Hydroisomerization of n-Butane over Platinum-Promoted Cesium Hydrogen Salt of 12-Tungstophosphoric Acid

The hydroisomerization of n-butane was carried out in a fixed-bed gas-flow reactor over Pt-promoted Cs2.5H0.5PW12O40 (denoted as Cs2.5). Two kinds of catalysts, a direct impregnation of Pt on Cs2.5 (denoted as Pt/Cs2.5), as well as a mechanical mixture of Pt/Al2O3 and Cs2.5 (denoted as Pt/Al2O3+Cs2.5), were used for the hydroisomerization. Pt/Al2O3+Cs2.5 showed a higher stationary activity than Pt/Cs2.5 because the Pt particles supported on Al2O3 were much smaller than those supported on Cs2.5. The initial activity decreased with increasing H2 pressure over Pt/Al2O3+Cs2.5. This indicates that the hydroisomerization of n-butane over Pt/Al2O3+Cs2.5 proceeded through a bifunctional mechanism, in which n-butane was hydrogenated/dehydrogenated on Pt sites and was isomerized on acid sites of Cs2.5. For the hydroisomerization of n-butane over Pt/Al2O3+Cs2.5 the hydrogenation/dehydrogenation on Pt sites is a limiting step at a low Pt loading and the isomerization on solid acid sites is a limiting step at a high Pt loading. During the reaction, hydrogen molecules were dissociated to active hydrogen atoms on Pt sites, and then the formed active hydrogen atoms moved to the solid acid sites of Cs2.5 (spillover effect) to eliminate the carbonaceous deposits and suppress the catalyst deactivation. Because Cs2.5 has suitably strong and uniformly-distributed solid acid sites, Pt/Al2O3+Cs2.5 showed a higher stationary activity than Pt/Al2O3+H-ZSM-5 and Pt/Al2O3+SO4/ZrO2 for the hydroisomerization of n-butane at a low H2 pressure.

n-Butane is an important industrial chemical which can be obtained from the petroleum industry and Fischer-Tropsch synthesis process. The skeletal hydroisomerization of n-butane to isobutane is a large-scale industrial process. Isobutane is utilized in the butene alkylation to isooctane as well as in the produce of MTBE (methyl tert-butyl ether) or ETBE (ethyl tert-butyl ether). All of these products are well known as non-leaded high octane gasoline additives. The hydroisomerization of n-butane occurs through carbenium cation intermediates, which requires the presence of strong acid in the system. Bifunctional catalysts containing metals and solid acids are promising for the hydroisomerization of n-butane. The metal sites provide the hydrogenation-dehydrogenation function and the acid sites provide the isomerization function. In generally, Pt is the most effective metal catalyst, and thus the development of highly active solid acid catalysts is an important task for designing bifunctional catalysts in the hydroisomerization of n-butane.
How to combine the Pt catalyst with the solid acid catalyst is an interesting subject in catalyst design. For preparing bifunctional catalysts, although the direct support of the Pt on the solid acid by impregnation is a universal method, the mechanical mixing Pt/Al 2 O 3 or Pt/SiO 2 with the solid acid is a unique method. As early as fifty years ago, a mechanical mixture of Pt/SiO 2 and aluminum silicates had been used for investigating the reaction mechanism of saturated hydrocarbons isomerization over bifunctional catalysts [40]. In the recent years, mechanical mixed catalysts have been of interest for the hydroisomerization of n-butane because