Reaction Behaviors of Bagasse Modified with Phthalic Anhydride in 1‐Allyl‐3‐Methylimidazolium Chloride with Catalyst 4‐Dimethylaminopyridine Reaction Behaviors of Bagasse Modified with Phthalic Anhydride in 1-Allyl-3-Methylimidazolium Chloride with Catalyst 4-Dimethylaminopyridine

The modification of lignocellulose with cyclic anhydrides could confer stronger hydro- philic properties to lignocellulose, which could be used in many industrial fields. To elucidate the modification mechanism of lignocellulose, bagasse was phthalated com- paratively with its three main components in 1-allyl-3-methylimidazolium chloride (AmimCl) using 4-dimethylaminopyridine as catalyst and phthalic anhydride as acyla- tion reagent in the present study. From FT-IR and 2D HSQC analyses, the skeleton of bagasse and the fractions were not significantly changed during phthalation in AmimCl. 2D HSQC results suggested that the reactive hydroxyls in bagasse were partially phthalated, and the reactivity of the hydroxyls in anhydroglucose units followed the order C-6 > C-2 > C-3. Similarly, the reactivity order of hydroxyls in anhydroxylose units was C-2 > C-3. For lignin, the predominant diesterification occurred during the homo- geneous modification, and both aliphatic and aromatic hydroxyls were phthalated. The reactivity order of phenolic hydroxyls was S-OH > G-OH > H-OH, which was distinct from that without catalyst. In addition, it was found that the thermal stability of phthalated bagasse was affected by the disruption of cellulose crystallinity and the degradation of components. The thermal stability of the phthalated bagasse decreased upon chemical modification and regeneration.


