Bambusa emeiensis were provided by YIBIN PAPER INDUSTRY CO., LTD. (Sichuan, China). Moreover, the bamboo sawdust (BS) was derived from Bambusa emeiensis after being crushed in a pulverizer and sieved 20–80 mesh (0.18mm–0.85 mm). The composition of BS as follows 39.62% glucan, 13.69% xylan and 31.16% lignin.

ZrO₂, SnO₂, TiO₂ and Fe₂O₃ were obtained by sintering Zr(OH)₄, Ti(OH)₄, Fe(OH)₃ and Sn(OH)₄ in air, respectively. ZrO₂, SnO₂, TiO₂ and Fe₂O₃ were impregnated with 1 mol L^-1 sulfuric acid for 1 h (metal oxide: sulfuric acid = 1 g: 15 mL), and then calcined in a muffle furnace. Typically, the calcination temperature was 500 °C for Fe, Sn, Ti and 600°C for Zr, as well as the calcination time was 3 h for Zr, Fe and 4 h for Ti, but 5 h for Sn. The final prepared superacid catalysts were named as SO₄ ²⁻/Fe₂O₃, SO₄ ²⁻/TiO₂, SO₄ ²⁻/ZrO₂ and SO₄ ²⁻/SnO₂, respectively.

5.0 g of diatomaceous earth (Chengdu Kolon Chemicals Ltd.) and 28.8 g of 85% phosphoric acid (Chengdu Kolon Chemicals Ltd.) was mixed and stirred to form a viscous paste mixture, which was then calcined in a muffle furnace at 300 °C for 48 h to yield white solid. Afterwards, the resulting product was cooled to room temperature in dry environment and ground into powder to produce solid phosphoric acid “SPA”.

Phosphotungstic acid (PTA) was purchased in Chengdu Kolon Chemicals Ltd (Sichuan, China).

All catalysts have been passed through a sieve with 0.15 mm aperture.

The heterogeneous catalytic degradation reaction system was formed in an oil bath at 170°C for 18 h. 5 g bamboo sawdust was added to a mixture of 30 g of PEG 400 and 10 g of glycerol, and 1.5 g solid acid catalyst was added. In addition, the filtrate was filtered and the residue was washed with 1,4-dioxane/water (4:1, v/v) until the reacted detergent was colorless. After that, the detergent was recovered by rotary evaporation at 60°C and reused. The filter residue was dried at 105°C for 8 h and then calcined at 550°C before calculating the liquefaction conversion. The mass burned off is the mass of organic matter in the bamboo sawdust residue, and this calculation ignores the effect of solid acid loss. R % = m r − m c m 0 * 100 % (1)

Where the R represents the residue yield (%), m_r refers to the residue mass (g), m_c stands for the residue mass after burning (g) and m₀ is the bamboo sawdust mass (g).

The liquefaction effects of different solid acids on the same mass (1.5 g) of bamboo sawdust are shown in Table 1. Different kinds of solid acids show great differences on bamboo sawdust, among which the residue rates of loaded acid SPA, heteropolyacid PTA and SO₄ ^2-/M_xO_y are 17.72%, 36.7% and 56.96%–68.02%, respectively. Therefore, among these catalysts, SPA exhibits the greatest effect on the liquefaction of bamboo sawdust. However, from the NH₃-TPD profiles (Figure 1), it can be seen that the heteropolyacid PTA is the most acidic as it has obvious NH₃ resolution peaks in the high temperature region relative to SPA and SO₄ ^2-/M_xO_y. According to previous studies, the stronger the acidity, the better the effect on biomass liquefaction (Lin et al., 2022). Nevertheless, for PTA and SPA, two typical proton solid acids, the opposite result is observed. The “free P₂O₅” on the surface of SPA readily interacts with low molecular organic matter. As a result, SPA is susceptible to forming complexes with low molecular organic matter, which makes it easier to come into contact with the substrate. In addition, water is produced during the reaction, which hydrolyzes some of the solid acid on the surface of the SPA to produce liquid phosphoric acid (Santander et al., 2019). Then, the liquid phosphoric acid comes into contact with the reactants and catalyzes the liquefaction of bamboo sawdust. As a result, the residue rate is the lowest under SPA catalyzation. These results indicate that the loaded solid acid has obvious advantage in catalyzing the liquefaction of bamboo sawdust in polyol alcohol system.