they have some advantages compared with the directly impregnated catalysts. These advantages include the strong mechanical strength, the high molding ability, and so on [27,28,35]. We have found that the mechanically mixed catalyst Pt/Al 2 O 3 +Cs2.5 showed a higher activity and a higher stability than those of the directly impregnated catalyst Pt/Cs2.5 for the hydroisomerization of n-pentane and n-hexane [5][6][7]. In the present study, we investigated the catalytic performance of Pt-promoted Cs2.5 catalysts for the hydroisomerization of n-butane and also compared Pt-promoted Cs2.5 catalysts with Pt-promoted SO 4 /ZrO 2 and Pt-promoted H-ZSM-5 for the hydroisomerization of n-butane. The mechanically mixed catalyst of Pt/Al 2 O 3 with Cs2.5 (abbreviated as Pt/Al 2 O 3 +Cs2.5) was prepared as follows: after grinding the mixture of 2.0 wt% Pt/Al 2 O 3 with Cs2.5 (mass ratio = 1:1) in a mortar for 30 min, the powder was pressed into a disk at 40 kg cm -2 and then sieved to 24-60 mesh. Thus the Pt loading in Pt/Al 2 O 3 +Cs2.5 was 1.0 wt%. SO 4 /ZrO 2 (abbreviated as SZ) was prepared using a method reported in the literature [43]. Zr(OH) 4 , which was obtained by the hydrolysis of ZrOCl 2 with NH 4 OH, was treated with an aqueous solution of H 2 SO 4 (1 N). After filtering out the liquid (H 2 SO 4 solution), the resulting solid (SO 4 2--Zr(OH) 4 ) was calcined at 773 K for 3 h in air to form SO 4 /ZrO 2 . H-ZSM-5 (abbreviated as HZ) was obtained from Na-ZSM-5 (Tosoh Corporation, HSZ-820 NAA, SiO 2 /Al 2 O 3 = 23.2, surface area: 322 m 2 g -1 ) by the ion-exchange method. Na-ZSM-5 was treated with an aqueous solution of NH 4 NO 3 (1 N) to form NH 4 -ZSM-5, followed by drying at 373 K for 24 h. NH 4 -ZSM-5 was calcined at 773 K for 3 h to form H-ZSM-5.

Catalyst Syntheses
The mechanically mixed catalysts of Pt/Al 2 O 3 with SZ (abbreviated as Pt/Al 2 O 3 +SZ) and Pt/Al 2 O 3 with HZ (abbreviated as Pt/Al 2 O 3 +HZ) were prepared using a method similar to that of Pt+Cs2.5. After grinding the mixture of 2.0 wt% Pt/Al 2 O 3 with SZ or HZ (mass ratio = 1:1) in a mortar for 30 min, the powder was pressed into a disk at 40 kg cm -2 and then sieved to 24-60 mesh. Thus the Pt loadings in either Pt/Al 2 O 3 +SZ or Pt/Al 2 O 3 +HZ were 1.0 wt%.

Characterization
Scanning electron microscope (SEM) observations were carried out using a Hitachi S-3400N instrument with an EDX. Pt metal surface and Pt particle size were measured by a H 2 adsorption method. The H 2 uptake was estimated by the extrapolation to zero pressure of the linear part of the isotherms. The difference between the total amount of adsorbed hydrogen (H tot ) and the reversible part of adsorbed hydrogen (H rev ) gave the irreversible part of adsorbed hydrogen (H irr ), which was used for calculating the Pt metal surface and Pt particle size. Temperature-programmed desorption of ammonia (NH 3 -TPD) was observed using a BELCAT-B automatic monitor equipped with a TCD and a mass spectrometer for ammonia species detection. A part of a 0.05 g aliquot of the sample was pretreated at 673 K for 1 h under He flow (50 mL min -1 ). After the temperature was decreased to 373 K, ammonia was adsorbed onto the surface, followed by evacuation for 1 h at 373 K to eliminate the weakly adsorbed ammonia. Then, NH 3 -TPD was carried out from 373 K to 973 K (8 K min -1 ).

