Synthesis of Acetic Acid on Pd-H 4 SiW 12 O 40 -based Catalysts by Direct Oxidation of Ethylene

Journal of Natural Gas Chemistry 11(2002)51 56 Synthesis of Acetic Acid on Pd-H 4 SiW 12 O 40 -based Catalysts by Direct Oxidation of Ethylene Xinping Wang, Kegong Fang, Jianlu Zhang, Tianxi Cai State Key Laboratory of Fine Chemicals, Dalian University of Technology, Dalian 116012, China [Manuscript received April 16, 2002; revised May 16, 2002] Abstract: Synthesis of acetic acid by direct oxidation of ethylene on Pd-H 4SiW 12O 40-based catalysts was studied in a fixed-bed integral reactor and a pulse differential reactor. From the performance of the catalysts with different compositions and configurations, it is proposed that acetic acid is predominantly produced via an intermediate of acetaldehyde. This can be easily confirmed by comparing the product distributions in the integral and the differential reactors. The active sites for acetic acid formation are considered to exist mainly at the boundaries between the H 4SiW 12O 40 and the Pd particles. The Pd-based catalysts reduced by H 2/N 2 have higher activities than those reduced by hydrazine, as explained by the degree of Pd dispersion obtained from the characteristics of hydrogen chemical adsorption. It was found that the Pd-Se-SiW 12/SiO 2 catalyst with selenium tetrachloride as a precursor was more active than that with potassium selenite, and that the acetic acid yield can be greatly increased by adding a suitable amount of dichloroethane (C 2H 4Cl 2/C 2H 4 mole ratio=0.03) to the reactants. Key words: acetic acid, synthesis, ethylene, Pd-H 4SiW 12O 40/SiO 2, dichloroethane 1. Introduction Acetic acid is one of the most useful organic acids. It is currently manufactured by acetaldehyde oxidation, methanol carbonylation, naphtha oxidation and n-butane oxidation. Recently, a new process for acetic acid production by direct oxidation of ethylene was developed by Suzuki and colleagues [1 3], using a Pd- Se-H 4 SiW 12 O 40 /SiO 2 catalyst. This process has attracted a great deal of attention because of its advantages, like the simplicity of the process and the low cost of the equipment [4]. It was indicated in our earlier report that the reduction method of the Pd component has a great influence on the activity of the Pd-H 4 SiW 12 O 40 /SiO 2 catalyst, and the reduction method with H 2 is better than that with hydrazine [5]. In this paper, we studied the effects of Pd and H 4 SiW 12 O 40 (SiW 12 ) contents on the catalyst performance and attempted to clarify their individual role in the reaction. Furthermore, we also examined the mechanism of acetic acid formation on the catalyst by comparing the product distributions between the integral and differential reactors. In order to increase the yield of acetic acid in the reaction, different Se precursors for modifying the catalyst and the effect of dichloroethane in the reaction, were also been investigated. 2. Experimental 2.1. Catalyst preparation Pd/SiO 2 and Pd-SiW 12 /SiO 2 catalysts were prepared by the incipient wetness method. Silica (310 m 2 /g) was treated with 5 wt% chlorhydric acid, washed with distilled water and dried at 393 K for 3 h prior to use as the support. The silica support was impregnated with an aqueous solution of hydrochloropalladous acid (H 2 PdCl 4 ) or a solution con- Corresponding author. Tel: 0411-3631333-3247.

