Catalyst of Palladium Supported on H-Beta Zeolite with Nanosized Al2O3 for Isomerization of n-heptane

Transcription

1 Journal of the Japan Petroleum Institute, 55, (2), (2012) 99 [Regular Paper] Catalyst of Palladium Supported on H-Beta Zeolite with Nanosized Al2O3 for Isomerization of n-heptane Toshiyuki KIMURA, Chie SUEZAKI, Koji SAKASHITA, Xiaohong LI, and Sachio ASAOKA School of Environmental Engineering, The University of Kitakyushu, 1-1 Hibikino, Wakamatsu-ku, Kitakyushu , JAPAN (Received July 25, 2011) To investigate the synthesis of isoparaffin from heavy naphtha, n-heptane conversion was studied. Applying and improving the isomerization catalyst developed for n-hexane was examined for n-heptane. The developed catalyst, Pd/nano-sized (defined as 5-50 nm, and here called as ns) Al 2O 3/H-BEA zeolite, was effective for the isomerization of both n-hexane and n-heptane, but n-heptane was more easily decomposed in comparison with n-hexane. In the case of n-heptane, cracking product selectivity was extremely high at about 85 % at 300. However, improved high activity and selectivity were obtained by removal of residual chlorine from the catalyst, which decreased the number of acid sites acting as cracking sites formed by residual chlorine on ns Al 2O 3. If the content of ns Al 2O 3 combined to H-BEA zeolite changed in the catalyst, cracking selectivity remained constant at a lower level comparing with the non-combined zeolite. ns Al 2O 3 also reduced the acidity of strong acid sites on the zeolite particle surface. X-ray photoelectron spectroscopy showed that reduced Pd/ns Al 2O 3/H-BEA catalyst with removed residual chlorines had the highest Pd dispersion because the chloride anions act effectively for cationic Pd dispersion on ns Al 2O 3 which can adsorb chloride anions. Since ns Al 2O 3 was also highly dispersed onto the zeolite particle surface, acid sites were formed at the boundary. The catalyst has highly dispersed Pd metal because of the stable Pd was highly dispersed on the acid sites by removing chlorine. Keywords Isomerization, n-heptane, Boundary, Nano-sized oxide, Palladium species, Chlorine removal 1. Introduction Background Inorganic solids suitable for catalysts are porous, especially nanoporous. Therefore, control of the chemical properties of the catalyst at the order of nanometers is important for designing and preparing catalysts with nano-pores. Heavy naphtha consists mainly of heptane, which is also a gasoline fraction, but shipment of gasoline to market has some restrictions. Isomerization of heavy naphtha would convert n-heptane of lower octane value to isoheptanes of higher octane value. Thus, heavy naphtha can be transformed to environmentally-friendly fuel oil which contains less sulfur and almost no aromatics. However, commercial operation is not yet possible because the cracking reaction easily occurs compared to light naphtha fractions like n-hexane. Therefore, a pre-treatment method for removing cracking sites in the catalyst was studied for n-heptane isomerization. Technology has been developed to synthesize isoparaffin from n-heptane over Pt _ SO4 2- /ZrO2 catalysts 1)~8) To whom correspondence should be addressed. kimura@kitakyu-u.ac.jp and Pt or Pd/zeolite catalysts 9)~11). However, these zeolite catalysts cannot use some binders, so mechanical strength of the catalyst is low and metal dispersion is not so high. If hybrid catalysts are used only as a binder, the selectivity for isoparaffin on Pd/SiO2/H-BEA zeolite catalyst is low 9). We studied the n-heptane isomerization reaction over tri-component catalyst, which is considered to have both increased mechanical strength and improved catalytic performance. We have already reported on some tri-component catalysts 16)~22). A catalyst for isomerization from n-hexane to isoparaffin was developed 23). The catalyst consists of Pd on a composite of 35 wt% nano-sized (defined as 5-50 nm, and here called ns) Al2O3 and 65 wt% BEA zeolite (SiO2/ Al2O3 25). Pd has higher activity for isomerization and generation of multi-branched isoparaffin from n- hexane than more expensive Pt. Generally, Pt catalyst is more active than Pd catalyst. However, most of these catalysts do not use binder as ns Al2O3 and the state of the metal precursor is different. Platinum metals are aggregated in the oxidized state, for example in the case of using Pd nitrate. Our catalysts use Pd chloride with HCl, so Pd is dispersed as PdCl4 2- on the support. Therefore, Pd is aggregated in the state of perchlorate. In addition, by using ns Al2O3, Pd is more

