Tailoring the Synergistic Bronsted-Lewis acidic effects in Heteropolyacid catalysts: Applied in Esterification and Transesterification Reactions

In order to investigate the influences of Lewis metals on acidic properties and catalytic activities, a series of Keggin heteropolyacid (HPA) catalysts, HnPW11MO39 (M = TiIV, CuII, AlIII, SnIV, FeIII, CrIII, ZrIV and ZnII; for Ti and Zr, the number of oxygen is 40), were prepared and applied in the esterification and transesterification reactions. Only those cations with moderate Lewis acidity had a higher impact. Ti Substituted HPA, H5PW11TiO40, posse lower acid content compared with TixH3−4xPW12O40 (Ti partial exchanged protons in saturated H3PW12O40), which demonstrated that the Lewis metal as an addenda atom (H5PW11TiO40) was less efficient than those as counter cations (TixH3−4xPW12O40). On the other hand, the highest conversion reached 92.2% in transesterification and 97.4% in esterification. Meanwhile, a good result was achieved by H5PW11TiO40 in which the total selectivity of DAG and TGA was 96.7%. In addition, calcination treatment to H5PW11TiO40 make it insoluble in water which resulted in a heterogeneous catalyst feasible for reuse.

all, H 5 PW 11 TiO 40 was the most active, water-tolerant and acid-tolerant HPAs. Calcination treatment to H 5 PW 11 TiO 40 made it insoluble in water which confirmed its heterogeneous performance in both esterification and transesterification.
Success of this work might clarify the different effects of Lewis metals on total acidity of HPAs and provide more information on how to select proper HPAs according to different requirements.

