Fuel Processing Technology 91 (2010) 11461151

Contents lists available at ScienceDirect

Fuel Processing Technology

Optimizing ethanolic hot compressed water (EHCW) cooking as a pretreatment to glucose recovery for the production of fuel ethanol from oil palm frond (OPF)
Chun Sheng Goh, Hui Teng Tan, Keat Teong Lee , Abdul Rahman Mohamed
School of Chemical Engineering, Universiti Sains Malaysia, Engineering Campus, Seri Ampangan, 14300 Nibong Tebal, Pulau Pinang, Malaysia

a r t i c l e

i n f o

a b s t r a c t
The objective of this work is to study the effects of reaction temperature, reaction time and ethanol percentage on the pretreatment of oil palm fronds (OPF) using ethanolic hot compressed water (EHCW) to enhance sugar recovery in enzymatic hydrolysis. A central composite rotatable design was used to optimize the pretreatment process conditions. All variables (individual and interactive) were found to affect glucose recovery signicantly. A quadratic polynomial equation was modelled for glucose recovery by multiple regression analysis using response surface methodology (RSM). Using a 10 bar pressurized reactor, the optimum conditions for pretreatment of OPF were obtained at 180 C, 42 min and EtOH wt.% of 26.4%. At the optimum conditions, the predicted glucose recovery was 88.48%. Experimental verication of the optimum conditions gave glucose recovery well within the estimated value of the model. 2010 Elsevier B.V. All rights reserved.

Article history: Received 27 November 2009 Received in revised form 25 March 2010 Accepted 28 March 2010 Keywords: Ethanol Biofuel Lignocellulose pretreatment Response surface methodology Optimization

1. Introduction As one of the world's largest palm oil producers, Malaysia generates an enormous quantity of lignocellulosic biomass from the plantation and oil palm mill. Thus, second-generation bio-ethanol (SGB) from lignocellulose has become an attractive option due to the availability of low cost feedstock. It was predicted that lignocellulosic biomass availability in Malaysia is 47,402 dry kton per year. In other words 9549 ktons of SGB can be produced if this biomass is fully utilized [1]. Oil palm frond (OPF) contributes the largest portion of lignocellulosic residues in Malaysia's agriculture sector, accounting 44% from 38,256 dry kton of oil palm residues. Compared to other types of feedstock for bio-ethanol, OPF is essentially free, being an agricultural waste, low cost in terms of cropping practice, collection and storage. In terms of economic and environmental considerations, conversion of OPF to bio-ethanol would be a strategic and synergistic move for the palm oil industry. Lignocellulosic OPF consists of lignin, cellulose and hemicellulose. Cellulose and hemicellulose can be converted into fermentable sugars and subsequently fuel ethanol. In order to improve the accessibility of cellulolytic enzymes on OPF, the structure of lignocellulose must be broken down. The pretreatment stage is therefore critical to improve the recovery of fermentable sugars from lignocellulose. However, conventional lignocellulosic biomass pretreatment methods such as dilute sulfuric acid-based pretreatment have undesirable effects. The production of inhibitory components such as furfural from degrada-

Corresponding author. Tel.: + 60 4 5996467; fax: + 60 4 5941013. E-mail address: [email protected] (K.T. Lee). 0378-3820/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.fuproc.2010.03.029

tion of monosaccharides would signicantly reduce the recovery of bio-ethanol during fermentation. Moreover, acidic medium might corrode the reaction vessel. In this respect, several types of alternative methods have been proposed to provide better means of lignocellulose pretreatment. Recently, liquid water under elevated temperature and pressure, namely hot compressed water (HCW) has attracted considerable attention due to its acid-free characteristic. Likewise, HCW is responsible to solubilize mainly hemicelluloses fraction as liquid soluble oligosaccharides. Apart from that, the risk of product degradation and precipitation of lignin compounds is lower because high amount of solubilized products are obtained at a much lower concentration [2,3]. In line with the development of subcritical water technology, there are several additives proposed to enhance the properties of the medium. Ethanol is one of the cost-effective additives to improve the pretreatment of lignocellulose using HCW. However, ethanolic water does not remove lignin intensively, but exposing more surface area of the lignocellulose to allow the access of enzymes [4]. The in-feed ethanol used as additive could come from the nal product, and can be recovered and reused with a properly designed device due to its low boiling point. It was believed that partial removal of lignin and hemicellulose induces the formation of 10100 nm sized pores which permit the access of enzymes to the substrate [5]. The objective of this study is to optimize the pretreatment of OPF using ethanolic hot compressed water (EHCW) with the aim to signicantly improve the recovery of fermentable sugar. An experimental framework was designed using rotational Central Composite Design (CCD) to determine optimal EHCW pretreatment conditions for OPF. Effects of reaction temperature, reaction time and EtOH% were investigated to obtain the maximum recovery of glucose. The