Catalytic Reaction
The hydroisomerizations of n-butane was performed in a fixed-bed quartz tubular reactor (φ: 8 mm) at 573 K under atmospheric pressure. In a typical reaction conditions, the total flow rate was 20 mL min -1 , the catalyst amount was 1 g, and the feed gas contained 10% n-C 4 H 10 , 10% H 2 , and 80% N 2 . The catalysts were pretreated in a flow of H 2 (60 mL min -1 ) at 573 K for 1 h before reaction. During reaction, the products were analyzed with an on-line FID GC (Hitachi GC-163) equipped with an Al 2 O 3 /KCl fused silica capillary column.  The size of Pt/Al 2 O 3 particles was about 0.5 μm. The Cs2.5 became small powder particles adhered to the Pt/Al 2 O 3 particles due to the low mechanical strength of Cs2.5. The Pt/Al 2 O 3 particles and the Cs2.5 particles contacted closely with each other. Moreover, the mechanical strength was greatly strengthened after pressing the mixture into a disk at 40 kg cm -2 for use as a catalyst. Table 1 lists conversion and selectivity for the hydroisomerization of n-butane at 573 K over various catalysts. The data at 5 min could be regarded as the initial catalytic performance and the data at 5 h could be regarded as the stable catalytic performance over each catalyst. Pt/Al 2 O 3 showed a low conversion (2.8%) after 5 h on stream, which indicates that catalyst acidity is indispensable for the hydroisomerization of n-butane. Cs2.5 showed an initial conversion of 23.5% and an initial selectivity to isobutane of 76.5%. The solid acid Cs2.5 could catalyze the hydroisomerization of n-butane even without Pt [44]. However, the conversion after 5 h on stream over Cs2.5 was low (9.2%), due to the severe deactivation. The initially white Cs2.5 catalyst became black after reaction at 573 K for 5 h, indicating that carbonaceous deposits were formed on the catalyst surface, covering the acid sites of Cs2.5. This is the reason for the deactivation of the Cs2.5 catalyst. On the other hand, Pt/Cs2.5 showed an initial conversion of 66.1% and a stable conversion of 42.2%. The synergy between Pt and Cs2.5 was great, because both the initial conversion and the stable conversion were remarkably improved by introducing Pt in the solid catalyst Cs2.5. Moreover, Pt/Al 2 O 3 +Cs2.5 exhibited a higher stable conversion (64.8%) and a higher stable selectivity to isobutane (92.5%) than those over Pt/Cs2.5 for the hydroisomerization of n-butane after 5 h on stream. On the other hand, the mixture of Pt/Cs2.5 and Al 2 O 3 just showed a similar performance (not shown in Table 1) to that of Pt/Cs2.5 for the hydroisomerization of n-butane. Thus the method for combining Pt with Cs2.5 is important for increasing the catalytic activity of bifunctional catalysts in the hydroisomerization of n-butane.

Hydroisomerization of n-Butane over Various Catalysts
Scheme 1 shows the mechanism of n-butane hydroisomerization over a Brönsted acid catalyst. Since a heteropolyacid is a kind of typical Brönsted acid [1], it is very probable that the hydroisomerization of n-butane over Cs2.5 also takes place via this mechanism [44]. At first, a sec-butyl carbenium cation was formed by a step of proton addition, followed by a step of H 2 elimination. Then, the sec-butyl carbenium cation was transformed to a tert-carbenium cation by a shift of the methyl group. Finally, the tert-carbenium cation captured a H 2 molecule and eliminated a proton to form an isobutane molecule. All of these steps were carried out on the Brønsted acid sites. Scheme 1. Mechanism of n-butane hydroisomerization over a Brønsted acid catalyst. Scheme 2 shows the mechanism of n-butane hydroisomerization over a bifunctional catalyst containing metal and heteropolyacid [18,19]. At first, the n-butane molecule eliminated a H 2 molecule to form a n-butene molecule on the Pt sites. Then, the formed n-butene molecule moved to the solid acid sites to form a sec-butyl carbenium cation by obtaining a proton. Then, the sec-butyl carbenium cation was transformed to a tert-carbenium cation by a shift of the methyl group on the acid sites. Then, the tert-carbenium cation eliminated a proton to form an isobutene molecule. Finally, the isobutene molecule moved to the Pt sites to form an isobutane molecule by a process of H 2 addition.

Scheme 2.
Mechanism of n-butane hydroisomerization over a bifunctional catalyst.
