Simulation Study of the Correlation between Structure and Conductivity in Stretched Nafion Elshad Allahyarov* Department of Physics, Case Western ReserVe UniVersity, CleVeland, Ohio 44106, and Joint Institute of High Temperatures, Russian Academy of Sciences (IVTAN), Moscow 125412, Russia Philip L. Taylor Department of Physics, Case Western ReserVe UniVersity, CleVeland, Ohio 44106 ReceiVed: May 30, 2008; ReVised Manuscript ReceiVed: October 14, 2008 We have used coarse-grained simulation methods to investigate the effect of stretching-induced structure orientation on the proton conductivity of Nafion-like polymer electrolyte membranes. Our simulations show that uniaxial stretching causes the hydrophilic regions to become elongated in the stretching direction. This change has a strong effect on the proton conductivity, which is enhanced along the stretching direction, while the conductivity perpendicular to the stretched polymer backbone is reduced. In a humidified membrane, stretching also causes the perfluorinated side chains to tend to orient perpendicular to the stretching axis. This in turn affects the distribution of water at low water contents. The water forms a continuous network with narrow bridges between small water clusters absorbed in head group multiplets. In a dry membrane the side chains orient along the stretching direction. I. Introduction In their role as proton-conducting membranes, ionomers are an important component of many hydrogen fuel cells. In these materials, the interplay among the short-range interactions between the hydrophobic backbone polymer and the hydrophilic terminal groups and the long-range Coulomb interactions between the electrostatic charges on the terminal groups and the protons induces a nanophase separation into proton-rich and proton-poor domains. A general model for the phase morphology of ionomers has been proposed by Eisenberg et al., 1 according to which a few head groups combine to form multiplets that restrict the mobility of the backbone chain segments directly attached to them. A spherical geometry is often assumed for these multiplets, whose sizes are typically less than a nanom- eter. 2 The average distance between these multiplets is mostly dictated by the concentration of head groups relative to that of backbone monomers. These multiplets then form microdomains, and one observes a microphase separation that can serve to facilitate proton diffusion. Recent experimental data on the morphology of ionomers describe the aggregations as elongated objects embedded in a continuous ionic medium. 3 A goal of much ionomer research is to increase the proton conductivity, and hence to make membranes that can operate under very low humidity conditions, as a higher conductivity results in a higher output power density in fuel cells containing these membranes. Proton conduction itself is a complex process, which strongly depends on the thermal and mechanical history of the membrane. 4-6 The mechanical history involves the manufacture of the membrane, which is usually achieved by one of two common procedures: solution casting 7,8 or extrusion. 9 The former method is used to make membranes from a solution of dissolved ionomer by allowing the solvent to evaporate from the solution. While this technique is suitable only for small-scale laboratory production, it has the advantage producing isotropic membranes with no residual preferential orientation of their backbone within the plane of the membrane. Most commercially available ionomeric membranes are fabricated by extrusion of a molded sample. This leads to a preferred orientation of the ionomer backbone, 10 the extent of which depends on the draw speed of the extrusion. This inherent structural anisotropy of the backbone matrix is believed to be a reason for the susceptibility to tearing or cracking of the membrane in any swelling or drying processes. This creates technical problems in keeping the membrane taut in the changing temperature-humidity conditions encountered in fuel cells. An equally important issue concerns the effect of this anisotropy on proton conductivity. While it has generally been assumed that such effects are significant, the problem of predicting the consequences of mechanical strain remains largely unsolved. There have been several experimental studies in which membranes were uniaxially stretched in order to probe the effects of strain on internal morphology. Gebel et al. 11 analyzed the form of ionic domains in unstretched and stretched membranes and showed that mechanical stretching induces ordering in the ionomer backbones. Elliot et al. 12,13 detected the anisotropy in stretched membranes by performing scattering experiments. Barbi et al. 10 used membrane elongation measure- ments to confirm the classical ionomer domain model, in which inverted micelles are interconnected by channels. As expected, uniaxial stretching of recast Nafion causes a preferential orientation of the Nafion backbone in the direction of stretching, and this is morphologically similar to the anisotropy in extruded membranes. Cable et al., 14 for example, stretched Nafion and found the proton conductivity to be higher in the plane of the membrane than normal to it. A slightly different result was found by Lin et al., 15 who noted little change in transverse conductivity on stretching but found an improved fuel cell performance, as compared to Nafion 117 and unstretched recast Nafion, as a * To whom correspondence should be addressed. J. Phys. Chem. B 2009, 113, 610–617 610 10.1021/jp8047746 CCC: $40.75  2009 American Chemical Society Published on Web 12/30/2008