Introduction
Nowadays, the growing environmental concerns resulting from petroleum-based plastic materials and excessive use of fossil fuel have driven a strong global interest in renewable bio-based polymers and composites derived from natural resources [1,2]. Biomass resource is one of the most abundant, inexpensive, and currently underutilized products from agricultural and biorefinery industries [3]. Biomass resources, especially agricultural residues with great renewability, biocompatibility, and biodegradability, are directing the development of next generation of chemicals, materials, energy, and processes [4]. Therefore, it could be particularly advisable to shift society's dependence away from fossil resources to biomass resources.
Bagasse, sugar industry residues, is mainly composed of lignocellulose, which represents one of the most abundant renewable resources on earth. Chemical modification represents an efficient process to introduce new functionalities into lignocellulose. Because of the poor dissolubility of lignocellulose in common molecular solvent, the heterogeneous modification has been studied for decades. However, heterogeneous modification is always accomplished at low efficiency and nonuniformity, and accompanied with the occurrence of side reactions [5,6]. Even worse, the solvent or reaction medium for heterogeneous modification could not be recovered after modification [7], which caused environmental pollution and wastage of resources. It is a never-ending endeavor to explore homogeneous reaction systems to compensate the deficiency of heterogeneous systems.
Fortunately, it has been found that several dual solvent systems and ionic liquids (ILs) could dissolve lignocellulose [8]. Compared with other solvent systems, ILs exhibit negligible vapor pressure, theoretically making infinite recycling possible. In addition, the high electrochemical thermal stability, nonflammability, and designability made ILs widely used in diverse fields, such as solvent for synthesis and catalysis, separation, analytical chemistry, and biomass refinery [9][10][11]. The bridge between lignocellulose and ILs were built by Swaltloshi et al., who reported the nonderivatizating dissolution of cellulose in BmimCl [12]. In 2007, it was reported that woods with different hardness could be readily dissolved in various imidazium-based ILs under mild conditions [13,14], which opened a new window of opportunities in homogeneous functionalization of lignocellulose. Since then, the chemical modification of lignocellulose in ILs has been expanding fast. Due to the existence of abundant hydroxyls in lignocellulose, the modification of lignocellulose with carboxylic anhydride, acid chloride, and isocyanides could be achieved using ILs as reaction media [15].
The chemical modification of lignocellulose with linear carboxylic acid anhydrides or acid chloride can produce the corresponding carboxylic acid or HCl as a by-product. However, the modification of lignocellulose with dicarboxylic acid anhydride is an effective process and can attach carboxyl groups onto lignocellulose. The attachment of carboxyl group can greatly improve the compatibility between modified lignocellulose and other polymer materials by producing new chemical linkages and the effect of hydrogen bonds [16]. Besides, the introduction of carbonyl group could also increase the hydrophilicity [17], achieve ion exchange [18], and remove heavy mental ion [19]. Our previous works showed that the homogeneous chemical modification of the isolated cellulose from bagasse with succinic anhydride and phthalic anhydride without catalyst could be achieved, and the optimal reaction conditions (reaction temperature 90°C and reaction time 90 min) for preparing modified cellulose with maximum degree of substitution (DS) were also screened [20,21]. In 2010, Li et al. reported chemical modification of cellulose in BmimCl with catalyst DMAP could remarkably improve the DS of modified products [22]. However, the structural parameters of lignocellulosic derivatives were complex and difficult to control due to the complicated mixture of cell wall components. Therefore, it is very essential to elucidate the modification mechanism of lignocellulose to produce tunable materials directly from lignocellulose, although some simple investigations have been published. Especially, the application of catalyst DMAP could obviously increase the degree of substitution of lignocellulose modified with cyclic anhydride, which is beneficial to produce polymer composites based on esterified lignocellulose. As far as the authors are aware, there have been no reports of the detailed structural changes of lignocellulose during the homogeneous phthalation with catalyst DMAP. We therefore investigated the reaction behaviors of lignocellulose during the homogenous phthalation with catalyst DMAP.
In the present study, bagasse was modified with phthalic anhydride with catalyst DMAP in AmimCl. Three main components cellulose, hemicelluloses, and lignin were isolated from bagasse, and were phthalated under the same conditions to elucidate the modification mechanism. The physicochemical properties of phthalated samples were characterized by FT-IR, CP/MAS 13 C NMR, 1 H NMR, liquid-state 13 C NMR, and 2D HSQC NMR. The thermal stability of the unmodified, regenerated, and modified materials was also investigated using TG.

Materials
Bagasse was obtained from a local factory (Jiangmen, China). It was dried in sunlight and then cut into small pieces. The cut bagasse was ground and screened to prepare 40-60 mesh size particles (450-900 µm). The dried ground samples were dewaxed with toluene-ethanol (2:1 v/v) and then dried in a cabinet oven with air circulation at 50°C for 24 h [15]. Ionic liquid AmimCl was purchased from the Shanghai Cheng Jie Chemical Co., Ltd. and used as received. All other chemicals used were of analytical grade and purchased from Guangzhou Chemical Reagent Factory (Guangzhou, China).

Isolation of cellulose, hemicelluloses and lignin from bagasse
Cellulose and hemicelluloses were isolated according to the previous literatures [21,23]. Briefly, the extractive-free bagasse was delignified at 75°C for 2 h with sodium chlorite at pH 3.8-4.0, followed by the extraction with 10% NaOH with the ratio of solid to liquid at 1:20 g/mL for 10 h at room temperature (four times). The solid residues were filtered out, washed thoroughly with distilled water, then washed with ethanol and dried in an oven with air circulation at 50°C for 24 h to obtain cellulose. The filtrate from NaOH extraction (the first time) was neutralized, concentrated, and transferred into three volumes of 95% ethanol with agitation. The precipitates were filtered out, washed with 70% ethanol, and freeze-dried to obtain hemicelluloses.
Lignin was isolated from the extractive-free bagasse according to the previous literatures [24,25]. The obtained crude lignin with high content of carbohydrates was purified as follows: dried crude lignin (1 g) was dissolved in 2 mL of acetone/water (9:1, v/v) in a beaker. The obtained lignin solution was added dropwise into 200 mL of distilled water with stirring, and the resulted suspension was further centrifuged. The obtained solid residues were washed with distilled water (thrice, total 90 mL), and then freeze-dried. The dried residues were dissolved into 2 mL of 1,2-dichloroethane/ethanol (2:1, v/v). The obtained solution was added dropwise into 200 mL of anhydrous ether in a beaker with stirring at room temperature, and the resulted suspension was further agitated for 30 min. The obtained suspension was centrifuged, and the solid residues was washed with anhydrous ether (thrice, total 90 mL) and freeze-dried.