Methods
Material and reagent. All the chemicals were of AR grade, which were obtained commercially and used without further purification. Na 7 PW 11 O 39 was synthesized according to the ref 17. Instrument. Elemental analysis was carried out using a Leeman Plasma Spec (I) ICP-ES and a PE 2400 CHN elemental analyzer. IR spectra (4000-500 cm −1 ) was recorded in KBr discs on a Nicolet Magna 560 IR spectrometer. The IR spectra of adsorbed pyridine (Py-IR) were depicted by subtracting the spectra before and after exposure to pyridine. X-ray diffraction (XRD) patterns of the sample were collected on a Japan Rigaku Dmax 2000 X-ray diffractometer with Cu Kα radiation (λ = 0.154178 nm). SEM micrographs were recorded on a scan electron microscope (XL30 ESEM FEG 25 kV). The concentrations of esters were determined periodically on Shimazu GC-14C fitted with a HP-INNO Wax capillary column (30 m × 0.25 mm) and flame ionization detector 18 . Each of the catalytic reaction was repeated for three times.
Catalyst preparation. K 5 PW 11 TiO 40 was synthesized according to the procedure described previously 19 . Firstly, Ti (SO 4 ) 2 solution (6 mmol in 2 M H 2 SO 4 ) was added to the aqueous solution of Na 7 PW 11 O 39 (6 mmol). Then, adjusting the pH of the mixture to 5.6 by NaHCO 3 , followed by adding solid KCl (2.24 g) until white precipitate, K 5 PW 11 TiO 40 , was formed. After that, the precipitate was filtrated and recrystallized with water for three times. Other catalysts were synthesized as following: firstly, 29.4 g (100 mmol) Na 2 WO 4 and 1.29 g (9.1 mmol) Na 2 HPO 4 were dissolved in 80 mL deionized water at room temperature under vigorous stirring. Secondly, 20 mL metal salts aqueous solution (12 mmol) (chlorides for Sn; nitrates for Fe, Cr and Zn; sulfate for Al, Cu and ZrO 2 ·xH 2 O) was added dropwise to the former solution with continuous stirring. After that, 100 mL deionized water was added additionally. The pH was adjusted to 5.6 by HNO 3 , then KCl (3.39 g) was added until the precipitate formed. This precipitate was filtrated and recrystallized with water for three times giving potassium salt of KPW 11 M (M = Ti IV , Cu II , Al III , Sn IV , Fe III , Cr III , Zr IV and Zn II ).
2 g potassium salts of KPW 11 M (M = Ti IV , Cu II , Al III , Sn IV , Fe III , Cr III , Zr IV and Zn II ) were dissolved respectively in 1000 mL deionized water and the potassium cations were replaced by H + using strong-acid cation exchange resins (Type 732, 20 g) for several times to give HPW 11 M, until no K + can be detected by ICP analysis. Then the HPW 11 M solution was rotary evaporated at 50 °C to remove water and give powders. These powders were calcinated for 3 h at 200 °C to obtain insoluble products. The formation of HPW 11 M undergos the following equations 17 : where A is the absorbance (area in cm −1 ), ε is the extinction coefficient (m 2 /mol), W is the sample weight (kg), c is the concentration of acid (mol/kg or mmol/g) and S is the sample disk area (m 2 ), respectively. The amount of Brønsted and Lewis acid sites was estimated from the integrated area of the adsorption bands at ca. 1540 and 1450 cm −1 , respectively, using the extinction coefficient values based on the previous report 21 .
Esterification reaction. A 25 mL three-necked glass flask equipped with a water-cooled condenser was charged with glycerin (2.3 g, 25 mmol), different volume of acetic acid and certain amount of catalyst. Each mixture was vigorously stirred and reacted at desired temperature for the required reaction time. After the reaction, the mixture was rotary evaporated at 55 °C to remove the excess acetic acid. The products, the unreacted glycerol and the solid catalyst were left in the reactor. Then the catalyst was separated by centrifuging, washed with water, and calcinated at 100 °C for reuse. The liquid mixture was tested by GC. In addition, we have checked the mixture before and after rotary evaporator by GC which confirmed that there was no additional reaction because the time (1 min) was too short to allow any additional reactions occur. The conversion of glycerol and selectivity of glycerides were calculated by the following equations: moles of glycerol in moles of glycerol out moles of glycerol in 100 7 (%) = × % ( ) Selectivity moles of one product moles of all products 100 8 Transesterification reaction. 1.88 mL glycerol triacetin was added to a 25 mL three-necked glass flask with ethanol-cooled condenser (the temperature is − 3 °C) under vigorous stirring. After being preheated to 65 °C, the specified amount of methanol (molar ratio of oil/methanol was 1: 6) and 4 wt% of catalyst was added under stirring at 300 rpm to keep the system uniform in temperature and suspension. The reaction was maintained for 4 h at 65 °C and atmospheric pressure. After reaction, the mixture was rotary evaporated at 45 °C to remove the excess methanol, while the products, unreacted triacetin and the solid catalyst were left in the reactor. Then the catalyst was separated by centrifuging, washed with water and calcinated at 100 °C for reuse. The liquid mixture was measured by GC. In addition, we have checked the mixture before and after rotary evaporator by GC which confirmed that there was no additional reaction because the time (1 min) was too short to allow any additional reactions occur.