The desorption temperature is related to the acid strength of the solid acid. Generally, a higher desorption temperature indicates a more difficult desorption of NH₃ and therefore a higher acid strength of the solid acid (Liu et al., 2014). In the low-temperature region of 100°C–200°C, all types of solid acids show resolved peaks, which are mainly due to the physical adsorption of NH₃ on the surface of solid acid as well as its chemisorption on the weakly acidic sites. In the middle and high temperature region, compared with other solid acids, PTA solid acid has a strong absorption peak in the high temperature region around 550°C, and thus has the highest acid strength. In addition, the NH₃ resolution peak of SPA is higher than that of SO₄ ^2-/SnO₂ in the middle and high temperature region. Therefore, the order of acid strength of these three solid acids is PTA > SPA > SO₄ ^2-/SnO₂.

In Table 1, the residue rates of bamboo sawdust after liquefaction of the three types of solid acids were ranked as follows: SPA (17.72%) > PTA (36.72%) > SO₄ ^2-/SnO₂ (56.96%). The acidity of PTA is higher than that of SPA, but in contrast, the residual rate after PTA treatment is higher. This is probably because SPA is a loading acid that produces small molecules of water during the liquefaction of bamboo sawdust, which hydrolyzes part of the solid acid on the surface of SPA to produce liquid phosphoric acid (Santander et al., 2019). In addition, the liquid phosphoric acid comes into contact with the reactants and catalyzes the liquefaction of bamboo sawdust under their joint action with the lowest residual rate. However, the residual rate of bamboo sawdust liquefaction catalyzed by PTA is smaller than that of catalyzed by SO₄ ^2-/SnO₂, which may be attributed to the fact that PTA is more acidic than SO₄ ^2-/SnO₂.

As shown in the NH₃-TPD profiles (Figure 1), compared with SO₄ ^2-/ZrO₂, SO₄ ^2-/Fe₂O₃ and SO₄ ^2-/TiO₂, the SO₄ ^2-/SnO₂ solid acid has a resolved peak in the high-temperature region at 500°C, and thus the acid strength is the largest. Moreover, SO₄ ^2-/ZrO₂ and SO₄ ^2-/TiO₂ have obvious resolved peaks in the middle and low temperature regions, with similar peak areas. In contrast, SO₄ ^2-/Fe₂O₃ has no distinct resolved peak in the whole NH₃-TPD curve. Therefore, the acid intensity of SO₄ ^2-/M_xO_y solid acid was ranked as follows: SO₄ ^2-/SnO₂ > SO₄ ^2-/ZrO₂, SO₄ ^2-/TiO₂ > SO₄ ^2-/Fe₂O₃.

From Table 1, the liquefaction residue rate of SO₄ ^2-/SnO₂ solid acid was 56.96%, while the residue rates of SO₄ ^2-/ZrO₂, SO₄ ^2-/TiO₂ and SO₄ ^2-/Fe₂O₃ were 61.59%, 68.02% and 65.21%, respectively. These results indicated that the stronger the acidic site of solid acid, the better the liquefaction effect of bamboo sawdust in polyol system. In addition, the liquefaction residue rate of SO₄ ^2-/Fe₂O₃ was essentially the same as that of SO₄ ^2-/ZrO₂ and SO₄ ^2-/TiO₂, suggesting that the weak acid sites of the solid acid had no impact on the catalytic effect of bamboo sawdust in the polyol system.

