Studies to optimize the process of biofuel production from castor stalk

Impact on chemical composition and enzymatic digestibility

It is important, to the economics of an industrially viable process, to utilize the combination of a high-solids pretreatment at large particle size to improve the process economics by increasing sugar and ethanol yields while decreasing capital costs along with the recovery of ionic liquid [34], [35], [36].

However, utilizing high-solid loadings in this conversion process is still relatively new, and more detailed investigation is required to overcome certain challenges, such as high concentrations of inhibitors and equipment to handle mass transfer limitations that are not as apparent at the low- and moderate solids loadings [37]. Different biomass loadings were utilized to investigate the efficacy of the IL pretreatment. The temperature of the reaction mixture was kept relatively low at 100°C, taking into account the stability of the IL, and the pretreatment time was varied from 2 h to 10 h at the interval of 2 h. The biomass loading was increased from 5% to 15% (w/w) with the interval of 5.

A techno-economic analysis, done by Klein et al., has shown that increasing the biomass loading is more important than increasing the rate of IL recycling, as it brings several simultaneous advantages including lower capital cost, lower electricity use, and lower working capital required [38], [39], [40], [41]. Increasing the biomass loading to 40–50 wt% during IL pretreatment was shown to have a prominent effect on the overall economics of bio-refinery operation. However, Cruz et al. reported significant drop in the glucan digestibility of glucan present in the biomass at these biomass loadings [28], [42]. The aim of the current study was to assess the maximum biomass loading that can be achieved while retaining high pretreatment efficiency for a given amount of IL.

The increase in the glucan digestibility is an important marker of lignocellulose pretreatment quality. Glucose is regarded as a major product of bio-refineries for its use in fermentation process to produce biofuels and other chemicals [43], [44]. The pretreated CS was subjected to hydrolysis by commercial CTec2 enzyme at 40 FPU/g loading for 24 h at 50°C and it was observed that the pretreated CS was significantly more digestible than the untreated CS in all sets of experiments.

As expected with the increase in the biomass loading, the biomass digestibility was decreased as a general pattern (Fig. 1), similar to the relative glucan content present in the pretreated biomass (Table S1), Interestingly, the glucan recovery was more at 10% biomass loadings as compared with 5% biomass loadings in the pretreated biomass for the lesser pretreatment periods (up to 6 h). Lignin removal was most efficient in the lower biomass loadings while for high biomass loading of 15%, the lignin removal was significantly lower. In the pretreated biomass, the lignin removal was 18.4% and 14.5% respectively for 5% and 10% biomass loading after the pretreatment time of 10 h with a particle size of 0.3–0.6 mm. It is interesting to note that the cellulose conversion after 24 h of enzymatic hydrolysis for 10% biomass loading was similar to that of the 5% biomass loading with values of 96% and 97% respectively (Fig. 1). This aspect can possibly reduce the cost of the process by employing higher biomass loadings. It was observed that there was a significant drop in the biomass digestibility at 15% biomass loading, which could be explained by the less efficient removal of the lignin, as a result of the pretreatment along with the other physical parameters. This suggested that the optimum biomass loading was 10% where there was no significant loss of sugars observed along with the high yields of sugars.

Fig. 1:

Enzymatic digestibility of the biomass of particle size 0.3–0.6 mm (a), 0.6–1.0 mm (b) and 1.0–2.0mm (c), after pretreatment with [EMIM][Ac] for various time intervals with 5%, 10% and 15% biomass loadings and 100°C temperature.

Correlation with the changes in crystallinity and surface morphology of CS

In order to investigate the mechanism of IL action on the digestibility of the cellulose at various pretreatment parameters, it is important to study the effect of IL treatment on the crystallinity and surface morphology of CS, because these properties are directly related to the accessibility of the enzymes into the highly organized structure of the biomass.

In particular, there remains a critical need to efficiently and economically break down the complex and highly recalcitrant structure consisting of tightly held cellulose, hemicellulose and lignin components in order to enhance enzyme accessibility. IL treatment causes the swelling in the biomass and disrupts the overall structure.

