Ca Cations Impact the Local Environment inside HZSM-5 Pores during the Methanol-to-Hydrocarbons Reaction

The methanol-to-hydrocarbons (MTH) process is an industrially relevant method to produce valuable light olefins such as propylene. One of the ways to enhance propylene selectivity is to modify zeolite catalysts with alkaline earth cations. The underlying mechanistic aspects of this type of promotion are not well understood. Here, we study the interaction of Ca2+ with reaction intermediates and products formed during the MTH reaction. Using transient kinetic and spectroscopic tools, we find strong indications that the selectivity differences between Ca/ZSM-5 and HZSM-5 are related to the different local environment inside the pores due to the presence of Ca2+. In particular, Ca/ZSM-5 strongly retains water, hydrocarbons, and oxygenates, which occupy as much as 10% of the micropores during the ongoing MTH reaction. This change in the effective pore geometry affects the formation of hydrocarbon pool components and in this way directs the MTH reaction toward the olefin cycle.

The proton form of zeolite ZSM-5 was obtained by calcining a commercial NH4ZSM-5 zeolite (Si/Al = 25, Alfa Aesar) at 550 °C for 5 h. HZSM-5 catalyst modified with calcium was prepared via incipient wetness impregnation of the calcined zeolite with aqueous solutions of Ca(NO3)3 (Alfa Aesar, 99.0%), following the procedure described elsewhere 1 aiming at 1 wt. % of metal loading. To obtain the partially ion-exchanged Na, a calculated amount of sodium nitrate (≥ 99% Merck) was dissolved in 100 mL demineralized water. The salt solution and 2 g of NH4ZSM-5 were added to a round bottle ask, and the mixture was stirred at 60 °C for 3 h. After stirring, the mixture was transferred into a ThermoScientific Heraeus Megafuge 16 centrifuge. The liquid was removed by decantation and the solid was washed three times with demineralized water and centrifuged for 5 minutes (5000 rpm). The impregnated and exchanged samples were dried overnight at 110 °C and calcined at 550 °C for 5 h in static air. XRD patterns (Fig. S1) demonstrate that the original MFI topology is preserved in both samples. The relative intensity of Brønsted acid sites (BAS, signal at 4 ppm) was compared by 1 H MAS NMR analysis (Fig.   S2). Morphology of ZSM-5 zeolite was preserved upon metal modification (Fig. S3). Amount of extraframework Al was estimated using 27 Al MAS NMR (Fig. S4). The preparation procedure and physico-chemical properties of a reference Silicalite-1 sample are provided elsewhere. 2

Catalyst characterization
The elemental composition (Si/Al ratio and metal content) of the zeolite catalysts was determined by ICP-OES (Spectro CIROS CCD ICP optical emission spectrometer). The samples were dissolved in a 1:1:1 mixture of HF (40%), HNO3 (65%) and H2O prior to measurements.
The crystallinity of the zeolite samples was determined by powder X-ray diffraction (XRD). Ex situ XRD measurements were performed on a Bruker D2 powder diffraction system (Cu Kα radiation, scan speed 0.01°/s, 2θ range 5−60°) and ID15A beamline, ESRF. For the synchrotron experiments, photon wavelength was 0.0124 nm (100 keV). The incident beam was monochromated using a double-bounce bend Si(111) monochromator in the Laue geometry. The primary beam was focused to approximately 100 x 100 mkm 2 using a compound refractive lens transfocator. The detector distance and tilt were calibrated using NIST CeO2 standard, and diffraction patterns were integrated using a locally modified version of the PyFAI package to eliminate outliers. Correction for detector transparency, homogeneity (flood field) and spatial distortion were applied at the time of integration. Further details can be found here. 3 Prior to measurements, sieved (250 -500 μm) ZSM-5 catalyst (25 mg) was placed in a Kapton capillary (i.d. 1.83 mm, wall thickness 0.025 mm) sealed with wax. The integrated XRD patterns were analyzed by Rietveld refinement using the GSAS-II software. The patterns were refined in q-range of 0.4 -7.5 Å -1 . The scale factor, background, and the unit cell parameters (Pnma space group) were refined.
The morphology of the zeolite crystals was analyzed by scanning electron microscopy (SEM, FEI Quanta 200F scanning electron microscope at an accelerating voltage of 3 kV).
The acidic properties of zeolites were determined by IR spectroscopy of adsorbed pyridine as a probe molecule. Spectra were taken in the 4000 -1000 cm −1 range using a Bruker Vertex 70v spectrometer.
Samples were pressed into self-supporting wafers (10 − 20 mg, diameter 1.3 cm) and placed in an environmental cell. The wafers were pre-treated in O2:N2 (1:4 vol. ratio) flow at 550 °C (heating rate 10 °C⸱min -1 ) to remove contaminants followed by cooling to 150 °C under dynamic vacuum (p < 10 -5 mbar). Afterwards the samples were exposed to an excess of pyridine vapor until saturation. Then the process of pyridine desorption was followed in a temperature range of 150 -450 °C under dynamic vacuum, before the measurement the sample was first cooled to 150 °C and an IR spectra was recorded.
For the experiments of pyridine adsorption with water preadsorbed, the pre-treated wafers were subjected to water vapors for 10 -30 min at 2 kPa of water in 130 mL He ⸱min -1 until saturation. After that, the cell was degassed under dynamic vacuum, and the samples were exposed to an excess of pyridine vapor until saturation. The recording of IR spectra was carried out as ascribed above. For the quantification of Brønsted and Lewis acid sites, integral molar extinction coefficients of 0.73 cm⸱mol -1 and 1.11 cm⸱mol -1 were used. 4 The conventional thermogravimetric analysis (TGA) of spent catalysts was performed using a Mettler Toledo TGA/DSC 1 instrument. An amount of used catalyst (≈ 10 mg) was placed in an alumina crucible and then heated up to 800 °C with a ramp rate of 5 °C⸱min -1 in O2⸱He (20:40) 60 mL⸱min -1 flow.
Textural and adsorptive properties of zeolites were studied by Ar porosimetry at -186 °C using a Micromeritics ASAP2020 machine. Prior to measurements, the samples were pre-treated at 400 °C under evacuation. The microporous volume was calculated by the t-plot method using a thickness range from 3.5 to 4.5 Å.