Results and Discussion
Structural characterization of the catalysts. The elemental analyses of HPW 11 M were given in Table 1. These results confirmed the molar ratio of P: W: M = 1:11:1, which showed the formation of mono-metal substituted undeca-tungstophosphates. Four bands were shown at 1072(8ν as P-O), 977 (ν as W=O ), 891 (ν as W−O−W inter-octahedral), and 796 cm −1 (ν as W−O−W intra-octahedra) in the FTIR spectra of HPW 11 M (Fig. 1A), which were attributed to the stretching vibrational peaks of HPA Keggin anions 22 . Compared with PW 11 O 39 7− (1082, 957, 873, and 763 cm −1 ), some shifts (ν as W=O and ν as W−O−W ) occurred due to the substitution of W by M ions to form saturated HPAs (Scheme S1). The FTIR spectra of HPW 11 M were similar to that of H 3 PW 12 O 40 (1080, 982, 890, and 797 cm −1 ), which showed the replacement of tungsten by metal cations formed a dodecatungstophosphoric Keggin structure scussesfully. The IR spectrum gave a band at 580 cm −1 which indicated the existence of an M-O bond 23 . The XRD patterns of as-prepared catalysts also supported the results of FTIR (Fig. 1B). It was found that the characteristic peaks were similar to those of PW 11  7− clusters and the formation of good crystals with the Keggin structure. The morphology of H 5 PW 11 Ti was measured by SEM and EDAX (Fig. S1). It showed that as-prepared material displayed well-shaped crystalline particles with molar ratio of W: P: Ti = 11.07: 1.14: 1.13.
The acidic properties of the catalysts. The FT-IR spectra of pyridine absorption are a powerful tool for identifying the nature of acid sites 21 . As shown in Fig. 2, HPW 11 M presented typical bands at around 1540 and 1639 cm −1 24 corresponding to strong Brønsted sites. Compared with HPW 11 M, new bands at 1450 and 1610 cm −1 were assigned to the coordinated pyridine adsorption on the Lewis acid sites, while the band at 1489 cm −1 was originated from the combination of pyridine on both Brønsted and Lewis acid sites 25 . The results indicated that the Lewis acid sites were successfully introduced to HPW 11 M molecules; Therefore, HPW 11 M exhibited double acidic sites including Brønsted sites and Lewis ones. The total contents of Brønsted acid and Lewis acid were obtained from titration with n-butylamine 20 and the separated density was calculated from the strength ratio of Brønsted acid and Lewis acid in FT-IR spectra of pyridine absorption 14 (Table 1). It was found that mono-substituted HPW 11     Catalytic activity of HPW 11 M catalysts in esterification. Commonly, Brønsted acid is active mainly in esterification, whereas Lewis acid is more active in transesterification 26 . However, some research showed that Lewis acid is also active in esterification of glycerol. Zhu's group recently reported a highly-active silver-exchanged phosphotungstic acid catalyst Ag 1 H 2 PW 12 O 40 which was applied in glycerol esterification with acetic acid, with 96.8% conversion and 48.4, 46.4 and 5.2% selectivity to MAG, DAG and TAG, respectively, and the reaction conditions were 120 °C, 15 min with the molar ratio of 10:1. The high efficiency came from the contribution of Lewis metal Ag 14 . N. Lingaiah's group also reported the incorporation of Zn into the secondary structure of heteropolytungstate which could promote the conversion of glycerol to carbonate with urea 16 . In order to evaluate the acidic performance of these HPAs, the esterification of glycerol with acetic acid was conducted in the presence of the HPW 11 M catalysts (  14 . In our work, the selectivities to DAG and TAG were 55.4 and 6.7% by H 5 PW 11 TiO 40 , and 43.8 and 2.4% by H 5 PW 11 CuO 39 , respectively, which were obtained with the mild conditions : molar ratio of glycerol to acetic acid = 1:5, 75 °C, 3 wt% of catalyst, and 1 h. Extending the reaction time could enhance the selectivities to DAG and TAG (Fig. 3a). It was found that from 0.25 h to 4 h, the selectivity toward MAG obviously decreased to 3.3% while the total selectivity to DAG and TAG increased to 96.7%. For H 5 PW 11 ZrO 40 and H 5 PW 11 ZnO 39 with almost the same Lewis acidity, but the former tended to produce more DAG while the later was fond of MAG, From the above equations, we could know that water was by-product of the glycerol esterification reaction. Hence, the solid acid catalyst was easily destroyed by water in the system. So a water-tolerant solid acid catalyst was desirable. When adding extra water to the reaction system, catalytic activity of the two catalysts could be influenced (Fig. 4). In view of H 5 PW 11 TiO 40 , the water content ranging from 0 to 0.6 wt% did not play a significant role in the glycerol conversion, which meant that H 5    Main parameters of the esterification reaction catalyzed by H 5 PW 11 TiO 40 including temperature, reaction time, catalyst dosage and molar ratio of glycerol to acid were investigated (Fig. S3). It suggested that in order to obtain high yields of DAG and TAG, high molar ratio of glycerol to acid (1:5), long reaction time (4 h), and 75 °C were needed.

Catalytic activity of HPW 11 M catalysts in transesterification. Based on the suggestion by
Santacesaria 30 , Brønsted acid catalysts were active mainly in esterification while Lewis acid catalysts were more active in transesterification. In order to evaluate the efficiency of the HPW 11 M on transesterification, glycerol triacetin was selected to investigate the influence of their Lewis and Brønsted acidity ( Table 3). It is known that triglyceride transesterification with methanol is a consecutive reaction including three continuous steps 31 :     (Fig. S4). In order to obtain high yield of glycerol, high molar ratio of triacetin to methanol (1:6), long reaction time (4 h), and 65 °C were needed, with which we have got a 92.2% conversion and a 83.3% selectivity of glycerol, respectively.
It might be more important to convert low quality feedstocks to biodiesel. However, the presence of high FFAs might have adverse effects on the catalyst activity. H 5 PW 11 TiO 40 might catalyze the esterification of FFA and transeterification of triacetin simultaneously. Therefore, the conversion of triacetin with some FFA contents was catalyzed by H 5 PW 11 TiO 40 with Lewis acidity and Brønsted one (Fig. S5).