As a result, although the carbenium cation is a key intermediate in the hydroisomerization of n-butane over either a heteropolyacid catalyst or a bifunctional catalyst containing Pt and heteropolyacid, the path for forming the carbenium intermediate over a bifunctional catalyst is different from that over a heteropolyacid catalyst. Because the rate of formation of a carbenium species by adding proton to n-butene is much faster than that by adding proton to n-butane, in the hydroisomerization of n-butane the bifunctional catalysts (Pt/Cs2.5 and Pt/Al 2 O 3 +Cs2.5) showed much higher conversions than those observed over the monofunctional heteropolyacid catalyst Cs2.5 (Table  1). Figure 2 shows H 2 uptake by Pt/Al 2 O 3 +Cs2.5 and Pt/Cs2.5 at 298 K. The H 2 uptake was used as for calculating the Pt surface area, Pt dispersion degree, and Pt particle size for the samples. The total H 2 uptake contains the reversible H 2 uptake (physical absorption) and the irreversible H 2 uptake (chemical absorption). The irreversible H 2 uptake at 0 torr could be obtained from the total H 2 uptake at 0 torr and the reversible H 2 uptake at 0 torr. The obtained irreversible H 2 uptake at 0 torr was used for calculating Pt surface area and Pt dispersion degree of each sample. As shown in Figure 2, the Pt surface area and Pt dispersion degree over Pt/Al 2 O 3 +Cs2.5 were much larger than those over Pt/Cs2.5. Thus Al 2 O 3 is a good support for Pt as comparison with Cs2.5. The high Pt dispersion degree gave Pt/Al 2 O 3 +Cs2.5 a higher selectivity for isobutane than that over Pt/Cs2.5 in the hydroisomerization of n-butane (Table 1). Although the catalysts prepared by directly supporting Pt on solid acids (using the impregnation method) are usually used for the hydroisomerization of n-alkanes, the mechanical mixtures of solid acids with Pt/SiO 2 or Pt/Al 2 O 3 have also been applied for the hydroisomerization of n-alkanes for several purposes. Firstly, a mechanical mixture of Pt/SiO 2 with aluminum silicates had been used for investigating the mechanism of n-alkanes hydroisomerization over Pt-promoted acid catalysts [40]. Because Pt sites and solid acid sites achieved their functions independently, the mechanical mixed catalyst showed a catalytic activity similar to that seen over the directly impregnated catalysts [40]. Secondly, because the Pt-supported zeolites have low mechanical strength and poor molding properties, the mechanical mixtures of acidic zeolites with Pt/SiO 2 or Pt/Al 2 O 3 have been used to increase the mechanical strength and the molding ability of the bifunctional catalysts [33,35,45]. Thirdly, in the cause of solid superacid SO -ZrO 2 showed high catalytic performances for the hydroisomerization of n-butane [27,28,46]. Fourthly, in the case of the heteropolyacid Cs 2.5 H 0.5 PW 12 O 40 , the mechanical mixed catalyst Pt/Al 2 O 3 +Cs2.5 showed a higher activity than that of the impregnated catalyst Pt/Cs2.5 for the hydroisomerization of n-pentane and n-hexane [5][6][7]. In the present study, we found that Pt/Al 2 O 3 +Cs2.5 showed a higher catalytic performance for the hydroisomerization of n-butane in comparison with Pt/Cs2.5. The directly supported catalyst Pt/Cs2.5 had a low Pt surface area and a low Pt dispersion degree, probably because Pt 2+ interacted with PW 12 O 40 3during the impregnation process. Figure 3 shows the time courses of n-butane hydroisomerization over various catalysts at 573 K. Pt/Al 2 O 3 showed a very low conversion due to the lack of strong acid sites. Thus Pt/Al 2 O 3 can barely be used as an independent catalyst for the hydroisomerization of n-butane. The presence of strong acid sites in the catalyst is indispensable for the hydroisomerization of n-butane. Cs2.5 showed an initial conversion (after 5 min on stream) of 23.5%, but the conversion decreased to 9.2% after 5 h on stream. The carbonaceous deposits which formed by the polymerization of alkene intermediates and byproducts (such as n-butene, iso-butene, and so on) covered the solid acid sites and caused a serious deactivation of the Cs2.5 catalyst. The carbonaceous deposits are hydrocarbons with large molecular weights and high ratios of C to H. When Pt was