Conversion of rice husk biomass into reducing sugars: in uence of pretreatment with water and [C16MIM][Br-] ionic liquid


 Rice husk is an agricultural lignocellulosic-rich waste which can be used in the production of second-generation ethanol. A biomass pretreatment is necessary to release structural sugars because of the recalcitrance-related properties of lignin and silicon. The use of a long chain imidazolium ionic liquid (IL), 1-hexadecyl-3-methylimidazolium bromide ([C16MIM][Br−]), for rice husk pretreatment is reported for the first time. The influence of the [C16MIM][Br−]:water ratio and temperature on reducing sugars yield was investigated using an experimental design. Structural, chemical, and morphological modifications were observed by Scanning Electron Microscopy (SEM), X-Ray Diffraction (XRD), Fourier Transform Infrared Spectroscopy (FTIR) and Thermogravimetric Analysis (TGA). Higher temperature (160°C) and lower water content (1:5, IL:water) improved sugar yield. The validation assay (180°C, 1:20 IL:water ratio, 10% (w/v) of biomass, 12 h) resulted in 38% yield of reducing sugars and 60% dissolution of biomass, whereas untreated biomass showed only 0.18% yield of reducing sugars. The characterization analysis confirmed the changes in rice husk structure and morphology after pretreatment with IL. Rice husk pretreatment with [C16MIM][Br−] resulted in a higher concentration of reducing sugars compared to untreated biomass. Therefore, solid ionic liquids combined with water could be used as an alternative solvent for lignocellulosic biomass pretreatment.


Introduction
Biofuels are effective alternatives to substitute for fossil fuels, mainly in the transport sector which has been growing signi cantly. Nowadays, the most produced biofuels are the rst-generation ones (Hira and de Oliveira, 2009;Sims et al., 2010). However, the use food crops as raw materials is considered to be a great issue because the price of these feedstocks increases (Chen et al., 2017;Hira and de Oliveira, 2009;Sims et al., 2010) and as a result, human and animal necessities could be affected (Chen et al., 2017;Mungodla and Linganiso, 2019). Second-generation fuels from lignocellulosic feedstocks and agricultural wastes (Joshi et al., 2017) such as sugarcane bagasse (Nasirpour and Mousavi, 2018), wheat straw (Li et al., 2009), rice straw (Nguyen et al., 2010), rice husk (Ang et al., 2012), forest residues (Aid et al., 2016), and municipal solid waste (Chen et al., 2017) have attracted attention because these materials are renewable and non-food sources. They are considered to be environmentally friendly and present versatility, availability, and low cost (Sorn et al., 2019).
Currently, ionic liquids (ILs) have attracted attention as potential solvents for biomass pretreatment (Lynam et al., 2012;Yang and Fang, 2015;Yoon et al., 2011). Ionic liquids are organic salts with relatively low melting points (below 100 °C) (Baek et al., 2017;Zhao et al., 2009;Zhu et al., 2006) They normally consist of an organic cation and an organic or inorganic anion (Chen et al., 2017;Liu et al., 2012;Vancov et al., 2012). ILs present relevant properties such as non-ammability, thermal and chemical stability, negligible volatility, and high solvent power. Also, they could be considered to be green and environmentally friendly solvents (Baek et al., 2017;Gogoi and Hazarika, 2017;Lopes et al., 2013;Zhu et al., 2006). The 1-hexadecyl-3-methylimidazolium bromide ([C 16 MIM][Br -]) ionic liquid presents a long carbon chain. It has high viscosity, and it is solid at room temperature. It has a relatively high melting point (66.91 °C) (Zhu et al., 2009), and it can be synthetized with high yield (over 90%). [C 16 MIM][Br -] solubilized in water, over the critic micellar concentration, promotes the formation of an emulsion which increases both the interfacial area and mass-transfer e ciency. It is well known that an ionic liquid with a long alkyl chain reduces the surface tension more than one with a shorter alkyl chain, considering that both are at the same concentration in aqueous solution and have the same anion (Toh et al., 2006). Thus, [C 16 MIM][Br -] could provide lower surface tension compared to other bromides with a C 2 or C 4 alkyl chain, which are commonly applied in the lignocellulosic materials.  (Yokoo and Miyafuji, 2014). Rice husk pretreatment has been evaluated with different ionic liquids. Poornejad et al. (2014) and Lynam et al. (2012)

Substrate
Rice husk was obtained from the rice industry of Santa Maria, Rio Grande do Sul, Brazil. The biomass was ground and sieved into particles smaller than 150 µm and stored at room temperature. The moisture of the rice husk was 7.9% (w/w), and the samples were preserved in a desiccator.

