Active, Selective, and Recyclable Zr(SO 4 ) 2 / SiO 2 and Zr(SO 4 ) 2 / Activated Carbon Solid Acid Catalysts for Esteriﬁcation of Malic Acid to Dimethyl Malate

: The esteriﬁcation of malic acid using traditional homogenous catalysts su ﬀ ers from the di ﬃ culty in reuse of the catalyst and undesirable side reactions. In this work, Zr(SO 4 ) 2 / SiO 2 and Zr(SO 4 ) 2 / activated carbon (AC) as solid acid catalysts were prepared for malic acid esteriﬁcation with methanol. The conversion of malic acid over these two catalysts is comparable to that over H 2 SO 4 and unsupported Zr(SO 4 ) 2 · 4H 2 O catalysts; however; a 99% selectivity of dimethyl malate can be realized on these two supported catalysts, which is much higher than that of conventional H 2 SO 4 (75%) and unsupported Zr(SO 4 ) 2 · 4H 2 O (80%) catalysts, highlighting the critical role of AC and SiO 2 supports in tuning the selectivity. We suggest that the surface hydroxyls of AC or lattice O 2 − ions from SiO 2 donate electrons to Zr 4 + in Zr(SO 4 ) 2 / AC and Zr(SO 4 ) 2 / SiO 2 catalysts, which results in the increase in electron density on Zr 4 + . The enhanced electron density on Zr 4 + reduces the degree of H delocalization from crystal water and then decreases the Brønsted acid strength. Consequently, the reduced Brønsted acid strength of Zr(SO 4 ) 2 / AC and Zr(SO 4 ) 2 / SiO 2 catalysts suppresses the intermolecular dehydration side reaction. In addition, these two supported catalysts can be easily separated from the reaction system by simple ﬁltration with almost no loss of activity.


Introduction
Production of chemicals from biomass resources is a promising strategy for solving the problem of the depletion of fossil fuels. Malic acid as a biomass-derived molecule can be produced via the bio-fermentation of biomass. Converting renewable biomass-derived malic acid to various malic acid esters has attracted considerable attention because they are widely used in the production of spices, mosquito-repellent incenses, and intermediates of medicines [1][2][3]. Generally, the esterification of malic acid uses a traditional homogenous acid catalyst (such as sulfuric acid, para-toluene sulfonic acid, and phosphoric acid) to promote the reaction rate in industry. Although homogenous (liquid) acids show excellent catalytic activity for malic acid esterification, unfortunately, they present two main drawbacks: (1) These liquid catalysts are difficult to separate from the reaction mediums, and inevitably suffer from the problem of non-recyclability and environmental unfriendliness; (2) excess high-acid strength will cause undesirable side reactions, among which mainly include intramolecular dehydration and intermolecular dehydration of alcohols. In comparison, heterogeneous (solid) acid catalysts can overcome the un-recyclability of liquid acid and are, accordingly, strongly recommended Zr(SO4)2 loading and textural properties of SiO2, Zr(SO4)2/SiO2, AC, and Zr(SO4)2/AC samples are summarized in Table 1. The Zr amounts in Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts were detected by inductive coupled plasma optical emission spectroscopy (ICP-OES) analysis, and the active species Zr(SO4)2 loadings on these two catalysts were then calculated according to the Zr amounts. The actual Zr(SO4)2 loadings on Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts were ca. 30% (Table 1). The Brunauer-Emmett-Teller (BET) surface area, pore volume, and pore size were determined according to N2 physical adsorption isotherms ( Figure S1). The surface areas of Zr(SO4)2/SiO2 (139 m 2 /g) and Zr(SO4)2/AC (468 m 2 /g) catalysts are drastically smaller than those of SiO2 (214 m 2 /g) and AC (1053 m 2 /g) supports, respectively, but which are still higher than that of bulk Zr(SO4)2•4H2O (5-11 m 2 /g) [28]. Higher surface area is favorable for dispersing active sites and tends to give higher catalytic activity. On the other hand, the mesopore volume and size of Zr(SO4)2/SiO2 are obviously lower than those of the SiO2 support, which indicates that Zr(SO4)2•4H2O deposits on the mesopores of SiO2. For the Zr(SO4)2/AC catalyst, the micropore volume and size of Zr(SO4)2/AC are slightly lower than those of AC; however, the mesopore volume and size of Zr(SO4)2/AC are obviously lower than those of AC. These results manifest that Zr(SO4)2•4H2O prefers to deposit on the mesopores than the micropores of AC. As we know, the presence of mesopores in a catalyst is significantly important for the liquid phase catalytic reaction, because mesopores facilitate the mass transfer in the liquid phase, which is beneficial to the contact between the catalyst and reactant.  