Biodiesel represents an alternative to oil-based diesel. It is a renewable, high quality fuel with excellent lubricating characteristics, synthesized in the transesterification of lipids (mostly vegetable origin) and low-molecular alcohols^[1-4]. According to OECD 2015 market report^[5], global biodiesel production may rapidly increase to 39 billion liters in 2024. Glycerol is the by-product of biodiesel production and the global glycerol market is rising rapidly. Due to oversupply, the glycerol price is dropping significantly and, at the same time, the sustainability issues occur^[6].

Moreover, the glycerol co-produced in the biodiesel production is usually "crude glycerol" containing water, methanol and other impurities^[7]. The conversion of crude glycerol into the chemicals with added value can expand the glycerol market and increase the efficiency of biodiesel industry. As shown in Scheme 1, glycerol can be converted into many valuable chemicals through esterification, acetalization, dehydration, and hydrogenolysis^[8-13]. One well-commercialized process is esterification of glycerol with various amount of acetic acid to produce monoacetin, diacetin and triacetin (MA, DA and TA). DA and TA used as transport fuel additives can improve low-temperature and viscosity properties, making them the common components of bio-additives^[14].

Traditionally, the esterification reaction was conducted in the homogeneous system using strong acids (H₂SO₄, HCl, HI and PTSA)^[15]. However, environmental problems such as waste management, corrosion of equipment and toxicity caused by these catalysts are incompatible with the tight environmental legislation. Thus, the replacement of homogeneous catalysts by the heterogeneous, renewable solid acids became an important task. Some heterogeneous catalysts used in esterification reaction including zeolites (HUSY, HZSM-5, SBA-15), ion exchangers (Amberlyst-15, Amberlyst-16), metal oxides and heteropolyacids et al^[16-20].

InbaeKim^[21]tested the various solid acids as esterification catalysts and obtained the following order of efficiency: PrSO₃H-SBA-15 > Amberlyst-15 > HPMo/Nb₂O₅ > HPMo/SBA-15 > HUSY > SCZ > SiO₂-Al₂O₃. Hu et al^[22] used metal oxides of bismuth, antimony, tin, and niobium as efficient catalysts in the same esterification reaction, reaching 96.8% of glycerol conversion and 54.2% selectivity toward diacetin. Manayil^[23] investigated the reaction between acetic acid and aromatic alcohols, catalyzed by propylsulfonic acid functionalized silica PrSO₃H-SBA-15. It was confirmed that the mesoporous PrSO₃H-SBA-15 showed better catalytic performance than microporous zeolites in esterification with benzyl-alcohol^[24-25].

Mesoporous silica materials possess hexagonal array and uniform channel distribution^[26]. Functionalized mesoporous silica is a highly attractive material for heterogeneous catalysis and widely used as heterogeneous catalysts. To increase the acidity of mesoporous silica, the mesostructured SBA-15 was functionalized with propyl, arene, and perfluoro sulfonic acids by post-grafting^[27].

Water as by-product of esterification reaction may reduce the performance of heterogeneous catalyst greatly. Modification of mesoporous materials with hydrophobic groups and the formation of hydrophobic environment into the catalyst's pores is one of the methods for increasing the efficiency of esterification reaction^[28-29].

In this study, an array of co-functionalized SBA-15 with excellent hydrophobic properties and acid content were synthesized and their performance was evaluated using glycerol-acetic acidesterification as a model reaction. The effect of reaction conditions on the selective formation of diacetin and triacetin was also investigated. The water-tolerability was studied by adding excess amounts of water. The application of sulfonated mesoporous SBA-15 as heterogeneous catalysts for glycerol-acetic acid esterification is described in this manuscript for the first time.

1 Experimental 1.1 Materials

The mesoporous silica material was synthesized following the previously described procedure^[30-32]. 3-mercaptopropyltrimethoxysilane (MPTMS, 95%) was obtained from Aldrich Chemicals. Methyltriethoxysilane (MTES), phenyltriethoxysilane (PTES), n-propyltriethoxysilane and n-decyltriethoxysilane were purchased from Alfa Aesar. N-Octyltrimethoxysilane (OTMS) and perfluorooctyltrimethoxysilane (PFOTS) were purchased from Acros organics. Toluene, anhydrous methanol and ethanol were obtained from Rianlon Chemical Co., Ltd. Other reagents were analytical grade, and used without additional purification.