Ionic Liquid
[C 16 MIM][Br -] ionic liquid was prepared by quaternization between hexadecyl bromide (1bromohexadecane, Aldrich) and the corresponding imidazolium (methylimidazolium, Aldrich). The ionic liquid was obtained by mixing 0.139 mol of methylimidazolium with 0.139 mol of the alkyl bromide and adding a volume of ethyl acetate corresponding to 10% (w/v) of the total volume of reaction mixture. The reactions were conducted in a Te on autoclave with a stainless steel jacket under static isothermal conditions, and a owrate of nitrogen gas (20 mL/min) was injected for 3 min before closing the autoclave. The maximum quaternization yield (91%) occurred after 48 h at 50 °C. After cooling the autoclave, the reaction product was placed in a rotary evaporator for 1 h to remove the ethyl acetate. After the solvent elimination, the solid product was washed with ethyl acetate until a white crystal was obtained. The formation of the desired product was con rmed by elemental analysis (CHN) and 13 C MAS-NMR in a spectrometer, Bruker DPX-400 (400.13 MHz), using dimethyl sulfoxide (DMSO) as solvent. The results are presented in the supplementary information in Figure S1 and Table S1.

Rice Husk Pretreatment With [C 16 MIM][Br -] Ionic Liquid
Rice husk pretreatment was carried out in a Te on autoclave with a stainless steel jacket heated in an oven. A sample of 0.1 g of rice husk was mixed with 1 g of ionic liquid dissolved in different volumes of deionized water. The reaction mixtures were placed in an oven for 12 h, for solvothermal treatment at different temperatures, according to an experimental design (Section 2.6). Then, 10 mL of deionized water was added to dilute the cellulose dissolved in the reaction, and the mixture was centrifuged at 4000 rpm in an Eppendorf centrifuge (model 5804R) during 15 min. The ionic liquid aqueous solution was separated from the undissolved biomass. Washing and centrifugation steps were carried out 4 times. Finally, the pretreated rice husk (biomass) was dried in an oven at 60 °C for 24 h (Ang et al., 2012) and characterized as described in section 2.5. Biomass dissolution was determined by weight difference before and after the pretreatment.

Enzymatic Hydrolysis
Enzymatic hydrolysis was based on the methodology described by Nguyen et al. (2010) with some modi cations. The cellulase enzyme (Celluclast 1.5L ®) used was purchased from Novozymes (Denmark), and its activity was determined according to Ghose (1987). A sample of untreated/pretreated biomass (0.1 g) was mixed with 0.215 mL of Celluclast (initial activity of 8 UI -activity of endo and exocellulase) and 1 mL of deionized water. The mixture was incubated in orbital shaker at 120 rpm, 50 °C, for 72 h. Reducing sugars were quanti ed using the dinitrosalisylic acid (DNS) method (Miller, 1959), and the yield (%) was determined according to Equation (1):

Rice Husk Characterization
Rice husk was characterized before and after the biomass pretreatment (validation condition, section 2.6). The composition of cellulose, hemicellulose, lignin, and ash was determined according to methodology of National Renewable Energy Laboratory (NREL) ( Table 1). equipped with a Gemini column. An X-Ray Energy Spectrometer (EDS) attached to an SEM was used for the determination of oxygen (O), carbon (C) and silicon (Si) content (Table 1).
Biomass crystallinity and structural changes were examined by X-Ray Diffraction (XRD) (Rigaku Mini ex ® 600 Cu Kα radiation, λ = 1.5406 Å, 40 mA, 40 kV). The samples were scanned over a 2θ range between 10 and 80° at 0.02° per step with scan speed of 0.5 s. The biomass crystallinity index (CrI) was determined according to Equation (2) (Ang et al., 2012): where I 002 is the maximum intensity for the crystalline portion of the rice husk sample at about 2θ = 22.6°, and I am is the intensity attributed to the amorphous portion of rice husk sample at 2θ = ~18.7°.
Fourier Transform Infrared Spectroscopy (FTIR) was employed to analyze the chemical changes in biomass using a spectrometer (Shimadzu, IR Prestige 21 with the method of direct transmittance KBr pellets). The spectra were obtained with an average of 45 scans in the range between 4000 and 400 cm -1 with 2 cm -1 resolution.
Biomass decomposition was analyzed before and after the pretreatment (Assays 1 and 5 of the CCRD, section 2.6) by Thermogravimetric Analysis (TGA), Derivative Thermogravimetry (DTG), and Differential Thermal Analysis (DTA) using a Shimadzu Model DTG-60H instrument. Samples were heated from ambient temperature up to 1000 °C, at a rate of 10 °C/min, in an oxidizing atmosphere (synthetic air supplied at 40 mL/min). Alumina was used as the reference material in thermal analysis.