In summary, the liquefaction efficiency of bamboo sawdust under the polyol system of PEG400/glycerol is mainly related to the acid strength of the solid acid and the type of the solid acid with less correlation to the total amount of the acid in the solid acid: (1) the stronger the solid acid, the better the catalytic effect; and (2) the loading of solid phosphoric acid, the better the catalytic effect. This is because water molecules will be generated during the reaction process, and the water molecules will lead to the hydrolysis of solid phosphoric acid to generate liquid phosphoric acid, which will increase the reaction rate and make the bamboo sawdust liquefied more fully under the polyol system.

In order to investigate the effect of different solid acids on the liquefaction process of bamboo sawdust, the biomass fractions in the solid acid-catalyzed residue were determined. As shown in Figure 2, the relative content of cellulose in the residue increased in all catalyst groups compared with that of bamboo sawdust, with the most significant increases in the SPA and PTA groups. These results indicate that cellulose is the most difficult part to liquefy during the liquefaction of bamboo sawdust. Acid-soluble lignin was fully reacted in all catalyst groups, while the content of acid-insoluble lignin varied. In addition, the hemicellulose content varies among the catalyst groups, with complete reaction of hemicellulose in the SPA and PTA groups, and changes in the hemicellulose content of the other groups compared to the bamboo sawdust. In order to quantify the effect of solid acid types on the content of three major components in the liquefaction of bamboo sawdust, the relative contents of these components in bamboo sawdust (before liquefaction) and their residues (after liquefaction) were quantified (Table 2). The relative content of the three major elements (X) in the residue (after liquefaction) as a percentage of the pre-liquefaction bamboo sawdust was calculated according to the following equation (Niu et al., 2011): X = M 1 × R × 100 % (2) where M₁ represents the relative content of the three major elements in the residue and R refers to the yield of the liquefied residue.

Cellulose is most difficult to liquefy because of its crystalline structure, which makes it difficult for the liquefying agent to act on it. In Table 2, the cellulose content in bamboo sawdust was 58.86%, which decreased under all conditions by adding solid acid. The cellulose contents of bamboo sawdust decreased to 39.51%–47.23% catalyzed by solid acid SO₄ ^2-/M_xO_y. In comparison, the cellulose content in bamboo sawdust was 27.29% using PTA with higher acid strength as catalyst. Therefore, the increase in solid acid strength facilitated the liquefaction of cellulose in bamboo sawdust. When SPA was used as a catalyst, the cellulose content was only 13.82%, which was attributed to the hydrolysis of solid phosphoric acid on the surface of SPA to small molecule phosphoric acid, promoting the hydrolysis of cellulose (Espinosa et al., 2013). Comparative analysis of the data show that cellulose in bamboo sawdust is the most difficult to be liquefied under the polyol system of PEG400/glycerol, which is the main obstacle restricting the complete liquefaction of bamboo sawdust by solid acids.

The hemicellulose content of bamboo sawdust is 9.41%, which is relatively low. Moreover, the addition of various solid acid catalysts can promote the liquefaction of hemicellulose in bamboo sawdust. Notably, there is no hemicellulose in the residues after solid acid catalysis by PTA and SPA. Therefore, the hemicellulose in bamboo sawdust can be completely liquefied by solid acid catalysis with sufficient acidity in the polyol system.

The lignin in bamboo sawdust could be categorized into acid-soluble lignin and acid-insoluble lignin, with contents of 4.43% and 27.30%, respectively. The addition of various solid acids completely liquified the acid-soluble lignin in bamboo sawdust, indicating that the acid-soluble lignin in bamboo sawdust was more easily liquefied. Moreover, the solid acid catalysts could also promote the liquefaction of acid-insoluble lignin in bamboo sawdust. Among them, the liquefaction effect of SPA was the best, which could reduce the acid-insoluble lignin content in bamboo sawdust from 27.30% to 4.25%. The lignin content in bamboo sawdust was extremely low. Studies have shown that lignin is more easily liquefied by acid and this incomplete liquefaction may be due to the long liquefaction reaction time. Additionally, the decomposed small-molecule lignin was further condensed to large-molecule compounds (Xie et al., 2014), which remain in the residue. Therefore, acid-insoluble lignin is present in the residue.