The lower crystallinity index indicates a higher amount of amorphous cellulose present in the biomass [45]. As expected the crystallinity index was found to decrease after IL treatment compared to untreated CS. However, at the 15% biomass loading the reduction in CrI was not significant which explains the low digestibility of the pretreated biomass and a reason of decreased enzymatic digestibility of the pretreated biomass (Fig. 2).

Fig. 2:

Crystallinity index (CrI) of the biomass of particle size 0.3–0.6 mm (a), 0.6–1.0 mm (b) and 1.0–2.0 mm (c), after pretreatment with [EMIM][Ac] for 2, 4, 6, 8 and 10 h at 5%, 10% and 15% biomass loadings.

In scanning electron microscopy, the pretreated biomass showed a highly disrupted structure along with presence of non-uniform deconstructed form. At 15% biomass loading less deconstructed surfaces were observed, indicating less penetration of IL and presence of native structures (Fig. S1).

Impact on chemical composition and enzyme digestibility

Particle size is one of the crucial factors which directly impacts the contact and diffusion of ionic liquid into the lignocellulosic material and in turn, the solubilisation of lignocellulosic biomass. The effect of particle size on the pretreatment process was examined for three particle size ranges of castor stalk residue (0.3–0.6 mm, 0.6–1.0 mm and 1.0–2.0 mm) treated with ionic liquid at 100°C for 24 h. The three particle sizes show certain influence on sugar conversion and yield of regenerated biomass (Fig. 2a and b). The relative glucan content was more in the pretreated biomass as compared to the untreated biomass due to partial removal of lignin. It was observed that the sugar conversion decreases with increasing of biomass particle size as a general trend. CS of larger particle size was not effectively solubilized in the ionic liquid. As expected at the largest particle size, 1.0–2.0 mm, there was a significant drop in the digestibility as well as in the sugar recovery. However, at relatively smaller particle sizes of 0.3–0.6 mm and 0.6–1.0 mm the ionic liquid pretreatment efficiently resulted in the high yields of sugars. As expected the highest amount of glucan was observed in 0.3–0.6 mm particle size CS at the lowest of biomass loading of 5% (after 10 h of IL pretreatment). However, at larger particle size (1.0–2.0 mm) the reduction in the lignin content was significantly low, 1.4% at 10% biomass loading. Lignin removal was significantly affected by the particle size of the biomass. This may be caused by the fact that the pretreatment of larger size particles of biomass could have the limitation of heat and mass transfer due to higher temperature gradient and increase in the diffusion path of ionic liquid penetrating inside the particle. While for the CS of the smallest particle size, the lignin removal was maximum 18.4% at 5% biomass loading at 10 h of pretreatment time. At the particle size of 0.6–1.0 mm the lignin removal was not as significant as that at smallest particle size, however the enzymatic digestibility was similar for both particle sizes, suggesting possible impact of reduction in CrI and disruption of biomass in enhancing the accessibility of the enzyme.

Previous studies showed that the larger size particle possesses lower bulk density therefore it requires more amount of ionic liquid contacting larger volume of biomass to achieve the same effectiveness with that of smaller size particle [32]. Moreover, the higher crystallinity index of untreated biomasses with larger size particles (Fig. 1) could directly decrease the sugar conversion.

From the results, it is apparent that there is a minimum particle size up to which the effect of particle size acts as limiting factor. In castor stalk it was 0.6–1.0 mm particle size above which the cellulose digestibility is greatly affected. This is an important result in terms of the economics of the process of biofuel production, as it can save the energy requirement for grinding. Biomass of particle size 0.6–1.0 mm was as efficiently digestible as biomass of particle size 0.3–0.6 mm upto a biomass loading of 10% (w/w). The chemical composition also influences the physical and mechanical properties, as the lignin content increases the hardness of the cell wall increases [46]. The cellulose content of the cortex region is highest 35.76% and hemicellulose and lignin of the cortex are lowest, 7.23% and 6.09%, respectively. While lignin and hemicellulose of xylem are highest (13.43% and 7.23%) and the cellulose content is in between cortex and pith, 33.76%. The chemical composition of the pith is; cellulose (32.9%), hemicellulose (9.61%) and lignin (8.77). The composition not only varies at different positions on the stem, such as lower stem region, upper stem region or branch region but also with harvest seasons [18], [47]. The composition reported by Li et al. [18] was cellulose (33.19%), hemicellulose (11.12%) and lignin (15.48%).