Reuse of catalysts.
It is necessary to determine the nature of H 5 PW 11 TiO 40 in esterification or transeterification. The solubility of H 5 PW 11 TiO 40 in acetic acid and methanol was measured by Uv-Vis spectroscopy (Fig. S6). We could see that H 5 PW 11 TiO 40 was insoluble either in acetic acid or in methanol, which suggested that it performed as a heterogeneous catalyst in both esterification and transesterification reactions. The behavior was attributed to the calcinations treatment to H 5 PW 11 TiO 40 at 200 °C for about 3 h to form insoluble powders which were also determined by the SEM image and EDAX (Fig.  S1). In other word, the catalyst could be easily separated from the production mixture. The catalyst was separated by centrifuging and decanted from the bottom of the reactor, and then washed with ethanol and dried at 60 °C overnight for further reaction cycles. As shown in Fig. 5, there was no considerable change in the catalytic activity after five reaction cycles either in esterification or transesterification. The IR spectrum of the products in transesterification (Fig. S7) indicate that no characteristic peaks corresponding to H 5 PW 11 TiO 40 were observed in range of 790 to 1000 cm −1 , which demonstrated that the leaching of H 5 PW 11 TiO 40 was negligible. The leaching of H 5 PW 11 TiO 40 was about 0.10 wt% and 0.12 wt% during esterification and transesterification for one cycle, respectively. To further determine the leaching of H 5 PW 11 TiO 40 , the catalyst was seperated after reacting for 20 min (91.0% of glycerol conversion) and was allowed to react further for over 1 h at the same conditions. The result showed that the conversion of glycerol was only 92.4%, which meant that H 5 PW 11 TiO 40 acted as a heterogeneous catalyst.
The stability of HPAs during the reaction was determined by IR spectroscopy (Fig. S8). After the reaction, the catalysts still kept its Keggin structure. The peaks at 1072, 977, 891, and 796 cm −1 attributing to the characteristic bands of Keggin structure could also be observed. The SEM image of the catalyst after reaction (Fig. S9) showed no-change of the morphology during the reaction. Therefore, HPAs was stable and could be reused at least for five cycles.

Conclusion
This work demonstrated a variety of HPAs H n PW 11 M (M = Ti IV , Cu II , Al III , Sn IV , Fe III , Cr III , Zr IV and Zn II ) with both Lewis and Brønsted acidity. The influence of these Lewis metals on the acidic strength of dodecatungstates showed that (1) moderated Lewis metals gave more influence on total acidity including Ti and Cu; (2) compared to Ti x H 3−4x PW 12 O 40 , Ti in substituted place played a less important role on total acidity. Their catalytic activities in esterification of glycerol and transeterification of triglycerides were generally followed their acidic properties. Among all these HPAs, H 5 PW 11 TiO 40 displayed properities like Lewis acid sites, insolubility in polar solvents, water-tolerance, acid-tolerance, and stability which were benefit to its excellent performance. To the best of our knowledge, H 5 PW 11 TiO 40 gave the highest selectivity to desired DAG and TGA (96.7%) in the esterification of glycerol which was higher than the reported Ag 1 H 2 PW 12 O 40 . Moreover, in transeterification, high conversion of triacetin and high yield of glycerol were obtained by H 5 PW 11 TiO 40 . Furthermore, H 5 PW 11 TiO 40 did not suffer from leaching and deactivation in five reaction cycles either in esterification or transesterification reactions.
This study provides useful information on Lewis metal substituted HPAs. It could be a promising candidate for the catalytic esterification of glycerol and synthesis of valuable biofuel additives, meanwhile, it could also act as a potential catalyst for production of biodiesel using low quality feedstocks.