2 100 easily dispersed than Pt. Pd catalyst can more effectively isomerize n-paraffins to both mono-branched and multi-branched products. However, since the isomerization mechanism will depend on the carbon number of the n-paraffin, the cracking reaction must be suppressed for isomerization of n-heptane. The negligibly small amount of chlorine sometimes causes remarkable changes in the acidity of the nanoporous catalyst. The acid sites generated by the chlorination may act as undesirable or desirable active sites. If the chlorination causes undesirable catalysis reactions, chlorine removal from the catalyst is required. The conversion of n-paraffin in naphtha to isoparaffin was recently studied using catalysts consisting of Pd, nano-sized oxide and zeolite, and chlorine removal improved the catalyst performance. This factor was also examined for the catalytic isomerization of n-heptane Reaction Mechanism of Isomerization and Function of Metal The nature of the carbenium ion decisively affects on the rate and the selectivity of isomerization. The isomerization mechanism is considered to consist of several steps 13)~15) as follows: alkane becomes alkene by dehydrogenation at metal, alkene becomes carbenium ion by H addition at acid sites of the solid acid catalyst, carbenium ion skeletally isomerizes at the acid site, skeletal isomerized carbenium ion becomes skeletal isomerized alkene by deprotonating, and isomerized alkene becomes iso-type alkanes by hydrogenation at the metal sites. In the case of the di-component catalyst, carbenium ion is generated by H addition to alkene, so strong acid is not required for generation of the carbenium ion. The nature of the acid decisively affects the isomerization process with formation of carbenium ion being the rate-determining step. As alkene is an intermediate in the isomerization reaction mechanism to the isoalkane, alkene is usually not observed in the products. The metal in the di-component catalyst not only dehydrogenates the alkane but also hydrogenates alkene intermediates, so that the production of alkenes is restrained. In the ideal state, dehydrogenation and hydrogenation quickly reach equilibrium. As hydrogenation under hydrogen pressure at low temperatures can restrain the production of alkenes, the production of dimer is also restricted. As a result, coke formation is also suppressed. Accordingly, byproduct production is restrained and the catalyst life is increased. The action of metal and hydrogen is to prevent the decline of activity and to improve the selectivity. Clearly, the dispersion of metal in the catalyst influences the catalytic performance. Therefore, tri-component catalysts composed of zeolites with metal and nano-oxide are expected to be more effective with more suitable dispersion of metal than di-component catalysts. There are nine structural isomers of n-heptane: : isoparaffins; : multi-branched isoparaffins. Fig. 1 Chemical Equilibrium Composition of n-heptane Isomerization n-heptane, 2-methylhexane, 3-methylhexane, 3-ethylpentane, 2,2-dimethylpentane, 2,3-dimethylpentane, 3,3-dimethylpentane and 2,2,3-trimethylbutane. Inner C _ C bonds including quaternary or tertiary carbon are unstable and easy to decompose if the bonds are located at the β position. n-c3h8, C4H10 and i-c4h10 are produced by the cracking reaction. Therefore, it is important to restrain the cracking reaction for