocellulosic biomass that was incompletely hydrolyzed. For any fermentation process, inhibitor production can have a dramatic negative impact. Most existing technologies for removing inhibitors, like resin filtration systems, can only remove limited types of inhibitors. No single technology can thoroughly reduce all of the prominent inhibitors and incorporating all kinds of technologies is costly. The SCLPP not only reduces the production of certain common growth inhibitors—namely, furfural and HMF— below levels harmful to fermentation, but also completely circumvents the production of other phenyl-containing inhibitors. Although the SCLPP requires longer processing time, we were able to obtain beneficial effects on saccharifying and reduce, or remove inhibitors by combining several common pretreatment methods in a manageable order to obtain the highest butanol yield. Combined aforementioned results, the SCLPP presented in this study is innovative as a comprehensive approach combining various methods. No previous reports have demonstrated an approach both facilitating high lignocellulose degradation and eliminating microbial inhibitor production. The SCLPP performs efficient and complete hydrolysis saccharification with significantly reduced microbial inhibitors. A significantly higher hydrolysis rate was observed when cellulases were used to digest sugarcane bagasse. Detailed investigation suggests that the increase of hydrolysis is correlated with the species of fungus used in our fourth pretreatment step. We identified the best pretreatment procedure parameters to assess the choice of suitable hydrolysates for subsequent fermentation steps. Comparing the fermentation of pure glucose, a defined sugar mixture, or sugarcane bagasse hydrolysates

Fig. 5. Changes of different products, sugars and pH during fermentation using mixture sugars as fermentation substrates. Error bars indicate SD (n = 3).

produced by C. beijerinckii NCIMB 8052 using the SHF method indicated that there are no inhibitory effects associated with the hydrolysates used. Thus, the SCLPP offer a mild approach that does not produce significant amounts of fermentation inhibitors. Finally, using the SSF approach, we observed higher ABE production from sugarcane bagasse hydrolysates. 4. Conclusions In this study, the SCLPP was developed using sugarcane bagasse as an example substrate for ABE production. The goal of our design was to minimize the generation of microbial inhibitors from the treated cellulose substrates while keep the maximum yield of sugar. The SCLPP for ABE production is a feasible, comprehensive approach that combines several mild pretreatment methods. The SCLPP reduced the inhibitory effects and obtained high ABE production even in the absence of a detoxification step. The innovation of the SCLPP is comprehensively combining several common methods rather than developing a single, independent method. It has potential application to increase ABE production from agricultural residues rich in lignocellulose. Competing interests The authors declare that they have no competing interests. Authors’ contributions Haifeng Su participated in the conception, design, data collection and analysis, and drafted the manuscript. Furong Tan and

