Effects of Basic Promoters on the Catalytic Performance of Cu/SiO2 in the Hydrogenation of Dimethyl Maleate

Continuous hydrogenation of dimethyl maleate (DMM) toγ-butyrolactone (GBL), 1,4-butanediol (BDO) and tetrahydrofuran (THF) is a promising process in industry. In this study, Cu-M/SiO2 catalysts modified by basic promoters (M = Mg, Ca, Sr, Ba, La) were prepared, and characterized by physical adsorption of N2, in situ XRD, H2-TPR, CO2-TPD. With the addition of basic promoters, the basicity of Cu-M/SiO2 catalysts was improved. The particle size of CuO on Cu-M/SiO2 catalyst was increased after modified by Mg, Ca, Sr, Ba. However, the CuO particle was decreased on the Cu-La/SiO2 catalyst. The series of Cu-M/SiO2 catalyst was applied to the hydrogenation of DMM. The addition of basic promoters increased the selectivity of GBL during the hydrogenation for the basic promoters improved the dehydrogenation of BDO to GBL in alkaline sites. Furthermore, Cu-La/SiO2 presented a higher activity in the hydrogenation of DMM, due to its higher dispersion of Cu.


Introduction
Dimethyl maleate (DMM) as an important chemical intermediate is widely used in pharmaceutical, pesticide, chemical fiber, automobile and other industries [1][2][3]. Moreover, the hydrogenation products of DMM, such as γ-butyrolactone (GBL), 1,4-butanediol (BDO) and tetrahydrofuran (THF) are all high value-added chemical intermediates (the reaction pathways of DMM as shown in Scheme 1) [4][5][6][7]. With the frequent breakthrough of the oxidation technology of n-butane to maleic anhydride and the decrease of production cost, it will be of significance for the development of the downstream products of maleic anhydride. Esterification of maleic anhydride and methanol to DMM, and then continuous hydrogenation to BDO, GBL and THF, are considered to be a very promising technology. The key to this process is the catalyst for hydrogenation of DMM.

Introduction
Dimethyl maleate (DMM) as an important chemical intermediate is widely used in pharmaceutical, pesticide, chemical fiber, automobile and other industries [1][2][3]. Moreover, the hydrogenation products of DMM, such as γ-butyrolactone (GBL), 1,4-butanediol (BDO) and tetrahydrofuran (THF) are all high value-added chemical intermediates (the reaction pathways of DMM as shown in Scheme 1) [4][5][6][7]. With the frequent breakthrough of the oxidation technology of n-butane to maleic anhydride and the decrease of production cost, it will be of significance for the development of the downstream products of maleic anhydride. Esterification of maleic anhydride and methanol to DMM, and then continuous hydrogenation to BDO, GBL and THF, are considered to be a very promising technology. The key to this process is the catalyst for hydrogenation of DMM.  The catalysts for the hydrogenation of esters mainly include Ru [8][9][10][11], Pt [12], Pd [13,14], Cu [15][16][17][18], and so on. The noble metal catalysts, such as Ru, Pt and Pd [19,20], have the advantages of high activity, but their price is expensive. A supported Cu catalyst with a strong capacity for the The catalysts for the hydrogenation of esters mainly include Ru [8][9][10][11], Pt [12], Pd [13,14], Cu [15][16][17][18], and so on. The noble metal catalysts, such as Ru, Pt and Pd [19,20], have the advantages of high activity, but their price is expensive. A supported Cu catalyst with a strong capacity for the adsorption, activation of C=O bond, and low hydrogenolysis activity of C-C bond, is cheap and easy to prepare-and then, it is widely used in ester hydrogenation reactions [21]. However, unmodified Cu catalyst has the disadvantage of poor activity and difficult to control the selectivity of hydrogenation products. Improving the preparation method of the supported Cu catalyst is the main way to enhance its catalytic performance. The preparation methods of supported Cu catalysts are as follows: Impregnation [22,23], ion exchange [24,25], sol-gel [26,27], deposition-precipitation [28], ammonia evaporation (AE) [29][30][31]. The AE method was invented by the Japanese Ube and used in the hydrogenation of esters [29]. It is a new method, based on ion exchange and deposition-precipitation: Copper ammonia complex ion is formed by ammonia and copper species precursor, followed by ion exchange with the hydroxyl on the surface of SiO 2 , and the excess copper ammonia complex ions forming Cu(OH) 2 deposited on the surface of SiO 2 with the evaporation of ammonia. The dispersion of Cu in the Cu/SiO 2 catalyst prepared by the AE method is high, and it is difficult to sinter at high temperature. Moreover, there are two kinds of copper species: Cu + and CuO, both