Figure 2 shows the catalytic performance of H2SO4, Zr(SO4)2•4H2O, Zr(SO4)2/SiO2, and Zr(SO4)2/AC catalysts for conversion of malic acid. The conversion of malic acid over Zr(SO4)2/SiO2 and Zr(SO4)2/AC solid acid catalysts is comparable to that over H2SO4 and unsupported Zr(SO4)2•4H2O catalysts (Figure 2a) from 20 to 120 min; however, the selectivities of dimethyl malate Zr(SO 4 ) 2 loading and textural properties of SiO 2 , Zr(SO 4 ) 2 /SiO 2 , AC, and Zr(SO 4 ) 2 /AC samples are summarized in Table 1. The Zr amounts in Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts were detected by inductive coupled plasma optical emission spectroscopy (ICP-OES) analysis, and the active species Zr(SO 4 ) 2 loadings on these two catalysts were then calculated according to the Zr amounts. The actual Zr(SO 4 ) 2 loadings on Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts were ca. 30% (Table 1). The Brunauer-Emmett-Teller (BET) surface area, pore volume, and pore size were determined according to N 2 physical adsorption isotherms ( Figure S1). The surface areas of Zr(SO 4 ) 2 /SiO 2 (139 m 2 /g) and Zr(SO 4 ) 2 /AC (468 m 2 /g) catalysts are drastically smaller than those of SiO 2 (214 m 2 /g) and AC (1053 m 2 /g) supports, respectively, but which are still higher than that of bulk Zr(SO 4 ) 2 ·4H 2 O (5-11 m 2 /g) [28]. Higher surface area is favorable for dispersing active sites and tends to give higher catalytic activity. On the other hand, the mesopore volume and size of Zr(SO 4 ) 2 /SiO 2 are obviously lower than those of the SiO 2 support, which indicates that Zr(SO 4 ) 2 ·4H 2 O deposits on the mesopores of SiO 2 . For the Zr(SO 4 ) 2 /AC catalyst, the micropore volume and size of Zr(SO 4 ) 2 /AC are slightly lower than those of AC; however, the mesopore volume and size of Zr(SO 4 ) 2 /AC are obviously lower than those of AC. These results manifest that Zr(SO 4 ) 2· 4H 2 O prefers to deposit on the mesopores than the micropores of AC. As we know, the presence of mesopores in a catalyst is significantly important for the liquid phase catalytic reaction, because mesopores facilitate the mass transfer in the liquid phase, which is beneficial to the contact between the catalyst and reactant. (DM) over Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts are both 99% at 120 min, which are 0.24 and 0.19 times higher than those of the conventional H2SO4 (75%) catalyst and unsupported Zr(SO4)2•4H2O (80%) catalyst, respectively, highlighting the critical role of AC and SiO2 supports in tuning the selectivity. Besides, the yield of DM over Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts is obviously higher than those over H2SO4 and Zr(SO4)2•4H2O catalysts ( Figure 2c). Consequently, pyridine adsorption IR and XPS were carried out on the supported and unsupported Zr(SO4)2•4H2O catalysts to investigate the function of AC and SiO2 support. The main side product was 2-methoxy dimethyl succinate (DS) whose structure is graphed in Equation (2) over the Zr(SO4)2•4H2O catalyst by the GC-MS (gas chromatography with mass spectrometry) result. We speculate that DS may have come from the intermolecular dehydration of The main side product was 2-methoxy dimethyl succinate (DS) whose structure is graphed in Equation (2) over the Zr(SO 4 ) 2 ·4H 2 O catalyst by the GC-MS (gas chromatography with mass spectrometry) result. We speculate that DS may have come from the intermolecular dehydration of malic acid with methanol. Thus, the esterification main reaction (Equation (1)) competed with the intermolecular dehydration side reaction (Equation (2)) over the Zr(SO 4 ) 2 ·4H 2 O catalyst. As we know, esterification and intermolecular dehydration reaction rates both strongly rely on the properties of Catalysts 2020, 10, 384 5 of 13 acid catalysts. The mechanism for the esterification of acid with alcohol over an acid catalyst is as follows: A proton bonds with O from carbonyl to enhance the electron-deficiency of C from carbonyl, which is beneficial to the nucleophilic attack of alcohol [33]. The mechanism for the intermolecular dehydration of alcohol over the acid catalyst is that the proton bonds with the O from hydroxyl to facilitate dehydration [34,35].
