Statistical optimization for deconstruction of poplar substrate by dilute sulfuric acid for bioethanol production

ABSTRACT In this study, response surface methodology was used to study the effects of H2SO4, concentrations, substrate loading and residence time on liberation of reducing sugars (RS), total sugars (TS) and total phenolic compounds from poplar leaves and twigs. Box–Behnken design with three variables and three levels showed maximum release of total phenolic compounds (57.39 mg/ml) corresponding to 0.8% H2SO4 concentration, 15% substrate level with residence time of 4 h. Under these conditions, the TS and RS released up to 161.20 and 5.24 mg/ml, respectively. Analysis of the pretreated substrate by Fourier transform infra-red and X-ray diffraction revealed the effectiveness of pretreatment conditions. Second-order polynomial equation using analysis of variance was employed for analyzing the results. The proposed model was found very significant with F-value of 48.39. The R2 and adjusted R2 values also revealed the accuracy of predicted response. GRAPHICAL ABSTRACT


Introduction
Lignocellulosic biomasses are the most abundant natural resources for biofuels production. The bioresource is available in the form of softwood, hardwood, grasses, agricultural and plants wastes (1). High percentages of cellulose (35-50%) and hemicellulose (20-35%) which could be converted into fermentable sugars by action of enzymes render these biomasses very attractive feedstocks. However, lignocellulosic materials contain 10-25% lignin, the main layer of plant cell wall, which makes it more rigid and provides recalcitrant structure which inhibits cellulose exposure to the environmental stresses.
The hard and rigid structure of lignocellulosic biomass could be deconstructed by employing various pretreatment techniques. Diversity of pretreatment techniques involves physical, biological, chemical or combinations of them. Every treatment has significant impact on conversion of lignocellulosic biomass (2,3). In acid treatment, the most commonly used acids are H 2 SO 4 , HCl and HNO 3 which hydrolyze the hemicellulose contents of lignocellulosic biomasses (4)(5)(6)(7).
Response surface methodology (RSM) is a statistical and mathematical modeling technique used to analyze the influence of different variables and their interactions on yield. Various biotechnological processes are being widely optimized by this technique (8)(9)(10). Conditions' optimization by routinely used one factor at a time approach is time taking and laborious which might lead to inaccurate findings at the end. Whereas, RSM provides interaction of multiple variables on response in shorter time. Scientists are therefore currently applying this approach to optimize various process conditions for maximum yield. In this study, various concentrations of dilute H 2 SO 4 were used to optimize pretreatment conditions of poplar waste (leaves and twigs) using three variables with three levels by Box-Behnken design (BBD) for bioethanol production.