Catalytic activity measurements: conversion and selectivity calculations
Performance of the catalysts was compared using throughput numbers and carbon-based selectivity.
Methanol throughput is defined as the amount of methanol in g converted per g of catalyst before the conversion of light oxygenates drops below 75%. Selectivity of the catalysts obtained by the integration of each product/group of products of interest before the conversion of light oxygenates drops below 75%. For throughput calculations, conversion was defined as the carbon-based fraction of light oxygenates (methanol and dimethyl ether) consumed during the reaction: Further procedure was reproduced from here.

Quantification of GC-switches
To evaluate the amounts of water and hydrocarbons retained and eluted from the working catalysts (Fig.   5b), we quantified these amounts with the steps below: First, we estimated theoretical amount of methanol and water, assuming 100% conversion of substrate and reaction ongoing according to the following scheme: We estimated that = 0.112 mg•min -1 assuming that P = 0.75 kPa and total flow is 10 mL•min -  We estimated the uncertainties for GC-switching experiments and compared them to theoretical numbers (Table S1). While oxygen balance is close to 100% for experiments of the catalyst, a certain amount of hydrocarbons appears to be irreversibly retained during each switch. We should note here that procedure for calculating oxygen balance is accurate because it mainly depends on the relative sensitivity factor of water. The quantification of hydrocarbons is much more complex as it involves many different molecules and varying sensitivity factors can lead to the lower overall calculation accuracy. We calibrated the GC for CH4, C2H4, C2H6, C3H6, C3H8, benzene, and toluene and used carbon numbers to estimate the relative sensitivity factors of other molecules. From these measurements we found that average numbers from integration are 0.0045 ± 0.0003 mmol for HCs and 0.0068 ± 0.0001 mmol for water for HZSM-5 catalyst (Fig. S14, Table S1).  We suggest that presence of adsorbed species over Ca 2+ centers hinders accessibility of reacting species to the acid sites. In order to check it, we carried out additional catalytic experiments with different Ca loading of parent ZSM-5. In addition to HZSM-5 and Ca/ZSM-5 (1wtCa/ZSM-5 below), we evaluated the performance of 2 more catalysts: 0.5wtCa/ZSM-5 and 1.5wtCa/ZSM-5, prepared using incipient wetness impregnation of HZSM-5. Properties of the Ca-modified catalysts are provided in Table S2.
Kinetic experiments demonstrate that overall the catalyst stability of 0.5wtCa/ZSM-5 and 1wtCa/ZSM-5 were significantly higher than that for HZSM-5 (Fig. S16a). The sample with the highest loading (i.e., more than 1 Ca 2+ per 2 Al sites) shows a strong decline in methanol conversion in both kinetic and TGA-MS experiments. To quantify the amount of adsorbates per Ca 2+ metal, we carried out TGA-MS experiments over these samples (Fig. S16b). We estimated that the amount of retained adsorbates is 2 molecules per mol of Ca.

MAS NMR of used catalysts
We performed 1 H-13 C NMR analysis of the used catalysts to study the hydrocarbon molecules occluded inside the catalyst pores during the MTH reaction. The samples were taken after the continuous flow experiments for 10 min TOS and 3 h TOS with 13 C labeled methanol to obtain higher signal intensity of the retained hydrocarbon species and the collected spectra are weight normalized. Different types of pulse sequences were implemented: CP MAS NMR which is sensitive to strongly adsorbed species, Hahn Echo MAS NMR to observe the mobile species, and 1 H -13 C{ 1 H} HETCOR MAS NMR 2D sequence to correlate signals from the protons and carbons simultaneously. Two main regions for (Figs. S17 -S18). 5 Used HZSM-5 sample demonstrate typical 13 C NMR spectra of retained alkylated aromatic species (characteristic peaks at 134, 131, 128, 22 and 18 ppm). 6 In contrast, the spectrum of used Ca/ZSM-5 is completely different, featuring sharp signals at 49 and 59 ppm, which can be assigned to adsorbed DME and methanol molecules as well as methoxy species. 7,8 More detailed assignment of chemical shifts is provided in Table 2 in the main text. As the Hahn Echo and HPDec experiments are quantitative, we can estimate relative ratio of hydrocarbon species to oxygenates by integration of following areas. The results are provided in Table S3. After 3 h on stream we can observe that oxygenate contribution increased for Ca/ZSM-5 catalyst compared to 10 min TOS, while the aromatics/olefins prevail for H and Na/ZSM-5 catalysts pointing out to formation and condensation of coke species (Figs. S17 -S18).