Synthesis of 5‐hydroxymethylfurfural from glucose over carbon‐based acid–base bifunctional catalyst

Carbon‐based acid–base bifunctional catalyst was synthesized by the functionalization of activated carbon (AC) with 3‐mercaptopropyl trimethoxysilane, 3‐aminopropyl trimethoxysilane, and H2SO4 via sulfhydrylation, ammonifiation, and sulfonation successively. The as‐prepared catalyst of NH2‐AC‐SO3H was characterized by infrared spectroscopy, X‐ray diffraction, thermogravimetric analysis and Brunauer–Emmet–Teller specific area analysis, and it was further used for producing fuel precursor 5‐hydroxymethylfurfural (5‐HMF) from glucose. The influences of the catalyst type, ratio of acid–base groups, catalyst amount, reaction temperature, and reaction time were investigated. It was found that NH2‐AC‐SO3H revealed excellent activity, and the yield of 5‐HMF reached 71.8% under the optimized reaction conditions with a 1.3:1 mass ratio of NH2‐AC‐SO3H to glucose at 140°C for 6 h in γ‐valerolactone. The reusability of NH2‐AC‐SO3H was also performed, and only slight decreases in the conversion of glucose and the yield of 5‐HMF were observed, still giving 65.3% yield of 5‐HMF after 4 runs.