introduced into Cs2.5 (i.e., Pt/Cs2.5 and Pt/Al 2 O 3 +Cs2.5), the deactivation was greatly repressed in the hydroisomerization of n-butane. Pt catalyzes the hydrogenation of the carbonaceous deposits (covered acid sites) by supplying hydrogen. In concrete, the remarkable effect of Pt in suppressing the catalyst deactivation was brought about by the activated hydrogen, which were formed on Pt, transferred to Cs2.5, and utilized to remove the carbonaceous deposits [5,6]. Pt/Al 2 O 3 +Cs2.5 showed a higher catalytic stability than that of Pt/Cs2.5 for the hydroisomerization of n-butane because the Pt dispersion degree in Pt/Al 2 O 3 +Cs2.5 was much higher than that in Pt/Cs2.5 ( Figure 2). The amount of the carbonaceous deposits on the used catalyst could be calculated by a temperatureprogrammed oxidation (TPO) method [47][48][49]. After the reaction was carried out over each catalyst for 5 h at 573 K, the reactor was cooled to room temperature in flowing N 2 gas. The catalyst was then treated in air flow (1.5 L h −1 ) by increasing the temperature at 2.5 K min −1 to change the carbonaceous deposits to CO 2 . The formed CO 2 could be detected by a TCD GC. Figure 4 shows the dependence of the rate of CO 2 (formed from carbonaceous deposits) on the calcination temperature over various catalysts after 5 h on stream at 573 K. An integration of the rate of CO 2 formation gave the amount of total CO 2 , from which the amount of the carbonaceous deposits (after 5 h on stream at 573 K) were calculated as 4.5, 0.6, and 0.4 wt% on Cs 2.5, Pt/Cs2.5, and Pt/Al 2 O 3 +Cs2.5, respectively. Moreover, according to the amount of CO 2 and H 2 O formed in the TPO measurement for various catalysts after 5 h on stream at 573 K, the H/C ratios were calculated as about 0.3, 0.6, and 0.7 for the carbonaceous deposits on Cs 2.5, Pt/Cs2.5, and Pt/Al 2 O 3 +Cs2.5, respectively. Two peaks with the highest rate at around 473 and 598 K were observed from the plot of Cs2.5. In general, the peak at lower temperature could be regarded as "soft coke" and the peak at the higher temperature could be regarded as "hard coke" [50,51]. Pt/Cs2.5 deposited both "soft coke" and "hard coke" on the surface, but the amount deposited on the Pt/Cs2.5 surface was much lower than that on the Cs2.5 surface. Thus Pt/Cs2.5 showed a much higher stability than that over Cs2.5 at 573 K ( Figure 3). Further, the amount of carbonaceous deposit on Pt/Al 2 O 3 +Cs2.5 (0.4 wt.%) was lower than that on Pt/Cs2.5 (0.6 wt.%) after 5 h on stream. In the TPO profile of Pt/Al 2 O 3 +Cs2.5, the peak of "hard coke" at 573-598 K was very small and just became a shoulder peak for the peak of "soft coke" at low temperature. Therefore, Pt/Al 2 O 3 +Cs2.5 showed a very high stability for the hydroisomerization of n-butane (Figure 3). Figure 5 shows the effect of Pt amount in the Pt/Al 2 O 3 +Cs2.5 catalyst for the hydroisomerization of n-butane at 573 K. Either the pressure of n-butane or the pressure of H 2 was 0.1 atm. The initial conversion greatly increased when a small amount of 2 wt.% Pt/Al 2 O 3 was added to Cs2.5, but it almost remained at a constant value when the amount of 2 wt.% Pt/Al 2 O 3 was more than 0.1 g. On the other hand, 0.5 g of 2 wt.% Pt/Al 2 O 3 was necessary for suppressing the deactivation to make the stationary conversion close to the initial conversion. Thus Pt has two effects in the hydroisomerization of n-butane. Firstly, Pt achieves a hydrogenation-dehydrogenation function, which greatly improves the initial conversion over Pt/Al 2 O 3 +Cs2.5. It seems that 0.1 g of 2 wt.% Pt/Al 2 O 3 is enough for increasing the initial conversion. Therefore, the hydrogenation-dehydrogenation on Pt sites is a limiting step when the amount of Pt/Al 2 O 3 is less than 0.1 g in Pt/Al 2 O 3 +Cs2.5, and the isomerization on Cs2.5 sites is a limiting step when the amount of Pt/Al 2 O 3 is more than 0.1 g in Pt/Al 2 O 3 +Cs2.5. Therefore, the speed of hydrogenation-dehydrogenation on Pt is very fast and a small amount of Pt can achieve the hydrogenation-dehydrogenation function. Secondly, Pt plays an important role for suppressing the catalyst deactivation and for maintaining the catalytic stability for the hydroisomerization of n-butane. The deactivation can not be completely suppressed with a small amount of Pt/Al 2 O 3 ( Figure 5). It needs a relatively large amount of 2 wt.