160

H. Su et al. / Bioresource Technology 187 (2015) 149–160

Gang Liu assisted the laboratory work, results interpretation and the manuscript revision. Zhijian Liu participated in discussion of the results and revision of the manuscript. All authors read and approved the final manuscript. Acknowledgements This work was supported by the earmarked fund for Modern Agro-industry Technology Research System (CARS-35) and the Applied Basic Research Programs of Sichuan province (No. 2014JY0065). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2015.03. 107. References Bamufleh, H.S., Alhamed, Y.A., Daous, M.A., 2013. Furfural from midribs of datepalm trees by sulfuric acid hydrolysis. Ind. Crops Prod. 42, 421–428. Bertaud, F., Sundberg, A., Holmbom, B., 2002. Evaluation of acid methanolysis for analysis of wood hemicelluloses and pectins. Carbohydr. Polym. 48, 319–324. Black, H., 1951. Determination of sodium carboxymethylcellulose in detergent mixtures by anthrone method. Anal. Chem. 23, 1792–1795. Brown, T.R., Brown, R.C., 2013. A review of cellulosic biofuel commercial-scale projects in the United States. Biofuel Bioprod. Bioref. 7, 235–245. Cai, B.-Y., Ge, J.-P., Ling, H.-Z., Cheng, K.-K., Ping, W.-X., 2012. Statistical optimization of dilute sulfuric acid pretreatment of corncob for xylose recovery and ethanol production. Biomass Bioenergy 36, 250–257. Cao, S., Aita, G.M., 2013. Enzymatic hydrolysis and ethanol yields of combined surfactant and dilute ammonia treated sugarcane bagasse. Bioresour. Technol. 131, 357–364. Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: a review. Bioresour. Technol. 99, 4044–4064. Chen, C., Boldor, D., Aita, G., Walker, M., 2012a. Ethanol production from sorghum by a microwave-assisted dilute ammonia pretreatment. Bioresour. Technol. 110, 190–197. Chen, W.-H., Ye, S.-C., Sheen, H.-K., 2012b. Hydrolysis characteristics of sugarcane bagasse pretreated by dilute acid solution in a microwave irradiation environment. Appl. Energy 93, 237–244. de Vasconcelos, S.M., Santos, A.M.P., Rocha, G.J.M., Souto-Maior, A.M., 2013. Diluted phosphoric acid pretreatment for production of fermentable sugars in a sugarcane-based biorefinery. Bioresour. Technol. 135, 46–52. Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2004. Butanol fermentation research: upstream and downstream manipulations. Chem. Rec. 4, 305–314. Guo, T., He, A.-Y., Du, T.-F., Zhu, D.-W., Liang, D.-F., Jiang, M., Wei, P., Ouyang, P.-K., 2013. Butanol production from hemicellulosic hydrolysate of corn fiber by a Clostridium beijerinckii mutant with high inhibitor-tolerance. Bioresour. Technol. 135, 379–385. Hansen, J., Møller, I., 1975. Percolation of starch and soluble carbohydrates from plant tissue for quantitative determination with anthrone. Anal. Biochem. 68, 87–94. Hayashi, S., Aono, R., Hanai, T., Mori, H., Kobayashi, T., Honda, H., 2003. Analysis of organic solvent tolerance in Escherichia coli using gene expression profiles from DNA microarrays. J. Biosci. Bioeng. 95, 379–383. Iiyama, K., Wallis, A.F.A., 1988. An improved acetyl bromide procedure for determining lignin in woods and wood pulps. Wood Sci. Technol. 22, 271–280. Iiyama, K., Wallis, A.F.A., 1990. Determination of lignin in herbaceous plants by an improved acetyl bromide procedure. J. Sci. Food. Agric. 51, 145–161. Janker-Obermeier, I., Sieber, V., Faulstich, M., Schieder, D., 2012. Solubilization of hemicellulose and lignin from wheat straw through microwave-assisted alkali treatment. Ind. Crops Prod. 39, 198–203. Jennings, E.W., Schell, D.J., 2011. Conditioning of dilute-acid pretreated corn stover hydrolysate liquors by treatment with lime or ammonium hydroxide to improve conversion of sugars to ethanol. Bioresour. Technol. 102, 1240–1245. Jönsson, L.J., Alriksson, B., Nilvebrant, N.-O., 2013. Bioconversion of lignocellulose: inhibitors and detoxification. Biotechnol. Biofuels 6, 16.