Homogeneous phthalation of bagasse and the isolated fractions (cellulose, hemicelluloses and lignin)
The homogeneous phthalation of bagasse and its isolated fractions was performed according to the previous literature [26]. The extractive-free ground bagasse (12 g) was first finely ball-milled for 4 h in a planetary mill (Grinder BM4, China, Beijing) at 400 rpm using two 500 mL ZrO2 jars (500 g, 1 cm diameter). Cellulose isolated from bagasse was also ball-milled for 4 h under the same conditions, while hemicelluloses and lignin were directly used in the present study.
Typically, the prepared materials (0.2 g) was added to 10 g of AmimCl at room temperature under nitrogen atmosphere for 5 min and heated at 90°C under stirring for 4 h to obtain a clear solution. Phthalic anhydride (the weight ratio of phthalic anhydride to material, 4:1) and DMAP (the weight ratio of DMAP to phthalic anhydride, 5%) were added to the solution. The flask was continuously purged with nitrogen atmosphere for 5 min. The reaction was performed at 90°C for 90 min with agitation. Subsequently, the resulted solution was precipitated with ethanol (99 wt%, 200 mL) under agitation. The suspension was further stirred for 24 h. The solid residues were filtered out, washed with ethanol (three times, total 600 mL) to remove unreacted phthalic anhydride, catalyst DMAP, and AmimCl, and free-dried for further characterization.

Determination of WPG and DS
WPG [15] was calculated according to Eq. (1): where M 0 and M 1 are the vacuum-dried weights of the bagasse or fractions before and after chemical modification, respectively.
The phthalation degree of phthalated samples was evaluated by DS, according to Eq. (2): where n OH ' is the substituted hydroxyls contents, and n OH is the hydroxyl content of unmodified sample.
The substituted hydroxyl content of the phthalated samples was determined based on the equivalent volume of known molarity NaOH and HCl by back-titration method [21]. A known weight of the sample was dissolved in 10 mL 0.1 mol/L NaOH. The excess of NaOH was backtitrated with 0.025 mol/L HCl by using automatic potentiometric titrator, and the final pH of end titration point was set as 7.0. The titration was repeated three times, and the average value of HCl volume was used as the calculations of the substituted hydroxyl content according to Eq. (3): where n OH ' is the substituted hydroxyls contents, m (g) is the dry weight of sample analyzed, c 1 (mol/L) is the molarity of NaOH, V 1 (mL) is the volume of NaOH, c 2 (mol/L) is the molarity of HCl, and V 2 (mL) is the equivalent volume of known molarity HCl.
Based on the assumption that the cellulose and hemicelluloses are composed of AGU and AXU, respectively, the theoretical hydroxyl content of unmodified cellulose and hemicelluloses were calculated from their macromolecular structure according to Eqs. (4) and (5), respectively. The total hydroxyl content of unmodified lignin was 6.31 mmol/g by 31 P NMR analysis. The hydroxyl content of unmodified bagasse was calculated from Eq. (6), according to the contents of cellulose, hemicelluloses, and lignin in the extractive-free bagasse: where n C (mmol/g) is the theoretical hydroxyl content of unmodified cellulose, 162 g/mol is the molar mass of AGU, and 3 is the number of hydroxyl groups on each AGU, and where n H (mmol/g) is the theoretical hydroxyl content of unmodified hemicelluloses, 132 g/mol is the molar mass of AXU, and 2 is the number of hydroxyl groups on each AXU, and n B ¼ n C · 44:85 þ n H · 33:13 þ n L · 19:24 (6) where n B and n L (mmol/g) are the hydroxyl contents of unmodified bagasse and lignin, respectively, and 44.85%, 33.13%, and 19.14% are the contents of cellulose, hemicelluloses, and lignin, respectively, in the extractive-free bagasse according to the standard NREL methods.

