Composition of raw sorghum samples
The composition of raw pith, rind and whole stem on a dry matter basis are shown in Table 1. The content of esterified pCA in pith was 2.21%, which was slightly higher than that in rind (2.08%). The content of cellulose (40.64%) and lignin (24.04%) in rind was slightly higher than that in pith (37.20% and 21.48%, respectively), while the xylan content of the rind (17.33%) was slightly lower than that of the pith (18.51%). The total carbohydrates including glucan, xylan, and arabinan in rind and pith reached 60.29% and 58.25%, respectively. It was worth noting that other components such as ash, extractives and lower content of sugars also exist in sorghum stems according to previous report [29], despite these constituents were not analyzed in this study. Variations in sorghum stem composition from previous researches were observed in this study, which may be due to the differences in the geographical location, fertilization, heterogeneity of feedstock of the samples and several other environmental factors [8, 29].
Comparison of NaOH-ethanol pretreatment of different sorghum stem parts
Box–Behnken models were designed to optimize various process parameters (NaOH loading A, ethanol content B, temperature C and time D) in NaOH-ethanol pretreatment of different sorghum stem parts. The following equations were derived for the analysis of the release of pCA and the recovery of xylan in residues.
Pith:
PCA release yield (%) = + 79.99 + 27.46A + 7.01B + 4.73C + 5.32D + 1.63AB - 4.27AC + 2.71AD + 1.70BC - 7.82BD - 1.13CD - 15.69A2 - 2.98B2 - 3.00C2 - 0.99D2
Xylan recovery yield (%) = + 80.39 - 9.88A + 3.26B - 0.35C - 0.4392D -0.54AB + 1.11AC - 0.95AD - 2.83BC + 4.83BD + 2.48CD - 3.13A² - 1.23B² - 2.51C² - 2.40D²
Rind:
PCA release yield (%) = + 85.52 + 33.62A + 6.54B + 3.57C + 7.38D + 5.03AB - 0.57AC - 1.27AD - 0.59BC - 7.92BD - 1.83CD - 22.48A2 - 5.14B2 - 4.22C2 - 2.50D2
Xylan recovery yield (%) = + 81.24 - 3.51A + 4.33B - 0.5442C - 1.56D + 3.00AB - 0.9225AC - 0.07AD + 0.945BC - 1.93BD - 0.41CD
Whole stem:
PCA release yield (%) = + 82.06 + 28.92A + 6.67B + 4.79C + 5.44D + 2.93AB – 4.34AC + 1.12AD + 0.4875BC - 7.45BD + 0.93CD - 16.88A2 - 3.46B2 - 2.45C2 - 1.24D2
Xylan recovery yield (%) = + 80.39 - 7.23A + 4.27B - 0.1708C - 0.56D -1.96AB + 1.65AC - 0.32AD + 2.14BC - 1.68BD - 1.69CD - 5.08A² - 1.20B² + 0.338C² - 0.5143D²
The effect of different variables on the release of pCA and the recovery of xylan in residues from different sorghum parts were investigated. Overall, NaOH concentration and ethanol content were the first and second most influential factors on the pCA release yield and xylan recovery yield, respectively. As shown in Table 2, the released contents of pCA from the sorghum stem were strongly depended on the NaOH concentration; however, some variation existed among different stem parts. At a low NaOH concentration (i.e. 0.5%), the release of pCA was very low, but significantly higher pCA was released from the pith than from the rind under the same conditions. As the NaOH concentration increased and the ethanol content exceeded 40% (v/v), the release of pCA from the rind slightly exceeded over that from the pith under the same conditions. NaOH concentration is also the key factor affecting the xylan dissolution, and more xylan was dissolved as NaOH concentration increased. In general, xylan was more resistant to be dissolved from the rind than from the pith in the same NaOH-ethanol conditions. This may be related to the difference in cell types and chemical composition in different sorghum stem parts. The pith is rich in parenchyma cells and the rind contains more vascular bundles. Parenchyma cells in pith have bigger lumens and thinner cell walls, and vascular bundles in rind are composed of tightly packed vessel elements [30]. On other hand, the easier dissolution of xylan from pith than those from rind under the same conditions may be partly attributed to the lower lignin content in the sorghum pith than in the rind [27]. Interestingly, there was a low linear correlation between the release of pCA and the recovery of xylan in residues for both pith and rind (Additional file 1: Fig. S1). Indeed, there is little chemically structural connection between pCA and xylan as most of pCA is attached to lignin with ester bond [31].