The acidity of Zr(SO 4 ) 2 ·4H 2 O, Zr(SO 4 ) 2 /SiO 2 , and Zr(SO 4 ) 2 /AC catalysts was examined by pyridine adsorption IR, as shown in Figure 3. The band at 1539 cm −1 is the characteristic peak of pyridinium ions, which are adsorbed on the Brønsted (B) acid sites, while the characteristic peak at 1445 cm −1 is assigned to pyridine coordinately bonded to Lewis (L) acid sites [36]. As shown in Figure 3, L and B acid sites both exist on the surface of Zr(SO 4 ) 2 ·4H 2 O, Zr(SO 4 ) 2 /SiO 2 , and Zr(SO 4 ) 2 /AC. According to the previous research, L acid sites of Zr(SO 4 ) 2 ·4H 2 O come from the electron-deficiency of Zr 4+ . B acid sites of Zr(SO 4 ) 2 ·4H 2 O originate from the interaction of crystal water with Zr 4+ , and the increase in electron-deficiency of Zr 4+ enhances the degree of H delocalization from crystal water and then increases the B acid strength [30,31,37]. The concentrations of B and L acid sites can be obtained by Equations (3) and (4) based on the integrated absorbance of B and L bands [36]. The concentrations of L and B acid sites of Zr(SO 4 ) 2 ·4H 2 O, Zr(SO 4 ) 2 /SiO 2 , and Zr(SO 4 ) 2 /AC are listed in Table 2. For all of these three catalysts, the concentrations of B acid sites on these three catalysts are significantly higher than those of L acid sites. As we know, the esterification and intermolecular dehydration reactions occurring at lower temperature were catalyzed by the B acid rather than L acid.
Catalysts 2020, 10, x FOR PEER REVIEW 5 of 13 malic acid with methanol. Thus, the esterification main reaction (Equation (1)) competed with the intermolecular dehydration side reaction (Equation (2)) over the Zr(SO4)2•4H2O catalyst. As we know, esterification and intermolecular dehydration reaction rates both strongly rely on the properties of acid catalysts. The mechanism for the esterification of acid with alcohol over an acid catalyst is as follows: A proton bonds with O from carbonyl to enhance the electron-deficiency of C from carbonyl, which is beneficial to the nucleophilic attack of alcohol [33]. The mechanism for the intermolecular dehydration of alcohol over the acid catalyst is that the proton bonds with the O from hydroxyl to facilitate dehydration [34,35].