Ju, X., Engelhard, M., Zhang, X., 2013. An advanced understanding of the specific effects of xylan and surface lignin contents on enzymatic hydrolysis of lignocellulosic biomass. Bioresour. Technol. 132, 137–145. Jung, Y.H., Kim, I.J., Kim, H.K., Kim, K.H., 2013. Dilute acid pretreatment of lignocellulose for whole slurry ethanol fermentation. Bioresour. Technol. 132, 109–114. Kim, D.-H., Kim, S.-H., Shin, H.-S., 2009a. Sodium inhibition of fermentative hydrogen production. Int. J. Hydrogen Energy 34, 3295–3304. Kim, Y., Hendrickson, R., Mosier, N., Ladisch, M., 2009b. Liquid hot water pretreatment of cellulosic biomass. In: Mielenz, J.R. (Ed.), Biofuels, vol. 581. Humana Press, pp. 93–102. Krastanov, A., Alexieva, Z., Yemendzhiev, H., 2013. Microbial degradation of phenol and phenolic derivatives. Eng. Life Sci. 13, 76–87. Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 48, 3713–3729. Lee, H.S., Rouseff, R.L., Nagy, S., 1986. HPLC determination of furfural and 5hydroxymethylfurfu-ral in citrus juices. J. Food Sci. 51, 1075–1076. Li, X., Ximenes, E., Kim, Y., Slininger, M., Meilan, R., Ladisch, M., Chapple, C., 2010. Lignin monomer composition affects Arabidopsis cell-wall degradability after liquid hot water pretreatment. Biotechnol. Biofuels 3, 27. Li, H.-Q., Li, C.-L., Sang, T., Xu, J., 2013a. Pretreatment on Miscanthus lutarioriparious by liquid hot water for efficient ethanol production. Biotechnol. Biofuels 6, 76. Li, J., Lin, J., Xiao, W., Gong, Y., Wang, M., Zhou, P., Liu, Z., 2013b. Solvent extraction of antioxidants from steam exploded sugarcane bagasse and enzymatic convertibility of the solid fraction. Bioresour. Technol. 130, 8–15. Liu, D., Wong, P.T.S., Dutka, B.J., 1973. Determination of carbohydrate in lake sediment by a modified phenol-sulfuric acid method. Water Res. 7, 741–746. Masuko, T., Minami, A., Iwasaki, N., Majima, T., Nishimura, S.-I., Lee, Y.C., 2005. Carbohydrate analysis by a phenol–sulfuric acid method in microplate format. Anal. Biochem. 339, 69–72. Pandey, A., Soccol, C.R., Nigam, P., Soccol, V.T., 2000. Biotechnological potential of agro-industrial residues. I: sugarcane bagasse. Bioresour. Technol. 74, 69–80. Pérez, J.A., Ballesteros, I., Ballesteros, M., Sáez, F., Negro, M.J., Manzanares, P., 2008. Optimizing liquid hot water pretreatment conditions to enhance sugar recovery from wheat straw for fuel-ethanol production. Fuel 87, 3640–3647. Poletto, M., Zattera, A.J., Forte, M.M.C., Santana, R.M.C., 2012. Thermal decomposition of wood: influence of wood components and cellulose crystallite size. Bioresour. Technol. 109, 148–153. Qin, L., Liu, Z.-H., Jin, M., Li, B.-Z., Yuan, Y.-J., 2013. High temperature aqueous ammonia pretreatment and post-washing enhance the high solids enzymatic hydrolysis of corn stover. Bioresour. Technol. 146, 504–511. Saha, A.K., Brewer, C.F., 1994. Determination of the concentrations of oligosaccharides, complex type carbohydrates, and glycoproteins using the phenol-sulfuric acid method. Carbohydr. Res. 254, 157–167. Shen, J., Kaur, I., Baktash, M.M., He, Z., Ni, Y., 2013. A combined process of activated carbon adsorption, ion exchange resin treatment and membrane concentration for recovery of dissolved organics in pre-hydrolysis liquor of the kraft-based dissolving pulp production process. Bioresour. Technol. 127, 59–65. Shimizu, K., Hayashi, S., Kako, T., Suzuki, M., Tsukagoshi, N., Doukyu, N., Kobayashi, T., Honda, H., 2005. Discovery of glpC, an organic solvent tolerance-related gene in Escherichia coli, using gene expression profiles from DNA microarrays. Appl. Environ. Microbiol. 71, 1093–1096. Theobald, A., Müller, A., Anklam, E., 1998. Determination of 5hydroxymethylfurfural in vinegar samples by HPLC. J. Agric. Food Chem. 46, 1850–1854. Tyagi, V.K., Lo, S.-L., 2013. Microwave irradiation: a sustainable way for sludge treatment and resource recovery. Renew. Sustain Energy Rev. 18, 288–305. VanWalsum, G.P., Allen, S., Spencer, M., Laser, M., Antal Jr., M., Lynd, L., 1996. Conversion of lignocellulosics pretreated with liquid hot water to ethanol. In: Wyman, C., Davison, B. (Eds.), Seventeenth Symposium on Biotechnology for Fuels and Chemicals, vol. 57/58. Humana Press, pp. 157–170. Viles, F.J., Silverman, L., 1949. Determination of starch and cellulose with anthrone. Anal. Chem. 21, 950–953. Yang, B., Wyman, C.E., 2008. Pretreatment: the key to unlocking low-cost cellulosic ethanol. Biofuels. Bioprod. Biorefin. 2, 26–40. Zhang, Y., Ezeji, T.C., 2013. Transcriptional analysis of Clostridium beijerinckii NCIMB 8052 to elucidate role of furfural stress during acetone butanol ethanol fermentation. Biotechnol. Biofuels 6, 66. Zheng, Y., Yu, X., Zeng, J., Chen, S., 2012. Feasibility of filamentous fungi for biofuel production using hydrolysate from dilute sulfuric acid pretreatment of wheat straw. Biotechnol. Biofuels 5, 50. Zhou, X., Xu, J., Wang, Z., Cheng, J.J., Li, R., Qu, R., 2012. Dilute sulfuric acid pretreatment of transgenic switchgrass for sugar production. Bioresour. Technol. 104, 823–827.