Profile development and dimensional change in the melt-spinning process of hollow fibers were studied using a finite element method, and a numerical simulation was compared with experimental results. The numerical simulation of a hollow-fiber spinning process was carried out by considering the changes in inner and outer boundaries. Initial dimensions of inner and outer radii were obtained by measuring the dimensions of extrudate at the extrudate swell point using a capturing device. Extrudate swell plays an important role in determining the initial dimensions of the inner and outer radii, but has less effect on the hollow portion at the die swell point. The effects of spinning variables on the hollow portion show that spinning temperature is the most critical variable in controlling the hollow portion, followed by mass throughput rate. Take-up velocity has relatively less effect. As the mass throughput rate and takeup velocity increase and the spinning temperature decreases, the hollow portion of asspun fiber increases.

Surface functionalization of nitrogen-containing species on carbon support has been shown to affect electrocatalytic activity and carbon-ionomer interaction in the fuel cell electrodes, but studies of the functionalization on mass transport are limited. We reported two schemes for grafting positively (nitrogen groups) charged species and a scheme for negatively (sulfonates) charged species on the surfaces of three carbon materials. The functionalization was characterized with BET, XPS, contact angle measurements, and zeta potentials. In full-cell tests, improvements over high current densities were observed in samples reacted with para-phenylenediamine or ammonia, whereas the performance decreased after functionalization with sulfonate groups. The improvement at high current densities exceeded the mass-activity improvement and was attributed to reduced mass-transfer polarizations. Furthermore, a statistical approach was developed to examine the changes in ionomer surface coverage using STEM images with fluorine EDS maps. Durability studies followed the DOE’s protocol of potential cycling between 1.0 V and 1.5 V. Faster rates of carbon loss were found to occur after all three surface functionalization schemes due to preferential oxidation of the weaker covalently bonded functional groups, but despite the increased rates of loss, after amination the most durable carbon support showed end-of-life performance comparable to the initial performance.

No abstract

No abstract

A novel reinforcing porous substrate that features unprecedented capability of offering proton conductivity is demonstrated for potential use in a reinforced composite proton exchange membrane. The unusual porous substrate (hereinafter, referred to as ''sPI substrate'') is composed of 3trihydroxysilyl propane-1-sulfonic acid (THSPSA)-based silicate coating layers and electrospun polyimide (PI) nonwoven fibers. The THSPSA coating layers bearing sulfonic acid groups endow the sPI substrate with strong affinity for water molecules and also afford appreciable proton conductivity. Another distinctive characteristic of the sPI substrate is the nanoarchitectured structure of the THSPSA coating layers (shell) that encapsulate the PI nonwoven fibers (core). As a result, the coreshell structured sPI substrate maintains a highly porous structure, which plays a crucial role in providing effective proton-conducting channels after the impregnation of a polymer electrolyte (herein, sulfonated poly(arylene ether sulfone) (SPAES)). Notably, owing to the assistance of the protonconductive sPI substrate, the sPI substrate-reinforced SPAES composite membrane presents higher proton conductivity than a PI nonwoven-reinforced SPAES composite membrane under various relative humidity (RH) conditions. This intriguing proton conductivity behavior is discussed based on an in-depth understanding of the unique core-shell structure and functionality of the sPI substrate and, moreover, is quantitatively interpreted by estimating theoretical proton conductivities predicted from series and parallel two-layer models.