Results and discussion
In this study, poplar leaves and twigs (1:1 ratio) dried powder was pretreated with dilute sulfuric acid. The experiments were conducted by two methods. In one method only dilute H 2 SO 4 (chemical) treatment was done, whereas in the second method dilute H 2 SO 4 treatment was followed by autoclaving at 121°C, 15 psi for 15 min (thermochemical). We presumed that maximum liberation of total phenolic compounds means that there is greater breakdown of lignin. As the main target was to get maximum degradation of lignin, total sugars (TS) and reducing sugars (RS) were also measured which represented the degradation of cellulose and hemicellulose content of poplar biomass. For this purpose, BBD of RSM with three   factors and three levels was applied for optimization of pretreatment conditions. The three factors used for pretreatment were sulfuric acid concentration (X 1 ), substrate loading (X 2 ) and residence time (X 3 ). The levels of these three variables are mentioned in Table  1.Three analyses that is, TS (mg/ml), RS (mg/ml) and total phenol (TP, mg/ml) of the pretreated filtrate were measured. The response was calculated by second-order polynomial regression equations (Equations (1)- (6)).
BBD results are mentioned in Table 2. It was found that maximum total phenolic compounds (57.38 mg/ml) and TS (161.1984 mg/ml) were liberated at 15% (w/v) of the substrate loading and 0.5% (v/v) sulfuric acid concentration with soaking time of 4 h at room temperature. When the same chemical treatment was followed by autoclaving, the results (Table 3) revealed that there was higher degradation releasing TS up to 303.064 mg/ml and RS up to 17.053 mg/ml. In this thermochemical treatment sugar production was higher as compared to the chemical treatment. However, total phenolic compounds released in the thermochemical treatment were lower than the chemical treatment alone. This means that thermochemical treatment was very effective in solubilization of hemicellulose content with lesser amounts of available phenolic compounds.
Regression equations for chemical treatments followed by autoclaving: Y (Reducing sugars, mg/ml) Y (Total phenolic compounds, mg/ml) All the data were statistically analyzed and the results were found significant as explained by regression equations. The proposed model was found very significant for removal of total phenolic compounds in both the treatments. For chemical pretreatment method, the Fisher's F-test value of 17.0, 119.36 and 62.57 were observed for TS, TPs and RS, respectively (Table 4). In case of thermochemical treatment, the F-test values were found 98.44, 27.74 and 14.51 for TS, TPs and RS, respectively ( Table 5). The P-value of .009, .000 and .001 were found very significant for TS, TPs and RS and well explained by the model ( Table 4). The coefficient of determination (R 2 value) predicts the goodness of fit of the model. The coefficient of determination values of 83.68%, 98.85% and 93.57% (chemical treatment) for TS, TPs and RS respectively revealed the accuracy of the model. For the thermochemical treatment, the R 2 values were 96.87%, 91.31% and 85.07% for TS, TPs and RS, respectively. Furthermore the credibility of the model was supported by adjusted R 2 values (54.30%, 96.82% and 81.99% for, TS, total phenolic compounds and RS, respectively). Figures 1 and 2 show the contour plots of TS, TP and RS released during pretreatment at different conditions. The H 2 SO 4 followed by steam under pressure treatment resulted higher degradation of hemicellulose content and consequently released more sugars. In our study, highest amount of RS in terms of glucose was released (17.053 mg/ml) with 1% H 2 SO 4 followed by steam at 121°C for 15 min. A previous study reported that 1.3-1.9 g/100 g of glucose was released from raw material treated at 200-220°C for 5 min (11). Maximum total phenolic compounds (57.38 mg/ml) were released with 0.8% (v/v) H 2 SO 4 at residence time of 4 h at room temperature. Kundu and Lee (12) reported that 2.47 g/L of total phenolic compounds were released from poplar treated with 50 mM oxalic acid at 170°C. Another study conducted on ethanol production from water hyacinth reported that H 2 SO 4 concentration of 1% was best for maximum release of RS (33.35 g/L) during pretreatment process (13). When switch grass was pretreated with dilute sulfuric acid, maximum sugars were produced at 1% concentration of sulfuric acid (14). Likewise, it had been reported that the optimized pretreatment conditions for conversion of cattail cellulose to sugars were: sulfuric acid concentration of 0.5%, temperature 180°C with 5 min exposure (15). The total phenolic compounds released in our study were higher as compared to others depicting that the sulfuric acid treatment is better as compared to oxalic acid pretreatment. The observed and predicted values from the model are shown in Figure 3 which indicate that most of our findings were very close to the values predicted by the model.
The pretreated substrate was structurally analyzed by Fourier transform infra-red (FTIR) spectroscopy to study effect of the pretreatments (Figure 4). Previous studies suggested that for chemical compositional analysis of wood, FTIR spectroscopy is a good established technique (16,17). Our most concerned peak area ranged from 1733 to 816 cm −1 . The peak at 1733 cm −1 corresponds to C9O linkage in xylans which was slightly decreased in acid treatment showing degradation of hemicellulose (18). The band at 1455/1418 cm −1 represented C-H deformation in lignin and carbohydrates and was highly extended in the treated substrate as compared to untreated (17,19). The absorption peak at 1373 cm −1 depicted C-H deformation in cellulose and hemicellulose (17,19). The peak observed at 1317.6 cm −1 illustrated C-H vibration in cellulose and C-O vibration in syringyl derivatives which are important components of lignin. The band noted at 1231.9 cm −1 corresponded to syringyl ring and C-O stretching in lignin and xylan (17,19). The peak at 1157.3 cm −1 in acid-treated poplar substrate followed by steam described C-O-C vibration in cellulose and hemicellulose. The band at 1028 cm −1 denoted C-O vibration in cellulose and hemicellulose. Figure 5 illustrates the X-ray diffraction (XRD) pattern of untreated, H 2 SO 4 treated and H 2 SO 4 followed by steam treated poplar biomasses. Crystallinity index depicted the crystalline structure of cellulose. The crystallinity index of untreated poplar biomass was 36.5% which increased in H 2 SO 4 treated (51.8%) and H 2 SO 4 followed by steam pretreatment (58.0%). This increased crystallinity indices of the treated biomasses revealed removal of lignin and hemicellulose contents as evidenced from previous studies (11,20). No change in positions of the peaks demonstrated stability in structure of cellulose. Various pretreatment conditions had been correlated with crystallinity indices which ultimately affect the enzymatic hydrolysis process (21,22).
After pretreatment the mass balance was measured and percent degradation was calculated. Results ( Figure 6) illustrated that maximum degradation   (80%) was observed with 0.6% H 2 SO 4 concentration with residence time of 6 h at room temperature. When the same treatment was followed by steam (121°C, 15 min) exposure, maximum degradation of 66.5% and minimum degradation of 14.7% were noted. Martin-Davison et al. (15) pretreated poplar hybrids at 200-220°C for 5 min and reported solid recovery rate of 74.9-67.3 g/100 g dry weight.