| INTRODUCTION
With the increase in the consumption of fossil energy and high attention to environmental problems, the utilization of lignocellulosic biomass is believed to be an effective way to reduce energy dependence on fossil. 1 The conversion of lignocellulosic biomass into platform compounds, especially 5-hydroxymethylfurfural (5-HMF) has attracted more and more concerns since it is an important intermediate for producing a series of high-value-added products and high-quality fuels via selective oxidation, dehydration, hydroxyl aldol condensation, and hydrogenation reactions. 2 5-HMF has been widely considered an alternative to petroleum, 3,4 and expected to play a key role in chemical, fuel, medical, food fields, and so forth.
Lignocellulosic biomass is mainly composed of cellulose, hemicellulose and lignin. Among them, cellulose is a carbohydrate polymer composed of D-glucose units linked by β-1,4-glycosidic bond. It can be further converted into glucose and fructose, which are widely employed to produce 5-HMF, 5 in which fructose has better reactivity and product selectivity than those of glucose. However, fructose is obtained by the isomerization of glucose in the presence of a base catalyst. Therefore, the production of 5-HMF from glucose is more competitive. 6 Unfortunately, poor conversion is often observed due to its stable sixmembered ring structure.
Homogeneous catalysts, including ionic liquids, 7 inorganic acids, 8 and metal chlorides, 9 are widely explored for the production of 5-HMF from glucose. The Sn-modified ionic liquid polymer was prepared by one-pot polymerization and further used as the catalyst for 5-HMF production, giving 99% conversion of glucose and 51.1% yield of 5-HMF, respectively. 7 HMF was synthesized from glucose using a combination of AlCl 3 and HCl as the homogeneous catalyst in a biphasic slug flow capillary microreactor. HMF yield of 53% was obtained and it could be further increased to 66.2% by adding 20 wt% NaCl in the aqueous phase. A systematic approach combined theory and experiment was also proposed, in which Brønsted acid and Lewis acid were used as cocatalyst, and high 5-HMF yield around 68% was observed in the presence of HCl. 9 High yield of 5-HMF was obtained from different biomass materials including glucose, starch, and food waste using a small amount of SnCl 4 catalyst in a cheap and green natural deep eutectic solvent (NADES) system, and the yield of 5-HMF derived from glucose was 64.3% in the NADES/ methyl isobutyl ketone biphasic system. 10 Although homogeneous catalysts exhibit high catalytic performance, common problems such as corrosion and pollution are often encountered. Solid acid, 5 solid base, 11 molecular sieve, 12 heteropoly acid, and its salt 13 have been used as heterogeneous catalysts for the conversion of glucose to 5-HMF based on the advantages of easy separation and recyclability. 14 For example, a mixed oxide of Al 2 O 3 -TiO 2 -ZrO 2 was evaluated for the conversion of glucose to 5-HMF, giving the maximum yield of 5-HMF at 63% in a biphasic system composed of tetrahydrofuran/H 2 O. 15 Recently, metal-organic frameworks (MOFs) are notable because of their high crystallinity, large specific surface area, and high porosity. The yield of 5-HMF was 29% in the presence of sulfo-modified MIL-101Cr compound in a tetrahydrofuran/water system, 16 and the yield of 5-HMF was 37% when using UiO-66-SO 3 H as the catalyst. 2 Graphitic carbon nitride supported UiO-66-type MOFs were synthesized via one-pot modulated hydrothermal method and employed for the production of 5-HMF from glucose, giving a yield of 54.9% in an isopropanol-mediated dimethylsulfoxide system. Moreover, the catalyst also revealed superior recyclability, still with 46.7% 5-HMF yield after five runs. 17 It is believed that the isomerization of glucose to fructose and the conversion of fructose to 5-HMF is promoted by basic groups 18 and acidic groups, 2 respectively. Therefore, heterogeneous acid-base bifunctional catalyst has been developed to produce 5-HMF from glucose. Acid-base bifunctional UiO-66-type MOFs catalyst was prepared by supported on graphene oxide modified by polydopamine, which displayed excellent performance for the conversion of glucose to 5-HMF with 55.8% yield. 19 Nitrogen-doped carbonaceous catalyst promoted the efficient conversion of glucose with a high yield of HMF at 62.8% due to the synergistic effect of acid-base bifunctional active sites. Moreover, the catalytic activity only decreased slightly during the recycle, and the yield of 5-HMF still reached 58.3% after 4 runs. 20 A sulfonated oxidized activated carbon (SO 3 H-OAC) catalyst with strong Brønsted base and acid sites was prepared and displayed excellent performance for the conversion of glucose to 5-HMF, achieving around 93% HMF yield. However, a biphasic THF/H 2 O-NaCl was required and about 20% activation loss was observed after 5 cycles. 5 These results indicate that the acid-base bifunctional heterogeneous catalysts have good catalytic performance and stability for the conversion of glucose to 5-HMF. However, tedious preparation process and/or damage in the structure are often encountered. For example, the structural frames of MOFs can be easily twisted and deformed due to the low thermal stability, thus leading to collapses and damages in the crystal structures. 21 To reduce the formation of by-products, research efforts have also been made on the reaction solvent. Polar aprotic solvents including water, dimethyl sulfoxide, methyl isobutyl ketone, tetrahydrofuran, valerolactone (GVL), and ionic liquids have attracted a great deal of attention. Among them, GVL is considered to be an ideal solvent due to its environmental friendliness and excellent performance in promoting HMF production. 22 For example, direct dehydration of glucose to 5-HMF was found to be the primary reaction during glucose decomposition, the selectivities of dehydration reactions to 5-HMF were 22%-30% at 175°C depending on GVL concentration. 23 Inorganic acid HCl catalyzed the conversion of glucose to 5-HMF in γ-GVL/H 2 O with an impressively high yield of 5-HMF when NaCl was used as the promoter, giving 62.45% 5-HMF yield. 8 In this work, carbon-based acid-base bifunctional catalyst was prepared by sulfhydration, ammoniation, and sulfonation reactions successively using cheap and available activated carbon (AC) as the support (Scheme 1). The prepared catalyst of NH 2 -AC-SO 3 H was characterized in detail and further employed for producing 5-HMF from glucose. The conversion of glucose was optimized by varying catalyst type, ratios of acid-base groups, catalyst amount, reaction temperature, and reaction time. Moreover, NH 2 -AC-SO 3 H was recovered and the cyclic catalytic performance was investigated.

