3.1. Composition of autohydrolysis liquor and spent solid of autohydrolyzed sunflower stalks
In the former work, it was reported that cellulose (36%) was the main component of sunflower stalk, followed by xylan (22%) and lignin (klason lignin: 26% and acid soluble lignin: 1%). The rest of the constituents of it were arabinan (0.7%), acetyl groups (3%), uronic acid (6%), protein (1%) and ash (3%) .
The autohydrolysis of sunflower stalk was accomplished at 160oC for 1 h according to the result of the previous study  and from the autohydrolysis of 100 kg sunflower stalk, 25 kg of the total dissolved substrate (ds) was obtained. The main components of the liquor were xylooligomers, some monosaccharides (mainly xylose followed by glucose and arabinose) and acetic acid, which were derived from either cellulose or hemicellulose fraction of the sunflower stalk. The resulting liquor was concentrated and the composition of concentrated autohydrolysis liquor (CAL) is presented in Table 2.
The concentration of xylooligomer in the CAL of the sunflower stalk was 27 g/L. In addition to xylooligomer, acetyl groups linked to oligomer, glucooligomer deriving from cellulosic components of the raw materials, arabinooligomer were present in the CAL, and their levels were 3.30, 1.35 and 1.50 g/L, respectively. The furfural production, dehydration product of the sugars, remained at low levels (Table 2) when the results were compared with a previous study done with dilute acid hydrolysis .
As expected, the pretreatment solubilized the hemicellulosic fraction of the sunflower stalk to a certain grade and changed the combination of dry matter. Table 3 indicates the composition of spent solid from autohydrolyzed sunflower stalk (SSAS). With the hydrolytic treatment of lignocellulosic materials, hemicellulose became partially soluble, the main component of SSAS was cellulose (43%) followed by lignin (32%) and hemicellulose (18%). The previous studies on the determination of the composition of spent solid of the hydrothermally treated sunflower stalk showed similar results, its major component was cellulose (47.9%) followed by lignin and hemicelluloses [32, 42].
The different studies reported similar results for the autohydrolysis of other lignocellulosic material sources. The autohydrolysis of agave bagasse at 160 ℃ for 50 min showed that it had a main component as cellulose (35%), followed by lignin (25%) and hemicellulose (6%) . The previous study presented that cellulose (78%), lignin (6%) and hemicellulose (1%) constituted the solid phase of autohydrolyzed wheat straw (at 190℃ for 15 minutes) . It was reported that the xylan content in the solid residues decreases with the increase of the pretreatment temperature [17, 19]. The differences in composition and structure of the sunflower stalk and autohydrolysis process conditions were thought to be the reasons for differences in results between this study and those previously reported.
3.2. Enzymatic hydrolysis
To get a loud yield of xylose, CAL was hydrolyzed by T. longibrachiatum xylanase from the GH10 family, the commercial endo-xylanase preparation. In general, unlike GH11 xylanases, GH10 xylanases produce low-DP oligosaccharides, mainly xylose monomer and xylobiose [44, 45]. Since the autohydrolysis solubilizes the hemicellulosic fraction of the sunflower stalk as oligomers, its CAL is an ideal substrate for the xylanases to produce xylose.
The use of the high concentration of the substrate may decrease the yield due to enzyme inhibitory activity that presents in the reaction mixture, while keeping the enzyme level high may cause an increase in product cost. Therefore, it is necessary to find the ideal substrate concentration and enzyme activity. The different concentrations of CAL of the sunflower stalk were treated at different levels of T.longibrachiatum xylanase for 46 hr at pH 4.6 and 50 oC. The experimental range and levels of the substrate concentration and enzyme activity investigated are given in Table 1. It was observed that all the hydrolysis progress curves of CAL (Figure 1) showed an increase in the first 24-hr and flatten after 24 hr. The highest reducing sugar was obtained at 11th run (75 mg ds/mL CAL and 240 U/mL) and the reducing sugar concentration of this run was 21 mg/mL at the end of the 24 hr. The lowest reducing sugar was obtained at the 10th run (14.65 mg ds/mL CAL and 150 U/mL) and it was 3.8 mg/mL after 24 hr hydrolysis period.
