(58d) Analysis of Kinetic and Diffusive Restraints on Methylbenzene Interconversion Pathways during MTO Using DFT and Kinetic Monte Carlo Simulations

Brønsted acid zeolite surfaces, alkenes, and arenes are methylated by a combination of methanol (CH₃OH) and dimethyl ether (CH₃OCH₃) during methanol-to-olefins (MTO) processes. Methylation reactions can occur via one of two distinct mechanisms: a sequential mechanism, in which the methylation agent first reacts with the zeolite surface to form a surface methyl:

CH₃OR + Z–H → Z–CH₃

preceding the methylation of a guest species:

C_nH_2n + Z–CH₃ → C_nH_2n+1+ Z–H

or a concerted mechanism, in which the methylation agent directly reacts with the alkene or arene:

CH₃OR + C_nH_2n + Z–H → C_nH_2n+1+ ROH + Z–H

During MTO, alkenes grow through repeated methylation reactions, crack, and sometimes egress in the “alkene-cycle”. Alternatively, alkenes can undergo hydride transfer and cyclization reactions to form arenes. During MTO arenes are methylated to form one of thirteen C₆–C₁₂ methylbenzene species which co-catalyze the formation of light alkene products via isomerization and dealkylation reactions in the “aromatic cycle” [1]. The transient nature of MTO reactions complicates kinetic studies, thus prompting the use of density functional theory (DFT) to provide insight into the mechanisms governing these reaction pathways. However, any assessment of these competing alkene and aromatic cycles in MTO must account for the dramatic differences in diffusivities of alkenes and arenes (and arenes of different sizes). Additionally, analysis of large reaction networks (~10³â€“10⁵ reactions) requires kinetic modeling.

Here, we present a novel kinetic Monte Carlo (KMC) package which models H-ZSM-5 crystals across experimentally relevant time and length scales to understand the role of transport during arene interconversion reactions (~100 reactions) previously mapped out using periodic DFT methods which account for dispersion. This KMC code uses general temporal acceleration schemes [2] which suppress rapid quasi-equilibrated reactions (such as diffusions of small molecules) without altering observed net rates. All methylbenzene interconversion pathways were analyzed via the concerted and sequential mechanisms with CH₃OH and CH₃OCH₃ as methylating agents and diffusion barriers (intersection-to-intersection) were obtained for all methylbenzene species via both sinusoidal and straight channels. Reactant, product, and transition state structures were manually generated, optimized, and then systematically reoriented and reoptimized to sufficiently sample the potential energy surface. The lowest energy state obtained from these systematic reorientations is used as the DFT-derived energy input for KMC simulations. These reaction and diffusion energies and barriers were then analyzed across distinct experimental regimes (arene methylation and methanol-to-olefins) for H-ZSM-5 crystals of distinct size and morphology.

All arene methylation interconversion pathways were analyzed with KMC and the results demonstrate that the product selectivity of arene methylation can be controlled by varying the input conditions. At benzene methylation conditions (373 K, 0.02 bar C₆H₆, 0.68 bar CH₃OCH₃, 0.1% conversion) [3], toluene is the primary product (99%) and formed via the sequential mechanism. At conditions more representative of MTO (623 K, 0.04 bar C₆H₆, 0.08 bar CH₃OR, 10% conversion) [4], KMC predicts that the majority of arene products desorb as C₇–C₉ species. Larger C₁₀–C₁₂ species are diffusion limited and once formed they must either de-methylate to C₉ species so they can egress as products or they become trapped in the zeolite pore. During MTH, it is likely that these trapped C₁₀–C₁₂ species will serve as co-catalysts to produce C₂–C₄ alkenes via the aromatic-cycleand eventually cause catalyst deactivation through formation of polyaromatic species that block access to Brønsted acid sites. The combination of KMC simulations and DFT-derived energies allows for rapid analysis of rates, surface species, and products of large and complex reaction pathways at multiple sites in zeolites—such as those of MTH. Additionally, the inclusion of diffusional barriers provides a more accurate prediction of the behavior of diffusion-limited C₁₀–C₁₂ species and the way they lead to catalyst deactivation, which can inform future studies investigating deactivation mechanisms.

[1] Ilias, S.; Bhan, A., ACS Catal., 2013, 3, 18–31

[2] Dybeck, E.C.; Paisance, C.P.; Neurock M, J Chem Theory Comput., 2017, 13, 1525–1538

[3] Hill, I.; Malek, A.; Bhan, A., ACS Catal., 2013, 3, 1992–2001

[4] Ilias, S.; Bhan, A.; J Catal., 2014, 311, 6–16