| Preparation of hybrid acid-base bifunctional catalyst 24
In a 100-mL flask, 2 g AC was added into 30 mL anhydrous toluene and ultrasonicated for 30 min followed by stirring at room temperature for 1 h. Then 2 mL deionized water, 2 mL ethanol, 0.4 mL formic acid, and 15 mL MPTMS were added, and the mixture was heated at 50°C for 4 h under magnetic stirring. After the reaction, the reaction mixture was cooled to room temperature and centrifuged. The solid was collected and dried, extracted by toluene and dried again, giving sulfhydrylated AC (AC-SH).
Typically, 0.3 g AC-SH was dissolved in 30 mL deionized water and 1.35 mL APTES was added to 150 mL ethanol, respectively, and the mixture was then stirred magnetically at 25°C for 24 h. After the reaction, the mixture was centrifuged, and the collected solid was washed by 70 mL ethanol and deionized water, followed by drying, thus giving aminated AC-SH (NH 2 -AC-SH).
NH 2 -AC-SH was dispersed in 40 mL 50 wt% H 2 SO 4 solution, and the mixture was stirred magnetically at 25°C for 4 h. The reaction mixture was then centrifuged, and the collected solid was washed by deionized water and ethanol three times separately, followed by drying, giving aminated and sulfonated AC-SH (NH 2 -AC-SO 3 H). The amount of acidic and basic groups in NH 2 -AC-SO 3 H was determined according to the reported procedure. 20

| Catalytic conversion of glucose into 5-HMF
Typically, 0.1 g NH 2 -AC-SO 3 H, 0.13 g glucose, and 5 mL GVL were added into a 50-mL-stainless steel autoclave, and the air in the autoclave was replaced with 1 MPa N 2 for three times. Then 1 MPa N 2 was charged and the reaction mixture was heated at 140°C for 4 h under magnetic stirring. After the reaction, the autoclave was cooled to room temperature and the pressure in it was gradually released. The reaction mixture was centrifuged and the liquid was analyzed quantitatively by highperformance liquid chromatography (HPLC). The solid catalyst was washed by deionized water and ethanol, S C H E M E 1 Preparation of heterogeneous catalyst of NH 2 -AC-SO 3 H. AC, activated carbon. followed by drying before reuse without further pretreatment.

| Analysis
The quantitative analysis of the reaction mixture was performed on an LC-100 PLUS HPLC equipped with a reversed-phase Sho-dex-C18-100-5 4E column (4.6 mm × 250 mm × 5 μm) and a UV detector. The mixture of methanol:aqueous acetic acid solution (10:90, v-v) was employed as the mobile phase at the flow rate of 1.0 mL/ min with column temperature at 40°C and wavelength at 280 nm. All experiments were repeated thrice and the reported yields were averaged data with a standard deviation of 0.3%-3.4%.

| Calculation
The yields of 5-HMF and levulinic acid (LA) were calculated as follows: where m 1 is the mass of 5-HMF or LA produced (g), m 2 the mass of glucose added (g), M 1 the molecular weight of 5-HMF or LA, 180 or 116 g/mol, and M 2 the molecular weight of glucose, 180 g/mol.