Experimental Design
A Central Composite Rotatable Design (CCRD) (Rodrigues and Iemma, 2014) was performed to evaluate the in uence of ionic liquid:water ratio (1:5 -1:30 (w/v)) and temperature (80 -160 °C) on the rice husk pretreatment (Table 2). Previous assays (data not shown) were performed to de ne the pretreatment time and temperature range. Since longer reaction periods showed no in uence on the biomass pretreatment, the sample mixture was pretreated for 12 h. The ratio of ionic liquid dissolved in deionized water was based on the work of Fu and Mazza (2011). The biomass loading in the pretreatment was xed at 10% (w/v). According to the CCRD results, a validation assay was performed. Three assays were carried out at a higher temperature (180 °C) and at an IL:water ratio of 1:20 (central point). Experimental data were analyzed using Statistica ® 7.0 (Stat Soft Ink., Tulsa, OK, USA) considering a signi cance level of 90% (p < 0.10).

Results And Discussion
3.    148.37 (+1) 120 (0) 120 (0) 80 The Pareto chart (Figure 1) shows the in uence of variables on reducing sugars yield according to the results in Table 2. The linear (L) and quadratic (Q) effects of temperature on reducing sugars yield were positive (p < 0.10), and the linear effect was greater. This positive linear effect of temperature indicates that increasing the temperature increases the yield. In addition, the linear effect of the IL:water ratio was negative (p < 0.10), suggesting that a lower ratio should be used for a higher yield of reducing sugars.
The second-order equation (Equation 3) describes the in uence of the IL:water ratio and temperature on reducing sugars yield after the pretreatment and enzymatic hydrolysis, and it was estimated from the coded matrix (Table 2).
RS Yield is the reducing sugars yield after enzymatic hydrolysis (%); C and T are the coded values for the IL:water ratio (w/v) and temperature (°C), respectively.
The prediction and signi cance of the empirical model were evaluated by analysis of variance (ANOVA) at a 90% con dence level (Table 3). The model was validated since the determination coe cient (R 2 ) was 97% and the calculated F value (F calc = 30.98) was higher than the tabulated F value (F tab = 3.45), indicating a satisfactory correlation between the independent variables and the response and showing that the model was predictive. Reproducibility of the assays could be demonstrated by central points (Assays 9-11). The response surface ( Figure 2) generated with the validation model (Equation 3) shows the reducing sugars yield as a function of temperature and the IL:water ratio. As observed in Figure 2, higher temperature improved the reducing sugars yield, while the IL:water ratio had no signi cant in uence within the evaluated range. A validation assay was performed, and 38.04 ± 2.1% yield of reducing sugars was obtained. This result was higher than Assay 8, the best condition of the CCRD. Moreover, this experimental value is in agreement with the value predicted by the coded model at the validation condition (39.93%). A control assay to evaluate the contribution of IL in the rice husk pretreatment was also carried out at 180 °C, 10% (w/v) biomass loading, using only water (20 mL, central point) for 12 h. In this condition, 23.76% yield of reducing sugars was observed, and this value was lower than that obtained in the presence of IL, which indicates that the IL acted during the pretreatment step by modifying the rice husk structure and improving the reducing sugars yield.
Temperature had a greater in uence on reducing sugars yield. Sriariyanun et al. (2015) also observed the positive in uence of high temperature pretreatment of rice straw on reducing sugars yield.  (Toh et al., 2006). Therefore, [C 16 MIM][Br -] interacts with the biomass in a different way compared with other ionic liquids in aqueous media, which present distinct alkyl chain compositions and properties. This is attributed to the in uence of water on the system (Ferrari et al., 2019).
Higher biomass dissolution (≈ 45%; Table 2) could be related to the high temperature pretreatment (148.37 and 160 °C; Assays 3, 4 and 8), whereas low temperature pretreatment (80 and 91.63°C) provided approximately 11% of biomass dissolution (Assays 1, 2 and 7). Validation assays performed at 180 °C, presented 60 ± 3.45% of biomass dissolution, con rming the effect of temperature. Hou et al. (2012) also observed higher dissolution at higher temperatures, which was associated with the removal of polysaccharides from rice straw.
Higher reducing sugar yields were also associated with greater biomass dissolution (Table 2). Ionic liquids could change the lignocellulosic matrix (section 3.2) at the same time that they hydrolyze some cellulose and hemicellulose to single sugars. Lignocellulose dissolution in ionic liquids increases with increasing temperature. This can be attributed to the hydrogen bond destabilization in the lignocellulosic matrix (Gogoi and Hazarika, 2019). Comparing the chemical composition of the validation assay with that of untreated biomass (Table 1), there is a clear reduction in the composition of cellulose and hemicellulose with the pretreatment, which suggests that the ionic liquid in uences the biomass dissolution. Thus, knowing that long chain ionic liquids have strong Van der Waals forces (Losada-Pérez et al., 2016) and high viscosity, it could be suggested that [C 16 MIM][Br -] ionic liquid helps the interaction with cellulose (Brandt et al., 2013) and increases biomass dissolution. Moreover, Table 2 shows that the untreated sample led to no biomass dissolution, and it had the lowest reducing sugars yield.

