William H. Green
Current Research

Chemical kinetic models are increasingly used as input for major business and public policy decisions, ranging from the design of chemical process units to energy tax proposals. Rapid advances in computer hardware and software now make it feasible to construct truly predictive kinetic models based on a detailed understanding of the molecular processes occurring in complicated systems, rather than the very empirical interpolative models typically used today. Our efforts are focused on the development, demonstration, and validation of new methods for predicting the course of chemical reactions, and on the application of these methods to important technological and policy problems.

We have recently developed a general algorithm for constructing compact but complete reaction schemes for multicomponent reacting mixtures, given a procedure for estimating all the reaction rate parameters involved. We are now working on methods for identifying the range of conditions over which these kinetic models are valid, putting error bars on the model predictions, and identifying the performance bounds of these complex systems.

In addition to empirical rate-estimation procedures, we are using and developing methods to calculate reaction rates from first principles, using quantum mechanics and transition state theory. These theoretical methods are already competitive with some experimental techniques, and we are exploring ways to further improve the accuracy of these calculations. For example, we have developed a technique for improving the accuracy of density functional calculations, a perturbative method for handling the coupling between floppy molecular motions and stiffer vibrations, and several algorithms for computing the pressure-dependence of gas-phase reactions. We have also developed new "adaptive chemistry" methods to couple detailed chemical reaction schemes with transport; the ultimate goal is to quantitatively model the multicomponent reacting flows found in real systems.

The true test of any model is how well it agrees with experiment. We are collaborating with experimentalists in areas such as combustion and catalysis. In addition to measurements made on the whole system, we measure some critical reaction rates individually using high precision techniques such as laser flash photolysis. Recently, we have had good success predicting the operable range for new low-emissions, high-efficiency HCCI engines.

Further details

< back to bio