Effects of Zn Addition into ZSM-5 Zeolite on Dehydrocyclization-Cracking of Soybean Oil Using Hierarchical Zeolite-Al2O3 Composite-Supported Pt/NiMo Sulfided Catalysts.

Zn-exchanged ZSM-5-Al2O3 (ZA) composite-supported Pt/NiMo (NM) sulfided catalysts were prepared using the conventional kneading method and were tested for dehydrocyclization-cracking of soybean oil. The effects of Zn addition on the activity and selectivity of products were investigated under moderate-pressure conditions of 0.5 and 1.0 MPa H2 in the temperature range of 420-580 °C. At the temperature 500 °C and higher, most of the sample soybean oil was converted at both the pressures of 0.5 and 1.0 MPa. At 1.0 MPa and 500 °C, the effects of Zn addition appeared and increased the yields of aromatics, while the catalyst without Zn produced larger amounts of products with more than C18. Further, at 0.5 MPa and 580 °C, the gas formation was inhibited in comparison to the cases of 1.0 MPa and the effects of the Zn addition also appeared and increased the yields of aromatics, while the catalyst without Zn produced larger amounts of products with more than C18. The Pt/NM/Zn(122)ZA test catalyst produced more than 63% of liquid fuels in the range C5-C18, and the yield of aromatics was 13%, the maximum value in the present study. The following reaction routes were proposed. The structure of triglyceride is converted by hydrocracking to three molecules of aliphatic acids and propane on the surface PtNiMo sulfide on Al2O3 support. The converted aliphatic acids are decomposed through decarboxylation to hydrocarbon fragments, which are further decomposed by cracking on the acid sites of the catalyst, the surface of NiMo sulfide, Al2O3, or ZSM-5. Finally, the formed C3 and C4 olefins are transformed to aromatics through the Diels-Alder reaction on the Zn species of ZnZSM-5. On the other hand, although gases were relatively small in amount, aromatic compounds were formed significantly, suggesting that cyclization might directly occur without conversion to gaseous hydrocarbons to some extent.

Introduction

While in most of developed class="Chemical">countries, more than hclass="Chemical">pan class="Chemical">alf of the primary energy source still depends on fossil fuels such as coal, petroleum, and natural gas, it becomes increasingly interesting to use renewable resources year after year. Among the renewable resources, much attention has been paid to carbon-neutral biomass to control recurring environmental issues. Solar energy and wind power of renewable energy sources will not give carbonaceous materials, which can complement the petrochemical industry, while biomass solely provides those carbonaceous matters. It becomes therefore very important to develop process technologies for generating from the biomass most of the matters now provided by the petrochemical industry. Among the biomass feedstocks, fats have such advantageous characteristics of liquids and structures with long alkyl chains that they can be transformed to fuels of gasoline, kerosine, and gas oil; aromatics; and hydrogen. Although transesterification with methanol provides biodiesel fuels, reactants such as glycerol, surplus methanol, and catalysts remain in the final stage of the reaction and have to be removed prior to their utilization.[1−4] Meanwhile, heterogeneous transesterification making fatty acid methyl ester (FAME),[5−7] catalytic cracking making gasoline, propylene, and aromatics for petrochemicals,[8−32] and hydrotreating of fats to biodiesel fuels with high cetane numbers and bio-jet fuels[16−18,33−51] have already been reported during the last decade. Among them, the hydrotreating of fats gives not only vehicle fuels, C3 fraction, and light gas oil with the higher durability for oxidation than that of FAME but also raw materials for petrochemicals, especially aromatics.[52−63] For example, after treatment of fats by ZSM-5 with mesopore and macropore at 300 °C and pressure up to and including 4.0 MPa, hydrotreatment of the oils obtained was performed using NiMo/Al2O3 at 340 °C and 4.5 MPa.[52,61] The two-step process using ZSM-5 with a 0.6 cm3/g pore volume at a high pressure gave 25–29 wt % n-paraffin and 8–15 wt % aromatics, whereas aromatics were not given by ZSM-5 with 0.35 cm3/g pore volume. When deoxygenation of fats was performed using CaO and hydrotalcite at 400 °C in nitrogen atmosphere,[53] more than 80% of hydrocarbon products were obtained. However, the selectivity of aromatics was less than 6%. When sampled vegetable oils were hydrotreated over Pt/Al2O3/SAPO-11 at about 380 °C and 30 bar, the product fuels included ca. 15 wt % aromatics.[42,54] Hydrotreating, which gives the stability in products, is performed at a relatively higher hydrogen pressure, and therefore, the yield of aromatics is usually low. In contrast, catalytic cracking of fats gives a higher yield of aromatic compounds.[8,55−59,62,63] For example, catalytic cracking of a fat using HZSM-5 and Mo/HZSM-5 at 650 °C produced more than 50% of aromatic compounds.[55] In another case, the catalytic cracking of a vegetable oil using HZSM-5 (Si/Al = 80) at 450–550 °C in a fixed-bed reactor produced more than 40% of organic liquid products consisting mostly of aromatics.[56] The catalytic cracking gave higher yields of aromatic compounds; however, the unavoidable deactivation occurs in the catalyst used.