Conclusively, the addition of solid acids promoted the liquefaction of the three main elements in bamboo in the polyol system of PEG400/glycerol. Acid-soluble lignin was the easiest to liquefy, and the addition of all types of solid acids resulted in its complete liquefaction. This was followed by hemicellulose and acid-insoluble lignin, which could be completely liquefied by solid acids with increasing acidic strength.

Cellulose is the most difficult to be liquefied, which may be due to the fact that cellulose molecules are aggregated in the form of micro protofibrils, in which the crystalline and amorphous regions alternate, and the amorphous portion of cellulose is easily hydrolyzed during acid hydrolysis, leaving the crystalline fraction behind (Li, 2004; Kuthi and Badri, 2014). Meanwhile, during the liquefaction process, the liquefaction products undergo condensation reactions with lignin, condensation products tend to remain on the crystalline cellulose in the residue (Xie et al., 2014). These factors make it difficult for the residual cellulose to come into contact with the liquefiers for further reaction.

In order to compare the changes of functional groups on the surface of residues after liquefaction with different catalysts, the liquefaction residues, bamboo sawdust and the mixture of bamboo sawdust and solid acid with 1.5 g of catalyst were characterized by IR spectroscopy, and the results were shown in Figure 3.

The infrared absorption peaks of the main components of BS (cellulose, hemicellulose and lignin) can be seen in the infrared spectra. The broad peaks between 3,300 and 3,500 cm-1 belong to the stretching vibration peak of -OH. Moreover, the absorption peak at 2,926 cm-1 is attributed to the vibration absorption peak of -CH3, while the absorption peak appearing at 1733 cm-1 should be attributed to the uncoupled C=O stretching vibration, which is also the absorption peak of characteristic functional groups of hemicellulose (Wen et al., 2011; Xie et al., 2014). Notably, the peak intensity is not high due to the small content in BS. The absorption peaks at 1,629 cm-1, 1,518 cm-1 and 1,457 cm-1 belong to the benzene ring skeleton vibration, corresponding to the lignin structure in bamboo sawdust (Zhang et al., 2010). Absorption peak at 1,384 cm-1 is attributed to C-H bending vibration, and the peak at 1,053 cm-1 belongs to C-O stretching vibration, which is also a characteristic absorption peak of cellulose and hemicellulose in the infrared (Latif et al., 2014). One of the peaks at 896 cm-1 belongs to β-glucoside bond vibration, which is also the infrared characteristic absorption peak of cellulose (Chen et al., 2011). Meanwhile, the mixture of BS and solid acid catalyst were characterized by IR spectroscopy. It was found that the addition of solid acid catalyst had no effect on the position and intensity of the IR characteristic absorption peak of BS (cellulose, hemicellulose and lignin). Therefore, the effect of solid acid-catalyzed liquefaction of BS could be evaluated by comparing the changes in the IR characteristic absorption peak of BS and liquefaction residue.

As shown in Figure 3, the IR spectral curves of pure BS, the mixture of BS and catalyst, and the residue of solid acid liquefaction are shown from top to bottom, respectively. Solid acids such as SO42-/ZrO2, SO42-/TiO2, SO42-/Fe2O3 and SO42-/SnO2 can catalyze the liquefaction of bamboo sawdust, but the residue rate is high. The liquefaction rate was low, and only part of cellulose, hemicellulose and lignin were liquefied (Figure 2; Table 2). As a result, the position of IR characteristic peaks was basically unchanged and the intensity did not change much, which was consistent with the detection results of residue compositions.