Characterization
FT-IR spectra were collected on an FT-IR spectrophotometer (Nicolet 510) using a KBr disk containing approximately 1% finely ground samples. Thirty-two scans were taken for each sample with a resolution of 2 cm -1 in transmittance mode in the range of 4000-400 cm -1 . 1 H NMR, 13 C NMR, 31 P NMR, and 2D HSQC spectra were recorded on Bruker AV-III HD 600 spectrometer (Germany TG of samples was performed on a simultaneous thermal analyzer (SDT Q600, TA Instrument). The sample weighted between 8 and 10 mg. The scans were first run from room temperature to 100°C, were equilibrated at 100°C for 5 min, and then were cooled to 40°C at the cooling rate of 20°C/min. The scans were again heated from 40°C to 700°C at the heating rate of 10°C/min under nitrogen flow.

Homogeneous phthalation of bagasse and its fractions
The reaction behaviors of bagasse during homogeneous phthalation in AmimCl were investigated comparatively with the isolated fractions under the same conditions. The contents of cellulose, hemicelluloses, and lignin in the extractive-free bagasse were determined as 44.85%, 33.13%, and 19.14%, respectively, according to the standard NREL methods [27].
As shown in Table 1, the DS of phthalated cellulose, hemicelluloses, and lignin were 19.88%, 13.67%, and 27.10%, respectively, indicating the reactivity of lignin was the highest among the three main components during bagasse homogeneous modification in AmimCl. The DS of phthalated bagasse (10.80%) was lower than phthalated fractions, which was probably due to the complicated linkages between bagasse fractions. Besides, the WPG of phthalated cellulose, hemicelluloses, and lignin were 17.04%, 5.98%, and -49.70%, respectively, suggesting that the significant degradation of lignin occurred during homogeneous phthalation. The WPG of phthalated bagasse was -10.51%, probably resulting from the degradation of lignin during bagasse homogeneous modification.

31 P NMR analysis
To investigate the detailed information of lignin hydroxyls during homogeneous modification, 31 P NMR spectroscopy was applied in the present study to quantify the hydroxyl contents of lignin samples [28,29]. The 31 P NMR spectra of unmodified lignin (L o , spectrum a) and phthalated lignin (L s , spectrum b) are shown in Figure 1. According to the reported literatures [30,31], the assignments of signals in 31  The hydroxyl contents of unmodified and phthalated lignin are present in Table 2, and the decreased percentages of lignin hydroxyls are shown in Figure 2. After phthalation, the aliphatic hydroxyl content decreased from 4.71 mmol/g in unmodified lignin to 3.33 mmol/g in phthalated lignin, and the total phenolic hydroxyl content was also reduced from 1.60 (L o ) to 1.27 mmol/g (L s ). These results indicated that both aromatic and phenolic hydroxyls were   phthalated during lignin modification in AmimCl. Besides, the total hydroxyl content of lignin decreased from 6.31 (L o ) to 4.60 mmol/g (L s ), while the carboxyl content slightly increased from 0.08 (L o ) to 0.09 (L s ) mmol/g. The result suggested that the predominant diester form [32] during lignin homogeneous modification in AmimCl. For lignin phenolic hydroxyls, the content of phenolic S-OH decreased from 0.16 to 0.09 mmol/g after phthalation; the content of phenolic G-OH decreased from 0.40 (L o ) to 0.25 (L s ) mmol/g; the decrease in the content of phenolic H-OH from 1.04 (L r ) to 0.93 (L s ) mmol/g was also observed. These results indicated phenolic S-, G-, and H-OH were phthalated during lignin modification in AmimCl.
As shown in Figure 2(a), the decreased percentage of aromatic hydroxyls was higher than that of phenolic hydroxyls, indicating the higher reactivity of aliphatic hydroxyls during lignin homogeneous phthalation. For phenolic hydroxyls, the decreased percentage of phenolic S-OH, G-OH, and H-OH were present (Figure 2b). Therefore, the phthalation reactivity of phenolic hydroxyls followed the order of S-OH > G-OH > H-OH.