| Characterization
Fourier infrared spectroscopy (IR) was measured on an NEXUS670 spectrometer in the range from 4000 to 500 cm −1 . The solid samples were ground with dried KBr powder, and compressed into a disc before analysis. X-ray diffraction (XRD) measurement was performed on a D8 Advance X-ray diffractometer with Cu target, Kα ray, and λ = 0.15406 nm at a scanning rate of 5°/min in the range from 10°to 80°. Thermogravimetric analysis (TGA) was performed by using a Perkin-ElmerTG-DSC7 spectrometer under N 2 atmosphere at a heating rate of 10°C/min in the range from room temperature to 800°C. Approximately 10 mg sample was used in each analysis and the gas flow rate was kept at 90 mL/min. Brunauer-Emmet-Teller (BET) specific area analysis was determined by N 2 adsorption-desorption isotherm at 77 K using one-point modified BET method on a Micromeritics ASAP 2460 analyzer after degassing under 10 −4 kPa at 100°C for 6 h. A back titration method was used to measure the amount of acid-base centers of the materials according to the reported procedure. 19 Typically, 0.1 g sample and 10 mL 0.01 M HCl (NaOH) solution was added successively in a conical beaker. The mixture was then stirred at room temperature for 30 min, filtered and rinsed with 25 mL distilled water for four times. The resulting filtrate was titrated with 0.01 M NaOH (HCl) solution using phenolphthalein as an indicator. group of AC reacted with MPTMS. The bands belonged to the vibrations of C─S and S─O also presented in the range from 690 to 810 cm −1 , 28 further confirming the reaction of MPTMS with AC. However, the vibration of S─H was not observed in the range from 2500 to 2600 cm −1 , probably due to its extremely weak absorption. 29 It can be seen from the IR spectrum of NH 2 -AC-SH (curve c), new band at 1304 cm −1 is caused by the stretching vibration of C─N. 30 The observation of these bands confirmed the reaction of APTES with AC-SH. In the IR spectrum of NH 2 -AC-SO 3 H (curve d), the stretching vibrations of ─SO 3 H, O═S═O, S─O, and C─S were observed at 1241, 1030, 806, and 620 cm −1 , 5 respectively, indicating that NH 2 -AC-SH was successfully sulfonated. In the IR spectrum of the recovered NH 2 -AC-SO 3 H (curve e), the band at 620 cm −1 almost disappeared, and the bands around 1241 cm −1 and those in the range from 690 to 810 cm −1 became weaker compared with those of the fresh catalyst ( Figure 1D), possible resulting from slight shedding of ─NH 2 and ─SO 3 H during the reaction process. 28

| XRD
The XRD patterns of AC, AC-SH, NH 2 -AC-SH, NH 2 -AC-SO 3 H, and its recovered sample (after 1 run) are presented in Figure 2. The peak around 21.5°in the patterns of all samples is attributed to the amorphous structure of AC (curves a-d). 31 In addition, a diffraction peak at 43.5°was also observed in all patterns, indicating a rather random orientation of the amorphous carbon composed of aromatic carbon flakes, and the carbon skeletons of AC of all samples are amorphous. 32,33 The sample derived from AC almost had the similar XRD patterns to that of AC (curves b-d), suggesting the microstructure of AC kept well during the introduction of acid-base groups. 34 Compared with the fresh catalyst NH 2 -AC-SO 3 H (curve d), almost no change could be observed in the XRD pattern of the recovered sample (curve e). These results lead us to the conclusion that the reaction process had little influence on the crystalline structure of the catalyst.

| TGA
The TGA curves of AC, AC-SH, NH 2 -AC-SH, NH 2 -AC-SO 3 H, and its recovered sample are shown in Figure 3. Although two weight loss stages were observed in the TGA curve of AC (curve a), only giving a total weight loss of 0.4 wt%. The loss below 120°C is associated with the removal of physically adsorbed molecular water, 35 and the slight weight loss observed at 200°C is attributed to the decomposition of oxygen-containing functional groups on the surface of AC. It also can be seen from Figure 3 that the TGA curves of the samples derived from AC (curves b-d) and the recovered NH 2 -AC-SO 3 H (curve e) almost had the same trend in the weight loss, which were significantly enhanced above 120°C compared with that of AC. In the range from 120°C to 150°C, relatively slow weight losses were assigned to the prepyrolysis of the samples. Obvious weight losses were then observed in the region from 200°C to 600°C, which must be  associated with the thermal decomposition of organic groups, such as ─SH, ─NH 2 , and ─SO 3 H. Hereafter, only a small amount of weight loss was observed above 600°C.
The derivative thermogravimetry (DTG) curves of AC, AC-SH, NH 2 -AC-SH, NH 2 -AC-SO 3 H, and its recovered sample (after 1 run) are shown in Figure 4. The maximum decomposition temperature of AC (curve a) is presented at 325°C, corresponding to the decomposition of the carbon structure. 36 The breakage of various bridge bonds with low bond energy occurred first, then oxygencontaining functional groups such as straight chains of ─OH and side chains of ─COOH were broken, giving volatile small molecules, 36 corresponding to the weight loss in the TGA curve above 200°C. Figure 4 revealed that four maximum thermal degradation peaks were observed in the DTG curves of the samples derived from AC (curves b-d). The thermal degradation peaks around 79°C are attributed to the removal of physical absorption water, 37 and the thermal degradation peaks in the range from 200°C to 420°C are mainly caused by the thermal decomposition of organic groups of ─SH, ─NH 2 , and ─SO 3 H. 38 The degradation peaks around 487°C are ascribed to the thermal decomposition of organic groups to small molecular substances, including C, CO, SO 2 , and H 2 O. 39 The results in Figures 3 and 4 also showed that the recovered NH 2 -AC-SO 3 H had similar TGA and DTG curves to those of the fresh catalyst, suggesting excellent thermal stability of carbon-based bifunctional catalyst, which would possibly bring excellent reusability.

