10.3.1 Nanoscale Membrane Structure

By and large, the neutron technique most used to study PEM materials has been SANS. It is understood that the nanostructure of PEMs plays a significant role in water uptake and transport, which are materials with properties vital to the performance of a working fuel-cell. However, in order to design a material with desired features, one must have a detailed understanding of the interplay between the nanostructure of the membrane and the overall performance properties. SANS is a versatile tool in elu­cidating the structure of a variety of membrane materials and can also be used to study transport, which will be discussed in the section on water transport.

As previously mentioned, Nafion® has been the most widely-studied PEM material to date and SANS has been employed extensively to study the nanoscale structure of this complex material [8, 10, 11, 14-23]. The earliest structural studies of Nafion® utilizing SANS and small-angle X-ray scattering (SAXS) revealed a broad peak at a Q value between 0.1 and 0.2 A 1, called the ionomer peak, which has been attributed to the correlation between the nanophase-separated ionic domains, termed clusters. The crystalline component contributes to the scattering at multiple length scales including peaks in the Q range 0.6-2.0 A-1, owing to the structure of the amorphous and local crystalline-lattice, in addition to a broad peak entered at lower values of Q (^0.05 A 1), which is related to the inter-crystalline scattering (known as the long period). Moreover, ultra-small-angle scattering reveals an upturn that can be associated with large-scale heterogeneities. The scattering for hydrated Nafion® over a wide range of length-scales can be seen in a review by Gebel and Diat [19, 24].

An example of the scattering using neutrons can be seen in Fig. 10.1, for Nafion® films cast from a dispersion, annealed between 80 and 180 °C, and equilibrated in liquid water. When annealed below the alpha relaxation temperature (Ta) of Nafion® (^ 100 °C), the lack of a peak at low Q values is evidence that there is no apparent long-range crystalline order in the film. Above Ta, however, one observes a peak due to long-range crystalline order and a shift in the crystalline peak to lower Q values with increasing annealing temperature, indicating that higher annealing-temperatures result in larger, more widely separated crystalline domains. An analysis of the scattering curves can be seen in the inset in Fig. 10.1. Clearly, the spacing of the crystallites increases with increasing annealing tem­perature. Moreover, the spacing between the ionic domains decreases with increasing annealing temperature. It is known that the crystalline structure plays an integral part in the mechanical stability and durability of fuel-cell membranes, but these data also reveal the relationship between the crystalline structure and water uptake. The decreased spacing between the ionic domains and the lower incoherent — background with increasing annealing temperature are evidence that these annealed films have a lower water-uptake. In these materials water retention must be bal­anced with annealing temperature in order to achieve desirable proton-conductivity and mechanical integrity. This is just one example of how SANS can be used to probe structure-processing-property relationships.

Over the decades, development and advancement of state-of-the-art scattering techniques were able to reveal the many scattering features, over multiple length — scales, of Nafion® and other perfluorosulfonic acid membranes. As a result, there has been a progression in the complexity and characteristics of the many mor­phological models that have been proposed to explain the observed scattering in effort to gain a deeper fundamental insight into how the structure is related to

material performance. The proposed models generated have included a variety of structural units and range from the earliest spherical cluster-network model to other models including lamellar, sandwich-like, fringed micelle, rod-like, and ribbon-like, as well as a model which includes cylindrical water channels. Each of these models was able to account for the scattering to some degree reasonably well, making it difficult to discern which most accurately describes the morphology. Of course, the models able to capture the many scattering-features over a large Q range, with physically relevant fitting parameters are of the highest value. For the details of the structural models of Nafion® and their respective parameters, the reader is directed to the extensive literature concerning this topic.

In general, scattering techniques provide an excellent way to characterize the global structure of fuel-cell membranes [19, 24]. For polyelectrolytes such as Nafion® there are scattering features that are ubiquitous and considered to arise from favourable structures for fuel-cell membrane application, although that con­ventional wisdom is now being called into question by more recent studies. Typically, polyelectrolyte fuel-cell membranes contain ionic moieties that are able to conduct protons or hydroxide ions in alkaline fuel-cell membranes. These ion — containing, polar-groups phase separate from the more hydrophobic components of the polymer and can form ion-conducting channels which are responsible for ion transport. Quite often these ionic domains give rise to a scattering peak in SANS and, if other hierarchical structures are present, other scattering features are observed. This is especially true when block copolymers are used to provide a structural basis for the membrane. A recent review by Elabd and Hickner [25] has evaluated the state-of-the-art block-copolymer membranes by leveraging the self — assembled nanostructure of block copolymers as a template for creating well — defined transport pathways for use in fuel cells.

