Jacqueline Keane (keane@ifa.hawaii.edu), NASA Astrobiology Institute

Pavel Senin (senin@hawaii.edu), University of Hawaii at Manoa

UniformSampler #1

Selected specie

UniformSampler #2

Intermediate sample(H, H2, C, CO, O, O2, N, N2)

Reactions networkH+O=OH, C+O=CO …

Grain surface active population

(C, CO, CO2, O, O2, O3, N2 …)

Cloud population(H, H2, C, CO, O, O2, N, N2)

Coreactant

UniformSampler #3

Product(s)

Grain surface buried population

(C, CO, CO2, O, O2, O3, N2 …)

UniformSampler #4

Configured parameters

Physical parametersAbundance, Temperature ...

Possible coreactants

Reaction processing

Chemical reactions on interstellar dust grains play a crucial role in interstellar chemistry by promoting the formation of organic products.

While many of the reaction rates are poorly understood and molecular formation routes are difficult to isolate in the laboratory, computer simulations of these reactions allows us to better understand the nature and evolution of interstellar molecules in molecular clouds. The work presented here is part of the University of Hawaii and NASA Institute of Astrobiology CASS 2006 initiative which aimed to involve graduate students from diverse regions of science into the field of Astrobiology.

This particular collaboration resulted in the development of computational software that implements a MCMC stochastic model of the surface chemistry. The model adopts “scoundrel approach” based on the Langmuir-Hinshelwood and Eley-Rideal mechanism.

While current work mimics in flow previously designed model and produces similar results, the presented software implementation written from scratch using Java and R. The software and source codes are available through the Google projects hosting using GPL and we believe that model will evolve much further through the contributor submissions.

When compared with original IDL implementation performance, current software version is faster, runs on virtually any platform which supports Java, has CLI and GUI interfaces. Software provides numerous options for set-up and tuning of simulations. Modular architecture and use of threads allow running multiple simulations simultaneously on multi CPU platform.

To assure the overall software equality we use automatic JUnit testing along with collection and analysis of various software metrics.

Currently we are working on the gas-phase chemistry code and planning to release complete gas-surface chemistry model soon. We are going to change gas-phase chemical composition and ambient environment by injecting a data produced by the gas-phase model between processing intermediate population chunks.

The sampling flow within the softwareMultiple independent samplers mimics

the real-life random process driven by the parameters provided.

The project homepagehttp://code.google.com/p/iclouds/

Hosted by Google, provides set of features:SVN, WiKi pages, Issue tracking,

groups (mailing list), developers blog.

The presented model deals with two processes: accretion of species from gas phase and on-surface chemistry.

The model implements the Monte Carlo accretion limited method by using two essential and simplifying assumptions rules the chemistry processing:

- the accretion timescale is much less than the timescales for reaction of the accreted atoms with other species;

- there is no desorption of grain mantle species and thus the surface chemistry is not coupled back to the gas-phase.

The reactions network considered here based on reactions that occur only between weakly bound sites: from one physisorbed to another physisorbed site as described by Langmuir-Hinshelwood chemistry. As for the species that are tightly bound to the surface the chemistry is ruled by Eley-Rideal mechanism.

The software produces two types of output: the resulting chemical grain composition and full grain chemical evolution log. While first is used to compare computational results with observational data, the evolution log used for studying dynamic chemistry flow.

This poster ….