| BET
The BET-specific surface areas of AC, AC-SH, NH 2 -AC-SH, NH 2 -AC-SO 3 H, and its recovered sample (after 1 run) are shown in Table 1. The surface area of AC reached up to 1357.5 m 2 /g (entry 1) while it was reduced to 428.5 m 2 /g after sulfhydration (entry 2). It can be ascribed to both amorphous silica and ─SH were introduced by sulfhydration. The former had a smaller specific surface area 40 while the latter possibly entered the pore channel of AC. Both of them have negative impacts on the specific surface area, thus resulting in a lower specific area of AC-SH. The results in Table 2 revealed that specific area of the samples was further reduced by ammoniation and sulfhydration, the specific surface area of NH 2 -AC-SO 3 H decreased to 163.9 m 2 /g, indicating that a small number of pores was occupied because of the introduction of ─SO 3 H and ─NH 2 . 5,41 Compared with the fresh catalyst, the specific surface area of the recovered catalyst decreased significantly after 1 run, only 82 m 2 /g was observed. The IR results in Figure 1 indicated the slight shedding of ─SO 3 H and ─NH 2 occurred during the reaction. Due to the entrance of the detached ─SO 3 H and ─NH 2 , the specific surface area of the recovered catalyst decreased.

| Screening of catalyst
The samples derived from AC were employed for the conversion of glucose into 5-HMF, as shown in Table 2. Almost no formation of 5-HMF in the presence of AC-SH or NH 2 -AC-SH. When AC-SO 3 H was used as the catalyst, FF was also generated besides 5-HMF, and the yields of 5-HMF and FF were 15.0% and 6.3%, respectively. FF was derived from glucose via isomerization, dehydration, and dealkylation successively, as shown in Scheme 2. 42 The yield of FF was obviously lower than that of 5-HMF due to higher activation energy required for the selective dealkylation of glucose. 6 All these results lead us to conclude that both acid group and base group play roles in the conversion of glucose  to 5-HMF, in which the isomerization of glucose to fructose followed by the dehydration of fructose to 5-HMF are promoted by basic group and acid group, respectively. 18 Therefore, NH 2 -AC-SO 3 H with a 1:2.5 molar ratio of ─SO 3 H to ─NH 2 was further employed. The results in Table 2 showed that the yield of 5-HMF was significantly increased to 43.2% but a poorer yield of FF at 4.4% was obtained. Therefore, it is reasonable to conclude that there is a synergistic effect between the acid and base groups, which is beneficial to improve the activity of the catalyst and the selectivity of the target product. 6 The decrease in FF yield may be caused by increasing the amount of the water generated during the reaction system, which promotes the degradation of FF, especially the condensation of FF with sugar intermediates. 43 Notably, almost no LA was generated, possibly due to the excellent performance of GVL in promoting 5-HMF production. 22