The compositions of the hydrolysis products at the end of 24-hr reaction are presented in Figure 2. The highest xylose (10.2 mg/mL) and glucose (1.1 mg/mL) were found at 11th run and the highest arabinose concentration (0.022 g/L) was found in the 3rd run. Since plant xylans are partially acetylated, the concentration of acetic acid increased with the increase in the substrate concentration. The highest acetic acid (1.9 mg/mL) concentration was found in the 13th run.
Since the solid residue, SSAS, had 18% hemicelluloses (Table 3), it was also hydrolyzed with T. longibrachiatum xylanase for 48 hr at pH 4.6 and temperature 50 °C. The result of the hydrolysis reaction (Figure 3A and B) was monitored by measuring the individual sugar concentration in the hydrolysate and the highest sugar concentration was determined as xylose and then glucose. The highest xylose yield was found after 24-hr hydrolysis period (19% or 3.0 mg/mL xylose) while the highest glucose level was found after the 48-hr hydrolysis period (2.5 mg/mL) (Figure 3A). The conversion of residual xylan to xylose increased with the hydrolysis time, but the selectivity (the ratio of xylose to other sugar) decreased due to the release of glucose and arabinose (Figure 3B). This is because xylanases from the GH10 family are less specific to xylan  and can also act on cellulose substrates .
It is not easy to compare the results we obtained in this study with the results obtained in previous studies due to differences in the type of biomass or hydrolytic reagent. The spent solid is mainly composed of cellulose followed by xylan, if cellulase was used as an enzyme, the yield of glucose is higher or if xylanase was used, the yield of xylose is higher at the end of the hydrolysis process. A study on the enzymatic hydrolysis of spent solid of hydrothermally processed corn bran with different xylanase activity from Eupenicillium parvum 4-14 and Aspergillus oryzae showed that A. oryzae derived enzyme had higher cellulase activity and it produced xylose and glucose with 6.96% and 31.04% yields, while E. parvum 4-14 derived enzyme had high hemicellulase activity and produced xylose and glucose with 45.43% and 16.77% yields . The study done on the spent solid of the hydrothermally treated corn bran with enzymes with xylanase different commercial xylanase showed that xylose and glucose were obtained as 8% and 1% yield, respectively, but when cellulase was used for the enzymatic hydrolysis, glucose yield was more than 25% . The previous studies on the production of sugars from enzymatic hydrolysis of spent solid of hydrothermally treated lignocellulosic biomass showed that cellulosic enzymes produced glucose with a higher yield than hemicellulosic sugars [3, 16, 29, 32].
The design of this research including response variables is given in Table 4, the percentage conversion ratio of xyloligomers to xylose in autohydrolysis liquors is expressed as xylose yield, the ability of xylanase to hydrolyze xylan is expressed as selectivity.
In this study, high xylose yield and selectivity were desired to keep other hydrolysis products low in the hydrolyzate. According to Table 4, the highest xylose yield is 59.09% (substrate: 50 mg ds/mL CAL, enzyme: 150 U/mL), the highest selectivity is 9.90 g/g, obtained, at run 9 (substrate: 25 mg ds/mL CAL, enzyme: 60 U/mL). The lowest xylose yield is 37.42% at run 7 (substrate: 50 mg ds/mL CAL, enzyme: 22.74 U/mL). Among the experiments with the same substrate concentrations, but different enzyme activity (runs 1, 2, 5, 6, 7, 8 and 12), run 7 was found to have the lowest xylose yield. At the same substrate concentration, increasing enzyme level from 22.74 to 150 U/mL increased the xylose yield. However, increasing the enzyme level to more than 150 U/mL did not make significant changes in the yield.
The quadratic models for xylose yield and selectivity are shown in the Eq. (2) and Eq. (3), in which Y1 and Y2 represent the xylose yield and selectivity, respectively, as the function of the substrate level (X1) and enzyme activity (X2).
To fit the experimental data and response function, the regression analysis was performed. The quadratic models for xylose yield and selectivity were appraised by ANOVA (Table 5). The p-values (<0.05) and F values (33.4 and 102.44) of both models showed their significance. Besides, the models did not show a lack of fit and the determination coefficients (R2) for the xylose yield (Y1) and the selectivity (Y2) were 0.96 and 0.99, respectively. The adjusted determination coefficient (Adj R2, 0.93 and 0.98) also verified the importance of the models. Table 5 presents Pred R2 values that are in acceptable accord with the Adj R2 values, showing a good adjustment between the determined and estimated values. It was seen that the experiments were precise and reliable, with lower values of the coefficient of variation (C.V. 3.41 and 1.34%). Adequate precision (signal to noise ratio) greater than 4 indicates adequacy of the model precision. The ratios of both the models (16.7 and 30.3) were found greater than 4 (Table 5).