When PTA with high acidity was used as catalyst, the IR absorption peak of hemicellulose at 1733 cm-1 basically disappeared compared with the BS spectrum, indicating that hemicellulose had been completely reacted. In addition, the absorption peaks of cellulose and hemicellulose at 1,053 cm-1 also disappeared, while new absorption peaks, i.e., the IR absorption peaks of cellulose, appeared at 1,080 cm-1, 1,053 cm-1 and 896 cm-1. Meanwhile, a new IR absorption peak was observed at 812 cm-1, which should be attributed to α-glycosidic bond vibrations (Wen et al., 2011), indicating an increase in cellulose content in the residue. This result is consistent with the detection of the residue compositions.

When SPA with better liquefaction effect was used as catalyst, the IR absorption peak at 1733 cm-1 of hemicellulose was found to be completely disappeared when compared with the BS IR spectrum, indicating that hemicellulose was completely liquefied. At the same time, a new peak reappeared at 1719 cm-1, which was attributed to the absorption peak of conjugated C=O in aromatic nuclei. After liquefaction, the surface lignin was enriched on the residue surface by the products of polycondensation. In addition, the IR characteristic absorption peaks of cellulose and hemicellulose at 1,053 cm-1 of bamboo sawdust catalyzed by SPA disappeared, and two new peaks at 1,187 cm-1 and 1,094 cm-1 were formed with high peak intensity. These two peaks are the infrared characteristic absorption peaks of cellulose (Wen et al., 2011), indicating that the content of cellulose in liquefaction residue is increased. On the other hand, the intensities of the IR characteristic peaks of 1,629 cm-1, 1,518 cm-1 and 1,457 cm-1 in the liquefied residue were reduced compared with those in the BS, indicating lignin degradation, which was consistent with the results of residue components.

Above infrared analysis results showed that with the increase of the acidity of the solid acid, the hemicellulose and lignin in bamboo material disappeared or weakened, while the characteristic absorption peak of residual cellulose was enhanced and the liquefaction effect was improved. All these indicated that the stronger the acid strength of solid acid, the better the liquefaction effect of bamboo sawdust under the polyol system of PEG400/glycerol.

Figure 4 shows the diffraction intensity curves of the liquefaction residue and BS under the condition of 1.5 g catalyst dosage. According to the XRD pattern of BS, the main diffraction angles 2θ° are 16.06°, 22.05° and 26.30°, among which 16.06° and 22.05° are the diffraction peaks of natural cellulose Ⅰ (Langan et al., 2001; Nishiyama et al., 2002). In addition, all samples have a narrow diffraction peak at 26.30°, which may be the absorption peak corresponding to the BS containing SiO2 (Tymchyshyn and Xu, 2010).

Compared with BS, the liquefaction residue of SO42-/ZrO2, SO42-/TiO2, SO42-/Fe2O3 has diffraction peaks of cellulose Ⅰ at 22.05° and 16.06°. It is noteworthy that the shape of the peak remains almost unchanged, but there is a decrease in strength, which may be due to the liquefaction of some cellulose in the bamboo. In addition, this may also be due to the presence of a large amount of solid acid catalyst in the residue.

With the increase of liquefaction efficiency, the diffraction absorption peak belonging to cellulose Ⅰ is disappeared at 16.06° when SO42-/SnO2 is used as catalyst, indicating that part of cellulose Ⅰ also reacts.

When PTA is used as a catalyst, the characteristic diffraction peaks of natural cellulose at 22.05° disappear, while the diffraction peaks of cellulose II at 21.96° and 20.80° appear (Langan et al., 2001), which may be attributed to the transfer of the metastable structure of natural crystalline cellulose Ⅰ from bamboo sawdust to thermodynamical stable cellulose Ⅱ (Isogai et al., 1989; Langan et al., 2001).

A similar phenomenon is observed in the SPA liquefaction residue when the residue rate is only 17.72%. The new diffraction absorption peaks are appeared at 21.96°, 20.80° and 22.96°. Meanwhile, the new characteristic diffraction peaks belonging to cellulose are sharp, indicating that the crystalline cellulose on the surface of the residue has been reacted, so the crystallinity of the remaining cellulose has been improved, which explains the difficulty of cellulose liquefaction in bamboo sawdust.