It seems that this hydrotreating would have advantages in transforming fats to fuels, and thus we used various types of class="Chemical">zeolites in catclass="Chemical">pan class="Chemical">alyst preparation to obtain gasoline directly by improving their pore structure and acidity at the hydrogen pressure as high as 3.0 MPa.[33] In the course of our study, hydrotreating of fats was performed under the following conditions: hydrogen pressures lower than 1.0 MPa and temperatures higher than 500 °C, which, however, were not common for hydrotreating and hydrocracking. This dehydrocyclization-cracking of fats is hypothetically drawn in the table of contents (TOC), where the catalysts used were the ZSM-5-Al2O3 (ZA) composite-supported Pt/NiMo (NM) sulfide catalysts with hierarchical structure, and a gasoline fraction including significant amounts of aromatics is obtained.[64] In this TOC, a molecule of fat includes four double bonds in alkyl long chains, which are assumed to yield 6 molecules of aromatics, 3 molecules of CO2, 1 molecule of C3H8, and 14 molecules of hydrogen. In this scheme, about 7 m3 of gaseous hydrogen could be prepared from 18 L of fat. However, the composite-supported zeolite catalysts did not produce hydrogen while fat was completely hydrocracked with 1 wt % Pt at 1.0 MPa of hydrogen and 580 °C; 12% of aromatics was obtained at 580 °C when K- and Na-added aluminas and HZSM-5 were mixed, and Pt and NiMo sulfides were supported.[64] Further, β-zeolite-Al2O3 composite-supported Pt/NiMo sulfide catalysts yielded aromatics even at 0.5 MPa of hydrogen.[65] On the other hand, it is known that the addition of Zn to zeolites could improve the formation of aromatics in the reforming of saturated hydrocarbons, as our research group has reported recently the precise reaction route from n-pentane to BTX on Zn-promoted ZSM-5.[66]

In this reported study, the dehydrocycliclass="Chemical">zation-cracking of class="Chemical">pan class="Species">soybean oil using Zn-exchanged ZSM-5-Al2O3 composite-supported Pt/NiMo sulfide catalysts was performed at a low hydrogen pressure of 0.5 MPa and the effects of Zn addition on the conversion of soybean oil and the formation of aromatics were investigated. Although the effect of the addition of Zn to ZSM-5 on the reaction remains unknown, it was found in the present study that the metal sulfide catalysts could convert soybean oil to valuable fuels and aromatics even at a low pressure 0.5 MPa hydrogen and a higher temperature of 580 °C and also that the addition of Zn would increase aromatics yields.

Results and Discussion

X-Ray Diffraction (XRD) Patterns for Crystal Structures and N2 Adsorption and Desorption Measurement for Pore Structures of Pt/NiMo/ZnZSM-5-Al2O3 Composite Catalysts

The XRD patterns of the test catclass="Chemical">alysts before use are shown in Figure , where characteristic class="Chemical">pan class="Chemical">signals for the zeolite crystals and Al2O3 were observed, while there were no such signals for the metal species. The results indicated that the zeolite crystals were kept in the composite supports and that the metal species of Pt, Mo, and Ni were well dispersed in the zeolite-Al2O3 composite catalysts. The amount of Pt addition was lower than those of Mo and Ni, and therefore, it may be difficult to detect the former by XRD. However, 1 wt % Pt would appear as XRD signals if it forms crystals. In contrast, the added amounts of Ni and Mo were large, and therefore, if there were crystals on the catalysts, they would appear as XRD signals. When Ni and Mo sulfides are supported on alumina, it is known that they are dispersed on alumina.[67] On the other hand, when they could not be dispersed on a support, they exactly appeared as XRD signals. When these signals are not detected in XRD, the Ni and Mo sulfides under discussion are likely to be dispersed on alumina. After the reaction, the XRD patterns of all catalysts tested were almost the same, as shown in Figures S1 and S2. These results are consistent with those reported previously,[33,64,65] where the zeolite-Al2O3 composites worked as appropriate supports for the molybdenum and nickel species, while the metal anionic species would not be retained on a single zeolite.