C.S. Goh et al. / Fuel Processing Technology 91 (2010) 11461151

1147

glucose recovery is dened as the percentage of the weight of glucose recovered to the weight of potential glucose in the biomass (cellulose). 2. Materials and methods 2.1. Materials OPF was collected from oil palm cultivation in the Engineering Campus of the Universiti Sains Malaysia, Penang, Malaysia. It was shredded and grinded to a particle size smaller than 1 mm and dried under the sun. Enzymes used for hydrolysis are Novozyme 188 (263 CBU/g) and Celluclast 1.5 L (798 EGU/g) (Novozyme, Denmark) obtained from Science Technics Sdn Bhd. 2.2. Characterization of OPF Determination of extractives in OPF was done according to the laboratory analytical procedure for determination of extractives in biomass [6]. Samples were washed with deionized water and dried overnight at 80 C prior to analysis. The analysis was carried out by two steps Soxhlet extraction. First, hot water extraction was carried out for 8 h to remove water-soluble compounds such as inorganic material, non-structural sugars and nitrogenous material. Sample was weighed to nearest 0.1 mg and added to the extraction thimble for extraction in a Soxhlet apparatus containing 200 ml of deionized water. Upon completion, the sample was carefully removed from the thimble and dried. This was followed by another Soxhlet extraction, but with ethanol for 24 h to remove chlorophyll, waxes and other minor components. 200 ml of 95% ethanol was used as extraction solvent. After that, lignin was removed by using acetic acid and NaClO2. Initially, 2.5 g of biomass was heated with 150 ml water at 75 C. Then, 0.2 ml acetic acid and 2.0 g NaClO2 were added into the slurry. After 1 h, 0.2 ml acetic acid and 1.0 g NaClO2 were added every 1 h for 3 h. The slurry was ltered and washed with distilled water and acetone. The delignied residue (holocellulose) was weighed after drying for 24 h in an oven at 105 C [4,7]. Then 1 g of holocellulose was transferred into a ask with 25 ml of 17.5% NaOH aqueous solution and stirred for 40 min at 20 C. The resulting slurry was then ltered and washed using 40 ml of 10% acetic acid aqueous solution and 1 l of boiling water at two different stages. The remaining wet solid (-cellulose) was dried in oven (105 C) and weighed. Hemicellulose content was determined by subtracting the -cellulose content from the holocellulose content [4]. Lignin content was determined by using Klason method [8]. 0.5 g of the solid sample was mixed with 5 ml of 72% sulphuric acid and maintain at 30 C in a water bath for 1 h with sporadic shaking. After that, distilled water was added to the slurry until a weight of 148.67 g was reached. Subsequently the slurry was autoclaved at 121 C at 2 atm for 1 h. The cooled sample was separated through ltration. The remaining solid was heated in an oven at 105 C for 24 h. For the total sugar analysis, 0.5 g of OPF was dissolved in 10 ml of 64% (w/w) sulphuric acid and left to hydrolyse at room temperature for 2 h until complete dissolution of the sample. The mixture was then diluted to 55% sulphuric acid. After incubating for 12 h, water was added to reduce the concentration to 4% and the mixture was heated up to 120 C for 1 h. The mixture was ltered and washed with water. Calcium carbonate was added to the ltrate to precipitate the sulphates. The mixture was ltered again and the ltrate was analyzed with HPLC [9]. The composition of raw OPF and EHCW pretreated OPF is listed in Tables 1 and 2. The BET surface area and average pore size for raw and pretreated OPF were 23.41 m2/g, 79.01 and 41.39 m2/g, 99.15 respectively. According to IUPAC classication, the pore size of this OPF falls into the mesopore group.

