Date of Award

Spring 5-25-2019

Document Type


Degree Name

Doctor of Philosophy (PhD)


Micro and Nanoscale Systems

First Advisor

Daniela S. Mainardi


Experimental work has been extensive in the catalysis field; however, the cost to move nano-engineered catalysts from the laboratory to industry using only experimental research is prohibitively high. Macroscopic phenomena (i.e. melting, defect formation, miscibility) can be understood in terms of the nano-scale mechanism using molecular simulations. Thus, using molecular simulations in combination with experiments lowers the costs related to design, as simulations can be used to rapidly screen candidate materials, so that experimental efforts can be limited to the catalyst candidates deemed as most promising by simulations. This work employs molecular simulation tools to predict the relative effectiveness of Fischer-Tropsch catalysts. The study presented in this work has concentrated on 3d (Co, Fe, Ni) and 4d (Ru) metal clusters formed by pure and binary metal combinations of these elements, as they are the ones known to exhibit Fischer- Tropsch activity.

Regardless of how the Fischer-Tropsch process takes place on a given catalyst material, all Fischer-Tropsch mechanisms proposed to date begin with a first and crucial step, which is the adsorption of CO onto the catalyst surface. The CO adsorption onto the catalyst surface is followed by its dissociation to further formation of long-chained hydrocarbons. Thus, CO adsorption and dissociation energies are hypothesized in this work to be predictors of the effectiveness of a given Fischer-Tropsch catalyst, aiding in the search and design of the most efficient material for this process.

First principle calculations on pure nanoclusters based on CO adsorption and dissociation on iron, cobalt, nickel, and ruthenium without support provides a starting reference for the study of this material as a Fischer Tropsch catalyst. These calculations were carried out using the Generalized Gradient Approximation (GGA) functional RPBE, with the double numerical polarization (DNP) basis set. Our results show that cobalt and iron based bimetallic nanoclusters in a core-shell arrangement containing 14 total atoms with the 10:4 Co to iron ratio and vice versa have stronger cohesive energy than the pure 14 atom clusters of the respective elements, as well as any other bimetallic combinations. The bimetallic cluster in a core-shell arrangement containing 14 atoms in total with the 10:4 Co to iron ratio shows the best CO adsorption and its bond breaking for later release. Similarly, 13 atom clusters with icosahedron symmetry were identified as the most stable symmetry at this theory level. Our study reveals that pure ruthenium and Co clusters consisting of 13 atoms show the best performance in CO adsorption and CO bond breaking compared to all the bimetallic (core-shell arrangement) nanoclusters. An initial predictor that can be used to anticipate potentially effective catalysts was identified as a percentage difference, based on the difference between the CO adsorption energy and the CO dissociation energy. A greater catalysts performance is expected when that percentage difference is maximized. The percentage differences calculated for the ruthenium (46%) and Co (38%) clusters confirm these findings.

Results obtained on pure systems for the first step in the Fischer Tropsch reaction mechanism (adsorption of CO onto the catalyst surface) indicate that not only the nature of the support but the crystallographic plane of the surface of the support that is exposed to the catalyst nanoparticle have an effect on the energy barrier for this reaction to take place. Nanoclusters on support such as silica and rutile with different miller planes were studied with the GGA functional with plane wave basis sets. Co and ruthenium clusters on rutile <110> plane were found to have increased performance with percentage differences of 50% and 60% respectively. Our results indicate that CO adsorbs more strongly on a hollow site of Ru cluster supported on the <110>-terminated rutile support than in any other investigated case. For CO adsorption on supported Co nanoparticles, the silica support is preferred rather than the rutile. The <100>-terminated silica support works best with Co cluster for CO adsorption, followed by the <111>, and finally <110>- terminated silica support. The percentage difference calculated on supported single metal systems shows that the order of preference (best to worst) of potential catalyst seems to be: Ru/rutile <110>, Co/rutile <110>, Ru/rutile <100>, and Fe/rutile <100>.