% Pt/Al 2 O 3 (about 0.5 g) to eliminate the catalyst deactivation over Pt/Al 2 O 3 +Cs2.5 for the hydroisomerization of n-butane.  Figure 6 shows the time courses of the hydroisomerization of n-butane over Pt/Al 2 O 3 +Cs2.5 at 573 K under various H 2 pressures. The partial pressure of n-butane was 0.1 atm and the partial pressure of H 2 was changed from 0 to 0.5 atm. Under a N 2 atmosphere (P H 2 = 0), the initial conversion was very high, but the deactivation was serious. With increasing H 2 pressure, the deactivation was gradually suppressed, but the initial conversion decreased. Thus H 2 has two effects in the hydroisomerization of n-butane: decreasing the initial conversion and suppressing the catalyst deactivation. Actually, H 2 plays two roles in the hydroisomerization of n-butane over bufunctional catalysts. One role of H 2 is the hydrogenation of isobutene to isobutane on Pt sites (the last step in Scheme 2). According to Scheme 2, the first step is the dehydrogenation of n-butane to n-butene for the hydroisomerization of n-butane over a bifunctional catalyst. Under a high H 2 partial pressure, the equilibrium of the dehydrogenation step shifts to n-butane, which causes the decrease of the concentration of n-butene intermediates in the reaction system. As a result, because the presence of H 2 is not favorable for the dehydrogenation step in the hydroisomerization of n-butane (the first step in Scheme 2), the initial conversion decreases with increasing H 2 pressure over Pt/Al 2 O 3 +Cs2.5. Another role of H 2 in the hydroisomerization of n-butane over bifunctional catalysts is the suppression of catalyst deactivation. As discussed above, the carbonaceous deposits that formed on the catalyst surface cause the catalyst deactivation. H 2 molecules can form active H atoms on Pt sites during the reaction. The formed active H atoms shift (spillover) to the acid sites and hydrogenate the carbonaceous deposits on the acid sites of Cs2.5. Therefore, the catalyst deactivation can be suppressed by increasing H 2 pressure in the hydroisomerization of n-butane over Pt/Al 2 O 3 +Cs2.5.  Figure 7 shows the time courses of the n-butane hydroisomerization over Pt/Al 2 O 3 +Cs2.5 at 573 K under various n-butane pressures. The partial pressure of H 2 was 0.1 atm and the partial pressure of n-butane was changed from 0.1 to 0.7 atm.

Hydroisomerization of n-Butane over the Pt/Al 2 O 3 +Cs2.5 Catalyst
The initial conversions over Pt/Al 2 O 3 +Cs2.5 showed almost the same values under various n-butane pressures. This implies that the initial rate is proportional to the n-butane pressure (i.e., first order in nbutane). Thus the "bimolecular mechanism" (by alkylation-cracking a C 8 intermediate) is not important for the hydroisomerization of n-butane over Pt/Al 2 O 3 +Cs2.5 [6,43]. On the other hand, the conversion after 5 h on stream decreased with increasing the n-butane pressure. The increase of n-butane pressure means the decrease in the ratio of H 2 to n-butane in the feed gas, which caused the catalyst deactivation during the reaction. Therefore, it needs a high H 2 pressure to maintain the catalyst stability for the n-butane hydroisomerization over Pt/Al 2 O 3 +Cs2.5 under a high partial pressure of n-butane.  Table 2 lists the results obtained in the hydroisomerization of n-butane with various bifunctional catalysts. We compared the Pt/Al 2 O 3 +Cs2.5 catalyst with the catalysts of Pt/Al 2 O 3 +SZ and Pt/Al 2 O 3 +HZ in this study because both Pt-promoted solid super acid SZ (SO 4 2--ZrO 2 ) and Ptpromoted acidic zeolite HZ (H-ZSM-5) have been reported as effective catalysts for the hydroisomerization of n-butane [24][25][26][27][28][29][30][31][32][33][34][35]. Pt/Al 2 O 3 +Cs2.5, Pt/Al 2 O 3 +SZ and Pt/Al 2 O 3 +HZ had the same Pt dispersion because the Pt metal was supported on Al 2 O 3 for all three catalysts. Therefore, the difference of solid acids (Cs2.5, SZ, and HZ) determined the difference of catalytic performances for the hydroisomerization of n-butane over various