The acidity of Zr(SO4)2•4H2O, Zr(SO4)2/SiO2, and Zr(SO4)2/AC catalysts was examined by pyridine adsorption IR, as shown in Figure 3. The band at 1539 cm −1 is the characteristic peak of pyridinium ions, which are adsorbed on the Brønsted (B) acid sites, while the characteristic peak at 1445 cm −1 is assigned to pyridine coordinately bonded to Lewis (L) acid sites [36]. As shown in Figure  3, L and B acid sites both exist on the surface of Zr(SO4)2•4H2O, Zr(SO4)2/SiO2, and Zr(SO4)2/AC. According to the previous research, L acid sites of Zr(SO4)2•4H2O come from the electron-deficiency of Zr 4+ . B acid sites of Zr(SO4)2•4H2O originate from the interaction of crystal water with Zr 4+ , and the increase in electron-deficiency of Zr 4+ enhances the degree of H delocalization from crystal water and then increases the B acid strength [30,31,37]. The concentrations of B and L acid sites can be obtained by Equations (3) and (4) based on the integrated absorbance of B and L bands [36]. The concentrations of L and B acid sites of Zr(SO4)2•4H2O, Zr(SO4)2/SiO2, and Zr(SO4)2/AC are listed in Table 2. For all of these three catalysts, the concentrations of B acid sites on these three catalysts are significantly higher than those of L acid sites. As we know, the esterification and intermolecular dehydration reactions occurring at lower temperature were catalyzed by the B acid rather than L acid.  Absortance (a.u.)  Because Zr(SO 4 ) 2 ·4H 2 O reacts with NH 3 at higher temperature, NH 3 -temperature programmed desorption (NH 3 -TPD) is not suitable for the determination of acid strength of the catalyst. As in previous discussion, the B acid strength intensely depends on the electron-deficiency of Zr 4+ in Zr(SO 4 ) 2 ·4H 2 O; therefore, we applied the XPS technique to characterize the electron density on Zr 4+ , and to then determine the B acid strength of supported and unsupported Zr(SO 4 ) 2 ·4H 2 O. As shown in Figure 4a, the binding energy peaks at ca. 186 and 184 eV in Zr(SO 4 ) 2 ·4H 2 O, Zr(SO 4 ) 2 /SiO 2 , and Zr(SO 4 ) 2 /AC samples are ascribed to Zr 4+ 3d 3/2 and Zr 4+ 3d 5/2 , respectively [28]. As shown in Figure 4b,c, the binding energy peak at 532.5 eV in AC is attributed to surface hydroxyls [38]; the binding energy peak at 532.5 eV in SiO 2 is assigned to O 2− ions [39,40]. that of unsupported Zr(SO4)2•4H2O. The O1s binding energy of Zr(SO4)2/AC (533.1 eV) is higher than those of pure AC (532.5 eV) and bulk Zr(SO4)2•4H2O (532.6 eV); similarly, the O1s binding energy of Zr(SO4)2/SiO2 (532.9 eV) is also higher than that of pure SiO2 (533.8 eV) and bulk Zr(SO4)2•4H2O (532.6 eV). These results illustrate that surface hydroxyls of AC or O 2− ions from SiO2 donate electrons to Zr 4+ in Zr(SO4)2/AC and Zr(SO4)2/SiO2 catalysts, resulting in the decrease in electron density on oxygen and the increase in electron density on Zr 4+ , which reduces the degree of H delocalization from crystal water and then decreases the B acid strength. The reduced B acid strength of Zr(SO4)2/AC and Zr(SO4)2/SiO2 catalysts suppresses the intermolecular dehydration side reaction, thereby greatly enhancing the selectivity of dimethyl malate.   Binding energies of the Zr 3d and O 1s regions of various samples are listed in Table 3. Zr 3d 5/2 and Zr 3d 3/2 binding energies of Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC are 0.4 eV lower than those of Zr(SO 4 ) 2 ·4H 2 O, suggesting that the electron density on Zr 4+ of supported Zr(SO 4 ) 2 ·4H 2 O is higher than that of unsupported Zr(SO 4 ) 2 ·4H 2 O. The O 1s binding energy of Zr(SO 4 ) 2 /AC (533.1 eV) is higher than those of pure AC (532.5 eV) and bulk Zr(SO 4 ) 2 ·4H 2 O (532.6 eV); similarly, the O 1s binding energy of Zr(SO 4 ) 2 /SiO 2 (532.9 eV) is also higher than that of pure SiO 2 (533.8 eV) and bulk Zr(SO 4 ) 2 ·4H 2 O (532.6 eV). These results illustrate that surface hydroxyls of AC or O 2− ions from SiO 2 donate electrons to Zr 4+ in Zr(SO 4 ) 2 /AC and Zr(SO 4 ) 2 /SiO 2 catalysts, resulting in the decrease in electron density on oxygen and the increase in electron density on Zr 4+ , which reduces the degree of H delocalization from crystal water and then decreases the B acid strength. The reduced B acid strength of Zr(SO 4 ) 2 /AC and Zr(SO 4 ) 2 /SiO 2 catalysts suppresses the intermolecular dehydration side reaction, thereby greatly enhancing the selectivity of dimethyl malate. The optimization of reaction parameters is very important for commercial feasibility. Figure 5 shows the influence of the catalyst amount of Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts on the conversion of malic acid. Catalyst amount is defined as the weight ratio of a supported catalyst system to reactant malic acid. The esterification of malic acid was executed at different catalyst amounts while other parameters were kept constant. The conversion of malic acid by Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts increased along with the enhancing of the catalyst amount from 2 wt.