Substrate
Poplar leaves and twigs were obtained from canal area of University of the Punjab new campus, Lahore, Pakistan. The leaves and twigs were washed to remove the dust particles and oven dried at 70°C till consistent weight. After that the material was chopped and ground to powder form (2 mm) by beater mill and subsequently used for the pretreatment process.

Pretreatment of the substrate
Pretreatment method used in this study has been described in an earlier report (23). Ten grams of pulverized poplar leaves and twigs (1:1 ratio) were soaked in different concentrations of H 2 SO 4 at the ratio of 1:10 (solid:liquid) for various times at room temperature. After that the samples were subjected to pressurized heat treatment as per experimental design. After treatment the samples were filtered and solid residues were washed up to neutrality.

FTIR spectroscopy analysis
The treated and untreated biomass was analyzed by Agilent technologies Cary 630 FTIR. The absorption spectra were recorded in the range of 4000-400 cm −1 .

XRD analysis
The pretreated and untreated poplar biomass was analyzed for crystallinity index by Bruker diffractometer using D8 Advance model. The analysis conditions used were: Cu tube, 40 kV voltage and 35 mA current. The peak patterns were obtained within 2θ range of 0-50 o . The crystallinity index was measured by ratio of intensity difference in peak position.

Analytical methods
The filtrate obtained after pretreatment was analyzed for estimation of RS by DNS method (24). Total sugar content in the filtrate was estimated as described by Dubois et al. (25). Total phenolic compounds released during pretreatment in filtrate were estimated by the method of Carralero et al. (26).
Experimental design BBD with three factors and three levels was used to optimize the pretreatment conditions in this study. The independent variables used were H 2 SO 4 concentration (X 1 ), substrate concentration (X 2 ) and time (X 3 ), and their levels are mentioned in Table 1. This design is most suitable for quadratic response surface and generates second-order polynomial regression model. The relation between actual and coded values was described by the following equation: where x i and X i are the coded and actual values of the independent variable, X o is the actual value of the independent variable at the center point and ΔX i is the change of X i . The response is calculated from the following equation using Minitab software (17th version).

Conclusion
In conclusion, results of this study suggested that 0.8% (v/ v) H 2 SO 4 concentration, 15% substrate and 4 h residence time at room temperature effectively liberated maximum total phenolic compounds from the poplar biomass. This maximum release of phenolic compounds indicated that H 2 SO 4 effectively deconstructed the biomass structure and most of lignin was broken down and thus provided more opportunities for enzymatic attack which subsequently enhanced saccharification process.

Disclosure statement
No potential conflict of interest was reported by the authors.