From the above XRD analysis results, it can be inferred that with the increase of solid acid acidity, the amorphous part of cellulose will be liquefied firstly during the catalytic process, leaving the crystalline part of natural cellulose, due to the alternation of crystalline and amorphous regions of natural cellulose. Then, when the acidity of the solid acid is further enhanced, part of the natural crystalline cellulose will be liquefied. Part of it is converted into the thermodynamically more stable cellulose II, which is the reason why bamboo sawdust is difficult to be completely liquefied.

Figure 5 shows the TG curve of liquefaction residue and the BS to remove the inorganic salts as well as the remaining solid acid from the residue at a catalyst dosage of 1.5 g. Based on the data in Figure 3, the weight loss of all the samples can be calculated and the results are shown in Table 3. From Figure 5, it can be seen that the weight loss zone of bamboo sawdust can be divided into three parts. In stage a, the weight loss rate is 4.85%, which is ascribed to the evaporation of water from the residue. In stage b, the weight loss rate is as high as 90.81%, which is mainly due to the pyrolysis of lignin, cellulose and hemicellulose according to the compositional analysis in Figure 2. In stage c, the weight loss rate is only 4.34%, which is primarily ascribed to the pyrolysis of charcoal residue after bamboo pyrolysis.

The weight loss of SO₄ ^2-/MxOy solid acid, SPA and PTA liquefaction residue is basically the same as that of the bamboo sawdust at stage a in the weight loss zone, which is mainly due to the evaporation of water from the residue, which is mainly due to the evaporation of water from the residue.

Compared with the weight loss of bamboo sawdust in the b-zone stage, the significant weight loss temperature range of all solid acid residues in this stage is shortened, with significantly faster weight loss and lower weight loss rate (Figure 3), which is inversely correlated with the acid strength. This is attributed to the liquefaction of cellulose, hemicellulose and lignin in bamboo cuttings catalyzed by solid acid. Notably, the greater the acid strength, the more complete the liquefaction, and the more complete the destruction of the main structure of bamboo sawdust. Therefore, in stage b, the temperature interval of weight loss is significantly shortened, and the rate of weight loss is significantly accelerated, while the weight loss rate is reduced.

For stage c, the TG curves of all catalyst liquefaction residues and bamboo cuttings are relatively flat, and the weight loss rate is relatively slow, which might be due to the pyrolysis of the more thermally stable organic matter in the liquefaction residues.

As for stage d, it can be seen from Figure 5 that only the liquefaction residue of SO42-/SnO2, SPA and PTA have thermogravimetric regions in the high-temperature region at this stage. The relationship is as follows: the range of weight-loss temperature, weight loss rate and weight loss rate of the residue in stage d increase with the decrease of the liquefaction residue rate, which indicates that the three elements in the liquefied bamboo sawdust produce products that are difficult to be pyrolyzed with the increase of the acidity of the solid acid. By analyzing the weight loss curve in stage d, it is found that the weight loss rate of SPA liquefaction residue is 35.84%. From Figure 2, it can be seen that the composition of SPA liquefaction residue is 21.96% acid-insoluble lignin and 78.04% cellulose. Therefore, it can be judged that the material of heat weight loss at this stage is mainly cellulose. Meanwhile, XRD results show that the cellulose in the residue is mainly composed of cellulose Ⅱ with relatively high crystallinity. Previous studies have shown that cellulose is less prone to weight loss during low-temperature pyrolysis when the crystallinity of cellulose is high (Mukarakate et al., 2016) cellulose is less prone to weight loss during low-temperature pyrolysis when the crystallinity of cellulose is high. This explains well why the liquefied residues of SO42-/SnO2, PTA and SPA increase by an interval of weight loss at 700 °C.