1.2 Synthesis of catalysts 1.2.1 Synthesis of PrSH-SBA-15

Mercaptopropyl grafted mesoporous SBA-15 was synthesized using previously described post-grafting method. The 1 g of SBA-15 was dispersed into 30 mL deionized water, various amounts of MPTMS were added and the mixture was heated at 100 ℃ for 24 h under reflux. The products were washed with MeOH and vacuum-dried for 12 h and obtained products were denoted as PrSH-SBA-15.

1.2.2 Synthesis of co-functionalized catalysts (Octyl/Propyl/Decyl-PrSO₃H-SBA-15-SH, PrSO₃H-PFO-SBA- 15, PrSO₃H-Ph-SBA-15, PrSO₃H-Me-SBA-15)

Several PrSO₃H-functionalized SBA-15 and its co-derivatized catalysts were prepared by previously described method. 3 g of PrSH-SBA-15 was dispersed in 30 mL toluene, followed by the addition of 4 mmol of OTMS (PFOTS or PTES or MTES). The mixture was heated at 130 ℃ for 24 h (under reflux). The resulting materials were filtered, washed with MeOH and vacuum-dried for 12 h.

1.2.3 Oxidation of -PrSH to -PrSO₃H

Complete oxidation of -Octyl/Decyl/Propyl/PFO/Ph/Megrafted-PrSH to -PrSO₃H were conducted with hydrogen pero- xide. 1 g of the synthesized thiol-propyl functionalized materials was added to 30 mL of 30% aqueous H₂O₂ and the mixture was stirred at room temperature for 24 h. Afterward, the products were filtered, and washed with MeOH and vacuum-dried for 24 h, and isolated as white powders.

1.3 Material Characterization

The pore size, pore volume and the surface area of derivatized mesoporous silica were determined by N₂ adsorption-desorption on a TristarII 3020 instrument (Micromeritics, USA). Before adsorption, the samples were degassed at 423 K for 4 h. The powder X-ray diffraction data were acquired on a PANalytical Empyrean apparatus (Netherlands) with 0.02° step in the small-angle. Fourier transform infrared (FT-IR) spectra were recorded at room temperature in the 4000~400 cm^-1 on a Nexus 870 spectrometer (Nicolet Instruments Co., USA). Simultaneous thermal analyzer Netzsch Model STA 449 F3 (Netzsch, Germany) was utilized for recording thermogravimetry/differential scanning calori- metry (TG/DSC) profiles of all samples under N₂ atmosphere.

The amount of the acidic sites in catalysts was quantified via acid-base titration. The sample was dispersed in 50 mL of 2 mol/L NaCl (50 mL), left for 24 h, and titrated with 0.01 mol/L solution of NaOH.

1.4 Glycerol-acetic acid reaction

Typically, the reaction mixture consisting of 2 g of glycerol, a corresponding amount of acetic acid and catalysts were added into the round-bottom flask. The qualitative and quantitative analysis of the resulting solution was performed on an Agilent 7890/5975A GC-MS and Agilent 7890A gas chromatograph with HP-5 column (0.32 mm×30 m×0.32 μm).

The conversion percentage of glycerol and selectivity for glycerol-acetic acid esterification products were calculated as:

$ \begin{array}{l} Conversion\ of\ glycerol\left( {mol\% } \right)\ = \ \frac{{Initial\ glycerol\left( {mol\% } \right) - Final\ glycerol\left( {mol\% } \right)}}{{Inital\ glyceroly\left( {mol\% } \right)}} \times 100\% \ \\ Selectivity\ for\ esterification\ products\left( {mol\% } \right) = \frac{{Desired\ product\left( {mol\% } \right)}}{{Total\ obatined\ products\left( {mol\% } \right)}} \times 100\% \end{array} $

2 Results and discussion 2.1 Catalysts Characterization

The structural properties of modified and non-modified mesoporous SBA-15 were initially assessed by powder XRD (Fig. 1). All functionalized materials exhibited three diffraction peaks characteristic for parent SBA-15, confirming the maintenance of a hexagonal, mesoporous structure in catalysts. The same conclusion can be derived from TEM images of corresponding samples (Fig. 2).