Figure 1

XRD patterns of Pt/NM/ZnZSM-5-Al2O3 composite catalysts.

XRD patterns of class="Chemical">Pt/class="Chemical">pan class="Chemical">NM/ZnZSM-5-Al2O3 composite catalysts.

Pore structures of the class="Chemical">Pt/class="Chemical">pan class="Chemical">NiMo/ZnZSM-5-Al2O3 composite catalysts were estimated by the Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) methods in N2 adsorption and desorption measurement. Table tabulates the results. As the content of ZSM-5 or ZnZSM-5 was 25 wt %, the effects of these microporous materials on pore structures were rather weak, and the values of BET surface areas and total pore volumes were very close to those obtained by the BJH method. Thus, the above-mentioned catalysts exhibited very similar surface areas and pore volumes, which were characteristic of those of alumina. After the reaction, the pore structures were almost the same as shown in Table S1.

Table 1

Surface Areas, Pore Volumes, and Pore Diameters by the BET and BJH Methods

				BJH
catalyst	BET SAa (m2/g)	total PVa (cm3/g)	average PDa (nm)	SAa (m2/g)	PVa (cm3/g)	PDa peak (nm)
Pt/NM/ZA	217	0.47	8.8	225	0.46	10.6
Pt/NM/Zn(11)ZA	227	0.48	8.5	236	0.47	10.6
Pt/NM/Zn(30)ZA	237	0.48	8.1	205	0.45	10.6
Pt/NM/Zn(122)ZA	225	0.48	8.5	209	0.45	10.6

SA: surface area; PV: pore volume; PD: pore diameter. In catalyst names, N, M, Z, and A indicate Ni, Mo, ZSM-5, and alumina, respectively.

SA: surface area; PV: pore volume; class="Disease">PD: class="Chemical">pore diameter. In catclass="Chemical">pan class="Chemical">alyst names, N, M, Z, and A indicate Ni, Mo, ZSM-5, and alumina, respectively.

class="Chemical">NH3-temclass="Chemical">perature-class="Chemical">programmed desorclass="Chemical">pan class="Chemical">ption (TPD) profiles and the amount of NH3 desorbed are shown in Figure . One broad peak and a long gradual slope were observed between 200 and 600 °C, indicating that the catalysts analyzed consisted of Al2O3, which would make peaks broader. It is likely that, as the amount of NH3 desorbed was about twice the moles of Al constituting ZSM-5, a significant amount of NH3 would be adsorbed on the Al2O3 component of these catalyst systems.[33,64] In Pt/NM/Zn(122)ZA, the broad peak between 200 and 300 °C was shifted to the higher-temperature range and the shoulder was observed between 500 and 600 °C. When NH3-TPD data were divided into weak acid sites and strong acid sites at 350 °C, the amount of the strong acid sites tended to increase with increasing amount of Zn added, suggesting that Zn, Si, and Al species would interact with each other and make new acid sites other than those of HZSM-5 itself on Zn sites deposited on the surfaces of ZSM-5.

Figure 2

NH3-TPD curves and amount of NH3 desorbed in Pt/NM/ZnZSM-5-Al2O3 composite catalysts. The total amount of NH3 desorbed is given on the right-hand side of the figure. The amounts of weak acid sites detected in the range of 100–350 °C (left) and strong acid sites detected in the range of 350–650 °C (right) are given in parentheses.

class="Chemical">NH3-Tclass="Chemical">pan class="Disease">PD curves and amount of NH3 desorbed in Pt/NM/ZnZSM-5-Al2O3 composite catalysts. The total amount of NH3 desorbed is given on the right-hand side of the figure. The amounts of weak acid sites detected in the range of 100–350 °C (left) and strong acid sites detected in the range of 350–650 °C (right) are given in parentheses.

