Computationally Intensive Research Project

Reliable Electronic Structure Prediction of Molecular Properties

Sotiris S. Xantheas,¹ David A. Dixon,² Kirk A. Peterson,³ Angela K. Wilson,⁴ Benjamin P. Hay,¹ Lai-Sheng Wang,⁵ Toshiko Ichiye,³ Jun Li,¹ Henry F. Schaefer,⁶ Cheuk-Yiu Ng⁷

¹Pacific Northwest National Laboratory, ²University of Alabama, Tuscaloosa, ³Washington State University, ⁴University of North Texas, ⁵Washington State University Tri-Cities, ⁶University of Georgia, ⁷University of California, Davis

FY07 Allocation - 100,000

Abstract

The field of environmental science is dominated by issues of scaling in space and time. The goal of environmental science is to understand the current state of the environment based on our knowledge of the past and to use this information to be able to make forward predictions. For example, given current practices for manufacturing, what will be their long-term environmental impact? Given potential environmental remediation strategies, what will these lead to? The use of remediation strategies should be based on a solid understanding of their long-term impact on the environment in order to avoid unforeseen consequences such as the past widespread release of chlorofluorocarbons (CFC's) into the atmosphere. Although the interest is clearly in the results at large spatial and temporal scales, detailed insight into behavior at the molecular scale is key to the understanding of (i) how humans have impacted the environment, (ii) how to remediate anthropogenic impacts on the environment, and (iii) how to minimize future anthropogenic impacts. Of particular interest are reliable, accurate values for thermodynamic and kinetic properties of molecular systems that can be used in molecular design, in process design, for example, studies of combustion, and in remediation models. Computational chemistry is a key technology for addressing the complex environmental cleanup problems facing the Department of Energy's nuclear production sites, as well as the problems associated with other polluted sites in the United States and the prevention of further pollution. The issue of accuracy as the result of a simulation is very important. For instance, a factor of 2 to 4 in catalyst efficiency may determine whether a chemical process is economically feasible or not and a factor of 4 in a rate constant at room temperature (25°C) corresponds to a change in the activation energy on the order of just less than 1 kcal/mol. Given a 50:50 starting mixture of two components, a change in the free energy, ΔG, of less than 1.5 kcal/mol leads to a change in the equilibrium constant by a factor of 10, leading to a 90:10 mixture at 25°C. The requirement for such accuracy means that we must be able to predict thermodynamic quantities such as bond dissociation energies (De or D00) and heats of formation (ΔHf) of small molecules and molecular clusters to better than 1 kcal/mol and activation energies to within a few tenths of a kcal/mol - a daunting computational task. Our goal is to develop and test the procedures needed to provide accurate predictions of such molecular properties. We then apply such tools to the prediction of a variety of molecular properties and systems including: transition metal compounds, combustion related compounds including the propargyl potential energy surface and alkane and alkoxy radical bond energies, main group chemistry including highly reactive compounds with high oxidation states and mono-, di- and tri-phosphate compounds, host/guest complexes for the design of separation systems such as clathrate hydrates and metal ion receptors, and iron-sulfur proteins.