% to 10 wt.%; however, Catalysts 2020, 10, 384 7 of 13 by further increasing the catalyst amount, the conversion of malic acid remains unchanged. This result indicates that the increase in catalyst amount from 2 wt.% to 10 wt.% enhanced the contact chance between the reactant and catalyst, which brings about the enhancement of conversion. By further increasing the catalyst amount, the conversion of malic acid is governed by other reaction conditions such as the mole ratio of methanol to malic acid, which is controlled by chemical reaction equilibrium. It is very interesting for the similar behavior observed for these two tested heterogeneous catalysts on Figure 5. For these two heterogeneous catalysts, the active species Zr(SO 4 ) 2 loadings are very similar (~30%), as shown in Table 1. In addition, AC or SiO 2 may only act as the role of structural promoter to disperse the active center. Thus, Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts show similar catalytic behavior when the active species Zr(SO 4 ) 2 loadings on these two heterogeneous catalysts are similar. The optimization of reaction parameters is very important for commercial feasibility. Figure 5 shows the influence of the catalyst amount of Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts on the conversion of malic acid. Catalyst amount is defined as the weight ratio of a supported catalyst system to reactant malic acid. The esterification of malic acid was executed at different catalyst amounts while other parameters were kept constant. The conversion of malic acid by Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts increased along with the enhancing of the catalyst amount from 2 wt.% to 10 wt.%; however, by further increasing the catalyst amount, the conversion of malic acid remains unchanged. This result indicates that the increase in catalyst amount from 2 wt.% to 10 wt.% enhanced the contact chance between the reactant and catalyst, which brings about the enhancement of conversion. By further increasing the catalyst amount, the conversion of malic acid is governed by other reaction conditions such as the mole ratio of methanol to malic acid, which is controlled by chemical reaction equilibrium. It is very interesting for the similar behavior observed for these two tested heterogeneous catalysts on Figure 5. For these two heterogeneous catalysts, the active species Zr(SO4)2 loadings are very similar (~30%), as shown in Table 1. In addition, AC or SiO2 may only act as the role of structural promoter to disperse the active center. Thus, Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts show similar catalytic behavior when the active species Zr(SO4)2 loadings on these two heterogeneous catalysts are similar. As we know, the esterification reaction is reversible and excess alcohol shifts the equilibrium to the right side of the reaction (Le Chatelier's Principle) in order to obtain a high yield of the ester product [41]. The molar ratio of methanol to malic acid was varied from 7:1 to 20:1 when the other parameters were kept constant. As shown in Figure 6, excess methanol significantly enhances the final conversion of malic acid. The conversion of malic acid over Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts reaches the maximum (~97%) at a 16:1 molar ratio of methanol to malic acid; however, by further increasing the molar ratio of methanol to malic acid, the conversion of malic acid is no longer enhanced. Figure 7 shows the reusability of Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts. These two supported catalysts can be easily separated from the reaction system by simple filtration. The recyclability of Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts were examined in a new reaction cycle without any treatment, such as calcination or washing. These two catalysts were reused in five consecutive cycles with very little loss of activity, further enhancing the feasibility of these catalysts in the esterification of malic acid. We have also checked the concentration of Zr ions in the solution and the Zr(SO 4 ) 2 loadings in the Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts after the reaction by using ICP-OES. There were negligible Zr elements detected in the solution after the first, second, third, fourth, and fifth cycles (Table S1), and the Zr(SO 4 ) 2 amounts in the Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts before and after five cycles of reaction were almost not changed (Table S2), which indicates that the leaching of Zr(SO 4 ) 2 did not occur.