Table 1 Composition of raw OPF. Components Moisture content Lignin Cellulose Hemicellulose Glucose Xylose Water extractive Ethanol extractive Ash content Composition (%) 10.26 12.96 30.19 24.26 40.56 17.40 5.12 5.84 11.36

2.3. Design of experiments A set of experiments (20 experiments) was designed to optimize the EHCW pretreatment of OPF using Central Composite Design (CCD) and Response Surface Methodology (RSM). Three independent variables investigated were temperature, residence time and EtOH wt.%. The CCD consists of ve levels including eight factorial points (2n = 8), six axial points (2n = 6) and six central points (nc = 6). The selection of levels for the independent variables was made based on a study reported elsewhere and some preliminary experiments [4,5]. The experiments were designed to examine linear, quadratic, cubic and cross-product effects of pretreatment condition variables. The range and levels of the independent variables studied are listed in Table 3. The value of was xed at 1.6818 for rotatable design. All variables at zero level constitute to the center points and the combination of each of the variables at either its lowest (2.0) level or highest (+2.0) level with the other variables at zero level constitutes the axial points. The experiment matrix and the experiment response (glucose recovery) are presented in Table 4. In order to minimize the effects of uncontrolled factors, the experiment sequence was randomized. 2.4. Statistical analysis Regression analysis and analysis of variance (ANOVA) were carried out using Design Expert software version 6.06 (STAT-EASE Inc., Minneapolis, USA) [10]. The experimental response y (glucose recovery, %) obtained was analyzed via RSM with the following second-order polynomial Eq. (1): y = 0 + i
n = 1 i xi

+ ii xi + ibj

= 1 ij xi xj

Table 2 Comparison of lignin, cellulose and hemicellulose content after pretreatment. Standard Raw OPF 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Lignin (%) 19.23 17.82 10.79 33.81 44.60 30.50 13.22 35.77 40.53 38.91 39.13 29.40 30.67 33.71 41.16 37.05 37.68 37.29 38.72 39.41 37.08 -Cellulose (%) 44.78 41.54 50.88 51.55 39.68 39.87 49.21 45.75 42.39 44.52 43.49 46.21 46.44 35.15 45.82 39.73 41.76 42.11 41.52 42.17 40.94 Hemicellulose (%) 35.99 40.64 38.33 14.64 15.73 29.63 37.57 18.48 17.08 16.57 17.38 24.39 22.89 31.14 13.01 23.22 20.56 20.60 19.76 18.43 21.98

1148 Table 3 Levels of the EHCW pretreatment condition variables. Variable Coding Unit Levels 2 Residence time Temperature EtOH A B C min C wt.% 30 100 10 1 42.2 120 26.2

C.S. Goh et al. / Fuel Processing Technology 91 (2010) 11461151

0 60 150 50

1 77.8 180 73.8

2 90 200 90

4.8) in 100 ml Erlenmeyer ask at 48 C for 48 h. 15 IU of glucosidase Novozym 188 and 15 FPU of Celluclast 1.5 L were loaded into the slurry for enzymatic hydrolysis. High-performance liquid chromatography (HPLC) was employed to analyse the content of glucose. 2.7. Glucose recovery High-performance liquid chromatography (HPLC) was employed to determine the glucose content of the slurry after enzymatic hydrolysis. Shimadzu liquid chromatograph with refractive index detector was used with LiChroCart Purospher STAR NH2 column which has an efciency of N50.000 N/m and a pore size of 12 nm. The temperature was set at 40 C. Acetonitrile and water were used as mobile phase in a ratio of 75:25 at a ow rate of 1 ml/min. The recovery of glucose is dened as a percentage of the weight of glucose obtained to the weight of potential glucose (cellulose) in raw OPF. 3. Results and discussion

where xi and xj are the coded independent variables and 0, i, ii and ij are intercept, linear, quadratic and interaction constant coefcients, respectively. Response surfaces and contour plots were developed using the tted quadratic polynomial equation obtained from regression analysis, holding one of the independent variables at a constant value corresponding to the stationary point and changing the other two variables. 2.5. EHCW pretreatment EHCW pretreatment was carried out using a batch reactor with a temperature controller. The stainlesssteel reactor has a total volume of 500 ml, with an electric heater and magnetic stirrer. The heating rate is approximately 5 C/min. Initially, dried oil palm frond (OPF) was charged into the reactor. Deionized water and ethanol were added at designated EtOH wt.%. Liquidsolid ratio was xed at 5:1. The reactor was kept sealed and pressurized to 10 bar to keep the water in liquid form. The reactor temperature was controlled with a programmable PID temperature controller. Once the mixture temperature reached the selected value, the magnetic stirrer was switched on at maximum speed to avoid mass transfer limitations. After designated reaction time, the reactor was quickly cooled down with ice to a temperature lower than 100 C and slowly depressurized. The liquid and solid fractions were separated through vacuo-ltration. In order to compare the effectiveness of the EHCW pretreatment, a compared set was prepared by treating the prehydrolyzed biomass with 3% (v/v) sulphuric acid at 120 C for 30 min [11]. 2.6. Enzymatic hydrolysis After ltration, the solid fraction from EHCW was washed with deionized water. The solid is then dried under the sun. 1 g of dry solid fraction was mixed with 50 ml of 0.05 M sodium acetate buffer (pH
Table 4 Experimental design matrix and results. Standard Experimental variables Reaction time, A (min) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 3% (v/v) acid 42.2 77.8 42.2 77.8 42.2 77.8 42.2 77.8 30.0 90.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0 60.0 30.0 Reaction temperature, B (C) 120 120 180 180 120 120 180 180 150 150 100 200 150 150 150 150 150 150 150 150 120 EtOH, C (wt.%) 26.2 26.2 26.2 26.2 73.8 73.8 73.8 73.8 50.0 50.0 50.0 50.0 10.0 90.0 50.0 50.0 50.0 50.0 50.0 50.0 0.0 Glucose recovery (%) 21.36 73.36 88.97 64.93 60.24 63.80 83.49 27.83 67.92 51.13 36.55 59.16 74.26 61.34 36.94 40.81 36.08 39.34 40.21 35.59 74.75