The regression coefficients of the model were determined by analysis of variance (ANOVA) in Additional file 1: Table S1. All models were significant at the P < 0.0001 level, which shows that all models were valid and do not lack fit, so this indicates that the model can be used to predict response. Several optimized solutions for model prediction were selected based upon the constraints: maximum xylan recovery in residues from different sorghum stem parts with the prerequisite of approximately 95% pCA release yield (not maximized). As shown in Table 3, the actual value was close to predicted value, which verified the reliability of the model. Compared with pith, the NaOH concentration in rind was higher, and the treatment time was shorter. The possible reason is that the structure of the rind is denser, so higher NaOH concentration was needed to release pCA, and in turn shortening the processing time.
According to the above optimization using response surface methodology, the following conditions were selected for NaOH-ethanol pretreatment in order to obtain the maximum release of pCA and the maximum recovery of xylan in residues for the pith, rind and whole stem in further experiments: 1.63% NaOH, 70% ethanol, 66 °C, 3.18 h; 1.90% NaOH, 70% ethanol, 69.8 °C, 1.00 h and 1.46% NaOH, 70% ethanol, 70 °C, 2.19 h, respectively. NaOH pretreatment was also carried out in same conditions (except ethanol) to further understand the significance of ethanol in NaOH-ethanol pretreatment. The solid recovery, pCA release and content of each component in solid residues are shown in Table 4. In general, the solid recovery, pCA release and the glucan and xylan content of NaOH-ethanol pretreated residues were relatively higher than the corresponding NaOH pretreated residues, but the lignin content was lower than the NaOH pretreated residues. These results suggested that ethanol addition promoted the release of pCA and delignification to a certain extent, but obviously prevented the deconstruction of xylan and cellulose. Compared to NaOH pretreatment, the release yields of pCA in pith, rind and whole stem by NaOH-ethanol pretreatment were increased by 8.16%, 8.38% and 8.39%, respectively (Table 4). Meanwhile, the delignification rates were increased by 5.57%, 11.47% and 11.04%, respectively (Fig. 1). The glucan recoveries of pith, rind and whole stem after NaOH-ethanol pretreatment reached 89.78%, 94.85% and 93.19%, which increased by 10.76%, 7.97% and 7.29% compared to NaOH pretreatment. The xylan recoveries of pith, rind and whole stem after NaOH-ethanol pretreatment reached 76.80%, 88.46% and 85.01%, which increased by 47.75%, 15.11% and 35.97% compared to NaOH pretreatment (Fig. 1). Despite more delignification by NaOH-ethanol pretreatment, the solid recovery of the NaOH-ethanol pretreated pith, rind and whole stem were 6.71%, 3.31% and 6.50% higher than those of NaOH pretreatment, respectively, which was mainly due to the improved recovery of glucan and xylan (Table 4). Particularly, the dissolution of xylan in pith was more serious in NaOH pretreatment condition compared to NaOH-ethanol pretreatment, indicating that addition of ethanol in NaOH solution could effectively minimize the deconstruction of xylan in pith (Fig. 1). Overall, the addition of ethanol to the NaOH pretreatment released almost all of esterified pCA while retaining most of cellulose and hemicellulose in the solid residue.