of which have been proved to have a synergistic effect, which can further improve the activity of ester hydrogenation. Therefore, the Cu/SiO 2 catalyst prepared by the AE method has the advantages of good dispersion and high activity, and has become a hot spot in recent years. Additives are an important factor affecting the catalyst activity, which changes the electronic structure of catalyst to improve the catalytic activity and selectivity, and also improve the dispersion of the active component and stability [32][33][34][35][36]. Yin et al. introduced NiO to improve the dispersion of metal Cu, and increased the TOF(turn of frequency) value; therefore, the introduction of NiO effectively improved the catalytic performance [32]. He et al. prepared Cu/SiO 2 catalysts promoted by different content of B, and found that the addition of an appropriate amount of B could improve the dispersion of Cu species and prevent the agglomeration of copper particles at high temperature [34]. With the increase of B content, the content of Cu + showed a rising trend, when Cu/B = 6.6, the content of Cu + was the highest, and the activity and stability of the Cu-B/SiO 2 catalyst was much higher than that of the conventional Cu/SiO 2 catalyst [34].
In industry, BDO, GBL and THF are co-produced by the hydrogenation of DMM with the yield of THF the highest [31], and the yield of different products was controlled by varying the reaction conditions according to the market. How to further improve the activity of supported Cu catalyst and regulate the selectivity of DMM hydrogenation products controllably is a hot research topic. Guo et al. used Cu-B/γ-Al 2 O 3 catalysts to obtain 100% DMM conversion at 533 K, 5 Mpa, H 2 /DMM 120 and LHSV (Liquid hourly space velocity) 0.36 h −1 [7]. Wang et al. enhanced the DMM conversion through adsorption precipitation compared to impregnation [37]. What is more, Chen et al. reported that Cu/SBA-15 could reach 100% DMM conversion in dimeathyl maleate hydrogenation [38]. However, the study on the catalytic performance of Cu/SiO 2 catalyst modified with alkaline additives in DMM hydrogenation is quite incomplete. In this paper, the basicity of the SiO 2 vector was modulated by adding the basic additives (Mg, Ca, Sr, Ba, La), and the effect of different alkali metals on the dispersion of Cu in Cu/SiO 2 catalyst prepared by the AE method was explored. Furthermore, the effects of Cu dispersion and the basicity of the supports on the catalytic performance of hydrogenation of DMM were studied.
The surface area, pore volume and pore size of SiO 2 , Cu/SiO 2 and Cu-M/SiO 2 were characterized by N 2 physical adsorption, as shown in Table 1. After loading active component Cu and metal promoter M, the surface area and pore volume of SiO 2 were decreased, due to the large loading amount of Cu and M. However, the average pore diameters of Cu/SiO 2 and Cu-M/SiO 2 were both larger than that of Catalysts 2019, 9, 704 3 of 10 pure supporter SiO 2 , which should be due to the preparation method (AE) of Cu/SiO 2 and Cu-M/SiO 2 for the dissolution of partial SiO 2 during the preparation under high temperature (363 K) and basic condition. Then, Si(OH) 4 formed by the dissolution of SiO 2 in the basic solution was further reacted with neutral Cu(OH) 2 (H 2 O) 4 in solution to form phyllosilicate [31,39]. In order to investigate the effect of adding alkaline additives on the dispersion of Cu species on the catalyst, in situ XRD characterizations of the calcined catalysts were carried out, and the result was shown in Figure 1. The peak at 2θ = 22.5 • is attributed to the amorphous SiO 2 , and the peaks at about 2θ = 43.2 • and 50.5 • are attributed to Cu. By comparing in situ XRD patterns of Cu-M/SiO 2 and Cu/SiO 2 , it was found that the peaks of CuO of Cu-M/SiO 2 (M selected from Mg, Ca, Sr or Ba) at 2θ = 43.2 • and 50.5 • were much sharper than the peaks of Cu/SiO 2 with no alkaline additives, which indicated that adding alkaline additives as Mg, Ca, Sr or Ba led to much larger Cu particles. However, there was no obvious Cu diffraction peaks at 2θ = 43.2 • and 50.5 • in the in situ XRD patterns of Cu-La/SiO 2 modified by La, suggesting the Cu particles were too small to detect by in situ XRD. This conclusion is consistent with the result of Cu-M/SiO 2 samples detected by XRD in Figure S1. We think that the effect of the reduction process has little effect on the dispersion of Cu/SiO 2 catalyst. Therefore, the Cu dispersion of Cu-La/SiO 2 is higher than that of unmodified Cu/SiO 2 . The addition of La increased the dispersion of Cu species on SiO 2 , resulting in an increase in the active Cu species per unit area.