heptane isomerization. As shown in Fig. 1, chemical equilibrium is reached in the isomerization of n-heptane. In this study, approach conversion to chemical equilibrium was adopted. Conversion to equilibrium [%] (produced amount of each isomer/equilibrium value of each isomer) Experimental Catalyst Preparation Cataloid AP-1 (JGC Catalyst and Chemicals Ltd.) was used as the ns Al2O3 precursor, which consists of 71.0 wt% alumina, 11.0 wt% acetic acid, and 18.0 wt% water and has an average particle size of 5.4 nm. For comparison with ns Al2O3, ns silica gel (Aerosil 200, Nippon Aerosil Co., Ltd.) was used as ns SiO2, which has an average primary particle size of ca. 12 nm. To prepare catalysts with concerted pore structure, two solid powders of ns Al2O3 or ns SiO2 and H-BEA zeolite (SiO2/Al2O3 25) were mixed at 35 : 65 with water, and then the powder mixture was kneaded mechanically. After kneading, the composites were extruded finally to catalyst pellets, followed by drying at 120 for 3 h, and calcination at 550 for 3 h. Calcined composites were impregnated with PdCl2 aqueous solution acidified with HCl. After aging in a closed vessel at room temperature for one night, the impregnated pellets

3 101 were dried at 120 for 3 h and calcined at 550 for 3 h. The calcined catalysts were reduced under hydrogen flow (GHSV: 5000 h -1 ; 450 ; 3 h). For chlorine removal, the reduced catalysts were washed with boiling deionized water, until chlorine leaching could not be observed, then the washed catalyst were dried at 120 for 3 h Reaction Test and Analysis The catalytic reaction performance was examined using feedstock of n-heptane n-c7h16 in a continuousflow reactor with a fixed bed catalyst volume of 0.5 ml for reaction tests under hydrogen pressure. The reaction condition was liquid hourly space velocity (LHSV) of h -1, molar ratio of H2/n-C7H16 15/1, total pressure of 0.12 MPa, and temperature range of Hydrogen was fed again at 450 to the catalyst bed at 80 ml/min for 3 h as a pretreatment for catalyst reduction before reaction. Then, the reactor was cooled down in H2 and heated again to the required temperature, and n-heptane was fed at ml/min for reaction. The product samples were taken as both liquid and gas. Gas chromatography was used to measure the composition of reaction products. For liquid samples, the column was Capillary TC-1 60 m with a FID detector, and Chromatopac C-R8A was applied for data processing Catalyst Characterization Elemental compositions of surface layer of Pd/ns oxide/h-bea zeolite catalysts were measured by X-ray photoelectron spectroscopy (XPS: SHIMADZU AXIS ULTRA-DLD) using a KRATOS equipped with a mono Al source operating at 450 W. The spectra were acquired at room temperature and narrow scans with rather high 40 ev pass energy for the samples. The spectrometer energy scale was calibrated with Ag 3d5/2, respectively. The binding energies and atomic concentrations of the catalysts were calculated from the XPS results using the total integrated peak areas of the Al 2p, Si 2p, Pd 3p, O 1s and C 1s regions. 3. Results and Discussion Comparison with n-hexane and n-heptane To confirm the ease of decomposition of n-heptane on 1.0 wt% Pd/35 wt% ns Al2O3/64 wt% H-BEA zeolite (SiO2/Al2O3 25) catalyst, the reaction tests for n-heptane were carried out in comparison with n-hexane by changing reaction temperature. Figures 2 and 3 show the conversions of n-paraffins, the selectivities for isoparaffin and cracked products, and confirmed that n-heptane is more easily decomposed than n-hexane. In the case of n-heptane, cracking product selectivity was extremely high at about 85 % at 300. The cracking reaction of paraffin is cleavage of the C _ C bond and is endothermic. Therefore, cracking reaction catalyst: 1 wt% Pd/35 wt% ns Al 2 O 3 /64 wt% H-BEA zeolite (SiO 2 / Al 2 O 3 25); LHSV: 2.2 h -1 ; H 2 /n-hexane: 15/1 mol/mol. : conversion; selectivities : mono-branched isoparaffins, : multi-branched isoparaffins, : cracked