52 Xinping Wang et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2 2002 taining H 2 PdCl 4 and tungstosilica acid (SiW 12 ) for 12 h and dried in air at 393 K for 2 h. Then, the sample was reduced in a liquid phase by hydrazine at 368 K, or in a flowing gas of 30% H 2 /N 2 (50 ml/min) at 573 K for 1 h, with temperature programmed heating from 323 K to 573 K at 1 K/min. Pd-Se-SiW 12 /SiO 2 catalysts were obtained by incipient wetness impregnation using a soluble selenium salt. 2.2. Activity measurements The catalytic reaction was carried out in a stainless steel fixed-bed flow reactor (i.d. 8.0 mm) at 423 K under 0.5 MPa pressure. The feed gas containing C 2 H 4, O 2, and H 2 O (gas) in N 2 was introduced into the reactor loaded with 2 cm 3 catalysts (40 20 meshes) at a total flow rate of 107 cm 3 min 1. The outlet gas and liquid products gathering at the steady state were analyzed by gas chromatography with TCD (using a GDX-502 packed column) and FID (using a GDX-103 packed column) respectively. All of the catalytic results were an average over the reaction time of 4 5 hours. The selectivity is defined as fractions of the sum of the products on the ethylene basis. To investigate the mechanism of acetic acid formation on the catalyst, the reaction was also carried out in a differential reactor (i.d. 3.5 mm) using 20 mg of 60 80 mesh catalyst, which was mixed with 200 mg of quartz sand of the same mesh. During the reaction, 1 ml of the reactant gas (containing O 2 and water vapor) was pulsed into the carrier gas (N 2 ) in front of the catalyst, and the products were analyzed online by gas chromatography with FID. In the reactions using the differential reactor, the deep-oxidation product CO 2 could not be detected, due to restrictions of the FID. 2.3. Catalyst Characterization The dispersion of Pd on the catalyst was measured by a Micromeritics CHEMISORB 2000 using H 2 as the adsorbent, assuming that, on average, two hydrogen atoms were adsorbed on one Pd atom. 3. Results and discussion 3.1. Effect of Pd and SiW 12 loading on the activity On all catalysts containing SiW 12, a small amount of ethanol was detected which was almost unchanged by the catalyst composition. When SiW 12 /SiO 2 was used as the catalyst in the reaction, no obvious amount of oxidation products were detected, except for a small amount of CO 2. However, on the Pd- SiW 12 /SiO 2 catalyst, as shown in Table 1, with an increase in the Pd content both the ethylene conversion and the acetic acid (AcOH) yield per hour continually increased. Table 1. Dependence of the catalytic performance of the Pd-SiW 12 /SiO 2 on Pd loading Pd/SiO 2 Conversion of Selectivity (%) Yield of mass ratio ethylene (%) AcOH AcH CO 2 AcOH (g/(l h)) 0.005 1.6 61.7 26.4 13.7 39.4 0.010 3.8 73.7 12.7 13.6 114.4 0.015 5.0 76.6 9.9 13.5 154.7 0.020 7.7 82.7 5.7 11.6 257.4 0.025 8.9 85.4 4.4 10.2 305.7 0.030 9.8 85.2 3.3 11.5 335.5 0.035 10.7 85.4 2.1 12.5 368.6 Reaction conditions: feed gas N 2 /O 2 /C 2 H 4 /H 2 O volume ratio=13/7/50/30, GHSV=3,000 h 1, p=0.5 MPa, T=423 K, catalyst 2 ml (SiW 12 /SiO 2 mass ratio=0.3). These results indicate that the Pd in the catalyst is the primary active component for catalyzing the ethylene partial oxidation. From Figure 1 we notice that the selectivity to acetaldehyde (AcH) is almost inversely proportional to the Pd loading in the catalyst. This is accompanied by an increase in the selectivity to acetic acid, while the total selectivity to acetic acid and acetaldehyde was nearly constant.