As we know, the esterification reaction is reversible and excess alcohol shifts the equilibrium to the right side of the reaction (Le Chatelier's Principle) in order to obtain a high yield of the ester product [41]. The molar ratio of methanol to malic acid was varied from 7:1 to 20:1 when the other parameters were kept constant. As shown in Figure 6, excess methanol significantly enhances the final conversion of malic acid. The conversion of malic acid over Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts reaches the maximum (~97%) at a 16:1 molar ratio of methanol to malic acid; however, by further increasing the molar ratio of methanol to malic acid, the conversion of malic acid is no longer enhanced.  Figure 7 shows the reusability of Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts. These two supported catalysts can be easily separated from the reaction system by simple filtration. The recyclability of Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts were examined in a new reaction cycle without any treatment, such as calcination or washing. These two catalysts were reused in five consecutive cycles with very little loss of activity, further enhancing the feasibility of these catalysts in the esterification of malic acid. We have also checked the concentration of Zr ions in the solution and the Zr(SO4)2 loadings in the Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts after the reaction by using ICP-OES. There were negligible Zr elements detected in the solution after the first, second, third, fourth, and fifth cycles (Table S1), and the Zr(SO4)2 amounts in the Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts before and after five cycles of reaction were almost not changed (Table S2), which indicates that the leaching of Zr(SO4)2 did not occur.  Tables S3 and S4, respectively. The conversion of malic acid achieved ca. 98% and the selectivity of DM reached 99% over the Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts, when the malic acid amount expanded 10 times (10 g). Besides, the malic acid conversion was 97% and diethyl malate selectivity was 99% with the application of Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts for malic acid esterification by ethanol, which indicates that the applicability of Zr(SO4)2/SiO2 and Zr(SO4)2/AC catalysts is extended to esterification by other alcohols.  Table S3 and Table S4, respectively. The conversion of malic acid achieved ca. 98% and the selectivity of DM reached 99% over the Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts, when the malic acid amount expanded 10 times (10 g). Besides, the malic acid conversion was 97% and diethyl malate selectivity was 99% with the application of Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts for malic acid esterification by ethanol, which indicates that the applicability of Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC catalysts is extended to esterification by other alcohols.

Catalyst Preparation
AC needed to be pretreated before use as a support. Typically, AC (10 g) was immersed in the 3 M HNO 3 under vigorous stirring at 70 • C for 4 h and was then washed with deionized water until neutral pH. Subsequently, the treated AC was dried at 110 • C in air overnight. The acid treatment can modify the surface properties of AC in favor of active component loading. Acid treatment for AC will introduce hydroxyl groups onto the AC surface, which can strengthen the interaction between the active species and AC support, thus enhancing the active species amount.

Catalytic Reaction
Catalytic esterification of malic acid with methanol was performed in a 50 mL three-necked flask. Certain amounts of malic acid (7.5 mmol, 1 g) and methanol (123.6 mmol, 5 mL) were added in the vessel. The typical reaction was carried out at reflux temperature with a 0.1 g catalyst under magnetic stirring of 600 rpm. Then, the post-reaction solution was analyzed by high-performance liquid chromatography (HPLC, Agilent, Santa Clara, CA, USA) and gas chromatography with mass spectrometry (GC-MS, Agilent, Santa Clara, CA, USA). The HPLC was equipped with a column (Aminex HPX-87H Ion Exclusion column 300 mm × 7.8 mm) and an ultraviolet detector. GC-MS was equipped with a flame ionization detector (FID) and HP-5MS column (30 m × 0.25 mm, 0.25 µm). As the decomposition temperature of malic acid is 140 • C, which is much lower than the injection temperature (280 • C) of GC, GC-MS was not suitable for determining the conversion of malic acid. Therefore, the conversion of malic acid was determined by HPLC through the external standard method; the selectivity of dimethyl malate and side products was determined by GC-MS with n-pentane as the internal standard.

