The influence of conditions for treatments using ethanol/water as solvents on the removal of each component was analyzed. Figure 1 shows the effects of time and temperature on the removal of hemicellulose, lignin, and ash during the treatment with 25% ethanol/water. Because cellulose was hardly extracted under the conditions applied in this experiment as shown in the next section, its removal is not presented here. For the treatment at 160°C and short time (≤2 h), the removal of all components was well below 30 wt%, and there was no significant difference in the removal rates among them (Figure 1A). This suggested that each component contains fractions that were readily dissolved in the solvent. For ash, the removed fraction should include alkali and alkaline earth metallic species (AAEMs) in the form of salts.

Figure 1B compares treatments at different temperatures at the fixed time of 4 h. The removal of hemicellulose became marked with increase in the temperature, reaching 97.3 wt% at 200°C. This aligns with the general trend that hemicellulose is the most easily extracted component under hydrothermal conditions (Yu et al., 2007). In contrast, the removal rates of lignin and ash decreased with temperature. The decrease in lignin removal is attributed to the re-polymerization, condensation and precipitation of dissolved organosolv lignin and possibly hemicellulose (Wang et al., 2018). A similar phenomenon is thought to occur with ash, although the degree of decrease in removal is smaller (from 14.4 to 7.9 wt%), compared to lignin (from 30.7 to 17.6 wt%). Di Francesco et al. (2021) reported that treatment of woody biomass (Populus) with a 20 vol% ethanol aqueous solution at 220°C for 4 h reduced the residual lignin to 3 wt%. Thus, higher temperatures may further improve lignin removal; however, higher temperatures may also lead to partial removal of cellulose. Because the removal of cellulose indeed increased to 3.7 wt% at 200°C, this study did not examine the treatment at higher temperatures.

Figure 2 shows the influence of reaction temperature and ethanol concentration on the removal of each component. The cellulose removal in this test was equal to or below 3.7 wt%, confirming that cellulose extraction was suppressed under employed conditions. The highest removal of hemicellulose was obtained in the absence of ethanol at 200°C, reaching 97.8 wt%. At all temperatures, increasing the ethanol concentration reduced the hemicellulose removal. The removal was below 20 wt% in ethanol-only conditions. Therefore, water-only system was concluded to be appropriate for hemicellulose extraction.

Ash removal also tended to decrease with increasing ethanol concentration, although the removal level was low, 13.7 wt% even at the highest case, indicating that the amount of ash soluble in ethanol/water system is limited at temperatures up to 200°C. Since silica is insoluble in water, most of the removed ash is inferred to consist of other metal species. However, silicon exists in rice husk in various forms, such as silicic acid that is soluble to water (Vassilev et al., 2013). Vassilev et al. (2012) analyzed the water solubility of ash metal species in various biomass types, finding that 2% of silicon was leached to water in average. The soluble portion of potassium was as high as 71% in average, the second element most easily leached into water after chlorine.

Unlike other components, lignin had an optimum ethanol concentration (40–60 vol%) to be dissolved, with the highest removal of 55.9 wt%. This result can be explained by the solubility parameter. In general, polymers and macro-molecules like lignin exhibit higher solubility in solvents with solubility parameters (δ values) similar to their own. Ni and Hu (2003) experimentally demonstrated that the solubility of Alcel lignin in ethanol/water mixed solvents at room temperature reached its maximum at an ethanol concentration of 71.3%, corroborating this finding through δ value calculations for both the solvent and lignin. Renders et al. (2016) also reported that lignin solubility in ethanol/water was highest at 50% ethanol concentration during catalytic reductive fractionation of woody biomass. Considering that lignin structure and solvent temperature also affect δ values, the results obtained in this study are deemed reasonable.

Next, the influence of pressurized CO₂ on the removal of components was investigated. Figure 3 compares the removal of components between treatments with CO₂ and N₂. It was clear that the addition of CO₂ had little effect. The aim of adding CO₂ was to promote the extraction and removal of components such as hemicellulose by providing acidic environment. The pH of water can be lowered to around three by dissolving CO₂ (Mohammadian et al., 2023). Under the experimental conditions of this study, if all the acetyl groups in the rice husk were released as acetic acid, the concentration would be 33 mmol/L, corresponding to a pH of 3.1. Thus, the pH achievable by acidification using CO₂ can be achieved solely by the acetyl groups. Pasquini et al. (2005) similarly attempted pre-treatment of sugar cane bagasse with ethanol/water under pressurized CO₂ atmosphere, but the purpose was delignification by the use of CO₂ as solvent under supercritical conditions. Little influence of CO₂ in the present study is attributed to the acidification of the solvent even in the absence of CO₂ by organic acids like acetic acid generated from hemicellulose, which does not allow for CO₂ to contribute to further acidification. Moreover, solubility of CO₂ in ethanol/water decreases with temperature (Dalmolin et al., 2006), limiting its contribution to the liquid phase reactions under employed conditions. Nevertheless, taking the global demand for developing CO₂ utilization technologies into consideration, the absence of any negative effects of CO₂, compared to N₂, is valuable information. Since the acidification of water by CO₂ dissolution is more pronounced at lower temperatures (Toor et al., 2011), the effect of pressurized CO₂ may be obtained if the treatment is conducted at sufficiently low temperatures while suppressing the generation of organic acids.