products. Fig. 2 n-hexane Isomerization catalyst: 1 wt% Pd/35 wt% ns Al 2 O 3 /64 wt% H-BEA zeolite (SiO 2 / Al 2 O 3 25); LHSV: 2.2 h -1 ; H 2 /n-hexane: 15/1 mol/mol. : conversion; selectivities : mono-branched isoparaffins, : multi-branched isoparaffins, : cracked products. Fig. 3 n-heptane Isomerization proceeds faster at higher temperature. The order of C _ H cracking reactivity is tertiary carbon secondary carbon primary carbon. Table 1 shows the selectivities for isomers and cracked products of n-hexane isomerization at various contact time. As shown in Table 1, in the case of n-hexane, the cracking reaction was parallel to the isomerization reaction. 2-Methylpentane and 3-methylpentane were difficult to decompose even with the C _ H bond of the tertiary carbon. The β position C _ C bonds of the cracked C _ H bond were terminal or absent because of allyl resonance. The cracking reaction can initiate only at the C _ H bond of secondary

4 102 Table 1 Conversion, Yields and Selectivities in n-hexane Isomerization at Various Contact Time Contact time [h] Conversion to equilibrium [%] Di-branched to equilibrium [%] Yields (selectivities) [%] isoparaffins 53.6 (94.7) 69.6 (95.6) 79.6 (96.1) 90.6 (96.7) 95.3 (96.9) mono-branched 49.6 (87.7) 63.3 (87.0) 70.3 (84.9) 76.6 (81.7) 74.4 (75.6) di-branched 4.0 ( 7.0) 6.3 ( 8.6) 9.3 (11.2) 14.0 (15.0) 20.9 (21.3) 2-MP 30.3 (53.6) 39.4 (54.0) 43.3 (52.3) 46.6 (49.7) 45.7 (45.0) 3-MP 19.3 (34.1) 23.9 (33.0) 27.0 (32.6) 30.0 (32.0) 29.4 (29.9) 2,2-DMB 1.7 ( 3.0) 2.9 ( 3.9) 4.0 ( 4.8) 6.5 ( 7.0) 9.9 (10.1) 2,3-DMB 2.3 ( 4.0) 3.4 ( 4.7) 5.3 ( 6.4) 7.5 ( 8.0) 11.0 (11.2) cracked 3.0 ( 5.3) 3.3 ( 4.5) 3.2 ( 3.8) 3.1 ( 3.3) 3.1 ( 3.1) MP: methylpentane; DMB: dimethylbutane. catalyst: 1 wt% Pd/ 35 wt% ns Al 2 O 3 /64 wt% H-BEA zeolite (SiO 2 /Al 2 O 3 25); reaction temperature: 250 ; H 2 /n-hexane: 15/1 mol/mol. catalyst: 1 wt% Pd/35 wt% ns Al 2 O 3 /64 wt% H-BEA zeolite; reaction temperature: 250 ; LHSV: 2.2 h -1 ; H 2 /n-paraffin: 15/1 mol/ mol). : n-heptane conversion; : isoheptane yield. Fig. 4 Effect of SiO 2 /Al 2 O 3 Ratio in BEA Zeolite on n-heptane Conversion and Isoheptane Yield 2-position carbon. Therefore, n-hexane is less likely to undergo the cracking reaction. In the case of n- heptane, carbenium ions of the cracking precursor with inner β position C _ C bonds were more easily produced without regard to allyl resonance from isoheptanes than n-hexane. Therefore, the cracking reactivity is higher for paraffins with larger carbon number. The cracked products contained large amounts of C3 and C4, and lesser amounts of C1 and C2. As shown in Fig. 3, in the case of n-heptane, the cracking reaction was successive. As the cracking reaction proceeds, isoparaffins are harder to obtain from n-heptane. Selective formation requires restraint of the cracking activity Effect of SiO2/Al2O3 Ratio of BEA Zeolite The cracking reaction is well known to occur on strong acid sites. Acid sites of zeolite can be controlled by the SiO2/Al2O3 ratio. Therefore, the effect of the SiO2/Al2O3 ratio of zeolite component in 1.0 wt% Pd/ns Al2O3/H-BEA zeolite catalyst was examined on skeletal isomerization of n-heptane. The catalysts contained 16.2 to 30 molar ratio of SiO2/Al2O3. Figure 4 shows the n-heptane conversion and isoparaffin yield. Yield of cracked products was still high on SiO2/Al2O3 of 16.2 and increased with higher SiO2/ Al2O3 ratio, whereas conversion peaked on SiO2/Al2O3 of 25. The cracking reaction could not be restrained by controlling the SiO2/Al2O3 ratio of the zeolite component, and SiO2/Al2O3 of 25 was more suitable for isomerization of n-heptane Removal of Residual Chlorine in Pd/ns Al2O3/ H-BEA Catalyst Excess chlorine in the acidic Al2O3 with PdCl2 aqueous HCl solution was considered to cause the cracking reaction. The mechanism of chlorination and chlorine removal is illustrated in Fig. 5. The calcined composites were impregnated with acidic PdCl2 aqueous HCl solution and calcined at 550 resulted