Controlling Bond Scission Pathways of Isopropanol on Fe- and Pt-Modified Mo2N Model Surfaces and Powder Catalysts

transition-metal-carbides
transition-metal-nitrides
temperature-programmed-desorption
d-band-model
biomass
ACS Catalysis (2024)
Published

2024

Doi

William N. Porter, Hilda A. Mera, Wenjie Liao, Zhexi Lin, Ping Liu, John R. Kitchin, and Jingguang G. Chen, “Controlling Bond Scission Pathways of Isopropanol on Fe- and Pt-Modified Mo2N Model Surfaces and Powder Catalysts”, ACS Catalysis 14(3), 1653-1662 (2024). https://doi.org/10.1021/acscatal.3c04700

Read the paper

This summary was written by a language model from the paper and has not been checked line by line. Rely on the published paper for details.

Question

Upgrading biomass-derived oxygenates requires breaking some bonds while keeping functional groups, and alcohol dehydrogenation to carbonyls is attractive because it also yields H2. Molybdenum nitride (Mo2N) had been identified as a promising material for such reactions, but there was little understanding of how adding a second metal to Mo2N changes the selectivity among C−O, C−H and C−C scission. The paper uses isopropanol as a probe molecule to ask how Fe and Pt overlayers on Mo2N steer the balance between dehydration (propylene) and dehydrogenation (acetone), and whether trends on model surfaces carry over to supported powder catalysts (Porter et al. 2024).

Approach

Polycrystalline Mo2N films, made by nitriding Mo foil in ammonia, were modified with Fe or Pt overlayers by vapor deposition, with coverage tracked by Auger spectroscopy. Temperature programmed desorption (TPD) after 0.5 L isopropanol exposure at 170 K quantified CO, propylene and acetone yields per surface site, and high-resolution electron energy loss spectroscopy (HREELS) identified surface intermediates. Powder γ-Mo2N, 1.4 wt% Fe/Mo2N and 5 wt% Pt/Mo2N (equal atomic fractions) were tested in a flow reactor at 413 K (1 to 2% conversion, for selectivity) and 473 K (for conversion), and characterized by CO chemisorption and in-situ XANES at the Fe K and Pt L3 edges. Spin-polarized PBE calculations in VASP (no dispersion correction, which was found to describe acetone desorption enthalpies worse) modeled the N-terminated γ-Mo2N(111) surface and a Pt monolayer on it. The Fe monolayer reconstructed upon isopropanol adsorption, so DFT results were reported only for Mo2N and Pt/Mo2N (Porter et al. 2024).

Key findings

On unmodified Mo2N, dehydration dominated: propylene selectivity was about 59% but total activity was the lowest of all surfaces, 0.016 ML. Fe and Pt overlayers both suppressed propylene formation. On Fe/Mo2N, C−C scission to CO rose at interfacial coverages (73.7% CO selectivity at 0.7 ML), while acetone formation grew only above a monolayer, reaching 79.9% selectivity at 2.0 ML Fe; the authors infer that C−H scission needs ensembles of adjacent Fe atoms. Pt/Mo2N gave acetone at 284 K, 146 K below the peak on Mo2N, the highest dehydrogenation activity among the partially covered surfaces (0.017 ML at 1.3 ML Pt), and strong C−C scission, with 2.1 ML Pt reaching the highest total activity of 0.088 ML. HREELS showed isopropanol adsorbs molecularly at 170 K, hydroxyl groups on Mo2N consistent with C−O scission, and rapid decomposition on Pt/Mo2N by 350 K (Porter et al. 2024).

The powder catalysts reproduced these trends. Initial conversion at 473 K followed Pt/Mo2N > Fe/Mo2N > Mo2N, matching total TPD activity, although Fe/Mo2N deactivated toward the Mo2N level. At 413 K, Mo2N gave the most propylene (0.42 mmol/h, 60% dehydration selectivity), while Fe/Mo2N was slightly more dehydrogenation selective (45.2% versus 40%) and Pt/Mo2N was 93.4% selective to acetone. CO chemisorption gave 2.6% Fe and 65.4% Pt dispersion. XANES showed Pt mainly metallic with a stronger white line than Pt foil, suggesting Pt−Mo2N electronic interaction, while Fe was a mixture not well fit by reference spectra (Porter et al. 2024).

DFT explained the selectivity switch through oxygen affinity. On Mo2N(111), isopropanol binds through O (−0.67 eV) and prefers O−H scission to an isopropoxy (−1.44 eV) over C−H scission (−0.08 eV), and the alkoxide then breaks its C−O bond toward propylene. On Pt/Mo2N(111), alkoxide formation is uphill (0.14 eV) and α C−H scission to a C- and O-anchored intermediate is preferred (−0.22 eV), leading to acetone. Surface Mo has a high d-band center (−1.12 eV) and binds O at −3.79 eV, while the Pt overlayer has a d-band center of −2.65 eV, resembles Pt(111), and binds O at only −1.31 eV. Calculated charges of +1.39 e on Mo versus +0.12 e on Pt help explain why Mo cannot stabilize the carbon-anchored intermediate (Porter et al. 2024).

Significance

The work shows that UHV model surface studies of metal-modified nitrides can predict activity and selectivity trends of the corresponding supported powders, and that metal modification turns Mo2N from a dehydration catalyst into a dehydrogenation catalyst, with Pt/Mo2N being highly selective to acetone. It also ties the selectivity to an oxygen binding energy and d-band center descriptor, extending ideas developed for metal-modified carbides to nitrides (Porter et al. 2024).