Dehydrocyclization-Cracking of Soybean Oil over Pt/NiMo/ZnZSM-5-Al2O3 Composite Catalysts

Figure shows the effects of temperature on class="Chemical">converclass="Chemical">pan class="Chemical">sion of soybean oil via dehydrocyclization-cracking reaction using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts. The solid lines indicate the results obtained at 1.0 MPa of hydrogen pressure, and the dotted lines indicate the results at 0.5 MPa of hydrogen. The conversions increased with increasing temperature and reached 100% for all catalysts tested at 580 °C. At a lower temperature of 420 °C, the conversions obtained at 0.5 MPa were higher than those at 1.0 MPa, which would be due to the following reasons: the catalyst surfaces might be reduced at the higher hydrogen pressure, while the activity has not been lost at the higher temperature. Further, it would be mainly because the relative partial pressure of the soybean oil on the catalyst surfaces decreased with hydrogen pressure increasing to 1.0 MPa. As it is assumed that the reaction follows the Langmuir–Hinshelwood rate equation as known in the usual hydrotreating reaction,[67] the retarding term of hydrogen in the denominator of the L–H equation would affect and decrease the rate at the higher pressure of hydrogen. The effects of Zn addition differed depending on the pressure. At the higher hydrogen pressure of 1.0 MPa, the conversion at the lower temperature increased with increasing Zn exchange ratio, suggesting that Zn would activate hydrogen molecules[66] to increase the activity. In contrast, at the lower pressure of 0.5 MPa, the catalyst without Zn exchange revealed the highest conversion at the lower temperature, suggesting that the acid sites of ZSM-5 rather than the reduced Zn sites would directly act as the key active sites.

Figure 3

Effect of temperature on the conversion of soybean oil via dehydrocyclization-cracking reaction using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts.

Effect of temperature on the class="Chemical">converclass="Chemical">pan class="Chemical">sion of soybean oil via dehydrocyclization-cracking reaction using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts.

The reaction profiles of class="Chemical">all catclass="Chemical">pan class="Chemical">alysts are summarized in Tables , 3, and S2–S5. These tables tabulate the conversion, branched hydrocarbon-to-straight-chain hydrocarbon (iso-/n-) ratios, research octane number (RON) values for C5–C14, olefin/paraffin ratio (O/P), cetane number, selectivity of gaseous fraction (C1–C4), gasoline fraction (C5–C11), kerosene fraction (C12–C14), diesel fraction (C15–C18), C19 and CO2 + CO, aromatics yield, and the material balance between feed and product. When the reaction profiles at 500 °C were compared with respect to Table and Figure , the conversion approached 100% and the selectivity of gaseous products was kept lower than 40% at 500 °C even at the higher pressure 1.0 MPa. When Zn was added, the fraction with more than C18 (C19 in the tables mentioned above) decreased, the selectivity of gaseous products remained unchanged and the yield of aromatics slightly increased, compared to the catalyst without the addition of Zn.

Table 2

Conversion, Iso/n Ratio, RON, Olefin/Paraffin Ratio, Cetane Number, Selectivity of Products, Aromatics Yield, and Material Balance at 1.0 MPa and 500 °Ca

catalyst	conv. (%)	iso-/n (C5–C14)	RON (C5–C14)	O/P (C2–C3)	CN (C15–C18)
Pt/NM/ZA	98	1.40	89	0.80	76
Pt/NM/Zn(11)ZA	83	1.32	84	0.76	72
Pt/NM/Zn(30)ZA	94	1.39	89	0.80	74
Pt/NM/Zn(122)ZA	100	1.42	87	0.87	74

In catalyst names, N, M, Z, and A indicate Ni, Mo, ZSM-5, and alumina, respectively.

Table 3

Conversion, Iso/n Ratio, RON, Olefin/Paraffin Ratio, Cetane Number, Selectivity of Products, Aromatics Yield, and Material Balance at 0.5 MPa and 580 °Ca

catalyst	conv. (%)	iso-/n (C5–C14)	RON (C5–C14)	O/P (C2–C3)	CN (C15–C18)
Pt/NM/ZA	100	0.68	77	0.77	74
Pt/NM/Zn(30)ZA	100	0.69	77	0.66	73
Pt/NM/Zn(122)ZA	100	0.76	81	0.66	73

In catalyst names, N, M, Z, and A indicate Ni, Mo, ZSM-5, and alumina, respectively.