Journal of Natural Gas Chemistry Vol. 11 No. 1 2 2002 53 Figure 1. Catalytic performance variation with Pd loading on the Pd-SiW 12 /SiO 2 catalyst(siw 12 /SiO 2 =0.3). The reaction conditions are the same as in Table 1. Thus, we can suppose that AcOH is mainly produced from the AcH intermediate on the H 4 SiW 12 O 40 /SiO 2 catalyst. On the other hand, for Pd loading ranging from 0.005 to 0.035 grams of Pd per gram of SiO 2 support, the catalysts showed almost no change in selectivity to CO 2, though the AcOH yield per hour sharply and linearly increased. This experimental phenomenon led us to consider that CO 2 is predominantly formed from a deep-oxidation of ethylene and not from a secondary oxidation of AcH and AcOH. Table 2 illustrates that although acetic acid could be produced on the Pd/SiO 2 catalyst, the combustion reaction producing CO 2 was predominant. Furthermore, the selectivity to AcOH was substantially increased by adding SiW 12 to the catalyst. At the same time, the AcOH yield per hour was markedly enhanced as well. For example, the addition of 0.2 g SiW 12 per gram of SiO 2 support increased the selectivity of AcOH from 30.8% to 80.9%, accompanied by an increase of the AcOH yield from 50.9 g/(l h) to 248.1 g/(l h), while the selectivity to CO 2 decreased from 67.7% to 13.7%. These results indicate that the role of SiW 12 in the Pd-SiW 12 /SiO 2 catalyst is not only to inhibit the combustion reaction occurring on the Pd particles, but also to produce new active sites on which ethylene can be converted to AcOH when combined with the Pd component. It is worth noting that no obvious difference in catalytic activity could be observed between the Pd-SiW 12 /SiO 2 catalysts prepared by impregnating the Pd/SiO 2 catalyst with a SiW 12 solution and that prepared by the coimpregnation method mentioned in the experimental section. So the promoting action of the SiW 12 on the Pd-SiW 12 /SiO 2 catalyst cannot attributed to the difference in dispersion and configuration between the Pd/SiO 2 catalyst and the Pd-SiW 12 /SiO 2 catalyst. Table 2. Catalytic performance of Pd-SiW 12 /SiO 2 with regard to SiW 12 loading SiW 12 /SiO 2 Conversion of Selectivity (%) Yield of mass ratio ethylene (%) AcOH AcH CO 2 AcOH (g/(l h)) 0 4.1 30.8 1.5 67.7 50.9 0.1 5.2 72.8 5.0 22.2 151.3 0.2 7.6 80.9 5.4 13.7 248.1 0.3 7.7 82.7 5.7 11.6 257.4 0.4 6.2 82.6 6.6 10.8 209.7 *Mass ratio of Pd/SiO 2 =0.0420, the reaction conditions are the same as in Table 1.

54 Xinping Wang et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2 2002 On the Pd-SiW 12 /SiO 2 catalyst, the AcOH yield per hour gradually increased with the SiW 12 loading, reached a maximum of 257.4 g/(l h) at the SiW 12 loading of 0.3 g per gram of SiO 2, and then declined due to the covering of the active sites by an overloading of SiW 12. 3.2. The main route of AcOH formation from ethylene on the Pd-SiW 12 /SiO 2 catalyst In order to examine the speculation that AcOH is mainly produced from the AcH intermediate on the Pd-SiW 12 /SiO 2 catalyst, several reaction runs were carried out in a differential reactor. As shown in Table 3, when the reaction was carried out in the differential reactor and ethylene was used as the feed gas, a selectivity of AcH as high as 98.4% was obtained, which was remarkably different from the 8.1% obtained in the integral reactor. When AcH instead of ethylene was pulsed into the differential reactor under similar conditions, all reactants were converted to AcOH. However, when ethanol was pulsed, 50.4% of it was converted to ethylene, while 49.3% was converted to AcH. These results indicate that, on the Pd- SiW 12 /SiO 2 catalyst, AcOH