Computationally Intensive Research Project
Nucleation, Growth and Evaporation Rates of Aerosols
MSC News
Additional Information
Theresa L. Windus,1 Mark S. Gordon,1 Shawn M. Kathmann2
1Iowa State University, 2Pacific Northwest National Laboratory
FY07 Allocation - 1,200,000
Abstract
Aerosols are a critical component in atmospheric radiative forcing and can affect climate in multiple ways -- including scattering and absorption of radiation and forming cloud droplets (cloud condensation nuclei). It is now recognized that the spatial and temporal distributions of aerosols due to industrial activities are as important in determining overall climate changes as the influence of greenhouse gases. However, it has also been recognized that there is insufficient understanding of how aerosols affect climate with significant uncertainty in formation rates, chemical composition, and properties. This gap in knowledge severely limits our ability to determine human effects on the climate. Understanding the nucleation, growth and evaporation rates of aerosols as well as their chemical properties is essential to improve climate models and overall global climate prediction as well as the general chemical ecosystem in the atmosphere.
In our research we will use several methods to determine multiple properties of aerosols. In particular, we will use Dynamical Nucleation Theory Monte Carlo (DNTMC) methods developed by Kathmann, Schenter, and Garrett, combined with force field and ab initio methods, to sample the configuration space of molecular clusters using clusters of various composition and size leading to aerosol formation. In addition, we will use a representative subset of the ensemble to determine molecular properties, such as vibrational spectra and excited state energies, to produce data that can be integrated with experiment to get a more complete picture of the aerosols that affect climate. Oxidation, radical and radiation reactions will also be examined since these have been determined experimentally to affect the lifetimes of aerosols. This project will have multiple overlapping phases. The DNTMC model has been added into NWChem and will allow any method in NWChem to compute energies with the DNTMC module. The first phase of computations will be to compare the force field model available in the original DNTMC code with ab initio methods for non-reacting sulfuric acid/water clusters to determine the fidelity of the force field models. The next phase will be to use effective fragment potentials (EFP) -- available in the GAMESS code -- to efficiently examine larger sulfuric acid/water clusters. During this phase EFPs will also be incorporated into NWChem. In the third phase, we will expand the computations beyond the sulfuric acid/water clusters to the examination of other important chemicals present in aerosols such as nitric acid and soluble and insoluble organic materials. We will examine reactions in important cluster sizes to determine oxidation rates, radical reactions, and radiation (excited state) effects. The calculation of the thermodynamic and kinetic properties of a single cluster at a given composition will require many configurations (i.e. 1000s to 10,000) and the software will take advantage of multi-level parallelism to exploit many processors to address the sampling issues. These computations will require at least a thousand processors during a single computation and will therefore effectively use the MSCF resources to their fullest extent possible.
