Abstract
The catalytic production of propylene from propane is carried out commercially over supported Pt catalysts, but the limited yields of propylene, catalyst deactivation and the high costs of Pt have led to increasing efforts in the development of new catalysts that can carry out the non-oxidative conversion of propane to propylene. Low-cost refractory metal oxides such as α-Al2O3 and α-Cr2O3 which have shown some industrial promise were examined with first-principle theoretical calculations and compared herein. Mixed Cr-Al oxide complexes, namely, Cr/α-Al2O3 and amorphous Cr2O3/α-Al2O3 were also examined herein. Density functional theory calculations were used to examine the influence of catalyst structure, surface composition, and possible active sites on the overall reaction energetics and plausible reaction mechanisms to provide fundamental insights into catalytic dehydrogenation and aid in the development of new catalytic materials.
For α-Al2O3, the dominant surface structures are either clean Al2O3 surface or the fully hydroxylated surface depending on partial pressures of water and the actual reaction conditions. Both surfaces were examined in detail as well as the influence of oxygen vacancies on the reaction energetics for propane dehydrogenation. The theoretical results for the catalytic activation of propane on the clean α-Al2O3 surface indicate that the initial activation of propane proceeds via the heterolytic splitting of C-H bond over the Al-O site pair to form Al-propyl and O-H intermediates resulting in an activation barrier of 101.0 kJ/mol. The subsequent activation of C-H bond in propyl proceeds at an adjacent O site to form propylene and OH* with a barrier that was 93.6 kJ/mol. The strong binding of H to the O sites (-115.8 kJ/mol) makes it very difficult to remove H* via H2 recombination as the intrinsic activation barrier is 260.5 kJ/mol, thus resulting in an overall barrier for the activation of propane to propylene of 328.0 kJ/mol. Therefore, while the Al2O3 surface is very reactive in the activation of C-H bonds, it cannot carry the persistent activity for catalysis.
The results on the Al2O3 surface indicate that the surface will be likely ready to form a fully hydroxylated surface. The activation of the C-H bond of propane over the hydroxylated Al2O3 surface does not proceed via a direct heterolytic activation over Al-O pairs but instead via the homolytic splitting on two OH*, resulting in a higher initial C-H activation barrier of 183.3 kJ/mol than that on clean surface. The 2nd C-H splitting proceeds at the OH group binding propyl in the 1st step with a barrier of 115.8 kJ/mol. The barrier for H2 recombination was calculated though to be somewhat higher (337.7 kJ/mol) even though the O-H bond is weaker. This is due to a much less stable transition state. As such, the terminal OH species that result from hydroxylation do not improve catalytic activity of α-Al2O3. The presence of O* vacancies can influence propane dehydrogenation as well. Theoretical results, however show that the strong binding of propyl brought by O vacancy inhibits catalytic activity.
The activation of propane was explored over chromia and supported chromia complexes. Propane activation proceeds via the heterolytic C-H activation over Cr-O site pair. On α-Cr2O3, the metal Cr site binds both propyl and H more strongly, while O site binds them much more weakly which prevents the formation of a deep energy well that one cannot escape. In H2 recombination, the barrier is calculated to be significantly lower at 29.0 kJ/mol, so then the highest point in energy profile is the transition state of 2nd C-H activation (248.4 kJ/mol), and the total barrier for turnover cycle is lower (260.5 kJ/mol) than those on α-Al2O3 facets. Therefore, the weaker H binding on O sites helps this surface to gain better activity.
On monomer grafted Cr/α-Al2O3 complex, binding properties of Cr sites are similar to α-Cr2O3; while the adjacent O site binds H less weakly (0.0 kJ/mol vs 86.8 kJ/mol) affected by the fully hydroxylated α-Al2O3 base. Through the same favored pathway on a single Cr-O pair, the moderate binding properties lead to moderate energies in the activation of C-H bonds, and also a moderate barrier in H2 recombination. So there is no deep well or high peak in the energy profile, and the total barrier (221.9 kJ/mol) is lower than previous cases. The dimer and the trimer Cr/α-Al2O3 complexes have similar binding properties and catalytic activities to the monomer case.
On amorphous Cr2O3/α-Al2O3 complexes, bindings on Cr sites are not as strong as on α-Cr2O3 and Cr/α-Al2O3, and bindings on O sites are similarly moderate as on Cr/α-Al2O3, so both Cr and O sites have moderate binding properties. As a result, the activation of C-H bonds and the recombination of H2 both have feasible barriers, so the overall barriers on amorphous Cr2O3/α-Al2O3 complexes (210 – 240 kJ/mol) are also moderate, similar to the grafted Cr/α-Al2O3 complexes.
By comparing all the examined cases, we see that the strong binding of the reaction species lead to either deep energy wells in C-H activation or high energy peaks in H2 recombination as on α-Al2O3 surfaces, and thus raise the overall barrier; while weak bindings on α-Cr2O3 result in shallower energy well in the 1st C-H activation and lower activation energy for H2 recombination, but it also leads to higher barrier for the 2nd C-H activation, so the resulted overall barrier are also quite high. On the mixed Al-Cr oxide complexes, the moderate binding strengths for reaction species lead to energy profiles without deep wells or high peaks, and thus the overall barriers for the turnover cycle are lower.
Bader charge analysis was used to understand the reactivity on the different surfaces examined. The results show that α-Al2O3 surfaces have very strong electron affinities that prevent the loss of adsorbates which retain negative charge in the surface. The α-Cr2O3 surface is a weak electron acceptor overall, but the electron acceptability of the O site and the Cr site are quite polarized which acts to increase the barrier of the 2nd C-H activation step. The mixed Al-Cr oxide complexes are weak electron acceptors as well, but the electron acceptability between O and Cr sites is more balanced. Therefore, the mixed Al-Cr oxide complexes hold the charges from reaction intermediates moderately in C-H activations, and also release them with low barriers in the desorption of products, so their electronic properties determine the moderate binding properties and thus higher activities for the turnover cycle of propane activation.