3.1. Regression analysis The glucose recovery obtained ranges from 21.36 to 88.97%. Experimental error was determined from runs 15 to 20 at the center point of the design. A regression analysis was performed to t the response function and predict the outcome of glucose recovery with a simple equation. The model is expressed by Eq. (2) which takes their actual values. Glucose recovery = 201:1938 + 2:8289 Reaction time + 1:5304 Reaction temperature + 1:3288 2 Ethanol%+ 0:0243 Reaction time + 0:0041 Reaction temperature + 0:0188 Ethanol% 0:0319 Reaction time Reaction temperature 0:0236 Reaction time Ethanol%0:0127 2 Reaction temperature Ethanol% Table 5 summarizes the analysis of variance (ANOVA) result. The model incorporates statistically tested signicant terms or regressors. The Prob N F value indicates the probability equals the proportion of the area under the curve of the F-distribution that lies beyond the observed F value. In other words, when theProb N F value is small, the particular term is considered to signicantly affect the measured response (glucose recovery in this case) of the system. In these cases the Prob N F less than 0.05 indicated that the particular term was statistically signicant. The analysis conrmed the signicant terms of A (reaction time), B (reaction temperature), C (EtOH wt.%) and their higher terms as well as interactions: A2, B2, C2, AB, AC and BC for glucose recovery. The coefcient of determination, R2 for the model was 98.65%. This indicates that only 1.35% of the total variability was not explained by the regressors in the model. The high R2 value species that the model obtained will be able to give a convincingly good estimate of response of the system in the range studied. The lack of t test which is not signicant for the model developed shows that the model satisfactorily ts the data. The predicted values versus actual values for glucose recovery were shown in Fig. 1. From the graph, the predicted values obtained were not far away from the experimental values, indicating that the model developed was successful in capturing the correlation between the pretreatment condition variables to the glucose recovery. 3.2. Model analysis From the result of regression analysis, glucose recovery was found to be affected by the main factors: reaction time (A), reaction temperature
2 2

C.S. Goh et al. / Fuel Processing Technology 91 (2010) 11461151 Table 5 ANOVA for model regression. Source Model A B C A2 B2 C2 AB BC AC Residual Lack of t Pure error
a

1149

SS

d.f.a 9 1 1 1 1 1 1 1 1 1 10 5 5

Mean square 765.67 200.82 522.69 89.64 862.80 188.06 1638.48 2287.23 801.17 645.94 9.44 13.87 5.01

F value 81.12 21.28 55.38 9.50 91.41 19.92 173.59 242.32 84.88 68.43 2.77

Probability N F b0.0001 0.0010 b0.0001 0.0116 b0.0001 0.0012 b0.0001 b0.0001 b0.0001 b0.0001 0.1440 signicant

6891.03 200.82 522.69 89.64 862.80 188.06 1638.48 2287.23 801.17 645.94 94.39 69.35 25.04

not signicant

SS, sum of squares; d.f., degree of freedom.