Correlation with the changes in crystallinity and surface morphology of CS

Our studies indicate that smaller particle size and lower biomass loading result in lowering the crystallinity of the biomass, coupled with higher sugar conversion. This shows that the decrease of crystallinity disrupted the crystalline structure of cellulose leading to more accessibility of enzymes into the biomass leading to a higher content of sugar product. As shown in Fig. 2, at 5% biomass loading the particle size of 0.3–0.6 mm and 0.6–1.0 mm showed CrI=0.26 and 0.29, respectively while the particle size range of 0.1–0.2 mm showed CrI=0.38 after 10 h of pretreatment time, suggesting ineffective penetration of IL into the crystalline form.

The pretreated biomass with particle size 0.6–1.0 mm showed a similar reduction in the CrI value to biomass of smaller particle size of 0.3–0.6 mm. This is consistent with the higher yields of sugars obtained for the particle size range of 0.6–1.0 mm with 10% loadings. The pretreated biomass of 1.0–2.0 mm particle size range showed quite high values of crystallinity indicating poor accessibility of enzymes into the highly organized complex of lignocellulose resulting in lower yields of sugars.

Surface morphological studies with particle size of 0.3–0.6 mm for all biomass loadings and for 0.6–1.0 mm up to 10% biomass loadings indicate mostly disrupted structures, thus facilitating the higher digestibility of the biomass, whereas for 1.0–2.0 mm particle size displayed a less disrupted surface morphology (Fig. S1).

Impact on chemical composition and enzyme digestibility

Previous studies showed that prolonged pretreatment at high temperatures possesses the risk of degrading the dissolved cellulose into undesirable by-products [21], therefore it is necessary to optimize the duration of IL pretreatment. To investigate the effect of reaction time on CS pretreatment, castor stalk was stirred in IL at 100°C for five different residence times viz. 2, 4, 6, 8 and 10 h. As expected, an increase in the digestibility was observed with increasing the pretreatment time (Fig. 1).

At particle size of 0.3–0.6 mm, the lignin removal was significantly higher, even at lesser duration of pretreatment at 5% and 10% biomass loadings, while at 15% biomass loading the lignin removal was more at longer pretreatment duration (Table S1–S3). Long pretreatment time intervals were reported to favor lignin extraction for CS of particle size 0.6–1.0 mm, while for largest particle size the lignin removal was very less. As the pretreatment time was increased the diffusion of ionic liquid into the biomass was improved and in turn increased the dissolution and extraction of lignin from the biomass. However, the increase in the cellulose conversion was not significant in the biomass of larger particle size (1.0–2.0 mm). At smaller particle size there was a successive increase in the digestibility. The biomass of particle size of 0.6–1.0 mm showed similar cellulose conversion yields as in the biomass of particle size 0.3–0.6 mm at the pretreatment time periods of 4 h and above. After 6 h of pretreatment time and 10% of biomass loading the biomass with particle sizes of 0.3–0.6 mm and 0.6–1.0 mm respectively showed 95% and 94% of cellulose conversion along with the glucan loss of 8.2% and 8.4%. This resulted the possibility of scaling up of the IL based pretreatment method due to high yields of sugars at 10% biomass loading and lesser glucan loss (comparable to 5% biomass loadings) at the end of 6 h heating time.

Correlation with the changes in crystallinity and surface morphology of CS

The crystallinity index decreases with the increase in pretreatment time. In the smallest particle size of biomass 0.3–0.6 mm, the effect of pretreatment time was the most significant, while for the biomass of larger particle size, it was least effective. In the biomass of particle size 0.6–1.0 mm, the CrI decreased significantly up to 10% of biomass loading supporting the high yields of glucan conversion (Fig. 1). In the surface morphology of the pretreated biomass, the more disrupted structures were observed for biomass of particle size 0.3–0.6 mm and 0.6–1.0 mm with increase in the pretreatment time.