in oxidation of chlorine to hydrochloric acid, and formation of chlorinated alumina which is a Lewis acid with cracking activity 24),25). The acidic sites of the catalyst could be reduced by washing with boiling deionized water. The chlorine can be removed by hydrolysis of the chlorinated alumina. Table 2 shows the effect of chlorine removal on the catalytic performances of the non-treated and treated catalyst ( treated means residual chlorine removed). Chlorine removal remarkably improved the catalytic performances, especially with higher conversion and suppression of cracked products. Chlorinated alumina cracking activity was also present in the non-treated catalyst. Therefore, the following studies were performed using the treated catalysts Effect of ns Oxide Species of Tri-component Catalyst Three different types of ns oxides were combined into tri-component catalyst, none, ns Al2O3 and ns SiO2, and catalytic performances were compared for the skeletal isomerization of n-heptane at the same conversion by changing contact time. As shown in Table 3, the

5 103 Fig. 5 Scheme of Chlorination and Chlorine Removal Table 2 Effect of Chlorine Removal from the Catalyst on Conversion, Yields and Selectivities in n-heptane Isomerization Catalyst Non-treated Treated Conversion [%] Yields (selectivities) [%] isoparaffins 40.6 (77.0) 57.2 (89.5) mono-branched 35.9 (68.1) 50.6 (79.1) di-branched 4.7 ( 8.9) 6.6 (10.4) 2-MH 17.5 (33.2) 25.9 (40.5) 3-MH 17.4 (33.0) 23.7 (37.0) 3-EP 1.0 ( 1.9) 1.0 ( 1.6) 2,2-DMP 2.0 ( 3.8) 2.3 ( 3.6) 2,3-DMP 2.7 ( 5.1) 4.3 ( 6.8) cracked 11.4 (21.6) 4.9 ( 7.7) aromatics 0.7 ( 1.4) 1.9 ( 2.8) MH: methylhexane; EP: ethylpentane; DMP: dimethylpentane. catalyst: 1 wt% Pd/ 35 wt% ns Al 2 O 3 /64 wt% H-BEA zeolite (SiO 2 /Al 2 O 3 25); reaction temperature: 250 ; LHSV: 2.2 h -1 ; H 2 /n-heptane: 15/1 mol/mol. cracking reaction on Pd/ns Al2O3/BEA zeolite catalyst was more suppressed in comparison with Pd/BEA zeolite and Pd/ns SiO2/BEA zeolite catalysts. Cracking of hydrocarbon may be driven by formation of an intermediate of carbenium ion on solid acid sites. Usage of ns Al2O3 as a binder for high silica zeolite particle leads to formation of acid sites on the surface of the zeolite catalyst, to produce carbenium ions on the sites, and to promote the cracking reaction 19). The basic structure of zeolite is a chain of tetrahedrons consisting of silicon cations, aluminum cations and oxygen anions. Oxygen anions are shared and bound three-dimensionally with cations. Silicon cation is tetravalent and balances electrically with tetravalent oxygen, but aluminum cation is trivalent and has residual electric charge of 1. The electronegative character of aluminum cation results in the solid acid sites. The structure of aluminum cation at the zeolite surface is octahedral. This acid site is strong and promotes the cracking reaction 22). Therefore, the ns Al2O3 binder makes the acid sites milder because the binder has anionic sites Effect of ns Al2O3 Content The effect of ns Al2O3 content on the catalytic performances was investigated. As shown in section 3. 4., ns Al2O3 suppresses the cracking reaction by reducing the acidity of the strong acid sites at the zeolite surface. Therefore, the content of ns Al2O3 should affect the cracking activity. As shown in Fig. 6, Pd/H-BEA zeolite catalyst without ns Al2O3 had the highest catalytic activity. Because the catalyst volume was the same in this experiment, the activity of Pd/H-BEA zeolite catalyst is only due to high content of zeolite in the catalyst. On the other hand, the conversion and cracked product yield on Pd/ns Al2O3/H-BEA zeolite catalyst remained constant at the range of 8 % to 35 % ns Al2O3 content. To clarify the effect of ns Al2O3 on the zeolite component, rate constants obtained as catalytic activities were normalized to the relative activities by zeolite content as shown in Fig. 7. In parallel, the cracking selectivities with ns Al2O3 content are shown in Fig. 8.