Catalysts 2019, 9, x FOR PEER REVIEW 3 of 10 temperature (363 K) and basic condition. Then, Si(OH)4 formed by the dissolution of SiO2 in the basic solution was further reacted with neutral Cu(OH)2(H2O)4 in solution to form phyllosilicate [31,39]. In order to investigate the effect of adding alkaline additives on the dispersion of Cu species on the catalyst, in situ XRD characterizations of the calcined catalysts were carried out, and the result was shown in Figure 1. The peak at 2θ = 22.5° is attributed to the amorphous SiO2, and the peaks at about 2θ = 43.2° and 50.5° are attributed to Cu. By comparing in situ XRD patterns of Cu-M/SiO2 and Cu/SiO2, it was found that the peaks of CuO of Cu-M/SiO2 (M selected from Mg, Ca, Sr or Ba) at 2θ = 43.2° and 50.5° were much sharper than the peaks of Cu/SiO2 with no alkaline additives, which indicated that adding alkaline additives as Mg, Ca, Sr or Ba led to much larger Cu particles. However, there was no obvious Cu diffraction peaks at 2θ = 43.2° and 50.5° in the in situ XRD patterns of Cu-La/SiO2 modified by La, suggesting the Cu particles were too small to detect by in situ XRD. This conclusion is consistent with the result of Cu-M/SiO2 samples detected by XRD in Figure S1. We think that the effect of the reduction process has little effect on the dispersion of Cu/SiO2 catalyst. Therefore, the Cu dispersion of Cu-La/SiO2 is higher than that of unmodified Cu/SiO2. The addition of La increased the dispersion of Cu species on SiO2, resulting in an increase in the active Cu species per unit area.  H 2 -TPR tests of Cu-M/SiO 2 and Cu/SiO 2 were carried out to investigate the changes in the interaction between the active Cu species and the support forces upon the addition of alkaline additives ( Figure 2). After the basic additive (except La) was added, the reduction peak of Cu-M/SiO 2 was shifted slightly towards the low temperature. According to the results of in situ XRD analysis, the addition of basic additives (except La) decreased the dispersity of Cu species and increased its particle size, which also reduced the interaction between the Cu particles and the SiO 2 support, leading to a decrease in the reduction temperature of the catalyst when the alkaline additive (except La) was added. However, for Cu-La/SiO 2 catalyst with La additive, the reduction temperature of H 2 -TPR was slightly higher than that of unmodified Cu/SiO 2 , which suggested that the addition of La enhanced the interaction of Cu species with the SiO 2 support, thus making Cu species more difficult to be reduced. However, for Cu-La/SiO 2 with La promoter, the reduction temperature of H 2 -TPR was slightly higher than that of unmodified Cu/SiO 2 , which indicated that the addition of La enhanced the interaction between Cu species and carriers and made the Cu species more difficult to be reduced. Similar phenomena have also been found in La modified Cu/SiO 2 investigated by Zheng et al., due to the formation of a Cu-O-La bond at the interface between LaO x and Cu species, which improved the stability of the catalyst [40]. In addition, with the characterization of in situ XRD, smaller CuO particles in Cu-La/SiO 2 were detected, and this also led to the stronger interaction between Cu and the carrier, thus increasing the reduction temperature of Cu-La/SiO 2 . What is more, the ratio of different catalysts' peak area is 1. H2-TPR tests of Cu-M/SiO2 and