Figure 4 presents the diagnostic plots to be used to assess the sufficiency of the models. As seen from Figures 4A and 4D, the predicted and determined values for both responses are in agreement with each other sufficiently. A normally distributed was observed in the normal % probability plots of residuals for both responses without deviation of the (Figure 4B and 4E). All the data points lied within the limits (±3) in the internally studentized residuals plots (Figure 4C and 4F).
Figure 5 presents the response surface plots to predict the xylose yield and selectivity over the independent variables (substrate concentration and enzyme activity). The maximum xylose yield (59%) was obtained at 36 mg ds/mL CAL substrate level and 196 U/mL enzyme activity (Figure 5A). The maximum selectivity (9.8 gg-1) was get at 25 mg ds/mL CAL substrate concentration and 60 U/mL enzyme activity (Figure 5B).
Decreased xylose yield and increased selectivity with the increase in the substrate concentration, while increased xylose yield and decreased selectivity with the increase in enzyme level were observed (Figure 2 and 5). Both the factors displayed (enzyme and substrate level) significant linear (p<0.05) effects on xylose yield. Substrate concentrationxenyzme level showed positive significant effects on both responses. The quadratic coefficient of enzyme level presented significant negative effects on both responses, but that of substrate concentration showed significant effect on the selectivity and non significant effect on the yield (Table 5).
Based on the two models, a graphical optimization was conducted by overlaying the contour plots of the models (Figure 6). The optimum operating condition with high levels of xylose efficiency and selectivity was chosen by applying the following criteria: xylose yield>50 and selectivity> 9 gg-1, and 60 mg ds/mL CAL substrate concentration and 234 U/mL enzyme level were chosen as the optimal working conditions foreboded through the program. To confirm the result, the enzymatic hydrolysis was conducted in duplicate at the optimum conditions. The change in xylose, glucose, acetic acid, xylose yield and selectivity along the hydrolysis process, conducted at the optimum conditions, are offered in Figure 7.
As appeared from Figure 7A, xylose concentration increases up to 24 hr. The maximum xylose concentration was measured at 48 hr as 9.3 mg/mL; after this point, its concentration decreased. The maximum glucose concentration was measured at 72 hr as 1.7 mg/mL. Xylose yield was found as 70% for 24 hr and 72% for the 48 hr hydrolysis period while selectivity was found maximum at 24 hr as 8.2 g/g and decreased with time (Figure 7B). Since the reaction taking longer than 24 hr did not cause an important rise in xylose yield, 24-hr hydrolysis seems suitable for the production of xylose.
Table 6 presents the results of the optimization work and the results of estimated conditions suggested by the statistical program used for the optimization. Xylose yield obtained in the experiment is higher than the predicted value while selectivity is slightly lower than the predicted value.
The differences in lignocellulosic material, pretreatment conditions and enzyme source make it difficult to compare the data obtained in this study with the results obtained from previous studies. In the previous study about the enzymatic hydrolysis of palm waste autohydrolysis liquor (treated at 121ºC for 20-80 min) with T. viride xylanase, xylose yield was found as 25.64% . Another study about hydrolysis of the autohydrolysis liquor of corn husk, obtained at 190ºC for 10 min, with the enzyme cocktail including endoxylanase, beta xylobiase and arabinofuranosidase reported the xylose yield as 35% . A study on the production of xylose by the coupling of the hydrothermal treatment with the enzymatic post-hydrolysis of corn cob found the xylose yield as 80% . As a result of hydrolysis of the liquor obtained as a result of autohydrolysis of tobacco stalks at 160 ° C for 1 hour, 79.8% xylose yield was obtained with T. longibrachiatum xylanase . The hydrolysis of the autohydrolysis liquor of corn bran, obtained at 165ºC for 40 min, with the enzyme blend from Aspergillus oryzae and Eupenicillium parvum xylanase reported the xylose yield as more than 80% . The results obtained from this work seen that the xylose can be obtained with high yield by the coupling of the hydrothermal treatment and enzymatic hydrolysis methods.