The author characterized fresh and partially deactivated samples, with the exception of the UV-Raman experiments. The author performed all the catalytic tests and he contributed in the planning of the experiments, interpretation of the results and preparation of the manuscript.

Catalyst deactivation during the methanol-to-hydrocarbons (MTH) reaction was investigated using five different commercially prepared microporous catalysts, including Mordenite, ZSM-22, ZSM-5, zeolite Beta and SAPO-34. The reaction was carried out in a fixed bed reactor at a constant feed rate per gram of catalyst. Deactivated and partially deactivated catalysts were obtained at increasing reaction times. The whole of the catalyst beds were characterized using nitrogen adsorption, thermogravimetric analysis, a dissolution-extraction protocol, and UV-Raman spectroscopy, focusing primarily on methods suitable for the quantification of the coke. The results illustrate that topology is the dominant parameter that influences not only catalyst lifetime and product distribution, but also the nature of the species causing the deactivation. For all catalyst topologies, when the entire catalyst bed is examined together, the micropore volume and BET surface area decrease more rapidly than total coke from TGA increases at short reaction times. In the materials with the more restricted access to the internal voids, such as ZSM-22 and SAPO-34, the loss of activity is to a large extent due to species which are soluble in dichloromethane and give rise to distinct features in the Raman spectra. For the Mordenite and Beta catalysts, which have larger pores comprising three dimensional networks, and to some extent for the ZSM-5 catalyst employed, the accumulation of more coke species which are insoluble in dichloromethane, presumably on the external surface of the zeolite crystals, is observed. This is linked to the appearance of more pronounced D and G bands in the Raman spectra, indicative of extended carbon species.

Zeolites representing seven different topologies were subjected to life-time assessment studies as methanol to hydrocarbons (MTH) catalysts at 400 °C, P(MeOH) = 13 kPa and P(tot) = 100 kPa. The following topologies were studied: ZSM-22 (TON), ZSM-23 (MTT), IM-5 (IMF), ITQ-13 (ITH), ZSM-5 (MFI), mordenite (MOR) and beta (BEA). Two experimental approaches were used. In the first approach, each catalyst was tested at three different contact times, all giving 100% initial conversion. The life-time before conversion decreased to 50% at each contact time was measured and used to calculate critical contact times (i.e. the contact time needed to launch the autocatalytic MTH reaction) and deactivation rates. It was found that the critical contact time is strongly correlated with pore size: the smaller the pore size, the longer the critical contact time. The second experimental approach consisted of testing the catalysts in a double tube reactor with 100% initial conversion, and quenching the reaction after 4 consecutive times on stream, representing full, partial, and zero conversion. After quenching, the catalyst bed was divided into four segments, which were individually characterised for coke content (temperature-programmed oxidation) and specific surface area (N adsorption). The axial deactivation pattern was found to depend on pore size. With increasing pore size, the main source of coke formation changed from methanol conversion (1D 10-ring structures), to partly methanol, partly product conversion (3D 10-ring structures) and finally mainly product conversion (3D 12-ring structure). As a result, the methanol conversion capacity changed little with contact time for ZSM-5, while it increased with increasing contact time for the catalysts with smaller pore sizes, and decreased with increasing contact time for pore sizes larger than ZSM-5.

The deactivation of zeolite catalyst H-ZSM-5 by coking during the conversion of methanol to hydrocarbons was monitored by high-energy space- and time-resolved operando X-ray diffraction (XRD) . Space resolution was achieved by continuous scanning along the axial length of a capillary fixed bed reactor with a time resolution of 10 s per scan. Using real structural parameters obtained from XRD, we can track the development of coke at different points in the reactor and link this to a kinetic model to correlate catalyst deactivation with structural changes occurring in the material. The "burning cigar" model of catalyst bed deactivation is directly observed in real time.