The inhibitors including acetic acid, levulinic acid, furan (furfural and 5-hydroxymethyl furfural (HMF)) and ferulic acid in pretreated liquid fractions were analyzed. The data showed that acetic acid and ferulic acid were the main inhibitors produced during pretreatments (Table 5). In comparison, the ferulic acid released in the NaOH-ethanol pretreated liquid fraction was slightly more than that in the NaOH pretreated liquid fraction, but the content of acetic acid was less than that in the NaOH pretreated liquid fraction. No levulinic acid, furfural and HMF were detected in both pretreated liquid fractions.
Enzymatic hydrolysis of solid residues by Cellic® CTec2
The enzymatic hydrolysis yields of raw and pretreated materials with Cellulase Cellic® CTec2 and β-glucosidase were shown in Fig. 2. The glucose enzymatic hydrolysis yield of NaOH-ethanol pretreated pith, rind and whole stem were 84.29%, 71.22% and 80.48%, respectively, which were 88.95%, 194.54% and 160.20% higher than that of untreated pith, rind and whole stem, respectively (Fig. 2A). Similarly, the xylose enzymatic hydrolysis yields of NaOH-ethanol pretreated pith, rind and whole stem were 78.33%, 62.70% and 75.53%, respectively, which were 59.89%, 87.78% and 117.23% higher than untreated pith, rind and whole stem, respectively (Fig. 2B). The enzymatic hydrolysis efficiency of the NaOH-ethanol pretreated residues was also obviously higher than that of NaOH pretreated residues. The glucose enzymatic hydrolysis yields of NaOH-ethanol pretreated pith, rind and whole stem were 7.39%, 15.45% and 12.18% higher than that of the NaOH pretreated pith, rind and whole stem, respectively (Fig. 2A). The xylose enzymatic hydrolysis yields of NaOH-ethanol pretreated pith, rind and whole stem were 6.43%, 10.86% and 17.61% higher than NaOH pretreated pith, rind and whole stem, respectively (Fig. 2B). This may be attributed to more effective in delignification by NaOH-ethanol pretreatment compared to that of NaOH pretreatment, because lignin is one of the major factors inhibiting enzymatic saccharification [32]. This result was consistent with the previous finding reported by Huang et al. [28]. They described that the introduction of ethanol into alkaline peroxide pretreatment enhanced the delignification of bamboo and thus improved its enzymatic hydrolysis efficiency. In addition, the improvement of the enzymatic saccharification efficiency of the NaOH-ethanol pretreated residues might be also partly brought by its higher release of pCA compared to NaOH pretreatment, since it has been reported that the presence of phenolic acids in residues has a negative effect on the enzymatic hydrolysis of lignocellulose [33]. In summary, alkali organosolv pretreatment not only reduces the over-degradation of cellulose and hemicellulose, but also enhances the delignification of biomass, resulting in enhanced enzymatic digestion of biomass [34, 35].
Effect of xylanase on enzymatic hydrolysis of solid residues by Cellic® CTec2
Since relatively high xylan was retained in NaOH-ethanol pretreated residues, the saccharification of cellulose may be still impeded by the highly reserved xylan in the lignocellulosic matrix [36]. Therefore, the effect of xylanase (Sigma X2629-100g) on enzymatic hydrolysis of solid residues was studied by adding xylanase to enzymatic hydrolysis system. Three different dosages of xylanase (5 BXU, 10 BXU and 15 BXU/g substrate) were selected for the enzymatic hydrolysis of NaOH-ethanol pretreated solid residues, and the results were shown in Fig. 3. After the addition of xylanase, the enzymatic hydrolysis yields of glucose and xylose were gradually improved with the increase of xylanase dosage (Fig. 3). This was attributable to the increased accessibility of cellulose to cellulase when xylan was degraded by xylanase. As shown in Fig. 4, when the dosage of xylanase was 15 BXU/g substrate, the glucose enzymatic hydrolysis yields of NaOH-ethanol pretreated pith, rind and whole stem reached 97.93%, 88.71% and 91.98%, respectively, which increased by 16.18%, 24.56% and 14.29% compared to those without xylanase (Fig. 4A), and the xylose enzymatic hydrolysis yields also reached 99.34%, 88.11% and 94.79% (Fig. 4B), which increased by 26.82%, 40.53% and 25.50% compared to those without xylanase. Similarly, the glucose enzymatic hydrolysis and xylose enzymatic hydrolysis yields were also increased when xylanase was added during the enzymatic hydrolysis of NaOH pretreated sample. The glucose enzymatic hydrolysis yields of the NaOH pretreated pith, rind and whole stem reached 90.36%, 78.86% and 84.00%, respectively, which increased by 15.12%, 27.83% and 17.09% compared to those without xylanase, and the xylose enzymatic hydrolysis yields were enhanced to 92.43%, 78.22% and 86.41%, respectively, which increased by 25.58%, 38.30% and 34.55% compared to those without xylanase (Fig. 4A and Fig. 4B). These results confirmed the promoting effect of xylanase on enzymatic hydrolysis of solid residues, which was consistent with the previous report of negative effects of retaining xylan on enzymatic hydrolysis of cellulose [37]. In comparison, the percentage increases in enzymatic saccharification by adding xylanase is more significant in rind than pith, probably due to higher recovered xylan in rind.