Physico-chemical properties of parent and modified materials are listed in the Table 1. The surface area and the pore diameter of parent material SBA-15 were 542 m²g^-1 and 9.7 nm, respectively. The corresponding values for functionalized Octyl-PrSO₃H-SBA-15 were 387.4 m²g^-1 and 8.8 nm, respectively. Functionalization with MPTMS and n-octyl groups apparently reduces the surface area, pore diameter and pore volume. The incorporation of functional groups occupies the inner surface of mesoporous silica pores.

To certify the successful incorporation of the organic functional groups on themesostructured SBA-15, FT-IR spectra of catalysts used in the experiment were recorded (Fig. 3). The signals at 3480 cm^-1 are associ- ated with O-H and Si-O of silica. Comparing with the parent SBA-15, the characteristic bands around 1244 and 1100 cm^-1 were attributed to the asymmetric vibration of S=O and asymmetric stretching vibration of SO₃H group, confirming the complete sulfonation and oxidation of the materials^[33]. The strong C-H vibration of the alkyl groups in n-octyl were observed at 2938 cm^-1 indicating the successful incorporating of alkyl groups. All these results reveal that the -SO₃H group and alkyl groups were successfully attached to the walls of SBA-15 molecular sieve.

The thermal stability of mesoporous materials were measured via TG, and the weight losses were illustrated in Fig. 4. TG analysis of catalysts shows two weight losses. The first weight loss occurs below 100 ℃ as a result of water desorption and the second loss peak at 500 ℃ is corresponding to the cleavage of chemically bound organic groups. These results suggested the excellent thermal stability of the series of catalysts.

2.2 Comparative analysis of solid catalysts

The progress of glycerol-acetic acid reaction can be divided into three consecutive stages. First, equal amounts of glycerol and acetic acid react to form monoacetin (MA). In the second step, MA reacts with an equal amount of acetic acid to obtain diacetin (DA) followed by the formation of triacetin (TA) in the third step. In the manufacturing process, MA is considered as the undesired reaction product. The acid-catalyzed three-step mechanism can be simplified by the following scheme:

$ {\rm{Gly}} + {\rm{AA}}\xrightarrow{{acid\ catalyst}} MA $

(1)

$ \text{MA}+\text{AA}\xrightarrow{acid\ catalyst}DA $

(2)

$ \text{DA}+\text{AA}\xrightarrow{acid\ catalyst}TA $

(3)

In our experimental setup, the glycerol-acetic acid esterification was performed at 110 ℃, with 1:9 molar ratio of glycerol to acetic acid. Initially, the commercial, tridimensional zeolites (HY and HZSM-5) and ion-exchange resin Amberlyst-15 were used as catalysts, and the results are shown in Fig. 5. Within the first 4 h of reaction, catalyst Octyl-PrSO₃H-SBA-15 showed superior properties compared to Amberlyst-15, HY, HZSM-5 and PrSO₃H-SBA-15. It indicated that the co-functionalized Octyl-PrSO₃H-SBA-15 has a superior conversion rate and selectivity towards DA and TA.

Furthermore, catalysts with various organo-functional groups were evaluated and the results shown in Fig. 6. It can be obviously seen that all the catalysts obtained high glycerol conversion. The diacetion and triacetin selectivity of Propyl-PrSO₃H-SBA-15, Decyl-PrSO₃H-SBA-15, PrSO₃H-PFO/PH/Me-SBA-15 is lower than Octyl-PrSO₃H-SBA-15 apparently. Alkyl groups with different chain length were incorporated on the proplysulfonic acid functionalized SBA-15 and n-octyl is the best choice. Octyl-PrSO₃H-SBA-15 obtained 100% glycerol conversion and 89.9% diacetin and triacetin selectivity.

