Multi-paradigm multi-scale simulations for fuel cell catalysts and membranes

Abstract

Dramatically improving the performance of fuel cell systems with their complex heterogeneous structures involving electrocatalysts, proton conducting membrane, reactant, and interfaces between them requires understanding the fundamental chemical, electrochemical, and physical phenomena at the heart of these complex materials and relating these fundamentals to the properties and performance of the membrane–electrode assembly. Our goal is to develop a predictive model that can be used to estimate the changes in performance upon changes in the design and which can be used to monitor performance of working fuel cells. Our strategy is to start with first principles quantum mechanics (QM) and to develop overlapping simulation methodologies in which QM is used to train a reactive force field that can be applied for large-scale (millions of atom) molecular dynamics simulations while retaining the accuracy of QM. The results of molecular dynamics are used to extract a coarse grain or mesoscale description useful in modeling properties at much larger scales. This model would enable the conception, synthesis, fabrication, characterization, and development of advanced materials and structures for fuel cells and for the associated hydrocarbon fuel reformers in an overall fuel cell system. We illustrate here some of the progress toward this goal.

Keywords:

• Polymer electrolyte membrane fuel cells
• Computational modeling

Acknowledgements

This work was partially sponsored by General Motors. The computation facilities of the MSC have been supported by grants from DURIP-ARO, DURIP-ONR, NSF (MRI), and IBM-SUR. In addition, the MSC is supported by grants from DOE (DE-FG01-04ER04-20 and DE-FC26-02NT41631), ARO-MURI, ONR-MURI, NIH, ONR (NO. 014-02-1-0665), Chevron-Texaco, and Beckman Institute.