catalysts. As shown in Table 2, Pt/Al 2 O 3 +Cs2.5 showed the highest conversion (64.8%) and the highest selectivity for isobutane (92.5%) among the various catalysts after 5 h on stream. Pt/Al 2 O 3 +SZ showed the highest initial conversion (77.6%), but the conversion rapidly decreased to 20.5% after 5 h on stream. Moreover, the selectivity for isobutane over Pt/Al 2 O 3 +SZ was low. Pt/Al 2 O 3 +HZ showed lower conversion and selectivity than those over Pt/Al 2 O 3 +Cs2.5. As a result, Pt/Al 2 O 3 +Cs2.5 is the best catalyst for the hydroisomerization of n-butane at a low H 2 pressure (0.1 atm). Brønsted acid sites in Cs2.5 contributed to the acid function of Pt/Al 2 O 3 +Cs2.5 for the hydroisomerization of n-butane since heteropolyacids are pure Brønsted acids without Lewis acid sites [1]. Figure 8 shows the NH 3 -TPD profiles of various catalysts. The NH 3 -TPD is a powerful tool for measuring the acidic strength of a solid acid. The NH 3 molecules desorbed from the weak acid sites at low temperatures and desorbed from the strong acid sites at high temperatures. Pt/Al 2 O 3 showed a weak and broad peak at about 473 K in the NH 3 -TPD profile, indicating that the acid sites in Pt/Al 2 O 3 are very weak. Although Pt/Cs2.5 and Pt/Al 2 O 3 +Cs2.5 showed the peaks with similar shape in the NH 3 -TPD profiles, the maximum temperature of NH 3 desorption from Pt/Cs2.5 was slightly lower than that from Pt/Al 2 O 3 +Cs2.5. Thus the acidic strength of Pt/Cs2.5 was slightly weaker than that of Pt/Al 2 O 3 +Cs2.5. This is the reason that Pt/Al 2 O 3 +Cs2.5 showed a higher initial conversion than that over Pt/Cs2.5 (Figure 3). The decrease of the acid strength of Pt/Cs2.5 probably because the ion exchange of Pt 2+ with H + in Cs 2.5 H 0.5 PW 12 O 40 occurred in the impregnation stage of the synthesis of Pt/Cs2.5. For a solid acid, the strongest acid sites provide its main character for the acid-catalyzed reactions. The peak at the maximum temperature in the NH 3 -TPD profile corresponds to the strongest acid sites and thus it determines the acidic strength of a solid acid. According to the peak position at the maximum temperature in the NH 3 -TPD profile of various samples (Figure 8), the acidic strength of various catalysts was in an order of Pt/Al 2 O 3 +SZ > Pt/Al 2 O 3 +Cs2.5 > Pt/Cs2.5 > Pt/Al 2 O 3 +HZ > Pt/Al 2 O 3 . Pt/Al 2 O 3 +HZ showed two peaks at around 473 K and 673 K in the NH 3 -TPD profile, while Pt/Al 2 O 3 +SZ showed two peaks at around 473 K and 873 K in the NH 3 -TPD profile. These results indicate that both Pt/Al 2 O 3 +HZ and Pt/Al 2 O 3 +SZ have two types of acid sites: weak acid sites and strong acid sites. The hydroisomerization of n-butane occurs through carbenium cation intermediates which are easy to be formed on the strong acid sites. Thus the strong acid sites provide the main effect for the catalytic performances over Pt/Al 2 O 3 +HZ and Pt/Al 2 O 3 +SZ. However, the weak acid sites also catalyze the reaction although the reaction rate is low. Because the products formed from the strong acid sites and the weak acid sites are different, the broadly distributed acid sites caused the decrease of the selectivity for isobutane over Pt/Al 2 O 3 +HZ and Pt/Al 2 O 3 +SZ (Table 2). On the other hand, either Pt/Cs2.5 or Pt/Al 2 O 3 +Cs2.5 showed only one peak in the NH 3 -TPD profile, implying that the strength of the acid sites were distributed uniformly on the surfaces of Pt/Cs2.5 and Pt/Al 2 O 3 +Cs2.5. The uniform acid strength gave Pt/Cs2.5 and Pt/Al 2 O 3 +Cs2.5 high selectivity for isobutane in the hydroisomerization of n-butane (Table 1). Figure 9 shows the time courses of various bifunctional catalysts for the hydroisomerization of n-butane at 573 K. Either the pressure of n-butane or the pressure of H 2 was 0.1 atm. For a bifunctional catalyst in the hydroisomerization of n-alkanes, the balance between acid and metal is very important for obtaining the optimum performance [52,53]. The acid strength is an important factor for controlling the activity and the selectivity in the hydroisomerization of n-alkanes. Strong acids usually exhibit high activity but also usually show low selectivity and serious deactivation.

