45666-AC5
Proton Transfer in Phosphoric Acid Based Fuel Cell Membranes
The ultimate goal of this project is to investigate the mechanism of proton transport (PT) in a typical fuel cell proton-exchange membrane (PEM) by use of a hybrid simulation method using Atom-centered Density Matrix Propagation (ADMP) and molecular dynamics implemented in ONIOM. The focus of this work is phosphoric acid (PA)-doped polybenzimidazoles. In Year One of this project, the focus was PA. This effort centered on determining the appropriate DFT functional and basis set to use for ADMP calculations. Results indicated that X3LYP/6-311+G(d,p) provided the best trade-off between computational cost and accuracy in modeling PT events in PA systems as compared to MP2/aug-cc-pVTZ calculations; however, use of the smaller basis set 6-31+G(d,p) in the X3LYP functional was considered to be an acceptable choice to reduce computational cost. At the MP2/aug-cc-pVTZ level of theory, the proton affinity of PA was determined to be 196 kcal/mol with a dipole moment of 0.461 D. Proton transfer in PA was investigated in a PA cluster consisting of a five PA molecules using X3LYP/6-31+G(d,p) in ADMP. Results at 450 K indicated that one PT event occurred every 100 fs. In additional studies, the COMPASS force field was used for molecular dynamics simulation of PA. COMPASS provided reasonable results for density, diffusion coefficients, and radial distribution functions (RDFs) for PA. During Year Two, the COMPASS force field was successfully used to construct amorphous cells of three benzimidazole polymers - poly[2,2'-(m-phenylene-5,5-bibenzimidazole] (PBI), poly(2,5-benzimidazole) (APBPI), and poly(p-phenylene-benzobisimidazole) (PBDI) to determine density and RDFs of these polymers doped with PA, water, and representative compositions of water and PA to study preferences for hydrogen bonding of the polymers with PA and with water including including an evaluation of competition for proton-accepting sites. In Year Three, results of MD simulations using the DREIDING and Universal force fields will be compared to COMPASS MD results and COMPASS predictions of bond length, angle, and torsion predictions obtained in Year Two. Preliminary results suggest DREDIING would be the preferred choice. The best force field will be then used in Year Three for ONIOM calculations in Gaussian with ADMP/X3LYP/6-31+G(d,p) at the high level of theory to investigate the PT mechanism in PA-doped polybenzimidazoles.