The acid content of the functionalized mesostructured catalysts were quantified using acid-base titration. The acidity of sulfonic acid mesoporous silica functionalized with electron-withdrawing groups like phenyl (in Sample PrSO₃H-Ph-SBA-15) or fluorinated groups (in Sample PrSO₃H-PFO-SBA-15) is close to the material containing only sulfonic acid groups (Table 1). The catalyst Octyl-PrSO₃H-SBA-15 had the highest content of acidic groups (1.30 mmol/g), which is consistent with its superior catalytic performance.

2.3 Effect of reaction conditions

The effect of reaction conditions (reaction temperature, reaction time, the amount of catalyst and the molar ratio of reactants) on glycerol conversion and selectivity towards DA and TA were investigated using Octyl-PrSO₃H-SBA-15 as catalysts. As shown in Fig. 7, reaction temperature, reaction time, the amount of catalyst and the molar ratio of reactants exhibited varying degrees of effect on catalytic performance. Among all reaction parameters investigated, the temperature had the highest influence on the efficiency and selectivity of this reaction.

2.4 Water tolerance

Esterification is an equilibrium reaction and the reaction rate depends on the local H₂O concentration near the acid site which may reversely hydrolysis reaction. Water tolerance of Octyl-PrSO₃H-SBA-15 and PrSO₃H-SBA-15 were synthesized and tested in this study by adding the increasing amount of water (5%(Weight percentage), 10%(Weight percentage), 15%(Weight percentage) and 20%(Weight percentage)) into the reaction mixture keeping all other reaction parameters constant (Fig. 8). For both catalysts, the addition of 20%(Weight percentage) water significantly reduces the glycerol conversion, compared to 15% water-doped system. The Octyl-PrSO₃H-SBA-15 catalyst reaches 99.5% conversion and 74.8% selectivity toward DA and TA at 20%(Weight percentage) of water, while PrSO₃H-SBA-15 has 94.6% conversion and 61.2% selectivity under the same experimental conditions.

The competitive adsorption of water and alcohol at the acid site may leads to the loss of catalytic activity. Octyl-PrSO₃H-SBA-15 has better water tolerance com- pared to PrSO₃H-SBA-15, which can be rationalized by the presence of hydrophobic groups bounded to the surface of pores which impedes the entrance of water into the pores and interaction with acidic groups. Therefore, increasing the hydrophobicity of catalyst can reduce the adsorption of water on the acid site. Co-functionalized catalyst with octyl chains was used to tune the hydrophobicity of the catalyst and to drive reactively formed water away from the acid site.

2.5 Reusability

The stability and reusability of Octyl-PrSO₃H-SBA-15 catalyst was also evaluated. After each glycerol-acetic acid reaction cycle, the solid catalysts were prepared for the next run by centrifugation and subsequent washing of solid residue with methanol. As shown in Fig. 9, Octyl-PrSO₃H-SBA-15 retained its catalytic activity for nine recycles, without significant decrease in conversion of glycerol and selectivity toward DA and TA. After nine recycles, the conversion and selectivity were 95.9% and 69.6%, respectively.

3 Conclusion

The glycerol-acetic acid esterification reaction was used as a model system for studying the catalytic performance of water-tolarant co-functionalized SBA-15 catalysts. The structure of catalysts was investigated by BET, FT-IR, TG, TEM and acid-base titration. The effect of reaction parameters such as the reaction temperature, reaction time, the amount of catalyst, and the molar ratio of reactants were discussed. Octyl-PrSO₃H-SBA-15 showed 100% conversion and 89.9% selectivity toward diacetin and triacetin with the molar ratio of glycerol to acetic acid 1:9 at 110 ℃ for 4 h. Water tolerance of novel catalysts (Octyl-PrSO₃H-SBA-15 and PrSO₃H-SBA-15) was thoroughly investigated. Octyl-PrSO₃H-SBA-15 showed superior performance and could be recycled for nine times without losing the conversion efficiency and selectivity toward DA+TA formation.

Acknowledgment: This work was supported by the National Natural Science Foundation of China (Grant No. 21673259) and the Natural Science Foundation of Jiangsu Province of China (No. BK2017 1241).