Comparison of Various Bifunctional Catalysts for the Hydroisomerization of n-Butane
As shown in Figure 9, the initial conversion was in the order Pt/Al 2 O 3 +SZ > Pt/Al 2 O 3 +Cs2.5 > Pt/Al 2 O 3 +HZ. This order coincided with the order of the acid strength of the various catalysts.
Moreover, Pt/Al 2 O 3 +SZ showed a serious deactivation under a low H 2 pressure (0.1 atm) and the activity after 5 h on stream was low. A high H 2 pressure is need for the hydroisomerization of n-butane over the Pt-promoted SZ (SO 4 2--ZrO 2 ) catalysts [25][26][27][28][29][30][31]. Actually, the deactivation of Pt/Al 2 O 3 +SZ could be suppressed under a high H 2 pressure of 0.7 atm. Moreover, because the conversion over Pt/Al 2 O 3 +Cs2.5 decreased with increasing H 2 pressure (Figure 6), Pt/Al 2 O 3 +SZ showed a higher stationary conversion than that over Pt/Al 2 O 3 +Cs2.5 under a high H 2 pressure of 0.7 atm. On the contrary, Pt/Al 2 O 3 +Cs2.5 showed a higher stationary conversion than that over Pt/Al 2 O 3 +SZ under a low H 2 pressure of 0.1 atm (Figure 9). Thus Pt/Al 2 O 3 +Cs2.5 is a good catalyst under low H 2 pressures and Pt/Al 2 O 3 +SZ is a good catalyst under high H 2 pressures for the hydroisomerization of n-butane. In comparison with Pt/Al 2 O 3 +SZ, using Pt/Al 2 O 3 +Cs2.5 as a catalyst for the hydroisomerization of n-butane decreases the cost of the process (because H 2 is much more expensive than N 2 ) and improves the safety of the operation (because a gas stream containing a large amount of H 2 is very dangerous).

Conclusions
By introducing Pt in Cs 2.5 H 0.5 PW 12 O 40 , the activity and the selectivity for isobutane in the hydroisomerization of n-butane were greatly increased. Because the Pt surface area and Pt dispersion degree of Pt/Al 2 O 3 are much larger than those of Pt/Cs2.5, the mechanical mixed catalyst Pt/Al 2 O 3 +Cs2.5 showed a higher stationary conversion than that over the directly supported catalyst Pt/Cs2.5. Moreover, Pt/Al 2 O 3 +Cs2.5 showed a higher initial conversion than that over Pt/Cs2.5 because the acid strength of Pt/Cs2.5 was lower than that of Pt/Al 2 O 3 +Cs2.5 (due to the ion exchange of Pt 2+ with H + in the stage of impregnation). Comparing with Pt/Al 2 O 3 +HZ, Pt/Al 2 O 3 +Cs2.5 showed a higher activity (due to the stronger acid strength of Cs2.5) and a higher selectivity for isobutane (due to the uniformly-distributed acid sites on Cs2.5). Pt/Al 2 O 3 +SZ showed the highest initial conversion among various catalysts, but the activity decreased rapidly under a low H 2 pressure (due to the excessively strong acidity of SO 4 2--ZrO 2 ). As a result, Pt/Al 2 O 3 +Cs2.5 is an excellent catalyst for the hydroisomerization of n-butane under a low H 2 pressure because Pt/Al 2 O 3 has highly dispersed Pt particles and Cs 2.5 H 0.5 PW 12 O 40 has properly strong and uniformed distributed solid acid sites.