Based on the experimental results above, a method to separate the four components, including silica, is proposed in Figure 4. The first step is to separate hemicellulose with water as solvent, which proved effective in isolating nearly all hemicellulose while having minimal impact on other components. For completing the treatment with short time and achieving high extraction, higher temperature is more suitable; however, saccharides derived from hemicellulose are thermally unstable and may convert into furans, organic acids, and humins when excess treatment conditions are applied (Mohammadian et al., 2023; Ramos, 2003). To confirm this, the recovery of saccharides extracted during water treatment was analyzed (Figure 5). At 160°C, the recovery of saccharides in solvent was almost same with the removal of hemicellulose, but as the temperature increased, a difference between them emerged. At 200°C, the removal of hemicellulose was 95 wt%, while the recovery of saccharides (mono-saccharides + oligo-saccharides) was only 12.0 wt%. Therefore, 180°C was chosen as the treatment temperature suitable for recovering saccharides in the solvent. Notably, xylooligosaccharides derived from lignocellulosic biomass have attracted great attention for the potential to be used in food, feed, health, and cosmetics applications (Santibanez et al., 2021; Amorim et al., 2019). The degree of polymerization (DP) of oligo-saccharides obtained from hydrothermal extraction is generally required to be reduced to lower level such as DP2–DP4 by further hydrolysis, depending on the applications. For instance, low-DP xylooligosaccharides have higher prebiotic potential (Sukri and Mimi Sakinah, 2018).

The first treatment also removed water-soluble ash, primarily composed of AAEMs, contributing to the purification of silica in the ash. The subsequent treatment uses 60 vol% ethanol/water to remove lignin. The delignification at this step is incomplete as indicated in Figure 2. However, harsh extraction conditions give damage to or remove ash and cellulose. Therefore, the fourth treatment was added for bleaching and obtaining high-purity pulp instead.

Ideally, the main components remaining in the solid residue after the second treatment should be cellulose and ash (silica). To extract both as chemical raw materials, incineration should not be employed for obtaining silica. Therefore, it was determined to extract silica as silicate (Na_2x Si _y O_2y+x ) using an alkaline aqueous solution. The silica in rice husk is amorphous and can be extracted through this reaction (2xNaOH + ySiO₂ → Na_2x Si _y O_2y+x + xH₂O) under mild conditions. In preliminarily test, when silica extraction was attempted using 0.3 M NaOH aqueous solution under CO₂, the extraction rate was only 26.6 wt%. Since silica extraction with alkaline aqueous solutions is strongly influenced by the solvent’s pH (Hsieh et al., 2009; Park et al., 2021), this result was attributed to the pH reduction caused by dissolved CO₂ (2NaOH + CO₂ → Na₂CO₃ + H₂O). Indeed, the pH of the solution after the reaction was 8.3. Thus, N₂ was employed as the filling gas for this treatment. The obtained sodium silicate can be used as a raw material for a wide variety of silica-based materials, as well as in applications such as binders and emulsifiers (Minju et al., 2021; Liu and Ott, 2020).

The final treatment applies alkaline hydrothermal processing under an oxygen atmosphere to remove residual lignin. As mentioned earlier, oxygen is effective for lignin depolymerization and extraction, while its lower oxidation ability, compared to other oxidative bleaching method, such as the one using hydrogen peroxide (More et al., 2021; Bazargan et al., 2020), allows minimal loss of cellulose when applying appropriate conditions.

The composition of solid residues and the yields of extracted components after sequential application of these four treatments to rice husk are shown in Figure 6. The total extraction rates after all treatments were 97.6 wt% for hemicellulose, 98.4 wt% for lignin, and 99.3 wt% for ash. Each stage selectively extracted the target component. Hemicellulose accounted for 66.5 wt% of the extracted components from the first treatment, lignin accounted for 78.1 wt% of the second extracts, and ash accounted for 77.5 wt% of the third extracts. Such a selective fractionation of the four components was attributable to the stepwise treatment under appropriate conditions. The loss of cellulose during the process leading to pulp was 22.8 wt%. The fourth treatment accounted for more than half portion of extracted cellulose.

