The proton exchange membrane (PEM) is a key component in proton exchange membrane fuel cells (PEMFCs). During standard fuel cell operation, the PEM degrades due to cyclic hygrothermal loads, resulting in performance loss or total failure. Improvement of current PEM materials and development of cheaper, more durable materials is essential to the commercialization of PEMFC technology, which may provide an attractive alternative energy source for transportation.

This dissertation investigates a new PEM material which is a blend of sulfonated perfluorocyclobutane (PFCB) and polyvinylidene fluoride (PVDF). Hereafter referred to as PFCB/PVDF, this polymer blend was developed by General Motors Company™ as a potential replacement for the current benchmark PEM, the DuPont™ product Nafion®. The PFCB/PVDF blend is less costly to manufacture than standard PEM materials and investigations into its long-term mechanical durability are ongoing.

Specifically, this document discusses the experimental and analytical work performed in the material characterization, constitutive expression development, and implementation of that expression into uniaxial and biaxial finite element geometries. Extension of the model to time-varying temperature and moisture conditions is also explored.

The uniaxial finite element model uses a non-linear viscoelastic viscoplastic (NLVE-VP) constitutive expression with parameters determined from uniaxial creep and recovery experiments at a single environmental condition. Validation tests show that this model accurately predicts results from uniaxial tension experiments, such as stress relaxation, force ramp, and multistep creep and recovery, to stresses of 8 MPa and strains approaching 15%, which is the maximum hygrothermal strain expected in an operating fuel cell.

The biaxial finite element model combines the NLVE-VP constitutive expression with the geometry of a pressure-loaded blister experiment, which better approximates fuel cell membrane constraints. Results from the biaxial model are compared to experimental results. The model accurately predicts strain in the blister test but predicts stresses that differ from those estimated from blister curvature.

Additionally, it is found that both the non-linear viscoelastic and viscoplastic parameters are functions of the operating environment. Future experimental work is needed to characterize that dependence before the constitutive model is used to simulate the response of the PFCB/PVDF blend to fuel cell operating conditions.