Biomass Characterization
The results indicated that the pretreatment in uenced the composition of biomass (Table 1). The cellulose, hemicellulose, and lignin content were reduced because the interaction between ionic liquid and the lignocellulosic biomass, as previously discussed. Moreover, the removal of silicon could be associated with the in uence of temperature and the interaction of ionic liquid and biomass (Lynam et al., 2012;Yang and Fang, 2015).

Scanning Electron Microscopy (SEM)
The morphological changes are shown in Figure 3. Untreated rice husk (Figure 3a) showed a packed and inaccessible morphology and non-porous material. The lignocellulosic matrix makes the access of enzymes di cult (Ebrahimi et al., 2017;Poornejad et al., 2013). Figure 3b shows the morphological modi cation after the pretreatment, where the packed structure was changed to a wave-shaped, more accessible, and disordered morphology which could improve enzymatic hydrolysis. These considerable changes in the morphology of pretreated material could be related to the reduction of cellulose, hemicellulose, and silicon content (Table 1).

X-Ray Diffraction (XRD)
The crystalline structure of rice husk was analyzed by X-Ray Diffraction (XRD). The enhancement of enzymatic hydrolysis is usually associated with lower crystallinity index values, which means a more accessible structure for enzymes (Gogoi and Hazarika, 2017;Yoon et al., 2011). Figure 4 shows an increase in the crystallinity degree after the ionic liquid pretreatment (CrI = 59.5%). This was con rmed by comparing the intense and sharp diffraction peak pro le located at 2θ = 22°. This highest crystallinity was a result of removing amorphous biomass such as hemicellulose.
Previous studies showed a similar crystallinity pattern that was also associated with the high temperature pretreatment and hydrolytic cleavage of glycosidic bonds (Ang et al., 2012;Ang et al., 2013;Gao et al., 2018;Gogoi and Hazarika, 2019;Johar et al., 2012;Kim et al., 2016;Yang and Fang, 2015). Also, a small shoulder peak located at ~21° and a shift from peak at 21.8° to 22.5° can be seen, which ensure lattice expansion (Cheng et al., 2012) related to the changes in the structure of the cellulose matrix after the pretreatment process. Moreover, the change in the crystallinity was con rmed by the presence of two additional peaks located at 26.5° and 34.5° which represent the quarter length of one cellobiose unit (Ouyang et al., 2015).

Fourier Transform Infrared Spectroscopy (FTIR)
The functional groups of untreated and pretreated rice husk were assessed by FTIR, and the spectra are presented in Figure 5. Both pro les were similar and associated with native rice husk. The FTIR spectrum showed bands located at 790, 898, 1085, 1423, 1511, 1637 and 2919 cm -1 (Ang et al., 2012;Yang and Fang, 2015). The C-O stretching bands at 1085 cm -1 correspond to cellulose/hemicellulose and lignin (Ang et al., 2012;Trinh et al., 2015). In addition, these bands are associated with the presence of the lignin-carbohydrate bonds in the biomass. After the pretreatment, the infrared-absorption intensity decreased as a result of the cellulose/hemicellulose dissolution caused during heating (Yang and Fang, 2015;Guo et al., 2008). Both untreated and pretreated biomass presented the C-O stretching bands located at 898 cm -1 and the C-H stretching located at 2919 cm -1 which appears in cellulose rich material (Yang and Fang, 2015). No signi cant changes were observed, suggesting cellulose retention, which is characteristic of rice husk. However, some bands were reduced after the ionic liquid pretreatment, such as the silicon bonds represented by bands located at 790 cm -1 , C-O stretching bands at 1085 cm -1 , C=O of acetoxy groups in hemicellulose located at 1734 cm -1 (Boukir et al., 2019; Equihua-Sánchez, Barahona-Pérez, 2019), and the lignin aromatic skeletal stretching located at 1423 and 1511 cm -1 (Boonsombuti et al., 2020;Yang and Fang, 2015). These results suggested the removal of silicon, hemicellulose, and lignin. This could be related to the structure destruction, which is in agreement with the XRD discussion.
In addition, a reduction of absorption band located at 1637 cm -1 was observed, which could be associated with the adsorbed water in the biomass (Ang et al., 2012).