FT-IR
The FT-IR spectra of unmodified cellulose (C o , spectrum a), regenerated cellulose (C r , spectrum b), and phthalated cellulose (C s , spectrum c) are shown in Figure 3. The absorbance at 3433,  2895, 1646, 1376, 1165, and 1057 cm -1 were associated with unmodified cellulose [20]. These peaks present in regenerated and phthalated cellulose indicated that the skeleton of cellulose was stable during cellulose dissolution and modification in AmimCl. Compared with the spectrum of unmodified cellulose, the spectrum of the phthalated cellulose (spectrum c) provides evidence of phthalation by showing the presence of three new bands at 1718, 1580, and 749 cm -1 [21], corresponding to the stretching of carbonyl group, antisymmetric stretching of carboxylic anions, and the out-of-plane C-H bending of ortho disubstituted benzene, respectively. In addition, the intensity of peak at 1057 cm -1 for C-O-C stretching in spectrum c slightly decreased indicating the degradation of cellulose macromolecules during dissolution and modification. This degradation was also reported in the previous literature [21].   skeleton of hemicelluloses remained unchanged after dissolution and modification in AmimCl. Compared with unmodified hemicelluloses, the new bands at 1713, 1571, and 752 cm -1 in phthalated hemicelluloses indicated the phthalation of hemicelluloses [33].
FT-IR spectra of unmodified bagasse (B o , spectrum a), regenerated bagasse (B r , spectrum b), and phthalated bagasse (B s , spectrum c) are shown in Figure 6. The bands were assigned according to the previous literature [37]. The peaks at 1605, 1515, 1375, and 897 cm -1 remained predominant in modified bagasse, suggesting the skeleton of bagasse fractions were unchanged during bagasse homogeneous modification in ionic liquid AmimCl [15]. The intensities of the peak at 1725 and 744 cm -1 increased after modification, providing the direct evidence of esterification between phthalic anhydride and bagasse. In conclusion, the phthalation of bagasse and its fractions occurred during homogeneous modification according to FT-IR analyses.

Solid-state 13 C NMR of bagasse
The poor dissolubility of bagasse samples in DMSO-d 6 made high-resolution liquid-state 2D HSQC NMR spectroscopy impossible, and CP/MAS 13 C NMR spectroscopy was applied for bagasse samples analysis, as shown in Figure 7. The signals in CP/MAS 13 C NMR spectra were assigned based on the reported literatures [15,21]. The signals at 191.2, 170.5, and 130.5 ppm correspond to benzene ring, carbonyl group, and carboxylic groups, respectively, in phthalated bagasse. Compared with unmodified bagasse, the intensities of the three signals in phthalated bagasse remarkably increased, indicating the attachment of phthaloyl group onto bagasse.
The signals between 60 and 105 ppm relate to the carbons of carbohydrates. The chemical shifts and intensities of signals in this region remained basically unchanged, indicating the carbon skeleton of bagasse polysaccharides (cellulose and hemicelluloses) was stable during homogeneous modification, which was consistent with the results from FT-IR. Due to the low resolution of CP/MAS 13 C NMR spectroscopy, the detailed information of bagasse fractions during homogeneous phthalation need to be further revealed with 2D HSQC NMR spectroscopy.

