Theoretical Chemistry, **Machine** VP

School of Chemistry,

University of Sydney

**Quantum Mechanical Simulation of Gas/Surface Dynamics**

Many industrial problems involve gaseous interaction with solid surfaces, such as thin film growth and catalysis. Understanding the dynamics of gaseous chemisorption on solid surfaces at a molecular level is crucial for practical applications. Traditional molecular dynamics techniques have been widely used for this purpose, however, the interaction potentials or forces between atoms used are often empirical. Recently, Car and Parrinello proposed a scheme which combines the first principles quantum mechanical calculations with the traditional molecular dynamics simulations. This is exciting since the interaction potential is now from first principles calculations. The main objectives of this project are to develop and extend this first principles molecular dynamics program and investigate gaseous interaction with solid surfaces.

**What are the basic questions addressed?**

The basic questions to be addressed are: what is the most stable atomic structure and its associated electronic structure? How do the atoms move starting from an assumed initial condition? In this project, we focus on the systems of hydrocarbons, hydrogen, oxygen and metal atoms chemisorption on diamond, silicon and copper surfaces.

**What are the results to date and future of the work?**

To date, we have carried out realistic simulations on two gas/surface systems: hydrocarbons chemisorption on diamond (111) surface and Palladium chemisorption on Copper (100) surface. For diamond system, 85% of the calculations are completed and the rest will be calculated in the next two months. For the Pd/Cu(100) system, several models have been tested including the bulk alloy and the surface calculations are being carried out. The future work will be the completion of the above two systems and the other systems outlined above.

**What computational techniques are used and why is a supercomputer
required?**

As stated above, we employ a first principles quantum molecular dynamics technique, ie. the Car-Parrinello approach, and the program ATAMI which was originally developed by a Japanese Computational Physics Group. Because of the ab initio in nature, it requires large memory and intensive CPU, therefore a conventional mainframe computer is far from satisfactory.