Re: Welcome to the world, eOn2
Posted: Thu Oct 14, 2010 2:58 am
The current project is based upon calculating the dynamics of solids by finding saddle points on the potential energy surface which give the rate at which atoms can move. This relies on an approximation called harmonic transition state theory, which gives rise to the Arrhenius form of a reaction rate = nu exp(-dE/kT), where dE is the energy of the saddle point for a reaction.
We can do better in a couple of ways. The first is to use the motion of the atoms as they escape from transition states to obtain true rates, without resorting to the harmonic transition state theory. For this, each client will run short trajectory from a high energy transition state and look for reactive trajectories to a new product state. This will give us exact dynamics.
The second improvement is in the evaluation of the interatomic forces. Currently, we are using empirical potentials which fairly accurately describe some idealized system, such as pure metals and pure oxides. To investigate more interesting properties such as surface chemistry, catalysis, defects in materials, and many other non-ideal systems, we need some quantum chemistry to treat electrons properly. This is a qualitative difference in the cost of the calculations. A single evaluation of the energy of a 100 atom system goes from about 1 ms with an empirical potential to 1 min with quantum chemistry (and the density functional level of theory).
Both types of job types are currently being implemented, and will be tested soon.
We can do better in a couple of ways. The first is to use the motion of the atoms as they escape from transition states to obtain true rates, without resorting to the harmonic transition state theory. For this, each client will run short trajectory from a high energy transition state and look for reactive trajectories to a new product state. This will give us exact dynamics.
The second improvement is in the evaluation of the interatomic forces. Currently, we are using empirical potentials which fairly accurately describe some idealized system, such as pure metals and pure oxides. To investigate more interesting properties such as surface chemistry, catalysis, defects in materials, and many other non-ideal systems, we need some quantum chemistry to treat electrons properly. This is a qualitative difference in the cost of the calculations. A single evaluation of the energy of a 100 atom system goes from about 1 ms with an empirical potential to 1 min with quantum chemistry (and the density functional level of theory).
Both types of job types are currently being implemented, and will be tested soon.