The final pulp obtained was of high purity, containing only 1.6 wt% lignin and 0.6 wt% ash (Figure 7). For comparison example, Dagnino et al. (2017) used a two-step treatment with dilute sulfuric acid and NaOH/water/ethanol to produce pulp from rice husk, which retained 6 wt% lignin and 24 wt% ash. In the present study, Na₂CO₃ aqueous solution was used as the solvent for the fourth treatment. Considering pH of the solution, NaOH potentially results in greater lignin removal. The performance of these salts is compared in Figure 7 using their solution at the same concentration. When used NaOH, the lignin content in the pulp decreased only slightly to 1.4 wt%, but the loss of cellulose increased to 26.6 wt%. Thus, lignin removal in the fourth treatment was mainly dependent on the presence of oxygen, and the advantage of using NaOH aqueous solution was considered minimal. Furthermore, NaOH aqueous solution is carbonated during processing of biomass (Kudo et al., 2014). Taking the regeneration and reuse of alkaline aqueous solution into account, Na₂CO₃ would be a reasonable choice for this treatment. As mentioned earlier, lignin in this treatment is oxidatively depolymerized and extracted. Consequently, organic acids such as oxalic acid, acetic acid, and formic acid are selectively obtained. However, many other organic compounds are also included in the extracts. Considering the low lignin content in the raw material of this step and the low concentration of lignin-derived compounds in the extract, the purpose of this treatment was set solely as cellulose purification. For this reason, compositional analysis of the extract was not performed in the present study. A method for valorizing the extracted organics is also discussed later.

In previous studies on biomass component fractionation, the main objective was to separate organic components, but there have been some reports that also considered the separation of ash. Zhang et al. (2015) applied a concept similar to that in the present study to separate rice husk into xylan, lignin, glucose-derived ethanol, and ash through hydrolysis, solvent extraction (butanediol), dilute acid hydrolysis, and then alcoholic fermentation, yielding 10, 22, 5, and 17 wt% from rice husk, respectively. Furthermore, Durand et al. (2024) recently proposed a sequential pre-treatment process for extracting biopolymers (cellulose and lignin) along with silica from rice husks, achieving lignin recovery at 47% with 82% purity and extraction of cellulose with 1% ash content and 82% purity. Compared to these prior studies, the originality of the present research includes pursuing the pre-treatment without expensive reagents including catalytic species and successfully separating each component while maintaining their structure as natural as possible. The introduction of oxidative alkaline treatment resulted in the production of very high-purity pulp. Hemicellulose-derived saccharides and cellulose were indeed obtained while moderately retaining their degree of polymerization (DP), ensuring their usability in broad applications. Further studies on this type of process and fine-tuning of treatment conditions are expected to enable even higher selectivity.

However, the proposed process has challenges to be solved: 1) Further optimization of processing conditions is essential to improve the yield of each extract. Many operational factors influence the yields. Since the performance of each treatment affects others, an overall process consideration is necessary. 2) Performing multi-step pre-treatment using solvents raises concerns about the significant energy required for heating the solution and materials. To address this, utilization of heat exchange is important. The residual heat from previous treatments can be used for the rice husk samples. While, heating of the solutions is recommended to proceed with heat exchange starting from the higher-temperature treatments. However, efficient heat exchange may be challenging with batch processing, as used in this study. A potential alternative is flow-through processing, where the extraction liquid flows through a fixed bed of rice husk samples. This system allows for easy heat exchange between the inflow and outflow solutions and enables all steps to be performed with a single reactor by changing the inflow solution and processing conditions. The flow-through extraction has recently gained attention for its ability to extract components like lignin in a natural structure-preserved form (Brandner et al., 2021; Xu et al., 2021). 3) From a cost perspective, solvent recycling is essential. For water, ethanol/water, and the aqueous solution used in the third step, their evaporation is necessary to recover the extracts and reuse the solution. If impurities like organic acids are present in the recovered solution and difficult to separate, it will be necessary to investigate their accumulation and impact on the treatments. An alternative to energy-intensive evaporation-based separation is the application of membrane separation technology. In the fourth step, the extract solution contains a mixture of organic compounds such as organic acids. Their recovery along with regeneration of the solution should be considered when they are valuable. Alkaline hydrothermal gasification is considered to be a method for regenerating the alkaline solution. It has been reported that the reaction using solid catalysts enabled to recover high-calorie gas consisted mainly of methane from lignin and Na₂CO₃ aqueous solution free from organic matter (Kudo et al., 2014). 4) For the practical implementation of the process, it is necessary to ensure that the internal pressure is sufficiently low. For example, in Japan, if the operating pressure exceeds 1 MPa, the system is considered to contain high-pressure gas, raising regulatory hurdles for process implementation. The saturated vapor pressure of water increases rapidly within the temperature range targeted in this study, exceeding 1 MPa at temperatures above 180°C. Therefore, it is essential to operate at the lowest feasible temperature and keep the pressure of any filling gas as low as possible.