The totalGlu. yields and totalXyl. yields of various pretreated materials after adding xylanase (15 BXU/g substrate) were calculated. In above condition, the totalGlu. yields of NaOH-ethanol pretreated pith, rind and whole stem were 87.92%, 84.14% and 85.72%, respectively (Fig. 4C), which were 20.03%, 21.45% and 17.49%, respectively, higher than those of NaOH pretreated samples. The totalXyl. yields were 76.29%, 77.94% and 80.58% (Fig. 4D), which were 58.77%, 29.66% and 49.17%, respectively, higher than those of NaOH pretreated samples. Taken together, the total reducing sugar yields (glucose and xylose) after enzymatic hydrolysis reached 84.06%, 82.29% and 84.09% for NaOH-ethanol pretreated pith, rind and whole stem, respectively, which were 29.56%, 23.67% and 25.56% higher than the corresponding NaOH pretreated samples. The substantial increase of total reducing sugar yields was attributed to the improvement in both solid recovery and enzymatic hydrolysis yields when using NaOH-ethanol pretreatment. It was worth noting that the totalXyl. yield of the NaOH pretreated pith was lower than that of the untreated pith, because nearly half of the xylan was degraded during the NaOH pretreatment.
Structural characterization of untreated and pretreated sorghum sample
FTIR spectrum analysis
The aromatic structure strength and functional group identification of the untreated and treated samples were investigated using FTIR techniques. FTIR spectra of the untreated and treated samples are shown in Additional file 1: Fig. S2. The wavelength assignments of the lignin, cellulose and hemicellulose related bands are summarized in Additional file 1: Table S2. Absorption spectra at 1732 cm-1 is due to the ester-linked acetyl, feruloyl and p-coumaroyl groups on hemicellulose and/or lignin [38]. The band of the pretreated material at 1732 cm-1 almost completely disappeared, indicating that almost all of the ester-linked phenolic acid was released into the liquid. The bands at 1106 cm-1, 1254 cm-1 and 1513 cm-1 are characteristic bands of lignin [39]. These lignin-related bands were significantly reduced after treatment, indicating that most of the lignin was removed during pretreatment. After pretreatment, the characteristic band of the β-anomer or the β-linked glucose polymer increased significantly at 899 cm-1, indicating a significant increase in the cellulose content in the treated residue [40]. In summary, FTIR spectroscopy results show that the pretreatment can break the ester bond of pCA to lignin and remove most of the lignin.
SEM analysis
In order to observe the changes in the substrate surface, SEM was applied to investigate the morphological features and surface characteristics of the raw and the pretreated pith, rind and whole stem. As can be seen from Additional file 1: Fig. S3, the surface of the raw material was relatively smooth. The surfaces of the pretreated residue were altered with evident coarse surface and porous areas. Available surface area of the cellulose fiber structure is essential for enzymatic hydrolysis of lignocellulosic materials [41]. Both NaOH-ethanol and NaOH pretreatment destroyed the recalcitrant structure of the lignocelluloses, increased the surface area and porosity of biomass, which accelerated the saccharification process.