is mainly produced from ethylene via the intermediate of AcH rather than via ethanol. Thus, the disparity in AcH selectivity obtained in the integral reactor and in the differential reactor when C 2 H 4 was used as the feed gas can be reasonably explained by the intermediate AcH produced at the front part of the catalyst bed in the integral reactor further reacting at the rear part, and most of it eventually being converted to AcOH before flowing out of the catalyst bed. Therefore, the selectivity to AcH is much lower than that obtained in the differential reactor. Table 3. Product distributions in the differential reactor and the integral reactor Reactor Reactant Selectivity (%) AcH AcOH CO 2 C 2 H 4 C 2 H 5 OH Integral a C 2 H 4 8.1 77.6 14.2 0.1 Differential b C 2 H 4 98.4 0 c <1.6 Differential b AcH 100 c 0 0 Differential b C 2 H 5 OH 49.3 <0.15 c 50.4 a: The reaction conditions are the same as in Table 1. b: The reactant was pulsed into the carrier gas before entering the reactor at 423 K. c: Data could not be taken. 3.3. Effect of preparation method and precursor The activity of the Pd-Se-SiW 12 /SiO 2 catalyst was strongly affected by the precursor of selenium as well as the reduction method. As shown in Table 4, the Pd-Se-SiW 12 /SiO 2 catalyst using selenium tetrachloride as the precursor and reduced by H 2 /N 2 had a very high yield of AcOH, i.e., 308.0 g/(l h), which exceeded the best result of 240.0 g/(l h) reported by Suzuki [1], while the catalyst with the same composition but reduced by hydrazine had a lower AcOH yield of 181.1 g/(l h). The regularity is practically the same as that demonstrated on the Pd- SiW 12 /SiO 2 catalyst, indicating that H 2 /N 2 is better than hydrazine for reducing the Pd component in the catalyst. To understand the regularity, the Pd- SiW 12 /SiO 2 catalysts prepared by the two methods were characterized in terms of the degree of Pd dispersion based on hydrogen chemisorption. Although the two catalysts have the same composition, the one reduced by H 2 /N 2 had a Pd dispersion which was about 3 times as high as that reduced by hydrazine, which can be well correlated with its relative activity in terms of AcOH yield per hour. On the other hand, the Pd-Se-SiW 12 /SiO 2 and Pd-SiW 12 /SiO 2 catalysts reduced by H 2 both exhibited higher selectivity to AcOH than those reduced by hydrazine. This phenomenon can also be attributed to the difference in Pd dispersion on both catalysts and the poor selectivity to AcOH of the Pd/SiO 2 catalyst. On the surface of the larger Pd particles, the ethylene combustion reaction predominantly occurred due to a lower dispersion and the poor contact between the SiW 12 and the Pd surface. A similar trend befell the Pd/SiO 2 catalyst. On the contrary, the one with a better Pd dispersion may possess more active sites at the boundaries between the SiW 12 and Pd particles, thus exhibiting higher activity for AcOH formation.

Journal of Natural Gas Chemistry Vol. 11 No. 1 2 2002 55 The activity of the Pd-Se-SiW 12 /SiO 2 catalyst noticeably declined when using potassium selenite as the precursor of the selenium component, even though it was reduced by H 2 /N 2. This can be ascribed to the alkalinity of the potassium selenite, which weakened the strong acidity of the SiW 12, or even destroyed the Keggin ion structure. Therefore, the method of using hydrogen reduction for the palladium component and the introducing of selenium with selenium tetrachloride are better than other methods reported in the literature [1,2] in which the Pd-Se-SiW 12 /SiO 2 catalyst was prepared using potassium selenite as a precursor and using hydrazine as the reducing agent. Reductant Table 4. Activity of Pd-Se-SiW 12 /SiO 2 and Pd-SiW 12 /SiO 2 catalysts prepared by different methods Precursor Yield of Selectivity (%) of Se AcOH (g/(l h)) AcOH AcH CO 2 NH 2 NH 2 SeCl 4 181.1 71.0 17.1 11.9 H 2 SeCl 4 308.0 80.5 8.2 11.3 H 2 K 2 SeO 3 118.4 65.9 13.2 20.9 NH 2 NH 2 K 2 SeO 3 240.0 a 86.4 8.1 5.1 H 2 257.4 b 82.7 5.7 11.6 NH 2 NH 2 88.9 c 64.6 10.4 25.0 a: Reported by Suzuki [1]. b: The Pd-SiW 12 /SiO 2 catalyst had a Pd dispersion (D Pd ) of 5%. c: The Pd-SiW 12 /SiO 2 catalyst had a D Pd of 1.8%. 