| Influence of acid-base group ratio
The effect of the acid-base group ratios (n:n) in NH 2 -AC-SO 3 H was investigated, and the results are shown in Figure 5. The yields of 5-HMF changed significantly by varying acid-base group ratios. However, the yield of FF only varied slightly in the range from 3.4% to 6.4%. When the molar ratio of ─SO 3 H to ─NH 2 was 1:1, giving yields of 5-HMF and FF at 18.3% and 3.6%, respectively. As increasing the content of ─NH 2 , both the yields of 5-HMF and FF were improved while the former was more obvious, and the yield of 5-HMF was increased to 43.2% with a 1:2.5 molar ratio of ─SO 3 H to ─NH 2 . The increase in 5-HMF yield can be attributed to the increase in the basic groups, which promotes the isomerization of glucose to fructose. 18 Thereafter, a drop in the yield of 5-HMF was observed inversely by further increasing the content of ─NH 2 . The yield of 5-HMF decreased to 26.8% when a 1:3 molar ratio of ─SO 3 H to ─NH 2 was employed. Moreover, the formation of LA was also detected, giving around a 1.6% yield. It has been reported that four types of reactions of dehydration, condensation, decomposition, and isomerization are involved in the conversion of glucose to 5-HMF. 12 Therefore, the formation of 5-HMF is usually accompanied by the formation of by-products, such as soluble polymers, insoluble black substances, FF, and LA. 44 The basic groups may not only promote glucose isomerization, but also favor the side reactions of 5-HMF, thus leading to the production of humin, LA, and levulinic acid esters, C 9 and C 15 alkanes via the polymerization, hydrolysis, and cross-condensation (Scheme 3). 45,46 As a result, a decrease in the yield of 5-HMF

| Investigation on catalyst amount
The effect of catalyst amount on the conversion of glucose to 5-HMF was investigated and the results are shown in Figure 6. As the amount of catalyst increased, the yield of 5-HMF first increased and then decreased while the FF yield varied slightly in the range from 4.4% to 6.2%. When a 1:1.3 mass ratio of glucose to NH 2 -AC-SO 3 H was used, the highest yield of 5-HMF at 43.2% was acquired, which must be caused by an increase in the catalytic active sites with increasing the catalyst amount. However, a negative effect was brought to the catalytic performance when the catalyst amount was further increased, and the yield of 5-HMF significantly dropped to 5.9% in the presence of a 1:1.5 mass ratio of glucose to catalyst. It could be ascribed to the limitation in the mass transfer between the excessive catalyst and substrate. In addition, the polymerization of 5-HMF was also enhanced by the excessive catalyst, generating by-product, such as humin, as shown in Scheme 3.

| Optimization of reaction temperature and reaction time
The effects of reaction temperature and reaction time on the conversion of glucose to 5-HMF are shown in Figures 7  and 8, respectively. As can be seen from Figure 7, POOR yields of 5-HMF were observed when lower reaction temperatures were employed. When the reaction temperature was raised from 120°C to 140°C, the yield of 5-HMF was improved from 17.8% to 43.2%. The significant improvement in the yield can be ascribed to difficulty in activate reactants at lower reaction temperature. Moreover, the isomerization and the hydrolysis processes involved in the conversion of glucose to 5-HMF are conducive to occur at higher temperature. 46 Thereafter, the product yield decreased contrarily when the reaction temperature was further raised, and the yield of 5-HMF was reduced to 18.6% at 160°C, possibly due to the promotion of side reactions at higher temperature. Additionally, it was observed that the color of the reaction solution changed as varying reaction temperature. A light yellow color of the reaction mixture was observed at lower temperature, indicating low content of oligomers resulted from side reactions. As the reaction temperature was raised, the color of the reaction mixture became dark brown, further confirming the promotion of byproducts, such as insoluble humic substance and soluble polymer. 44 When the reaction temperature was increased from 120°C to 160°C, FF yield was slightly enhanced from 3.8% to 8.5%, indicating that FF has better stability than that of 5-HMF at higher temperature. 47 The results in Figure 8 showed that the yield of 5-HMF was significantly increased from 23.9% to 71.8% when the reaction time was prolonged from 3 to 6 h, and the yield of FF was also increased from 5.3% to 9.0%. Thereafter, the yield of 5-HMF conversely decreased by further extending the reaction time. When the reaction time was extended to 7 h, the yield of 5-HMF decreased to 63.2% while a 0.8% yield of LA was observed. It can be attributed to the chemical instability of 5-HMF, which makes the side reactions such as condensation and hydration easy to occur with longer reaction time, leading to the formation of by-products, such as LA and levulinic acid ester. 45,48 Moreover, both the hydrolytic deactivation of the acid groups and the side reactions were accelerated in the presence of base groups when a longer reaction time was employed. 46 The results in Figure 8 also showed that the yield of FF decreased to 7.6% when the reaction time was prolonged to 7 h, suggesting that the production of 5-HMF and FF was inhibited by competitive side reactions. 49