(B), EtOH wt.% (C) and their respective higher-order terms (A2, B2 and C2). Signicant interactions terms were found to exist between the main factors (AB, AC and BC). Fig. 2 illustrates the three-dimensional response surface plot of glucose recovery at EtOH wt.% of 50% to investigate the interactive effect of reaction time and reaction temperature. It was obvious that reaction temperature has a massive interaction with reaction time, as indicated by the high F value as shown in Table 5. For shorter reaction time, higher temperature is required to obtain high recovery of glucose, owing to the fact that intrinsic rate constants are strong functions of temperature. From Table 2, by comparing runs 1, 2, 5, and 6 with runs 3, 4, 7, and 8, it can be clearly seen that the content of hemicellulose dropped tremendously at high temperature. It was therefore concluded that high temperature tends to increase the removal of hemicellulose. On the other hand, for a longer reaction time high temperature gives an adverse effect to the glucose recovery. As shown in Table 2, at low temperature cases, runs 3, 4, 7, and 8, the content of cellulose is reduced when the reaction time was prolonged. The adverse effect is probably due to the solubilization of cellulose and the degradation of glucose to furfurals in hot water at longer reaction time [12]. In addition, over cooking of biomass for long period and high temperature is suspected to burn the surface of biomass. Enzymatic digestion will be inhibited by the burnt surface [5]. The response surface plot of glucose recovery at 150 C with a different reaction time and ethanol % is shown in Fig. 3. Generally the trend of glucose recovery is rather similar to that of Fig. 2. The addition of ethanol into hot compressed water is believed to partially facilitate the removal of lignin and hemicellulose from lignocellulosic materials [5]. This would cause the structure of the pretreated OPF to be more porous as reported by Teramoto et al. [5]. The characterization of OPF shows that the BET surface area of pretreated biomass has larger values than the untreated one, which are 41.39 and 23.41 respectively. This shows that EHCW helps to increase the surface area per unit volume. It is suspected that the formation of pores by EHCW allows easier access of enzymes on the cellulose. By comparing hemicellulose content for runs 13, 14 and center points in Table 2, it is shown that the content of hemicellulose dropped with higher ethanol content. Thus higher recovery of glucose (N67%) can be obtained at a short reaction time with high percentage of ethanol in the medium. Despite that, high content of ethanol does not provide any advantages for longer reaction time. At concentrated EHCW medium, the recovery of glucose decreased after a long period of reaction. This is probably because at a longer reaction time, the initially exposed cellulose tends to be solubilized and hence degraded into furfurals. The fact that lower cellulose proportion found in run 8 (42.39% at 77.8 min) compared to run 7 (45.75% at 42.2 min) supports the inference. On the other hand, at low percentage of ethanolic water, the removal of lignin and hemicellulose is slower, and probably this slow rate prevents the quick degradation of cellulose and allows the glucose recovery to increase at longer reaction period.

Fig. 4 shows the interaction between ethanol concentration and temperature in a three-dimensional response surface plot for a 1 h reaction which reveals two important ndings. For the rst nding, at concentrated EHCW, temperature does not signicantly inuence glucose recovery. However, this is not true for low EtOH wt.% in HCW. The recovery of glucose was found to signicantly increase at higher temperature. The following explains this phenomenon. In EHCW, ethanol plays the role as the solvent for hydrophobic components, such as lignin. Thus, at low ethanol wt.%, there is insufcient solvent and therefore, higher reaction temperature will play a signicant role to increase the recovery of glucose. High reaction temperature is able to disrupt hemicellulose. This is proven by comparing the hemicellulose content of run 1 to run 3, and run 2 to run 4, where the hemicellulose content is largely reduced at higher temperature. However, at high ethanol wt.%, the presence of sufcient ethanol (solvent) will be able to solubilize more lignin to increase the recovery of glucose up to a point that the positive effect of temperature is no longer signicant. Therefore at concentrated EHCW, the recovery of glucose remains almost stagnant with the increase of temperature. Another important nding is that the recovery of glucose resulted from diluted EHCW is higher than concentrated EHCW at higher reaction temperature. This is because in concentrated ECHW case, high temperature medium tends to solubilize and degrade cellulose which is largely exposed to the liquid due to the quick removal of lignin by large amount of ethanol. For comparison purposes, OPF was also subjected to the more common pretreatment method using acid. The conversion of cellulose

Fig. 1. Predicted versus actual glucose recovery.

1150

C.S. Goh et al. / Fuel Processing Technology 91 (2010) 11461151

Fig. 2. (a) Three-dimensional response surface plot of glucose recovery: effect of residence time and reaction temperature, EtOH wt.% = 50%; (b) interaction plots between residence time and reaction temperature.

