Research Structure and Reactivity Solvation Variable Spin Molecules Phosphorus Project Dossiers

Interests and Future Plans

The development of theoretical chemistry methods that are as robust for condensed phases as those already available for the gas phase continues to be at the theoretical frontier. An alternative to approaches involving the explicit inclusion of hundreds to thousands of surrounding atoms/molecules is to treat the embedding medium as a dielectric continuum with additional terms to account for non-electrostatic interactions between explicit and implicit regions. We have pioneered methods for accomplishing this using classical and quantal theories. We have moreover developed the methodology to handle equilibrium and non-equilibrium solvation regimes, the latter being particularly relevant for spectroscopy and reaction dynamics.

Concomitant with continuing methods development, we are investigating phenomena of fundamental biological, chemical, and environmental interest, e.g. conformational issues in sugars, solvent effects on uni- and bimolecular chemical reactions, partitioning of organic molecules between unlike media, and fate constants of environmental contaminants in aqueous media.

Accomplishments

Primarily in collaboration with Professor Don Truhlar of the University of Minnesota we have developed, implemented, and applied new techniques for including the effects of solvation in classical and quantal calculations. In particular, we have:

• developed and continue to distribute our licensed codes, AMSOL and OMNISOL, to calculate solvation free energies using semiempirical quantum mechanical and solvent-accessible molecular surface area continuum models, respectively. In addition to these two local codes, one or more of our solvation models either are or can readily be included in each of the following software packages: AMPAC, DGAUSS (using plug-in DGSOL), GAMESS (using plug-in GAMESOL), GAUSSIAN98 (using plug-in MN-GSM, which is available directly from Gaussian Inc.), SPARTAN, and ZINDO (using plug-in MN-ZM, which is incorporated directly into the parent code).

• developed classical and quantum mechanical solvation models to calculate all of the components of the free energy of solvation for a given solute. The models available presently cover water and any organic solvent for which a few key experimental properties are known. Our models, available for classical and quantal theories, in the latter case including semiempirical and ab initio molecular orbital theory and density functional theory, have been calibrated against 2000+ experimental data--an order of magnitude larger than any other test set described in the literature--leading us to conclude that at present our models are the most robust available. (See Publications 14, 18, 19, 22, 39, 46, 48, 49, 50, 54, 66, 70, 71, 74, 79, 81, 84, 98, 99, 100, 102, 110, 113, and 115.)

• developed analytic first derivatives for our continuum solvation free energies with respect to molecular geometry change. We are thus able to optimize geometries in solution with maximum efficiency and we are moreover able to characterize solution stationary points having no corresponding stationary points in the gas phase. (See Publications 48, 54, 71, and 114.)

• developed a two-response time model within the formalism of a configuration interaction model including all single excitations from a Hartree-Fock reference to analyze solvatochromic effects on molecular electronic spectra. This model provides the most accurate predictions to date for solvatochromic effects on the UV spectrum of acetone. (See Publication 126.)

  
  
  

  
  
  
• developed new models for the description of charge distributions within organic molecules. The charges from our semiempirical mapping technique, termed Class IV charges, are at least as good as charges derived from fitting to electrostatic potentials, but cost no more than standard Mulliken charges. Moreover, the method has been designed to accurately predict molecular dipole moments, making it useful for interpreting measurements of nonlinear optical properties, for instance. (See Publications 39, 42, 94, 111, and 117.)

• developed more efficient algorithms for the calculation of molecular surface areas as well as for the evaluation of radial integrals required by our solvation models. (See Publications 48, 54, and 71.)

• examined the utility of pairwise screening algorithms to replace complicated volume integrals with sums of analytical pairwise terms. This approach improves computational efficiency and should have broad application to molecular simulations, in addition to serving with our Generalized Born continuum model. (See Publications 54, and 71.)

  
  
  

  
  
  
• applied our models to successfully predict solvation effects on many different molecular equilibria, including acid-base, anomeric, conformational, and tautomeric equilibria. Among the molecules we have studied are biologically relevant solutes (dopamine, glucose, the nucleic acid bases, phosphate esters) and environmentally relevant solutes (chloroalkanes, chloroalkenes, and nitroaromatics). We have also successfully modeled solvent-solvent partitioning of organic solutes, a phenomenon of particular importance for drug delivery, environmental compartmentalization, etc. (See Publications 15, 17, 23, 25, 26, 30, 33, 43, 50, 58, 65, 72, 78, 87, 88, 89, 90, 106, 109, 110, 112, 116, and 118.)

• applied our models to successfully predict solvation effects on many different kinetic phenomena, including 1,2-hydrogen atom shifts in carbenes and nitrenium ions, electrocyclic organic reactions, hydrogen-atom abstractions in radiolytic destruction of aqueous organics, and detoxification pathways for organophosphorus-based chemical warfare agents. (See Publications 21, 38, 39, 84, 87, 88, 91, 92, 105, 113, 118, and 119.)

  
  
  

  

  Figure. Regional breakdown of aqueous solvation effects (kcal/mol) as they influence the rate of hydrogen atom abstraction from methanol by a H atom.

  
  
  
• provided new examples and in general described the use of quantum calculated solvated molecular properties in structure-activity relationships for the prediction of molecular toxicity/activity in solution. (See Publications 28, 47, and 104.)