| Recyclable performance of NH 2 -AC-SO 3 H
The recyclable performance of the acid-base bifunctional catalyst NH 2 -AC-SO 3 H is shown in Figure 9. The catalytic activity of NH 2 -AC-SO 3 H slightly decreased after 1 run, the yields of 5-HMF and FF decreased from 71.8% to 65.7% and 9.0% to 8.1%, respectively. The losses of ─SO 3 H and ─NH 2 groups during the reaction must be responsible for the change, 50 resulting in a drop in the active sites and catalytic performance, in agreement with the results of IR. Surprisingly, almost no loss in the yields of 5-HMF was observed for further reuse, 65.3% yield was still obtained after 4 runs. These results suggested that the leaching of ─SO 3 H and ─NH 2 groups mainly took place at the first run.
The amounts of ─SO 3 H and ─NH 2 in the fresh NH 2 -AC-SO 3 H and the recovered sample were also determined. The results in Table 3 revealed that the amounts of acid group and base group decreased in the recovered samples. The amounts of ─SO 3 H and ─NH 2 were reduced from 1.294 to 0.941 and 0.278 to 0.269 mmol/g after 1 run, respectively. Obviously, the acid group is more easier to leach during the reaction. It could be explained by the preparation process of NH 2 -AC-SO 3 H, in which the ammonifiation followed by the sulfonation. Thus it is reasonable to speculate that the acid group may be more distributed on the surface of the catalyst. Therefore, the amino group was more likely to enter the pore channel of the catalyst, leading to decrease in BET-specific surface, as shown in Table 2, while the acid group tended to enter the reaction mixture during the reaction. The results in Table 3 also show that the content of active groups decreased little after 4 runs compared with that of the recovered sample after 1 run, 0.936 and 0.265 mmol/g were still detected, further confirming the leaching of the active groups during mainly took place at the first run. These results suggest that base catalyzed the isomerization of glucose to fructose plays a crucial role in the bifunctional-catalyzed reaction system, which is responsible for the excellent reusability of NH 2 -AC-SO 3 H.

| Comparison with the conversion of glucose to 5-HMF
Since the reusable catalyst of NH 2 -AC-SO 3 H was employed in this work, the catalytic performance of glucose to 5-HMF in the presence of various heterogeneous catalysts was compared, and the results are shown in Table 4. The yields of 5-HMF in the range from 52.0% to 64.0% have been reported. Among them, a higher yield of 5-HMF at 64.0% was obtained when using NbW-SBA-15 as the catalyst. However, it was still much lower than that obtained in this work, in which a 71.8% yield of 5-HMF was observed. Thus it is reasonable to conclude that NH 2 -AC-SO 3 H had the advantages and prospects in application due to its high catalytic performance as well as acceptable stability.

| CONCLUSIONS
The acid-base bifunctional catalyst of NH 2 -AC-SO 3 H exhibited excellent catalytic performance for the conversion of glucose to 5-HMF. The catalyst type, the molar ratio of acid-base groups, catalyst amount, reaction temperature, and reaction time had significant effects on glucose conversion, and the yield of 5-HMF reached 71.8% under the optimized reaction conditions. After the reaction, NH 2 -AC-SO 3 H could be recovered by simple filtration and reused without further pretreatment. The yield of 5-HMF slightly decreased from 71.8% to 65.3% after 4 runs. The slight decrease in the catalytic activity was caused by the leaching of partial ─SO 3 H and ─NH 2 groups during the reaction process, which mainly took place at the first run. Additionally, the loss of ─SO 3 H was obviously higher that of ─NH 2 . The excellent stability of NH 2 -AC-SO 3 H is possibly ascribed to the slight leaching of ─NH 2 during the reaction.

ACKNOWLEDGMENTS
This work was supported by the key project of Hubei province (2022BCA081) and Hubei science and technology innovation project (2021BLB229).