Figure 4

Carbon number distribution of products in dehydrocyclization-cracking reaction of soybean oil using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts.

class="Chemical">Carbon number distribution of class="Chemical">products in dehydrocycliclass="Chemical">pan class="Chemical">zation-cracking reaction of soybean oil using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts.

In catclass="Chemical">alyst names, N, M, Z, and A indicate Ni, Mo, class="Chemical">pan class="Chemical">ZSM-5, and alumina, respectively.

In catclass="Chemical">alyst names, N, M, Z, and A indicate Ni, Mo, class="Chemical">pan class="Chemical">ZSM-5, and alumina, respectively.

On the other hand, when the reaction profiles were class="Chemical">comclass="Chemical">pared at 580 °C and the lower class="Chemical">pressure 0.5 MPa with resclass="Chemical">pect to Table and Figure , the class="Chemical">pan class="Chemical">conversion reached 100% and the selectivity of gaseous products kept lower, i.e., about 25%. Similar to the cases of the higher pressure 1.0 MPa, with the addition of Zn, the fraction with more than C18 (C19 in the tables above) decreased and the yield of aromatics increased, compared to the catalyst without the addition of Zn. The aromatics yield of Pt/NM/Zn(122)ZA at 0.5 MPa and 580 °C was the highest 13% among the catalysts tested. Even at 580 °C, the selectivity of liquid products with C5–C18 was higher at the hydrogen pressure 0.5 MPa than that at 500 °C and 1.0 MPa, reached approximately 60%, and increased with increasing amount of Zn added.

As seen in the generclass="Chemical">al information from class="Chemical">pan class="Chemical">all reaction data in Tables S2–S5 for all catalysts, the conversion and the aromatics yield increased and the ratios of iso-/n- and O/P of C2 and C3 decreased with increasing temperature. The decrease in the O/P ratio may be due to the cyclization of C2 and C3 olefins to aromatics.[66] The selectivity of CO and CO2 tended to increase with increasing temperature. The transportation fuels with C5–C18 were obtained with a selectivity of about 73% at 420 °C and 0.5 MPa of hydrogen in both absence and presence of Zn.

The class="Chemical">coke formation in the used catclass="Chemical">pan class="Chemical">alysts was analyzed by thermogravimetry-differential thermal analysis (TG-DTA) measurement, and the results are shown in Table . The weight loss at 400–600 °C was attributed to coke, as judged from the TG-DTA curves shown in Figure S3. The coke formation decreased with the addition of Zn, indicating that the presence of Zn would inhibit the coke formation probably because of the promotion of hydrogen transfer near Zn sites of ZSM-5. Further, the coke formation was not affected by the hydrogen pressure between 0.5 and 1.0 MPa, indicating that hydrogen could be provided to carbonaceous fragments even at the pressure as low as 0.5 MPa of hydrogen.

Table 4

Coke Formation Analysis of Used Catalysts by TG-DTA

	H2 press. (MPa)	200–300 °C (mg)	300–400 °C (mg)	400–500 °C (mg)	500–600 °C (mg)	totala (mg)
Pt/NM/ZA	1.0	0.21	0.43	1.41	2.08	4.13 (3.49)
Pt/NM/Zn(11)ZA	1.0	0.23	0.41	0.86	2.44	3.94 (3.30)
Pt/NM/Zn(30)ZA	1.0	0.08	0.17	1.07	2.30	3.62 (3.37)
Pt/NM/Zn(122)ZA	1.0	0.13	0.25	1.19	2.21	3.78 (3.40)
Pt/NM/ZA	0.5	0.27	0.45	1.23	2.18	4.13 (3.41)
Pt/NM/Zn(30)ZA	0.5	0.18	0.30	0.92	2.26	3.65 (3.18)
Pt/NM/Zn(122)ZA	0.5	0.21	0.41	1.25	1.97	3.85 (3.22)

The weight loss at 400–600 °C is shown within parentheses. In catalyst names, N, M, Z, and A indicate Ni, Mo, ZSM-5, and alumina, respectively.

The class="Disease">weight loss at 400–600 °C is shown within class="Chemical">parentheses. In catclass="Chemical">pan class="Chemical">alyst names, N, M, Z, and A indicate Ni, Mo, ZSM-5, and alumina, respectively.

