The development of accurate, computationally tractable methods to model complex molecular systems is a fundamental goal of modern theoretical chemistry. High-level quantum mechanical methods are often very reliable; however, the compute-intensive nature of the calculations severely limits the range of applications. Alternately, simple empirical models frequently lack sufficient accuracy (or transferability) to be predictive for chemical events such as ionic interactions or formation/cleavage of chemical bonds.

This talk will outline some recent developments toward the improvement of chemical models by the design of 1) algorithms to remove scaling bottlenecks in the calculation of the electronic structure and classical electrostatic problems, and 2) new force field models that include quantum mechanical many-body effects. Topics include methods to circumvent the cubic or higher scaling bottleneck of single-determinant wave function theories [1], new methods for improved accuracy in linear-scaling Ewald and solvation techniques [2], and a density-functional based many-body force field model for molecular simulations [2]. Preliminary applications of some of the methods to proteins and DNA in solution will be presented.

[1] J. Khandogin, A. Hu & D. York, J. Comput. Chem., in press;
     D. York et al., Phys. Rev. Lett. v89, 5011 (1998);
     D. York, in Combined Quantum Mechanical and Molecular Mechanical Methods, ACS Symp. Ser. (J. Gao M. Thompson, Eds.), Ch. 18, p275 (1998);
     D. York et al., J. Am. Chem. Soc. v118, 10940 (1996);
     T.-S. Lee et al., J. Chem. Phys. v105, 2744 (1996).

[2] D. York and M. Karplus, J. Phys. Chem. A v103, p11060 (1999);
     D. York, T.-S. Lee and W. Yang Chem. Phys. Lett. v263, p297 (1996).

[3] D. York & W. Yang, J. Chem. Phys. v104, 159 (1996);
     D. York, Int. J. Quant. Chem. QCS29, p385 (1995).

Back to Workshop Schedule

Back to Reactive Flow and Transport Phenomena