Speciation of ethylene on model skeletal cobalt catalysts as characterised using inelastic and quasi-elastic neutron scattering techniques

The adsorption of ethylene gas on a model skeletal cobalt catalyst, and the effect of co-adsorbed carbon monoxide on substrate mobility has been characterised using inelastic (INS) and quasielastic (QENS) neutron scattering techniques. Skeletal cobalt catalysts were synthesised by dealuminating powdered cobalt-aluminium alloy (66 wt.% Al, 33 wt.% Co) using concentrated base solutions (6 M and 17 M NaOH_(aq)). Two molarities of base solution were tested according to previous reaction conditions described in previous literature. 6 M NaOH_(aq) solution was shown to synthesise a high surface area catalyst. Different base-to-aluminium molar ratios of 25:1, 5:1, 2.5:1 and 1:1 were tested to minimise the volume of 6 M solution required for scaledup catalyst synthesis. The impact of temperature controls on the BET surface area of the resultant catalyst during synthesis was also tested. The cobalt-aluminium alloy was characterised using powder X-ray diffraction (PXRD), and showed dominant Co₂Al₉ and Co₄Al₁₄ phases. Microwave plasma atomic emission spectrometry (MP-AES) determined the Al:Co ratio in the cobalt-aluminium material as 2.1:1. The resultant skeletal cobalt catalysts were passivated and characterised using PXRD and BET surface area and porosity. Face centred cubic (fcc) and hexagonal close-packed (hcp) metallic cobalt phases were identified using PXRD, as well as peaks identified as cobalt oxide (CoO). BET surface area analysis of the four passivated skeletal cobalt catalysts were 24 m²g -1 for 1:1 molar ratio, 19 m²g ^-1 for 2.5:1 molar ratio, 23 m²g ^-1 for 5:1 molar ratio and 8 m²g ^-1 for 25:1 molar ratio. The 5:1 NaOH(aq)-topowdered cobalt-aluminium alloy molar ratio was selected for scale-up reaction. Skeletal cobalt catalyst underwent a temperature programmed reduction (TPR) and temperature programmed desorption (TPD) pre-treatment at 280 ºC. Hydrogen consumption during TPR was calculated at 0.96 mmol g_cat ⁻¹ using the recorded change observed in thermal conductivity (TCD). This hydrogen consumption was higher than that observed in the reduction of CoO and Co₃O₄ (13.3 and 16.7 mmol g⁻¹ respectively) and corresponded to a reduction of surface hydroxylates. The pre-treated catalyst was further characterised using XRD, BET surface area and porosity, and X-ray photoelectron spectroscopy (XPS). XPS analysis showed the presence of hydroxylates even after TPR/TPD pre-treatment indicating that even handling skeletal cobalt in a glovebox would cause the surface to readily hydroxylate.

Ethylene adsorption on skeletal cobalt catalysts was characterised using INS on the TOSCA and MAPS spectrometers; the characterisation of total elasticity of the catalyst, the ethylene dosed catalyst, and the ethylene-carbon monoxide co-adsorbed catalyst via QENS was carried out on the IRIS spectrometer. Ethylene adsorption on TOSCA was performed at 200 K and 300 K (room temperature) with and without carbon monoxide co-adsorption. INS spectra showed a poor ethylene uptake due to small changes in peak intensity between background catalyst spectra and ethylene-dosed spectra. Ethylene fully dehydrogenated to coke on adsorption to skeletal cobalt catalyst at 200 K. Vibrational modes at 988 and 1102 cm^-1 were assigned to chemisorbed hydrogen bound to a threefold metal site (Co₃H). C-H aliphatic vibrational modes at 300 cm^-1 were identified after adsorption at 300 K, indicating the presence of methyl groups. QENS experiments on the co-adsorption of carbon monoxide with ethylene showed inhibited substrate motion compared to a purely ethylene-dosed skeletal cobalt catalyst.

Sub-ambient ethylene chemisorption and TPD was performed. The sample underwent pretreatment at 280 ºC and was then cooled to -100°C (~200 K). The catalyst sample was then pulsed with six pulses of N5.0 pure ethylene gas and the sample was heated to 300°C. The TCD data recorded clearly showed a change in surface chemistry of the sample as denoted by the sharp changes in thermal conductivity recorded at 90°C, 180°C, and 250°C, each of which were preceded by smaller, more broad features at 45°C, 120°C, and 210°C respectively. Due to the large change in recorded thermal conductivity during the ramp, it is likely that the species evolved are hydrogen gas either from the evolution of surface hydrogen, or from the dehydrogenation of hydrocarbon substrates. This hypothesis is based on the large difference in thermal conductivity between argon (the calibration gas) and hydrogen when compared to the difference in thermal conductivity of other gaseous products that may have been evolved to argon (ethane, physisorbed ethylene etc). However, without a mass-spectrometer to record the mass fragments of evolved species, this change in surface chemistry can only be speculated. Additionally, the temperatures at which the speculated hydrogen evolved differ significantly from those reported by Weststrate et al in single crystal Co(0001) studies. It should be noted that the sharpness of TCD response produced was due to a non-linear ramp rate and, as such, the TCD response cannot be said to be truly representative of the response produced from a linear ramp TCD