UC Berkeley

The oligomerization of propene to produce higher molecular weight molecules was investigated as a model reaction pathway for the synthesis of liquid transportation fuels and fuel additives from C2 to C5 light olefins. In this scheme, light olefins could come from a variety of sources including the cracking of petroleum, as a byproduct in the production of hydrocarbons from synthesis gas during Fisher-Tropsch synthesis, or from the dehydration of alcohols created during biomass fermentation. Transformation of these light olefins into heavier molecules could allow for future production of transportation fuels from many carbon-rich sources, including natural gas, coal, and biomass, instead of the current system that relies almost exclusively on petroleum.

Microporous and mesoporous Brønsted acidic and exchanged nickel materials are the most common heterogeneous catalysts for the oligomerization of light olefins into heavier products. Much is unknown about the role of the catalyst in influencing the oligomer size and the degree of oligomer branching - both characteristics crucial to the production of high quality liquid fuels - making the selection and design of appropriate oligomerization catalysts challenging. It was therefore the goal of this dissertation to establish how the catalyst site, proximity of sites, and catalyst support influence the final product distribution of oligomers. The discussion begins with an examination of the role of the acid site density in the Brønsted acidic zeolite H-MFI on the activity and selectivity to propene dimers. An increase in the aluminum site density, represented by a decrease in the catalyst Si/Al ratio from 140 to 10, was determined to decrease the conversion of propene to heavier products from 75% to 10% at 548 K. Examination of the reaction pathways for oligomer formation using kinetic analyses and DFT simulations indicate that site density influences the relative rates of oligomer growth and desorption. Specifically, the high loading of hydrocarbons in zeolites with low Si/Al ratios limit oligomer growth beyond the dimer lowering the propene conversion, as fewer oligomers are formed, but also increasing dimer selectivity due to the smaller concentration of long oligomers required for secondary cracking reactions.

Regardless of the Si/Al ratio in H-MFI, the activity of the Brønsted acid sites for oligomer cracking and aromatic formation limit the control over the product distribution with these catalysts. To achieve better oligomer control and limit secondary oligomer reactions, heterogeneous nickel-exchanged aluminosilicates were explored. These materials can achieve near complete conversion of ethene to oligomers with > 98% selectivity at high olefin pressures; however, the manner in which these catalysts convert light olefins into heavier products is not understood. Therefore, to determine any potential benefit to using these catalysts over Brønsted acidic zeolites, the reaction mechanism, state of nickel sites, and influence of catalyst support were investigated to determine their roles in catalyst activity and oligomer branching.

A series of Ni-exchanged Na-X zeolites with various nickel loadings were successfully synthesized via aqueous ion exchange with nickel (II) nitrate and explored as propene oligomerization catalysts. Characterization of Ni-Na-X indicates that Ni remains Ni2+ both after synthesis and under reaction conditions, contrary to previous reports. Although all catalysts were > 98% selective to oligomers at 453 K and 1-5 bar propene pressure, the catalyst activity was determined to be a strong function of the nickel loading. At high nickel loadings, the catalyst is active immediately upon exposure to propene but deactivates rapidly to 0% conversion. As the nickel loading is decreased below 1 wt%, however, the catalyst exhibits low initial activity and instead activates with time on stream, before deactivating and reaching a non-zero steady-state activity after more than 2000 min of time on stream. Development of a reaction network and subsequent microkinetic model indicates that the activation period is caused by migration of Ni2+ cations from inaccessible positions of the zeolite to the supercage, where catalysis occurs. The subsequent catalyst deactivation is caused by complexation of nearby sites within the zeolite supercage leaving only isolated Ni2+ sites active at steady state.

Once an understanding of the time on stream activity profile was established, the role of the support on the catalyst activity and degree of dimer branching was examined. Exchanging the non-catalytic co-cation in the zeolite, Na+ in Ni-Na-X, for other alkali metal and alkaline earth co-cations was determined to influence both the propene oligomerization activity and dimer isomer distribution. Specifically, Li+, the smallest alkali metal co-cation, and Sr2+, the largest alkaline earth co-cation examined, led to the highest dimer branching and catalyst activity per Ni2+ cation in their respective groups. It was determined that this effect was caused by both larger cations expanding the zeolite lattice and alkali metal cations present in the zeolite supercage taking up otherwise open pore volume. This led to the conclusion that space around the Ni2+ cations in the supercage is what governs catalytic activity and dimer branching in these catalysts.

The realization that space around the Ni2+ site controls catalyst activity led to the exploration of larger mesoporous aluminosilicate structures as potentially more active propene oligomerization catalysts. To this end, Ni-exchanged MCM-41 and MCM-48 (pore size = 23 Å) and SBA-15 (pore size = 57 Å) were synthesized and examined as oligomerization catalysts. It was determined that the same principles established in zeolites for making an active catalyst, such as high Ni2+ dispersion, were still applicable to these larger-pored systems. As predicted, further increasing the space around the active site did increase the catalyst activity with the highest activity per Ni2+ site existing for the SBA-15 material. The decreased steric constraints from the support in these structures, however, led to increased trimer production as well as catalyst deactivation caused by heavy molecules depositing in the pores. The more open environment also resulted in less control over the degree of dimer branching causing all mesoporous catalysts to produce a 49/51 mixture of branched to linear dimers at 453 K and 1 bar propene pressure.