Reaction Routes in Dehydrocyclization-Cracking Using Pt/NM/ZnZSM-5-Al2O3 Composite Catalysts

Figure shows the reaction routes of dehydrocycliclass="Chemical">zation-cracking of class="Chemical">pan class="Species">soybean oil. The structure of triglyceride would be converted by hydrocracking to three molecules of aliphatic acids and propane on the surface PtNiMo sulfide on Al2O3 support. The converted aliphatic acids would be decomposed through decarboxylation to hydrocarbon fragments, which would be further decomposed by cracking on the acid sites of the catalyst. This reaction may be promoted on the surface of NiMo sulfide, Al2O3, or ZSM-5, which would be used as acid sites at temperatures of 500 °C and higher. Finally, C3 and C4 olefins would be formed and be transformed to aromatics through Diels–Alder reaction on Zn species of ZnZSM-5. Reduced Zn species may control the size of micropores in ZSM-5 and promote cyclization of olefins and successive dehydrogenation even at both the lower temperature and higher pressure.

Figure 5

Reaction routes of aromatics formation of soybean oil via dehydrocyclization-cracking using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts.

Reaction routes of aromatics formation of class="Species">soybean class="Chemical">pan class="Chemical">oil via dehydrocyclization-cracking using Pt/NiMo/ZnZSM-5-Al2O3 composite catalysts.

In generclass="Chemical">al, the hydrotreatment of a fat in class="Chemical">pan class="Chemical">hydrogen atmosphere progresses by hydrodeoxygenation, decarbonylation, and decarboxylation. Among these reactions, decarboxylation and decarbonylation consume less hydrogen. When the reaction progresses through decarboxylation, the yield of CO2 is expected to reach 15%. The amount summing the formed CO2 and the CO2 estimated from the formed CO was lower than 15 wt %, indicating that hydrodeoxygenation would occur to some extent. However, it was thought that more than half of the sample soybean oil would be converted through decarboxylation and decarbonylation using the Pt/NiMo sulfide.[64,65] Further, Zn-exchanged ZSM-5 might affect the formation of CO and CO2, as it is known that the basic components promote the decarboxylation and decarbonylation.[64,65] In our previous report, the test ZnZSM-5-Al2O3 composite catalysts promoted the aromatics formation through dehydrocyclization of n-pentane, and it was proposed that dehydrocyclization would progress selectively on reduced Zn sites through Diels–Alder reactions of C4 olefins with ethene, propene, and butene forming benzene, toluene, and xylene, respectively.[66] It is known that hydrocarbons with C1–C5 form aromatics on catalysts including ZSM-5.[66,68−71] In the present study, although gases were relatively small in amount, aromatic compounds were formed significantly, suggesting that cyclization might directly occur without conversion to gaseous hydrocarbons to some extent. As Zn was incorporated into ZSM-5 initially during the preparation of a catalyst, it seems that Zn would act as an active species for cyclization and aromatization of hydrocarbon fragments in the structure of ZSM-5. On the other hand, Pt was supported on the catalyst at the final preparation stage, suggesting that the interaction degree between Zn and Pt could be rather low.

When only class="Chemical">NiMo was used without class="Chemical">pan class="Chemical">Pt, the activity was rather low. The addition of 0.5 wt % Pt with NiMo was not enough to obtain the activity.[64] Therefore, 1 wt % Pt was used for the catalysts examined. Pt would provide enough hydrogen for a NiMo catalyst to convert the large molecule of a fat. Further, it seems that NiMo sulfide species could initially accelerate the conversion of fat structure to a fatty acid and that the carboxylic acid would successively be decomposed, as shown in Figure .

To decrease the class="Chemical">consumclass="Chemical">pan class="Chemical">ption of hydrogen, the decarboxylation route is preferred. When CO2 is formed selectively, however, the reverse reaction with hydrogen may occur forming CO and H2O. To avoid this reaction, basic compounds were added to the composite supports,[64] which increased decarboxylation and decarbonylation. Generally, basic carriers may reduce aromatics formation because of the decrease in the amounts of acid sites in zeolites. The addition of Zn increased the aromatics formation with the decarbonylation and decarboxylation routes maintained, suggesting that the addition of appropriate substance could achieve simultaneously the formation of aromatics and the decomposition of carboxylic acid without excess consumption of hydrogen.