Catalyst Characterization
The Zr amounts in the catalysts and the concentration of Zr ions in the solution after the reaction were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, PerkinElmer ICP-OES 7300DV, PerkinElmer, Waltham, MA, USA) measurements.
The structural information of the catalysts was determined by X-ray diffraction (XRD, Brucker D8 Advance, Madison, WI, USA), using Cu K α radiation (λ = 1.5405 Å) under 40 kV and 40 mA at room temperature.
The N 2 physical adsorption/desorption isotherms of samples were obtained by a Micromeritics ASAP 2020 M+C apparatus (Norcross, GA, USA). Pretreatment of samples was conducted in vacuum for 4 h at 200 • C. The specific surface area was calculated by the Brunauer-Emmett-Teller (BET) method, the mesopore volume and mean mesopore size were determined by the Barrett-Joyner-Halenda (BJH) method from the desorption branch of the isotherms, and the micropore volume and average micropore size were calculated by the t-plot method.
The acidity of the sample was determined by pyridine adsorption infrared (IR) on a Thermo Fisher Nicolet 6700 spectrometer (Thermo Fisher Nicolet, Waltham, MA, USA). Each finely ground sample (ca. 30 mg) was pressed into a thin wafer with a diameter of 1.3 cm, and the thin wafer was then placed into a quartz cell equipped with CaF 2 windows. Before pyridine adsorption, the sample was evacuated at 200 • C for 4 h and cooled to room temperature. Then, pyridine as a molecular probe was exposed to the disk at room temperature for 1 h, and the sample was then evacuated at 150 • C for 1 h to remove the physical adsorption species on the surface; ultimately, the infrared spectrum was recorded. The concentrations of Brønsted (B) and Lewis (L) acid sites can be obtained by Equations (3) and (4) based on the integrated absorbance of B and L bands [36]. X-ray photoelectron spectra (XPS) were obtained with an ESCALAB 250 Xi (Thermal Fisher Scientific, Waltham, MA, USA) spectrometer equipped with a hemispherical electron analyzer and mono Al K α 150 W X-ray source. The XPS spectra of AC and Zr(SO 4 ) 2 /AC samples were calibrated by setting the sp 2 hybrid of the C 1s peak to 284.3 eV, and the XPS spectra of SiO 2 , Zr(SO 4 ) 2 ·4H 2 O, and Zr(SO 4 ) 2 /SiO 2 samples were calibrated by setting the sp 3 hybrid of the C 1s peak to 284.8 eV.

Conclusions
In summary, we applied Zr(SO 4 ) 2 /SiO 2 and Zr(SO 4 ) 2 /AC as solid acid catalysts for malic acid esterification by methanol. These two solid acid catalysts exhibit outstanding catalytic activity, which is comparable to traditional H 2 SO 4 and unsupported Zr(SO 4 ) 2 ·4H 2 O catalysts. Moreover, a 99% selectivity of dimethyl malate can be realized on these two supported catalysts, which is much higher than that of the unsupported Zr(SO 4 ) 2 ·4H 2 O (80%) catalyst, highlighting the critical role of AC and SiO 2 supports in tuning the selectivity. Pyridine-adsorbed IR combined with XPS reveals that surface hydroxyls of AC or O 2− ions from SiO 2 donate electrons to Zr 4+ in Zr(SO 4 ) 2 /AC and Zr(SO 4 ) 2 /SiO 2 catalysts, resulting in the decrease in electron density on oxygen and the increase in electron density on Zr 4+ , which reduces the degree of H delocalization from crystal water and then decreases B acid strength. The reduced B acid strength of Zr(SO 4 ) 2 /AC and Zr(SO 4 ) 2 /SiO 2 catalysts suppresses the intermolecular dehydration side reaction, thereby greatly enhancing the selectivity of dimethyl malate. In addition, these two supported catalysts can be easily separated from the reaction system by simple filtration with almost no loss of activity.