Fig. 3. (a) Three-dimensional response surface plot of glucose recovery: effect of residence time and EtOH wt.%, reaction temperature = 150 C; (b) interaction plots between residence time and EtOH wt.%.

into fermentable glucose was found to be 74.75% by using 3% sulphuric acid for pretreatment. However, with the addition of ethanol into HCW (as reported in this study), the glucose recovery can signicantly increase to almost 90% (even without optimization). Furthermore, it must be noted that using ethanol instead of acid as additive to hot compressed water helps to reduce the risk in costs, plant safety and environmental sustainability, owing to the reusability of ethanol and its non corrosive nature. 3.3. Optimization of glucose recovery The optimization of glucose recovery was carried out based on the three variables which were in the range of experimental runs. The software predicted that optimized conditions for glucose recovery were obtained when the reaction temperature, residence time and EtOH wt.% were at 180 C, 42 min and 26.4, respectively with a predicted glucose recovery of 88.48%. The experimental glucose recovery was 90.39%. This shows that the experimental value obtained was in good agreement with the value calculated from the model.

4. Conclusion The result from the EHCW pretreatment of OPF shows high recovery of glucose at moderate temperature. A CCD was conducted to study the effects of reaction temperature, time and ethanol percentage on the pretreatment of OPF. The optimal values of the variables were as follows: 180 C, 42 min and EtOH wt.% of 26.4%. The predicted glucose recovery was 88.48 wt.% at the optimum conditions. The experimental value was well within the estimated value of the model. By operating the pretreatment at optimum conditions, it avoids using the longest possible time and highest ethanol content, subsequently reduces the overall production cost. Acknowledgements The authors gratefully acknowledge the nancial support from Ministry of Higher Education, Malaysia (FRGS Grant No. 6071162) and Universiti Sains Malaysia (Waste Management Cluster Research University (RU) Grant, Postgraduate Research Grant Scheme and USM Fellowship) that has made this work possible.

C.S. Goh et al. / Fuel Processing Technology 91 (2010) 11461151

1151

References
[1] C.S. Goh, K.T. Tan, K.T. Lee, S. Bhatia, Bio-ethanol from lignocellulose: Status, perspectives and challenges in Malaysia, Bioresource Technology 101 (2010) 48344841. [2] O. Bobleter, Hydrothermal degradation of polymers derived from plants, Progress in Polymer Science 19 (1994) 797841. [3] A.T.W.M. Hendriks, G. Zeeman, Pretreatments to enhance the digestibility of lignocellulosic biomass, Bioresource Technology 100 (2009) 1018. [4] Y. Teramoto, S.H. Lee, T. Endo, Cost reduction and feedstock diversity for sulfuric acid-free ethanol cooking of lignocellulosic biomass a pretreatment to enzymatic saccharication, Bioresource Technology 100 (2009) 47834789. [5] Y. Teramoto, S.H. Lee, T. Endo, Pretreatment of woody and herbaceous biomass for enzymatic saccharication using sulfuric acid-free ethanol cooking, Bioresource Technology 99 (2008) 88568863. [6] T. Ehrman, LAP 010 standard method for the determination of extractives in biomassAvailable from, http://cobweb.ecn.purdue.edu/~lorre/16/research/LAP010.pdf1994. [7] L.E. Wise, M. Murphy, A.A. D'Addieco, Chlorite holocellulose, its fractionation and bearing on summative wood analysis and on studies on the hemicelluloses, Paper Trade Journal 122 (1946) 3543. [8] A. Rodrguez, A. Moral, R. Snchez, A. Requejo, L. Jimnez, Inuence of variables in the hydrothermal treatment of rice straw on the composition of the resulting fractions, Bioresource Technology 100 (2009) 48634866. [9] T. Foyle, L. Jennings, P. Mulcahy, Compositional analysis of lignocellulosic materials: evaluation of methods used for sugar analysis of waste paper and straw, Bioresource Technology 98 (2007) 30263036. [10] D.C. Montgomery, Design and Analysis of Experiments5th edition, John Wiley & Sons, USA, 2001. [11] G. Guo, W. Chen, W. Chen, L. Men, W. Hwang, Characterization of dilute acid pretreatment of silvergrass for ethanol production, Bioresource Technology 99 (2008) 60466053. [12] T. Rogalinski, T. Ingram, G. Brunner, Hydrolysis of lignocellulosic biomass in water under elevated temperatures and pressures, Journal of Supercritical Fluids 47 (2008) 5463.

Fig. 4. (a) Three-dimensional response surface plot of glucose recovery: effect of reaction temperature and EtOH wt.%, reaction time = 60 min; (b) interaction plots between reaction temperature and EtOH wt.%.