6 104 Table 3 Effect of ns Oxide Species on Product Yields and Selectivities in Comparison at Almost Same Conversion ns Oxide none ns Al 2 O 3 ns SiO 2 Conversion [%] Yields (selectivities) [%] isoparaffins 62.7 (82.4) 69.9 (90.5) 63.9 (85.0) mono-branched 54.3 (71.4) 61.9 (80.2) 56.9 (75.7) di-branched 8.4 (11.0) 8.0 (10.3) 7.0 ( 9.3) 2-MH 29.3 (38.5) 32.5 (42.1) 30.2 (40.2) 3-MH 23.8 (31.3) 27.5 (35.6) 25.2 (33.5) 3-EP 1.2 ( 1.6) 1.9 ( 2.5) 1.5 ( 2.0) 2,2-DMP 3.7 ( 4.8) 3.1 ( 4.0) 2.3 ( 3.1) 2,3-DMP 4.7 ( 6.2) 4.9 ( 6.3) 4.7 ( 6.2) cracked 12.3 (16.2) 6.0 ( 7.7) 10.3 (13.7) aromatics 1.1 ( 1.4) 1.4 ( 1.8) 1.0 ( 1.3) MH: methylhexane; EP: ethylpentane; DMP: dimethylpentane. catalyst: treated 1 wt% Pd/64 wt% H-BEA zeolite (SiO 2 /Al 2 O 3 25)/35 wt% none or ns Al 2 O 3 or ns SiO 2 ; reaction temperature: 250 ; LHSV: h -1 ; H 2 /n-paraffin: 15/1 mol/mol. k : relative activity based on non-composed zeolite. catalyst: treated 1 wt% Pd/ns Al 2 O 3 /H-BEA zeolite (SiO 2 /Al 2 O 3 25); reaction temperature: 250 ; LHSV: 2.2 h -1 ; H 2 /n-paraffin: 15/1 mol/mol. : mono-branched isoparaffins; : multi-branched isoparaffins; : cracked products. Fig. 6 Effect of ns Al 2 O 3 Content in the Catalyst on Products Yields Figure 7 indicated that Pd/ns Al2O3/H-BEA zeolite catalyst can contain mesoporous space and mild acid sites at the boundary between zeolite and ns Al2O3. Pd can be highly dispersed at the acid sites of the catalyst. In addition, Fig. 8 shows that since cracked products were not increased at ns Al2O3 content of more than 8 wt%, the presence of only a low percentage is necessary for ns Al2O3 to suppress the cracking reaction. ns Al2O3 can be highly dispersed onto zeolite surface. ns Al2O3 content of 35 wt% was considered to be optimum based on the catalytic performance and cost performance. Pd/ns SiO2/H-BEA zeolite catalyst was investigated in a similar way. The results showed that catalyst activity decreased with increase of ns SiO2 content, i.e., Fig. 7 Effect of ns Al 2 O 3 Content in the Catalyst on Relative Activity per Zeolite decrease of zeolite content. Pd/ns SiO2/H-BEA zeolite catalyst can not form new and milder acid sites at the zeolite surface like Pd/ns Al2O3/H-BEA zeolite catalyst. As the bond structure at the boundary is almost Si _ O _ Si, there are no acid sites. Therefore, the catalytic performances of Pd/ns SiO2/H-BEA zeolite catalysts depend only on the characteristics of the zeolite Characterization of Tri-component Catalysts for Pd Dispersion XPS confirmed Pd dispersion of Pd/H-BEA zeolite, Pd/ns Al2O3/H-BEA zeolite catalyst and Pd/ns SiO2/ H-BEA zeolite catalyst (ns oxide/h-bea 35/65) in non-reduced, reduced, reduced and treated states. Figure 9 shows the Pd 3p peaks of intensity (based on Si) on non-reduced catalysts. The intensity of 100 was observed on Pd/H-BEA zeolite catalyst. On the other hand, intensities of about 300 and about 400 were observed on Pd/ns Al2O3/H-BEA zeolite catalyst and Pd/ns SiO2/H-BEA zeolite catalysts, respectively. Pd/