XRD analysis
The crystallinity index (CrI) was calculated according to X-ray diffractograms. XRD analysis of untreated and pretreated sorghum samples are shown in Additional file 1: Fig. S4. In raw materials, the CrI of the pith, rind and whole stem were 30.29%, 34.61% and 30.84%, respectively. After NaOH-ethanol pretreatment, the CrI of pith, rind and whole stem were increased to 63.01%, 43.27% and 54.41%, respectively. Similarly, the CrI of NaOH pretreated pith, rind and whole stem were also increased to 58.14%, 42.17% and 45.86%, respectively. Lignocellulosic crystallinity was considered as an important characteristic for enzymatic digestibility [42]. According to previous reports, the CrI of lignocelluloses was inversely related to the amorphous substances in cell wall where degradation of hemicelluloses and disordered fractions of cellulose, and delignification can all make the CrI increase [36]. Therefore, higher CrI of NaOH-ethanol pretreated residue may be due to the higher level of lignin removal and the recovery of cellulose compared to NaOH pretreated residue.
Overall mass balance
The process of pretreatment and enzymatic saccharification of sorghum samples (20 g) is shown in Fig. 5. NaOH-ethanol pretreatment of the pith, rind and whole stem was performed under the respective optimal conditions. Enzymatic saccharification of pretreated pith, rind and whole stem yielded 6.98 g, 7.54 g and 7.13 g of glucose, respectively, which accounted for 84.37%, 83.46% and 84.02% of the glucose in sorghum pith, rind and whole stem, respectively. Meanwhile, enzymatic saccharification produced 3.23 g, 3.04 g and 3.18 g of xylose, respectively, which accounted for 76.72%, 77.21% and 79.16% of the xylose in sorghum pith, rind and whole stem, respectively. After acidification, the NaOH-ethanol pretreated liquid fractions of pith, rind and whole stem contained 0.41 g, 0.40 g and 0.40 g of pCA, respectively. Macroporous adsorption resin D101 was used to recover pCA, and the adsorption rate and desorption rate reached 96% and 91%, respectively. After desorption from resin column, the recovered pCA were 0.36 g, 0.35 g and 0.36 g, which accounted for 81.45%, 84.13% and 84.51% of the esterified pCA in sorghum pith, rind and whole stem, respectively. Since the macroporous adsorption resin also has an adsorption capacity on other hydroxycinnamic acid derivatives such as ferulic acid in liquid fraction, further purification procedure will be need to obtain high purity of pCA. In this study, optimization using the response surface methodology revealed that the pith, rind and whole stem require different NaOH-ethanol pretreatment conditions for maximal pCA release and xylan preservation due to the difference in cell type and chemical composition of the pith and rind. From the perspective of pCA release yield and total sugar yield, however, there were no huge differences between the pith and rind if they were pretreated by respective optimal NaOH-ethanol conditions. Considering the separation cost of the different stem parts, whole sorghum stem can be directly used as feedstock for biorefinery.
Moreover, the hemicellulose obtained from the liquid fractions of the pretreated pith, rind, and whole stem were 0.74 g, 0.38 g and 0.51 g, respectively, which contained 42.64%, 44.29% and 40.97% of xylan, respectively. The hemicellulose was collected and subjected to enzymatic hydrolysis for production of xylooligosaccharides (XOS). EpXYN1 displayed the best enzymatic hydrolysis efficiency, and the total XOS (xylobiose to xylohexaose) yields on basis of xylan from the pretreated pith, rind and whole stem were 31.19%, 32.71% and 30.28%, respectively (Additional file 1: Table S3). The results suggested that the dissolved xylan during the NaOH-ethanol pretreatment process can also be utilized for XOS production.