Cu/SiO2 were carried out to investigate the changes in the interaction between the active Cu species and the support forces upon the addition of alkaline additives (Figure 2). After the basic additive (except La) was added, the reduction peak of Cu-M/SiO2 was shifted slightly towards the low temperature. According to the results of in situ XRD analysis, the addition of basic additives (except La) decreased the dispersity of Cu species and increased its particle size, which also reduced the interaction between the Cu particles and the SiO2 support, leading to a decrease in the reduction temperature of the catalyst when the alkaline additive (except La) was added. However, for Cu-La/SiO2 catalyst with La additive, the reduction temperature of H2-TPR was slightly higher than that of unmodified Cu/SiO2, which suggested that the addition of La enhanced the interaction of Cu species with the SiO2 support, thus making Cu species more difficult to be reduced. However, for Cu-La/SiO2 with La promoter, the reduction temperature of H2-TPR was slightly higher than that of unmodified Cu/SiO2, which indicated that the addition of La enhanced the interaction between Cu species and carriers and made the Cu species more difficult to be reduced. Similar phenomena have also been found in La modified Cu/SiO2 investigated by Zheng et al., due to the formation of a Cu-O-La bond at the interface between LaOx and Cu species, which improved the stability of the catalyst [40]. In addition, with the characterization of in situ XRD, smaller CuO particles in Cu-La/SiO2 were detected, and this also led to the stronger interaction between Cu and the carrier, thus increasing the reduction temperature of Cu-La/SiO2. What is more, the ratio of different catalysts' peak area is 1.  In order to further explore the acid-base property of the catalyst surface after adding basic promoters, the CO2-TPD of Cu/SiO2 and Cu-M/SiO2 catalysts were studied. As shown in Figure 3, both Cu/SiO2 and Cu-M/SiO2 catalysts have a significant CO2 desorption peak at 400~500 K in the In order to further explore the acid-base property of the catalyst surface after adding basic promoters, the CO 2 -TPD of Cu/SiO 2 and Cu-M/SiO 2 catalysts were studied. As shown in Figure 3, both Cu/SiO 2 and Cu-M/SiO 2 catalysts have a significant CO 2 desorption peak at 400~500 K in the CO 2 -TPD profiles, and the position and intensity of the desorption peak are very close. However, a CO 2 desorption peak of the Cu-M/SiO 2 catalyst is observed at 500~580 K, and the intensity of the peak is less than that of the main desorption peak at 400~500 K. In CO 2 -TPD, the higher the temperature of Catalysts 2019, 9, 704 5 of 10 desorption peak of CO 2 means the stronger the basicity of the catalyst; therefore, the addition of basic additives really increased the basicity of the catalyst surface. It is also found that the desorption peak of CO 2 from (b) to (f) catalyst shifts slightly towards low temperature, indicating that the basicity of the catalyst surface is Cu-Mg/SiO 2 > Cu-Ca/SiO 2 > Cu-Sr/SiO 2 > Cu-Ba/SiO 2 > Cu-La/SiO 2 . The actual order of basicity of the oxides is: MgO < CaO < SrO < BaO < La 2 O 3 [41], which is contrary to the basicity of the catalyst surface. The reason should be the basic promoters in the calcined Cu-M/SiO 2 not only exist in the form of oxides, but also may exist in some form similar to Cu-O-M or M-O-Si [42].