7 105 Fig. 8 Effect of ns Al 2 O 3 Content in the Catalyst on Cracking Selectivity fine line: Pd/H-BEA zeolite; black line: Pd/ns Al 2 O 3 /H-BEA zeolite; gray line: Pd/ns SiO 2 /H-BEA zeolite. Fig. 10 Pd 3p Region Intensity of XPS Spectrum in Reduced Catalysts fine line: Pd/H-BEA zeolite; black line: Pd/ns Al 2 O 3 /H-BEA zeolite; gray line: Pd/ns SiO 2 /H-BEA zeolite. Fig. 9 Pd 3p Region Intensity of XPS Spectrum in Non-reduced Catalysts ns SiO2/BEA catalyst had the highest Pd dispersion for non-reduced catalyst. Non-reduced Pd may be supported in different states. Since Pd was dispersed in the state of PdCl4 2- and was still present as PdCl2 after calcining on ns SiO2/H-BEA zeolite support, the nonreduced catalyst had the highest Pd dispersion. On the other hand, non-reduced Pd was dispersed in the state of PdO on ns Al2O3/H-BEA zeolite, because ns Al2O3 could adsorb chlorine on the surface. Therefore, chloride ions moved to the ns Al2O3 surface by calcining and Pd was aggregated as PdO. Figure 10 shows Pd 3p peaks of intensity (based on Si) on reduced catalysts, which were different from the non-reduced catalysts. Pd/ns Al2O3/H-BEA zeolite and Pd/ns SiO2/H-BEA zeolite had almost the same intensities. PdCl2 over ns SiO2/H-BEA zeolite and PdO fine line: Pd/H-BEA zeolite; black line: Pd/ns Al 2 O 3 /H-BEA zeolite; gray line: Pd/ns SiO 2 /H-BEA zeolite. Fig. 11 Pd 3p Region Intensity of XPS Spectrum in Reduced and Treated Catalysts over ns Al2O3/H-BEA zeolite seemed to be reduced to similar states of Pd metal by hydrogen. Therefore, the reduced Pd dispersions were the almost same. Figure 11 shows Pd 3p peaks of intensity (based on Si) on reduced and treated catalysts. Pd dispersity of Pd/ns Al2O3/H-BEA zeolite catalyst was 1.5 times and 4 times compared to those of Pd/ns SiO2/H-BEA zeolite and Pd/H-BEA zeolite catalysts, respectively. In addition, Pd 3p peaks of intensity on Pd/ns SiO2/ H-BEA zeolite and Pd/H-BEA zeolite catalysts were almost the same before treatment. However, as shown in Fig. 12, the peak on Pd/ns Al2O3/H-BEA zeolite cat-

8 106 for Pd dispersion through chlorine removal and ionexchange. The present study successfully developed a tri-component catalyst with high activity and reduced cracking for n-heptane isomerization. Acknowledgments Financial support from CREST, Japan Science and Technology Agency for this work is greatly acknowledged. Financial support for basic research for this work from JOGMEC is also greatly acknowledged. The technical cooperation of Ms. Sowaka Sudo is highly appreciated. References fine line: non-reduced; black line: reduced; gray line: reduced and treated. Fig. 12 Pd 3p Region Intensity of XPS Spectrum in Various States of Pd/ns Al 2 O 3 /H-BEA Zeolite Catalyst alyst increased about 2 times after treatment. Pd was aggregated by chloride ions adsorbed on ns Al2O3. The aggregated Pd around the chloride may be dispersed again by removal as chlorine. Reduced Pd/ns Al2O3/H-BEA zeolite catalyst has the highest Pd dispersion after removal of excess chlorine. ns Al2O3 can adsorb chlorine, so can effectively promote Pd dispersion. Since the acid sites at the boundary probably also act as Pd dispersed sites, Pd can be dispersed near the zeolite through ion-exchange by removal of chlorine. 