The initiation of the methanol-to-olefins (MTO) process is investigated using a multiscale modeling approach where more than 100 ab initio computed (MP2:DFT) rate constants for H-SSZ-13 are used in a batch reactor model. The investigated reaction network includes the mechanism for initiation (42 steps) and a representative part of the autocatalytic olefin cycle (63 steps). The simulations unravel the dominant initiation pathway for H-SSZ-13: dehydrogenation of methanol to CO is followed by CO-methylation leading to the formation of the first C–C bond in methyl acetate despite high barriers of >200 kJ/mol. Our multiscale approach is able to shed light on the reaction sequence that ultimately leads to olefin formation and strikingly demonstrates that only with a full reactor model that includes autocatalysis with olefins as cocatalysts is one able to understand the initiation mechanism on the atomic scale. Importantly, the model also shows that autocatalysis takes over long before significant amounts of olefins are formed, thus guiding the interpretation of experimental results.

The reactivity of acidic zeolites with close Al-pairs was investigated using density functional theory. Different spatial distances and relative orientations of the two Al atoms were considered. Adsorption of methanol and ammonia was computed and shown to be correlated. Additionally, reaction barriers for the stepwise and concerted mechanism of the dehydration of methanol to dimethyl ether were computed. These barriers were found to correlate well with the adsorption energy of ammonia as a descriptor. This correlation reduces some of the stronger deviations observed in apparent activation barriers, when computing intrinsic barriers referenced to an adsorbed species. Excluding nearest neighbors Al-pairs that violate Löwenstein’s rule, the effect of different Al-pair distributions is found to influence apparent activation barriers typically by less than 20 kJ/mol with a mean absolute deviation of 7 kJ/mol.

No abstract

We investigate the influence of acidity and confinement for different aluminum T-site substitutions in H-ZSM-5 using reactions related to the methanol-to-olefin (MTO) process as examples. We use density functional theory at the PBE-D3 level to study all 12 different T-sites existing in the MFI framework. We find that transition-state energies vary by about 20 kJ/mol with the commonly employed T12 site having some of the lowest barriers. A large part of the energetic differences can be ascribed to differences in dispersion forces due to the surrounding framework, as also evidenced by smaller and uncorrelated differences in calculated heats of adsorption of ammonia. Our analysis shows that taking the T12 site as a computational active site model will yield reaction barriers that are among the lowest of all T-sites available.

No abstract