Thermogravimetric Analysis (TGA)
TGA analysis was used to quantify hemicellulose, cellulose and lignin in the rice husk biomass. Lignocellulosic materials have a mixture of approximately 53 wt.% of hemicellulose and cellulose and 19 wt.% of lignin (Singh et al., 2013). These values are in agreement with the results found in the biomass characterization (Table 1). Thermogravimetric analysis in O 2 atmosphere was performed to evaluate the biomass decomposition. Figure 6 shows the TGA, Differential Thermal Analysis (DTA), and Derivative Thermogravimetry (DTG) pro les of untreated rice husk in oxidant atmosphere. Figure 6 shows four weight loss stages. The rst stage refers to the moisture content (9.20 wt.%) at the temperature range of 30-137°C, and an endothermic peak can be seen in the DTA curve in the same range with a maximum at 70 °C. The second stage (0.80 wt.%) is relatively constant at the temperature range of 137-181°C and refers to the chemisorbed water (Maschio et al., 1992). The third (≈ 53 wt.%) and the fourth (≈ 19 wt.%) stages, starting at 181 °C and ending at 578 °C, involve hemicellulose, cellulose, and lignin decomposition. The maximum weight loss was in the range of 360-515 °C. DTA curves showed two exothermic peaks with a maximum at 345 °C and 460 °C, referring to the cellulose/hemicellulose and lignin decomposition, respectively (Singh et al., 2013). The results are in agreement with previous studies (Carrier et al., 2011;Pijarn et al., 2010;Yang et al., 2006) and are summarized in Table 4. After the biomass pretreatment (Assays 1 and 5 of the CCRD, Figures 7a and 7b, respectively), four weight loss stages were observed. However, the rst and second stages were almost constant. These results were expected due to the drying process (moisture content and chemisorbed water ≈ 3%) at the temperature range of 30-150 °C. A low endothermic peak can be seen in DTA curve in this range with a maximum at 70 °C. The third stage (≈ 53 wt.%) involves hemicellulose and cellulose decomposition, and it started at a lower temperature (150 ºC) compared to the untreated rice husk ( Figure 6). The DTG pro le showed that both temperature and IL treatment changed and separated cellulose, hemicellulose, and lignin into fractions with different structural compositions (Figures 7a and 7b). Similar results were previously reported with different ionic liquids (Zhang et al., 2014). For both assays (1 and 5), one peak corresponding to the hemicellulose fraction was clearly observed at 250-300 °C, and another one, characteristic of cellulose, was observed in the range of 300-400 °C. This result is in agreement with the XRD analysis (section 3.2.3), considering that an intensi cation in the crystallinity index was observed (Moniruzzaman and Ono, 2013;Zhang et al., 2014). Moreover, a greater weight loss in hemicellulose was observed compared to the untreated biomass, suggesting an ease in the thermal breakdown of the ionic liquid pretreated biomass (Singh et al., 2015). DTA curves showed two exothermic peaks in the range of 240-390 °C, lower temperature compared to the untreated rice husk. No exothermic peak in the DTA pro le was detected at the temperature range of lignin decomposition.

Conclusions
This study evaluated the effect of [C 16 MIM][Br -] ionic liquid on rice husk pretreatment at different temperatures and IL:water ratios. Temperature showed a great in uence on the pretreatment within the studied range. This was con rmed by validation assays (180 °C, IL:water ratio 1:20) with 38.04 ± 2.1% yield of reducing sugars and 60 ± 3.45% biomass dissolution. Despite the biomass loss, the pretreated rice husk showed changes in its structure and morphology that improved enzymatic hydrolysis.