Detailed reaction behaviors of cellulose during homogeneous phthalation
The HSQC spectra of unmodified cellulose (spectrum a) and phthalated cellulose (spectrum b) are present in Figure 8. The primary correlations of cellulose were well assigned, as reported in the previous literatures [20,21]. The cross-peaks of cellulose were clearly observed in Figure 8  3.18 [C-C 5 (C 5 /H 5 )], and 103.44/4.33 [C-C 1 (C 1 /H 1 )] ppm; the two C-C 6 (C 6 /H 6 ) peaks were also distinctively located at 60.77/3.79 and 60.77/3.58 ppm, respectively. These results confirmed that the skeleton of cellulose was stable during homogeneous phthalation in AmimCl, consistent with the FT-IR analysis.
The cross-peaks at 131.49/7.62, 131.11/7.52, 129.43/7.78, and 128.94/7.61 ppm were associated with C 10 /H 10 , C 11 /H 11 , C 12 /H 12 , and C 13 /H 13 in the aryl group of phthalated cellulose, respectively. The presence of these peaks in phthalated cellulose confirmed the attachment of phthaloyl group onto cellulose. More importantly, two peaks from substituted C-6 in AGU C-C 6 ' (C 6 /H 6 ) appeared at 64.39/4.39 and 64.50/4.67 ppm, and peaks from substituted C-2 [C-C 2 ' (C 2 /H 2 )] and C-3 [C-C 3 ' (C 3 /H 3 )] in AGU were located at 74.32/4.60 and 75.30/4.83 ppm, respectively. These results suggested the successful phthalation of hydroxyls at C-6, C-2, and C-3 positions in AGU. The DS of hydroxyls on different positions could be evaluated upon the integral area of the characteristic substituted correlations. By integration, the DS of C 6 -OH, C 2 -OH, and C 3 -OH in AGU were 12.05, 2.54, and 1.84%, respectively. Therefore, the DS of hydroxyls in AGU followed the order C-6 > C-2 > C-3. The phthalated cellulose was previously prepared under the similar experimental conditions without catalyst in our laboratory, and the DS of hydroxyls at C-6, C-2, and C-3 positions were 6.30, 2.01, and 0% in AGU, respectively. By contrast, it could be concluded that the addition of catalyst DMAP could improve the uniformity and DS of phthalated cellulose.

Detailed reaction behaviors of hemicelluloses during homogeneous phthalation
The 2D HSQC spectra of unmodified hemicelluloses (spectrum a) and phthalated hemicelluloses (spectrum b) are present in Figure 9. According to the previous literatures [38,39],

Detailed reaction behaviors of lignin during homogeneous phthalation
The 2D HSQC NMR spectra of unmodified lignin (spectrum a) and phthalated lignin (spectrum b) are illustrated in Figure 10, and the structure of identified lignin subunits with   hydroxyls are shown in Figure 11, including aryl ether (β-O-4', A), phenylcoumaran (β-5', B), guaiacyl unit (G-unit, C), syringyl unit (S-unit, S), and p-hydrophenyl unit (H-unit, H). According to the previous publications [24,25], the correlations of these signals were well assigned. The cross-peaks from S-units C 2,6 /H 2,6 was at 104.68/6.72 ppm, and the signals for C α -oxidized S-units (S') appeared at δ C /δ H 107. 15 The relative quantities of the primary substructures were calculated according to the previous literatures [40,41]. The aromatic lignin S-, G-, and H-units were expressed as a fraction of 100%, and the relative molar quantities of aryl ether, phenylcoumaran were expressed as a percentage of the total aromatic S-, G-, and H-units. The detailed information of quantitative lignin substructures calculated from HSQC spectra are listed in Table 3.
After phthalation, the content of aryl ether increased from 57.80/100Ar (L o ) to 58.33/100Ar (L s ), and the content of phenylcoumaran also increased from 2.76/100Ar (L o ) to 3.13/100Ar (L s ). These increases were probably due to the decreased intensity of C 2,6 /H 2,6 correlations from aromatic units resulting from the condensation of lignin [41] during the homogeneous phthalation. The relative percentages of G-units decreased from 45.78% to 37.50%, while the increases in the relative percentage of S-and H-units were observed. This result suggested that the G-units were easily degraded during lignin homogeneous modification [42,43]. Besides, the increase in the S/G ratio from 1.08 (L o ) to 1.54 (L s ) further confirmed the degradation of Gunits during lignin homogeneous modification.

