Date of Award


Level of Access Assigned by Author

Campus-Only Dissertation

Degree Name

Doctor of Philosophy (PhD)




Brian Frederick

Second Committee Member

François Amar

Third Committee Member

Carl Tripp


Fast pyrolysis of lignocellulosic biomass gives a bio-oil composed of a complex mixture of oxygenated organic compounds. The oxygen content of the bio-oil decreases its energy density, so removal of oxygen is necessary. One method of increasing the energy density is through catalytic upgrading. Most traditional catalysts (noble metals, zeolites, and metal sufides) aren't efficient at removing oxygen from these compounds. We have investigated a new class of hydrodeoxygenation catalysts, which are based upon well studied selective oxidation catalysts, which is essentially the reverse mechanism. The hydrodeoxygenation catalysts are in the form of metal oxide bronzes. These bronzes are formed by heating the catalyst in hydrogen at high temperatures (-350 °C for ten hours). The resulting bronze is in the form HyWO3-z due to loss of oxygen (creation of oxygen vacancies) and absorption of hydrogen into the bulk. The state of the catalysts has been determined experimentally by thermogravimetric analysis (TGA). The same catalysts have been found to be active for the hydrogenation of acrolein at 50 °C and the hydrodeoxygenation of allyl alcohol above 250 °C. We have used density functional theory calculations (DFT) to determine the mechanism of hydrogen adsorption and vacancy creation on two metal oxide clusters: Mo3O9 and W3O9. On these same clusters, we have calculated the potential energy surface for the reduction of acrolein to allyl alcohol, propene, and 1-propanol. On the W3O9 cluster we used a micro-kinetic model, which qualitatively agrees with the product distribution seen in experiment for hydrogenation of acrolein to allyl alcohol and the hydrodeoxygenation of allyl alcohol to propene.

Files over 10MB may be slow to open. For best results, right-click and select "save as..."