Conclusions

The test class="Chemical">Zn-exchanged ZSM-5-Al2O3 class="Chemical">pan class="Chemical">composite-supported Pt/NiMo sulfide catalysts increased aromatic compounds compared to the catalysts without Zn and exhibited 11–12% of aromatics yields at 1.0 MPa and 500 °C in the dehydrocyclization of soybean oil. The Zn-exchanged catalysts also increased aromatics yields even at 0.5 MPa and 580 °C, and the Pt/NM/Zn(122)Z(24)60Al catalyst exhibited the low selectivity of gaseous products (26%), while exhibiting the high selectivity of transportation fuels with C5–C18 (63%) and the high yield of aromatics (13%). Zn-exchanged catalysts inhibited the coke formation, compared to the catalysts without Zn.

The class="Chemical">Zn-exchanged ZSM-5-Al2O3 class="Chemical">pan class="Chemical">composite-supported Pt/NiMo sulfide catalyst referred to here can be categorized as a hierarchical type. Although our catalysts may not be the best alternatives for giving aromatic compounds, it is evident that the coexistence of ZSM-5 with very active micropores and Al2O3 with mesopores, which can completely maintain Mo anionic species on the surfaces and diffuse the very large carbonaceous molecules such as fats,[67] would bring about the very good performance. Although dehydrocyclization-cracking can be regarded as a technology capable of processing large carbonaceous molecules, it is still required to increase the aromatics selectivity, whereas the more accurate construction of the catalyst materials used would lead to the further development of the hierarchical catalysts as an integral part of the dehydrocyclization-cracking process technology described above.

Experimental Section

Material and Catalyst Preparation

The elementclass="Chemical">al anclass="Chemical">pan class="Chemical">alysis of soybean oil (Wako, extra pure) was C: 76.64%; H: 10.95%; N: 0.30%, and O: balance. Its single components were linoleic acid 50–56%, oleic acid 21–28%, palmitic acid 9–12%, stearic acid 3–5%, and linolenic acid 7–10%. Soybean oil was used without further purification. A zeolite used was ZSM-5 (H type, SiO2/Al2O3 (mol/mol) 24, HSZ-822HOA, Tosoh). Zn-exchanged ZSM-5 was initially prepared by the conventional ion-exchange method using zinc nitrate hexadydrate and aqueous ammonia solution to control pH, and the detailed method was given in the previous paper.[66] Most of Zn added was incorporated into ZSM-5, and the exact amount of Zn was determined by X-ray fluorescence (XRF, Shimadzu EDX-720).

class="Chemical">Al2O3 (surface area, 250 m2/g; class="Chemical">pore volume, 0.70 cm3/g; average class="Chemical">pore class="Chemical">pan class="Chemical">size, 8 nm) was normally used for hydrotreating and was supplied from Nippon Ketjen. Alumina sol (Cataloid AP-1, 70 wt % of Al2O3) was used for a binder and was supplied from Shokubai Kasei. The conventional kneading method was used to make hierarchical Zn-exchanged ZSM-5-Al2O3 composite supports. ZnZSM-5 (25 wt %), Al2O3 (60 wt %), and alumina sol (15 wt %) as Al2O3 were mixed with ion-exchanged water and calcined at 500 °C for 3 h in air.[33,64] Metal-loaded catalysts were prepared by the conventional impregnation method using hierarchical composite supports. An aqueous solution of Mo and Ni species including ammonium heptamolybdate tetrahydrate (Nakarai Tesque, (NH4)6Mo7O24·4H2O) and nickel nitrate hexahydrate (Nakarai Tesque, Ni(NO3)2·6H2O) was used. After impregnation, catalyst samples were dried and calcined at 500 °C for 3 h under air stream. After calcination, catalysts included 16 wt % MoO3 and 3.3 wt % NiO (Ni/Mo ratio, 0.4). An aqueous solution of H2[PtCl6]·6H2O was successively used to add 1 wt % Pt to the NiMo catalysts. The name of a composite support was shown with Zn in the top with Zn/Al ratio (mol/mol, Al in ZSM-5) in parentheses, and the abbreviations Z and A for zeolite type ZSM-5 and Al2O3, respectively. The name of a metal-supported catalyst was shown as Pt/NM/support name, where N is Ni and M is Mo.