4. Conclusion The conversion of normal paraffin, n-heptane to the isoparaffin, isoheptanes, was studied on composite catalysts consisting of Pd, ns oxides and BEA zeolite. The optimum composite catalyst was Pd/ns Al2O3/ H-BEA zeolite catalyst. The composite had selective product distribution for skeletal isomerization, probably due to the metal dispersion at the boundary between ns Al2O3 and H-BEA zeolite. In addition, the catalyst could suppress the cracking reaction by ns Al2O3 highly dispersed at the zeolite particle surface. Only a low percentage of ns Al2O3 are necessary to suppress the cracking reaction. The catalyst was remarkably improved by chlorine removal, especially with higher conversion and suppression of the cracking reaction. Non-treated catalyst formed chlorinated alumina which is Lewis acid, and the acid sites had cracking activity. Reduced Pd/ns Al2O3/H-BEA zeolite catalyst had the highest Pd dispersion after removal of residual chlorine. Since ns Al2O3 can adsorb chlorine, it can act effectively 1) Baba, S., Shibata, Y., Kawamura, T., Takaoka, H., Kimura, T., Kousaka, K., Minato, Y., Yokoyama, N., Iida, K., Imai, T., Jpn Kokoku Tokky Koho JP (1993), JP (1993), JP (1993), JP (1993), JP (1993). 2) Hosoi, T., Kitada, S., Shimizu, T., Imai, T., Nojima, S., Shokubai, 32, 117 (1990). 3) Hosoi, T., Shimizu, T., Itoh, S., Takaoka, H., Imai, T., Yokoyama, N., Prep. Am. Chem. Soc., Div. Petrol. Chem., 33, 562 (1988). 4) Kimura, T., Shimizu, T., Imai, T., J. Jpn. Petrol. Inst., 47, (3), 179 (2004). 5) Okuhara, T., J. Jpn. Petrol. Inst., 47, (1), 1 (2004). 6) Kimura, T., Catal. Today, 81, 57 (2003). 7) Kimura, T., PETROTECH, 25, (2), 111 (2002). 8) Gosling, C. D., Rosin, R. R., Bullen, P., Shimizu, T., Imai, T., Petrol. Tech. Quart., Winter, 55 (1997/1998). 9) Li, X., Yang, J., Liu, Z. W., Asami, K., Fujimoto, K., J. Jpn. Petrol. Inst., 49, (2), 86 (2006). 10) Kondo, J. N., Yang, S., Zhu, Q., Inagaki, S., Domen, K., J. Catal., 248, 53 (2007). 11) Matsuda, T., Watanabe, K., Sakagami, H., Takahashi, N., Appl. Catal. A: General, 242, (2), 267 (2003). 13) Kumar, H., Froment, G. F., Ind. Eng. Chem. Res., 46, 4075 (2007). 14) Bouchy, C., Hastoy, G., Guillon, E., Martens, J. A., Oil Gas Sci. Tech., 64, 91 (2009). 15) Maesen, T. L. M., Calero, S., Schenk, M., Smit, B., J. Catal., 221, 241 (2004). 16) Kimura, T., Sakashita, K., Asaoka, S., Mater. Res. Innov., 15, (suppl 2), s101 (2011). 17) Asaoka, S., Ito, K., Minohara, S., Ali, M. A., Bamufleh, H. S., Prep. Pap. Am. Chem. Soc., Div. Petrol. Chem., 50, 372 (2006). 18) Ali, M. A., Asaoka, S., Petrol. Sci. Tech., 27, 984 (2009). 19) Ito, H., Jang, H., Sakashita, K., Asaoka, S., Pure Appl. Chem., 80, 2273 (2008). 20) Sakashita, K., Yoshino, M., Nishimura, I., Hayakawa, Y., Asaoka, S., Prep. Pap. Am. Chem. Soc., Div. Petrol. Chem., 54, 38 (2009). 21) Sakashita, K., Yoshino, M., Nishimura, I., Kimura, T., Asaoka, S., J. Jpn. Petrol. Inst., 54, (3),180 (2011). 22) Sakashita, K., Nishimura, I., Kimura, T., Asaoka, S., J. Jpn. Petrol. Inst., 54, (4), 248 (2011). 23) Kimura, T., Gao, J., Sakashita, K., Li, X., Asaoka, S., J. Jpn. Petrol. Inst., 55, (1), 40 (2012). 24) Tamele, M. W., Discuss. Faraday Soc., 8, 270 (1950). 25) Shorrock, J. K., Clark, J. H., Wilson, K., Chisem, J., Org. Proc. Res. Dev., 5, (3), 249 (2001).

9 107,,,,, Pd Al 2 O 3 /H-BEA,,,,,,,,,, 2,