Short-chained olefins and oxygenates are abundantly available from renewable or petroleum-based feedstocks, but further processing is needed to upgrade these molecules into useful fuels and products. Bimolecular coupling is an excellent way to add functionality and increase the carbon chain length of molecules. Understanding the various mechanisms and site requirements for this class of reactions is vital towards rational design of superior catalysts.
We first investigated the bimolecular condensation of ethanol to n-butanol. A variety of heterogeneous catalysts were screened for ethanol coupling and it was found that hydroxyapatite (HAP; Ca5(PO4)3OH) displayed an unusually high activity and selectivity (>80%) to butanol. Steady-state, gas-phase kinetics over HAP showed the reaction is autocatalytic with respect to acetaldehyde. Furthermore, the rate of hydrogen transfer between alcohols and aldehydes is much faster than the rate of aldol condensation. These results indicate that butanol is formed through a Guerbet pathway, in which ethanol dehydrogenates to form acetaldehyde, acetaldehyde undergoes aldol condensation to form crotonaldehyde, and crotonaldehyde is hydrogenated to form butanol. When the reaction was conducted in the presence of CO2, butanol formation rates were suppressed by a factor of six while acetaldehyde formation rates stayed constant, demonstrating that the active sites for the two reactions are different. By comparing the CO2 titration results with in-situ infrared spectroscopy, it was found that ethanol dehydrogenation is catalyzed by basic Ca-O sites on the surface of HAP while aldol condensation requires CaO/PO43- pairs.
The structural requirements for an aldol condensation catalyst were further probed by investigating the bimolecular coupling of acetone to methyl isobutyl ketone (MIBK) in the presence of hydrogen over a physical mixture of HAP and Pd/SiO2. The reaction is found to proceed by consecutive aldol addition to form diacetone alcohol (DAA), dehydration of DAA to mesityl oxide (MO), and hydrogenation of MO to MIBK. The products formed by feeding DAA and MO reveal that aldol addition of acetone is rapid and reversible, and that the subsequent dehydration of DAA is rate-limiting. Pyridine and CO2 titration show that aldol dehydration occurs over basic sites via an E1cB mechanism. A series of cation-substituted HAP samples were prepared by ion-exchange to further investigate the role of acid-base strength on catalyst performance. Characterization of these samples by PXRD, BET, ICP-OES, XPS, CO2-TPD, and Raman spectroscopy demonstrated that the exchange procedure used does not affect the bulk properties of hydroxyapatite. DFT calculations reveal that in addition to affecting the Lewis acidity/basicity of the support, the size of the cation plays a significant role in the chemistry: cations that are too large (Ba2+) or too small (Mg2+) adversely affect reaction rates due to excessive stabilization of intermediate species. Strontium-exchanged hydroxyapatite was found to be the most active catalyst because it promoted α-hydrogen abstraction and C-O bond cleavage of DAA efficiently.
The scope of this work was then extended to probe bimolecular reactions that involve C-N bond formation. Using the knowledge gained from previous studies, we developed a Ni-HAP catalyst that was active for propanol amination to propylamine. The catalyst is an order of magnitude more active than a typical Ni/SiO2 catalyst and displays a higher selectivity towards the primary amine. The reaction proceeds via dehydroamination, a process that involves sequential dehydrogenation, condensation, and hydrogenation. Kinetic and isotopic studies indicate that α-H abstraction from propoxide species limits the rate of the dehydrogenation step, and hence the overall rate of reaction. The rate of propanol dehydrogenation depends on the composition of the support and on the concentration of Ni sites located at the interface between the Ni nanoparticles and the support. The superior performance of Ni/HAP is attributed to the high density of basic sites on HAP which are responsible for stabilizing alkoxide intermediates and suppressing the disproportionation and secondary amination of amines.
Finally, we investigated an alternate pathway towards synthesizing alkylamines, namely by the coupling of olefins with ammonia over silicoaluminates. We show that hydroamination of isobutene to tert-butylamine requires strong Brønsted acid sites to stabilize the formation of the carbenium intermediate, which is the rate-limiting step. FTIR and TPD studies reveal that tert-butylamine strongly inhibits the active site, although its desorption can be assisted by the concurrent adsorption of isobutene. Small-pored zeolites with one-dimensional channels are inactive because tert-butylamine clogs the pore mouths and impedes diffusion of isobutene. However, if the pore size is larger than a critical diameter, the intrinsic rate and activation energy become invarient with zeolite topology.