Date of Award

Summer 7-13-2014

Document Type


Degree Name

Doctor of Philosophy (PhD)


Mechanical Engineering

First Advisor

J. Will Medlin

Second Advisor

John W. Daily

Third Advisor

Charles B. Musgrave


Many technological processes, such as catalysis, electrochemistry, corrosion, and some materials synthesis techniques, involve molecules bonding to and/or reacting on surfaces. For many of these applications, transition metals have proven to have excellent chemical reactivity, and this reactivity is strongly tied to the surface's adsorption properties. This thesis focuses on predicting adsorption properties for use in the design of transition metal surfaces for various applications.

First, it is shown that adsorption through a particular atom (e.g, C or O) can be treated in a unified way. This allows predictions of all C-bound adsorbates from a single, simple adsorbate, such as CH3. In particular, consideration of the adsorption site can improve the applicability of previous approaches, and gas-phase bond energies correlate with adsorption energies for similarly bound adsorbates.

Next, a general framework is presented for understanding and predicting adsorption through any atom. The energy of the adsorbate's highest occupied molecular orbital (HOMO) determines the strength of the repulsion between the adsorbate and the surface. Because adsorbates with similar HOMO energies behave similarly, their adsorption energies correlate. This can improve the efficiency of predictions, but more importantly it constrains catalyst design and suggests strategies for circumventing these constraints. Further, the behavior of adsorbates with dissimilar HOMO energies varies in a systematic way, allowing predictions of adsorption energy differences between any two adsorbates. These differences are also useful in surface design.

In both of these cases, the dependence of adsorption energies on surface electronic properties is explored. This dependence is used to justify the unified treatments mentioned above, and is used to gain further insight into adsorption. The properties of the surface's d band and p band control variations in adsorption energy, as does the strength of the adsorbate-surface coupling. A single equation, with only a single adsorbate-dependent fitting parameter as well as a few universal fitting parameters, is developed that can predict the adsorption energy of any radical on any close-packed transition metal surface. The surface electronic properties that are input into this equation can be estimated based on the alloy structure of the surface, improving prospects for high-throughput screening and rational catalyst design.

The methods discussed in this thesis are used to design a novel catalyst for ethylene epoxidation, which is experimentally synthesized and tested. Initial tests indicate that this catalyst may have improved selectivity over pure Ag