Accessibility of Reactants and Neighborhood of Mo Species during Methane Aromatization Uncovered by Operando NAP-XPS and MAS NMR

Abstract

One-step nonoxidative methane dehydroaromatization is a facile process to generate aromatics and CO-free hydrogen. Despite their high activity and aromatics selectivity, Mo/HZSM-5 catalysts suffer from a continuous deactivation, hampering their application, yet the cause is intensively debated. Employing a combination of characterizations including, but not restricted to, high-resolution electron microscopy, operando NAP-XPS, and MAS NMR spectroscopy, we endeavored in this contribution to get deeper insight into the nature of active sites and origin of catalyst deactivation. Our results indicated (i) an irreversible reaction-induced MoO x particle sintering, (ii) reversible buildup/removal of coke species, (iii) no quantitative correlation between the deactivation rate and the presence/loss of Brønsted acid sites, and (iv) that coke accumulation occurs almost exclusively on Mo instead of Brønsted acid sites. Deactivation is explained by partial blocking of Mo species by coke, which diminishes the accessibility of methane to active sites and successive narrowing and/or blocking of pores hindering the diffusion of larger reaction products (e.g., naphthalene) to the outer surface. Active sites for aromatics formation are referred to as highly dispersed Mo species (mononuclear and tiny subnanometer oxy- and/or oxycarbidic Mo clusters) located inside the micropores on/or close to Brønsted sites.

Solid-State NMR Probe Molecules for Catalysts and Adsorbents: Concepts, Quantification, Accessibility, and Spatial Distribution

“…Another possibility for showing proximity is the in situ generation of a solid probe. This was demonstrated by oxidation of gaseous TMP to solid TMPO on supported metal-oxide catalysts. , Thereby, BAS and LAS in close proximity to the oxidative sites were identified.…”

“…TMP was also used for probing various metal oxides that act as LAS (for details see Table ). Supported metal oxides, among them Cu, Co, and Mn, resulted in chemical shifts of TMP bound at LAS in the range δ 31P = −37 to −60 ppm. , A loading of SiO 2 -supported oxides of Ca, Pd, Pt, and Ru did not result in 31 P MAS NMR peaks of TMP adsorbed at LAS . There was no peak of TMP at LAS for SnO 2 or sulfated SnO 2 .…”

“…The easy oxidation of TMP to TMPO is a clear disadvantage of the probe molecule and is frequently observed also for other alkylphosphines. ,, Conversely, in an oxygen-free environment the oxidation can only happen at accessible oxygen donor sites MO x . This enables a quantification of such sites, as for example metal oxides of Ru, Cu, Co, Mn, or Mo, also in the presence of coke. , Deconvolution of such spectra and peak assignment is, considering the large chemical shift anisotropy, best achieved by performing multiple measurements and varying the sample spinning speed (see Figure ). At low loadings of SBA-15-supported MoO x , a direct correlation between metal oxide loading and the generated TMPO amount is observed.…”

“…The gaseous TMP is, upon oxidation, transformed into the solid probe molecule TMPO ( vide infra ), which is immediately deposited nearby the oxidative site. Because of low local TMPO concentration, a dimer formation is not expected, and thus a quantitative investigation of the BAS and LAS sites in the direct environment of the MO x site is possible. , …”

“…195−199 and oxidative MO x sites. 200,201 Trimethylphosphine oxide (TMPO) 31 P BAS; LAS 909.7 S 0.55 140,141,202 SQ; SQ Versatile probe for acid sites. 24,57,140,203−205 Quantification suffers from peak overlap (deconvolution errors) and dimer formation in excess loadings (unclear stoichiometry).…”