Characterization of Catalysts and Catalytic Reaction

XRD, class="Chemical">N2 adsorclass="Chemical">pan class="Chemical">ption and desorption, TG-DTA, and NH3-TPD were measured to characterize the catalysts prepared. The detailed methods have been reported elsewhere.[33,64] XRD patterns were measured to analyze the crystal structure of catalysts using Rigaku Ultima IV under the following conditions: sample, 0.10 g; 2θ = 10–70°; Ni-filtered single-colored Cu Kα1 radiation (λ = 0.15418 nm); current, 20 mA; voltage, 40 kV; scan speed, 1 °/min; scan mode, continuous; detecting slit, 2/3°; scattering slit, 2/3°; radiation slit, 0.45 mm; present time, 1 s.

To estimate the precise pore structures of catclass="Chemical">alysts, class="Chemical">pan class="Chemical">N2 adsorption and desorption was measured (BELSORP mini II, Nihon BEL Co. Ltd.). The total surface areas of the catalysts were estimated by the BET method. The surface areas (SA), pore volumes (PV), and pore diameters (PD) of mesopores larger than the pore size of 3.3 nm were estimated by the BJH method. To analyze the changes in the pore structure of catalysts, the spent catalyst was calcined at 500 °C for 12 h to remove residual coke after the reaction was performed. SAs, PVs, and PDs measured by the BET and BJH methods are tabulated in Table .

The amounts of class="Chemical">NH3 desorbed for catclass="Chemical">pan class="Chemical">alysts were measured to evaluate their acidic properties using a gas chromatography-thermal conductivity detector (GC-TCD, Shimadzu GC-8A) under the following conditions: weight of catalyst, 0.04 g; column, 140 °C; injection and detector temperature, 170 °C; current, 100 mA; carrier gas, 10 mL/min; and attenuation.[16] NH3 was adsorbed at 100 °C by introducing pulses of NH3 of 1 cc until further adsorption of NH3 does not occur.

To anclass="Chemical">alyze class="Chemical">pan class="Chemical">coke formation, TG-DTA (Shimadzu, DTG-60AH) of spent catalysts was performed under the following conditions: 10 mg of catalyst sample added in an Al pan; temperature range, 25–600 °C, and heating rate, 10 °C/min.

Dehydrocyclization-Cracking of Soybean Oil Using Pt/NiMo/ZnZSM-5-Al2O3 Composite Catalysts

A catclass="Chemical">alyst (1 g; 600–355 μm: 70 wt %, 355–125 μm: 30 wt %) was class="Chemical">packed into a class="Chemical">pan class="Chemical">conventional fixed-bed flow reactor (ID, 8 mm; length, 30 cm). Presulfiding of the catalyst was performed under the following conditions: 30 cc/min of 5% H2S/H2, 0.1 MPa, 400 °C, 3 h. After cooling, the reactor was pressurized by 0.5 or 1.0 MPa of H2 and the top of the reactor was preheated by a ribbon heater at around 210 °C. Then, 100% soybean oil was introduced and the temperature was increased at a heating rate of 5 °C/min. The dehydrocyclization-cracking was operated under the following conditions: 0.5 or 1.0 MPa; H2, 300 mL/min; weight hourly space velocity (WHSV), 6.4 h–1; 420–580 °C. Gas and liquid products including hydrocarbons, CO, and CO2 were separated in a gas–liquid separator. The volume of gaseous products and weight of liquid products were measured separately. Liquid products with C5–C18 hydrocarbons emerging at 420–580 °C were determined by GC-flame ionization detection (FID). The gaseous products were determined by GC-FID (Shimadzu GC-2014 for gaseous hydrocarbons) and GC-TCD (for CO and CO2). Diesel fractions of liquid products were determined by GC-FID (Shimadzu GC-2014), and gasoline fractions of liquid products were determined by GC-FID with PONA solution (Shimadzu GC-2010). The determined gaseous and liquid products were reported in our previous paper.[64] The database in the GC-PONA solution determined more than 95% of the products. The analytical conditions for these GC-FID and GC-TCD were described elsewhere.[33,64]

The product selectivity was estimated in Table acclass="Chemical">cording to eq where Wclass="Chemical">pan class="Chemical">gas is the weight of gaseous products and Wliquid is the weight of liquid products. Wgas was estimated from the gas flow rate in the outlet and by the analysis of GC. Wliquid was weighed directly. The total product weight in each fraction was estimated by the ratio of the fraction in GC for gas and liquid products and the values of Wgas and Wliquid.