In addition to the work on Nafion®, there is a rich body of literature in which SANS has been used to probe the structure of a variety of PEM materials including sulfonated polyimides [26], sulfonated polyetherketones [24, 27], sulfonated tri — fluorostyrenes [28], poly(styrenesulfonic acid)-grafted cross-liked polytetrafluoro — ethylene [29], and a host of other materials [30-34]. Ultimately, one seeks to understand the role that molecular-level structure and chemistry play in the development of material nanostructure and how this nanostructure is correlated with performance properties such as water content and transport, as well as ion con­ductivity. For example, in the work by Iwase et al. SANS (in conjunction with SAXS) was used to investigate the hierarchical structure of graft-type PEMs syn­thesized from cross-linked PTFE [29]. The structure was studied over a large range of length scales (0.6 nm to 1.6 pm) as a function of the degree of grafting, Xg. It was determined that the structure of these materials consisted of conducting layers of polystyrene sulfonic acid (the grafted domains) arranged in lamellar stacks on the surface of the PTFE crystallites. Within the conducting layers, they observed scattering features consistent with correlations between sulfonic acid domains. With less than 15 % grafting the grafted domains were found to reside mainly in the amorphous domains between the PTFE crystalline lamellae. Within this regime, the lamellar spacing increased with increasing grafting content up to a value of Xg of about 5 % and remained constant until 15 %. Above 15 % the grafting domains appeared to phase separate from the hydrophobic matrix and become contiguous, thus forming a highly conductive domain around the crystallites.

While X-ray scattering is certainly more widely accessible for structural char­acterization of membrane materials, neutrons offer the unique benefit of contrast variation, or contrast matching, in the structural determination of systems with complex architectures. Owing to the large differences in scattering-length (SL) between deuterium and hydrogen, one can use isotopic replacement in the polymer, or the solvent, to highlight the scattering from various structural components or phases. One example of this can be found in the work by Gebel et al. [35] in which they used various mixtures of D2O/H2O to swell N(CH3)+-neutralized forms of Nafion® as a way of elucidating the nature of the scattering entities in these hydrated films. By varying the ratio of D2O to H2O and normalizing by the scat­tering of Nafion® in pure H2O they were able to match out the structural component of the scattering due to Nafion® and to observe the counterion condensation at the interface between the hydrophobic components of the polymer and the hydrophilic water domains. This was the first measurement of condensation in a perfluoro- sulfonated ionomer. Using contrast variation to explore neutralized forms, different models could be applied to determine which structure accurately described the scattering curves. While this study was unable to determine the shape of the scattering particles (i. e., spherical or rod-like), it was determined that the features were aggregates of the polymer backbone surrounded by the electrolyte solution, as opposed to the scattering particles being cavities filled with the electrolyte solution.

Recently, a series of studies using in situ SANS, among other techniques, on block-copolymer electrolyte membranes consisting of a polymethylbutylene (PMB) block and polystyrenesulfonate (PSS) block have begun to call into question whether or not ionic aggregates are necessary for effective proton-transport, espe­cially in the presence of structures established by block-copolymer morphology [33, 36, 37]. The composition of the block copolymer was varied in order to tune the size of the domains. Also varied was the degree of sulfonation of the polystyrene block, within a particular composition. This body of work represents an excellent example of the application of neutron scattering to elucidate the structure-property relationships of PEM materials in environments that are application relevant. These in situ measurements were achieved using a specially designed sample chamber at the National Institute of Standards and Technology (NIST) Center for Neutron Research wherein the humidity and temperature of the environment surrounding the sample could be controlled. Moreover, the water reservoir within the sample chamber could be filled with various mixtures of H2O and D2O, allowing contrast — variation experiments to be performed. The scattering was measured over a range of relative humidity and temperature values using D2O. For one particular block copolymer composition, the scattering indicated the presence of a hexagonal phase over the entire range of relative humidity and temperature values studies (Fig. 10.2). At low temperatures (25 °C) and humidity (^25 %), the scattering arises from the block-copolymer morphology. However, at 95 % relative humidity and 40 °C, a shoulder at approximately Q =1.8 nm-1 was observed, which became more pro­nounced and intense upon further heating and humidification. It was acknowledged that this peak was similar to that of the ionomer peak observed in Nafion®, but was referred to as the ‘water peak’ as it was only visible upon hydration. The water domain-spacing was taken to be 2n/Qmax of the peak. For a given block-copolymer system under the same environmental conditions (relative humidity = 95 % and at 60 °C) the position of the water peak was shown to shift to higher values of Q with increasing levels of sulfonation. This result was attributed to a decrease in the average distance between sulfonate groups upon increasing sulfonation. A contrast — variation study was performed to determine the origin of the water peak. The water reservoir was filled with a volumetric mixture of D2O/H2O of 32/68, chosen to

match the scattering-length density (SLD) of the dry PSS block. The water peak was shown to disappear when the samples were humidified with this mixture, indicating that the peak arises due to the presence of a substructure within the PSS superstructure, most likely a heterogeneous distribution of water-rich and water — poor domains as is found in most polystyrene ionomer systems.

One of the most important observations of this study was the absence of the water peak when the size of the hydrophilic domains was below a critical thickness (Fig. 10.3). For PSS-PMB copolymers this critical thickness was on the order of 6 nm to 10 nm for a sulfonation level of about 47 mol percent. It was determined that the water-rich domains were effectively homogenized due to confinement effects. This SANS work has played a critical supporting role in determining the molecular and morphological origins for the enhanced water retention and proton transport observed in the study of these copolymer systems. These results have provided a new perspective in the strategy for developing materials for use in PEM fuel cells.