Catalysts 2019, 9, x FOR PEER REVIEW 5 of 10 CO2-TPD profiles, and the position and intensity of the desorption peak are very close. However, a CO2 desorption peak of the Cu-M/SiO2 catalyst is observed at 500~580 K, and the intensity of the peak is less than that of the main desorption peak at 400~500 K. In CO2-TPD, the higher the temperature of desorption peak of CO2 means the stronger the basicity of the catalyst; therefore, the addition of basic additives really increased the basicity of the catalyst surface. It is also found that the desorption peak of CO2 from (b) to (f) catalyst shifts slightly towards low temperature, indicating that the basicity of the catalyst surface is Cu-Mg/SiO2 > Cu-Ca/SiO2 > Cu-Sr/SiO2 > Cu-Ba/SiO2 > Cu-La/SiO2. The actual order of basicity of the oxides is: MgO < CaO < SrO < BaO < La2O3 [41], which is contrary to the basicity of the catalyst surface. The reason should be the basic promoters in the calcined Cu-M/SiO2 not only exist in the form of oxides, but also may exist in some form similar to Cu-O-M or M-O-Si [42].

Hydrogenation of DMM
The hydrogenation of DMM was carried out in a fixed bed reactor over unmodified Cu/SiO2 and basic metal modified Cu-M/SiO2 (M = Mg, Ca, Sr, Ba, La) to evaluate the influence of basic promoters to the catalytic performance of Cu-M/SiO2 catalysts. The conversion of DMM and the selectivity of the products were shown in Table 2.

Hydrogenation of DMM
The hydrogenation of DMM was carried out in a fixed bed reactor over unmodified Cu/SiO 2 and basic metal modified Cu-M/SiO 2 (M = Mg, Ca, Sr, Ba, La) to evaluate the influence of basic promoters to the catalytic performance of Cu-M/SiO 2 catalysts. The conversion of DMM and the selectivity of the products were shown in Table 2. The conversion rate of DMM reached 100% over the unmodified Cu/SiO 2 catalyst showed its good activity, and THF as the only hydrogenated product was observed for the weak alkalinity of the Cu/SiO 2 surface. What is more, when Cu/SiO 2 exhibits low activity in high LHSV, the THF selectivity Catalysts 2019, 9, 704 6 of 10 still maintains 100%. Therefore, we think high LHSV has little effect on selectivity. After the basic additive (except La) was added to Cu/SiO 2 catalyst, the catalytic activity decreased significantly for the conversion of DMM over Cu-M/SiO 2 (M = Mg, Ca, Sr, Ba) did not reach 100%, especially the conversion over Cu-Ba/SiO 2 just 64.21%. However, the catalytic activity of La modified Cu-La/SiO 2 was higher, and the conversion rate of DMM was also 100%.
The activity of Cu-M/SiO 2 (M = Mg, Ca, Sr, Ba, La) in the hydrogenation of DMM should be related to the dispersion of Cu on the surface of Cu-M/SiO 2 . The higher dispersion of Cu on Cu-La/SiO 2 led to the higher catalytic activity for the hydrogenation of DMM, and the other Cu-M/SiO 2 (M = Mg, Ca, Sr, Ba) with the lower dispersion of Cu than that of unmodified Cu/SiO 2 catalyst caused that DMM had not been completely converted under the same reaction condition. The addition of basic promoter to Cu/SiO 2 catalyst varied the selectivity of the hydrogenation products. There was just THF as the only product during the hydrogenation of DMM over the unmodified Cu/SiO 2 . After modified by basic promoter, the selectivity of THF decreased, and more GBL was obtained. In addition, a small amount of BDO was generated. One possible reason for the low selectivity of THF over modified catalysts was the covering of the part of the acid sites on the surface of SiO 2 by the basic additives leading to the weakening of the dehydration of BDO to THF under acidic conditions. Moreover, we have tested the Cu-La/SiO 2 catalyst for 100 h stability test. As shown in Table 3, We found the catalyst still maintain stable activity and selectivity.

Catalysts Preparation
Preparation of Cu/SiO 2 : 3.78 g Cu(NO 3 ) 2 ·3H 2 O was dissolved in 40 mL deionized water, and stirred at room temperature for 10 min until it is completely dissolved, then 11.5 mL ammonia (28 wt%) was added to obtain the solution of copper ammonia with stirring for 30 min. Added 4 g SiO 2 into the solution and stirred for 240 min. The mixture was quickly moved to the oil bath at 363 K to evaporate ammonia until the pH to 6.5, and then it was cooled to room temperature and filtered. The filter cake was washed by deionized water for three times followed with drying at 393 K for 480 min and calcination at 723 K for 240 min. The final catalyst was denoted as 25 wt% Cu/SiO 2 .