3.4. Effect of dichloroethane as an additive to the reactants on acetic acid formation We discovered that the yield and selectivity of acetic acid could be increased by adding a suitable concentration of dichloroethane into the ethylene feed. As shown in Table 5, when 0.03% of the additive in volume was added to the reactant, the acetic acid yield per hour increased from 318.5 g/(l h) to 368.3 g/(l h), and the selectivity increased from 84.3% to 88.8%. However, no better results were observed by increasing the concentration of dichloroethane in ethylene. Kluckovsky et al. [6], while studying the effect of dichloroethane on the selective oxidation of ethylene, determined that a trace amount of dichloroethane in ethylene could greatly increase the selectivity of the desired product, epoxy ethane. In the process of acetic acid formation via partial oxidation of ethylene, the chlorine atoms might suppress CO 2 formation by modifying the oxygen species, just as in the case of epoxy ethane synthesis. Studies to discern the mechanism of promoting acetic acid formation by dichloroethane are in progress. Table 5. Effect of dichloroethane as an additive to the reactants on acetic acid formation* C 2 H 4 Cl 2 /C 2 H 4 Conversion of Yield of Selectivity (%) (Vol%) C 2 H 4 (%) AcOH (g/(l h)) AcOH AcH CO 2 0 9.6 318.5 84.3 5.7 10.0 0.03 10.3 368.3 88.8 4.1 7.1 0.06 10.2 357.1 88.3 5.2 6.5 0.09 9.8 334.2 87.8 6.2 6.0 *On catalyst Pd-SiW 12 /SiO 2 being reduced by H 2 /N 2 with composition of Pd/SiW 12 /SiO 2 mass ratio= 0.02/0.3/1. Reaction conditions: N 2 /O 2 /C 2 H 4 /H 2 O volume ratio=13/7/50/30, GHSV=3,000 h 1, p =0.7 MPa, T =433 K.

56 Xinping Wang et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2 2002 4. Conclusions For the synthesis of acetic acid by direct oxidation of ethylene on Pd-H 4 SiW 12 O 40 -based catalysts, the following conclusions can be derived from our experimental results: (1) Comparing the product distributions from an integral and a differential reactor illustrated that acetic acid is mainly produced via the intermediate of AcH rather than via ethanol. The main reaction proceeds predominantly on active sites containing Pd and SiW 12, which exist at the boundaries between the SiW 12 and Pd particles. (2) The Pd based catalyst reduced by H 2 /N 2 has a higher Pd dispersion, giving the Pd particles better contact with the SiW 12, and is more active than that reduced by hydrazine. (3) The Pd-Se-SiW 12 /SiO 2 catalyst prepared by using selenium tetrachloride as a precursor is more active than that obtained by using potassium selenite. The addition of a suitable amount of dichloroethane (C 2 H 4 Cl 2 /C 2 H 4 mole ratio =0.03) to the reactant can also greatly increase the acetic acid yield. References [1] Suzuki T. JP Patent 7 089 896. 1995 [2] Sano K, Suzuki T, Azuma T et al. JP Patent 9 067 298. 1997 [3] Sano K, Suzuki T, Uchida H. Catal Catal, 1999, 41: 86 [4] Sano K, Suzuki T, Uchida H. Catal Catal, 1999, 41: 290 [5] Fang K G, Wang X P, Zhang J L, Cia T X. Chin Chem Lett, 2001, 12(2) : 125 [6] Jin S S. Organic Catalysis, Shanghai: Shanghai Sci Technol Press, 1986. 198