Thermal analysis
In the present study, the effect of dissolution, regeneration, and modification in AmimCl on the thermal stability of samples was investigated by comparing unmodified, regenerated, and phthalated samples. Figure 12 illustrates the TG curves of cellulose (A), hemicelluloses (B), lignin (C), and bagasse (D), respectively.
The onset and midpoint degradation temperature of regenerated cellulose were higher than those of unmodified cellulose, indicating the improved thermal stability after dissolution and regeneration in IL. This was probably because cellulose fractions with low molecular weight could not be regenerated from IL [15]. As shown in Table 4, the onset and midpoint degradation temperature of phthalated cellulose were lower than those of regenerated cellulose, suggesting the thermal stability of phthalated cellulose decreased after homogeneous modification. The decreased thermal stability of phthalated cellulose was probably due to partial hydrolysis, degradation of cellulose, and decrystallization of crystalline cellulose macromolecules [21].
Similar to cellulose, the thermal stability of regenerated hemicelluloses were slightly higher than that of unmodified hemicelluloses, probably resulting from the loss of hemicellulosic fractions with low molecular weights during dissolution and regeneration in AmimCl. Compared with regenerated hemicelluloses, the onset and midpoint degradation temperature of phthalated hemicelluloses decreased to 212 and 265°C, respectively, indicating the decreased thermal stability of phthalated hemicelluloses after homogeneous modification. This decrease was also possibly resulted from the degradation of hemicelluloses during homogeneous modification [44].    The onset and midpoint degradation temperature of regenerated lignin were both lower than those of unmodified lignin, as listed in Table 4, indicating the decreased thermal stability of lignin after dissolution and regenerating. Compared with regenerated lignin, the thermal stability of phthalated lignin after homogeneous modification was also decreased. This decrease was probably resulted from the degradation of lignin during dissolution, regeneration, and modification in AmimCl.
For bagasse, the thermal stability of bagasse samples decreased after dissolution and regeneration in ionic liquid AmimCl. This was probably resulted from the degradation of lignin during dissolution and regeneration. Compared with regenerated bagasse, the onset degradation temperature of phthalated bagasse decreased, while the midpoint degradation temperature of phthalated bagasse increased. These results suggested the thermal stability of lignocellulose was affected by the loss of fractions with low molecular weights and the degradation of phthalated samples. The former led to the increase in the thermal stability, while the later resulted in the decreased thermal stability of phthalated lignocellulose.

Conclusions
The phthalation of bagasse and its fractions occurred during homogeneous modification in AmimCl, and the reactivity of the three main components followed the order: lignin > cellulose > hemicelluloses. The FT-IR and NMR analyses suggested the skeleton of bagasse remained stable during dissolution, regeneration, and modification; the reactivity of hydroxyls in AGU followed the order: C-6 > C-2 > C-3; and the reactivity order in AXU was as follows: C-2 > C-3. For lignin, both aromatic and phenolic hydroxyls were phthalated; the reactivity of aromatic hydroxyls was higher than phenolic hydroxyls; and the reactivity order of lignin phenolic hydroxyls was as follows: S-OH > G-OH > H-OH. The thermal stability of modified lignocellulose was affected by the disruption of cellulose crystallinity and the degradation reaction of the modified components.