Preparation of Cu-M/SiO 2 : A certain amount of M nitrate (M selected from Mg, Ca, Sr, Ba, or La) containing M 0.2 g was dissolved in 40 mL deionized water. 4 g SiO 2 was added into the solution. The mixture was filtered after had been stirred for 600 min, followed by drying at 393 K for 480 min and calcination at 723 K for 240 min. The modified carrier was denoted as M/SiO 2 . 3.78 g Cu(NO 3 ) 2 ·3H 2 O was dissolved in 40 mL deionized water, and stirred at room temperature for 10 min until it is completely dissolved, then 11.5 mL ammonia (28 wt%) was added to obtain the solution of copper ammonia with stirring for 30 min. Added the above made M/SiO 2 into the solution and stirred for 240 min. The mixture was quickly moved to the oil bath at 363 K to evaporate ammonia until the pH to 6.5, and then it was cooled to room temperature and filtered. The filter cake was washed by deionized water for three times followed with drying at 393 K for 480 min and baking at 723 K for 240 min. The final catalyst was denoted as 25wt% Cu-5wt% M/SiO 2 .

Catalyst Characterization
N 2 physical adsorption: The surface area, pore volume and pore diameter of the catalyst were determined by N 2 physical adsorption at 77 K using Micromeritics ASAP 2020 (Micromeritics, Hangzhou, China). Firstly, the sample was heated to 573 K under vacuum condition for 6 h to remove the adsorbed species, and then N 2 physical adsorption isotherm was performed. The surface area of the sample was calculated by the BET equation, and the pore volume and pore size distribution of the catalyst were obtained by BJH theory.
In situ XRD: X-ray powder diffraction patterns of the samples were obtained in the scanning angle (2θ) range of 10-80 • under H 2 /Ar atmosphere at 573 K on a Thermo ARL SCINTAG X-TRA (Thermo, Hangzhou, China) using Cu Kα1 radiation (λ = 1.5406 Å) operated at 40 kV and 40 mA. Because of the in situ measurement system, the peak positions had some fluctuations compared with Standard XRD patterns.
The elemental content of the catalysts was detected using XRF (Thermo ARL ADVANT'X, Thermo, Hangzhou, China). H 2 temperature-programed reduction (H 2 -TPR)-0.1 g calcined catalyst was loaded into a quartz tube, and then a mixture of 5% H 2 /Ar was passed into the tube with the flow rate at 30 mL/min, while the exhaust gas was detected by a thermal conductivity detector (TCD). After the baseline was stable, the following heating program was carried out: retaining at 303 K for 10 min, and then rising to 1123 K for 82 min. CO 2 temperature-programed desorption (CO 2 -TPD)-0.1 g calcined catalyst was loaded into a U-tube, and treated at 723 K for 1 h in an Ar atmosphere to remove the adsorbed species. Then the temperature dropped to 373 K, and CO 2 was passed into the U-tube. After CO 2 adsorbed at 373 K for 30 min, the atmosphere was switched to Ar, and the exhaust gas was detected by a mass spectrum (MS). After the baseline was stable, the following heating program was carried out: rising to 773 K from 373 K for 40 min, and then retaining at 773 K for 10 min.

Hydrogenation Reaction
The hydrogenation reaction was carried out in a fixed bed reactor (Tianjin Pengxiang Technology Co. Ltd., Tianjin, China) with the diameter of 10 mm and the length of 500 mm and the constant temperature zone length ≥100 mm. The catalyst (20-40 mesh) was loaded into the middle of the constant temperature zone, and activated at 573 K with H 2 (30 mL/min) for 240 min. Then the temperature dropped to 513 K. H 2 was introduced up to the desired pressure for 5 MPa and the raw material (20 vol% DMM in methanol) was pumped into a vaporizer using a high-pressure pump (Series II), and then reacted with H 2 . The reaction products were quantitatively analyzed by gas chromatography (Agilent, GC, 7890A, Santa Clara, CA, USA) with DB-1 chromatographic column and FID (flame ionization detector) detector. In the reaction, methane and methol are also generated. However, methane was not