Theses Doctoral

Molybdenum, Tungsten and Nickel Compounds as Catalysts for the Dehydrogenation of Formic Acid

Neary, Michelle Catherine

Though petroleum fuels are currently a crucial part of our daily life, there is interest in developing energy sources that are more sustainable and better for the environment. One possible energy source is hydrogen, which burns cleanly to produce only water as a byproduct. However, hydrogen itself cannot be easily transported and, therefore, other storage mediums are necessary. One such storage medium that has been investigated in recent years is formic acid, which is a liquid at room temperature and easier to handle. A crucial aspect of using formic acid is the ability to release hydrogen on demand. Testing possible catalysts for this transformation has driven my research over the last five years.

Chapter 1 investigates the ability of a series of cyclopentadienyl molybdenum hydrides, CpᴿMo(PMe₃)₃₋ₓ(CO)ₓH (Cpᴿ = C₅H₅, C₅Me₅; x = 0, 1, 2, 3), to catalyze formic acid dehydrogenation. Though several compounds in the series CpᴿMo(PMe₃)₃₋ₓ(CO)ₓH have been structurally characterized before, we were able to characterize several more by X-ray diffraction. Since the compounds are structurally similar, differences in catalytic activity are governed by the electronics, which are determined primarily by the number of PMe₃ ligands relative to CO. The best catalysts are the hybrid compounds, CpᴿMo(PMe₃)₂(CO)H, due to the fact that they can be easily protonated by formic acid and readily release hydrogen to continue the catalytic cycle.

Additionally, I observed that methanol and methyl formate were being produced as side products. Since methanol is also a potential hydrogen storage medium, its production is of interest. In this case, the tricarbonyl compounds, CpᴿMo(CO)₃H, were most selective for formic acid disproportionation relative to dehydrogenation. This is likely due to their relative propensity to transfer a hydride ligand to formic acid rather than to become protonated by it. We also investigated the ability of formic acid to reduce ketones and aldehydes via transfer hydrogenation.

Because the phosphine-rich compounds were such effective catalysts, we sought to investigate the reactivity of other compounds with phosphine ligands towards formic acid. To this end, Chapter 2 focuses on studies involving Ni(PMe₃)₄, and Chapter 3 looks at Mo(PMe₃)₆ and W(PMe₃)₄(η²-CH₂PMe₂)H. Ni(PMe₃)₄ is indeed able to catalyze formic acid dehydrogenation. Density Functional Theory studies suggest that the mechanism involves formation of a formate-hydride followed by decarboxylation to produce a dihydride species. The ability of the PMe₃ ligand to induce decarboxylation also provides a route to synthesize Ni(PMe₃)₄ from Ni(O₂CH)₂•2H₂O and Ni(py)₄(O₂CH)₂•2py, which has been structurally characterized.

To expand on the nickel phosphine reactivity, a heteroleptic nickel phosphine complex employing the bisphosphine ligand 1,2-bis(diphenylphosphino)benzene (bppb), namely (bppb)Ni(PMe₃)₂, was synthesized, characterized and tested with formic acid. It also catalyzes dehydrogenation, but rearranges to Ni(PMe₃)₄ and the inactive compound, (bppb)₂Ni. The structural characterization of these and other (bppb)Ni compounds shows that the bppb ligand allowed for extreme flexibility in crystallization.

Chapter 3 reveals that Mo(PMe₃)₆ and W(PMe₃)₄(η²-CH₂PMe₂)H are likewise catalysts for formic acid dehydrogenation. However, the compounds produced along the way are also of interest. The known carbonate species, Mo(PMe₃)₄H₂(O₂CO), is formed from Mo(PMe₃)₆ and formic acid, and we have structurally characterized it. The tungsten carbonate species is also produced in the analogous reaction with W(PMe₃)₄(η²-CH₂PMe₂)H. Other compounds observed include W(PMe₃)₄H₂(O₂CH)₂ and W(PMe₃)₄H₃(O₂CH), the latter of which has also been characterized by X-ray diffraction. Finally, both Mo(PMe₃)₆ and W(PMe₃)₄(η²-CH₂PMe₂)H react with formic acid to make trimeric species, [M(PMe₃)₃(CO)(O₂CH)(μ-O₂CH)]₃ (M = Mo, W), which display an unusual anti/anti configuration of the bridging formate ligands.

Chapter 4 revisits some of the side products from Chapter 1 in more detail, particularly [CpMo(CO)₃]₂ and [CpMo(μ-O)(μ-O₂CH)]₂. The presence of semi-bridging and bridging ligands, respectively, makes it difficult to determine whether there is actually a metal-metal bond. Natural Bond Orbital (NBO) analysis reveals that there is indeed a Mo-Mo bond in [CpMo(CO)₃]₂, but not in [CpMo(μ-O)(μ-O₂CH)]₂. The Covalent Bond Classification method can be used to depict these and other compounds in a way that more accurately reflects the true bonding in the molecule.


  • thumnail for Neary_columbia_0054D_13505.pdf Neary_columbia_0054D_13505.pdf application/pdf 11.4 MB Download File

More About This Work

Academic Units
Thesis Advisors
Parkin, Gerard
Ph.D., Columbia University
Published Here
August 12, 2016