Macroscopic Access Resistances Hinders the Measurement of Ion-Exchange-Membrane Performances for Electro-Dialysis Processes

Ion exchange membranes (IEMs) are widely used in industrial sectors from batteries and energy harvesting, to water treatment and desalination. They are designed with a high selectivity and a low electrical resistance. However, the way to measure such performances independently from interpretation and device parameters (geometrical sizes, water flux, electrodes, measurement type, etc.) remains a challenge. Therefore, many performances announced in the current literature are not reliable. This study focuses on membranes prepared for reverse electro-dialysis applications, but could be generalized to other membrane based processes. Due to the rising energy demand and the severe environmental crisis, worldwide reduction of conventional fossil fuel consummation is needed. As a renewable and clean energy resource, the Blue energy is an osmotic energy released during the mixing of solutions of different salinity. Naturally, an irreversible mixing of river water and sea water happens at the estuary. A precise and engineering control of this mixing enables the blue energy harvesting. The worldwide salinity gradient energy could reach 70 GW, which is not negligeable. Since the beginning of blue-energy harvesting research, a lot of advances have been made about membrane design and nanofluidic diffusio-osmotic transport, but the efficiency gap of 6 orders of magnitude between nano-scale designs and industrial power plants remains unsolved. Just after synthesis, selective membranes are benchmarked on very small samples on the basis of the output power. It is usually calculated based on the membrane resistance per unit area, and the measured Donnan potential. This method of power per unit area based on membrane resistance is therefore a key point to compare several devices and their efficiency. A classical electrochemical cell has been used to place a Nafion selective membrane of varying area between two electrodes. The membrane resistance and power density have been experimentally measured. Homogeneous salt concentration or gradient of concentration have been used to measure different parameters. This systematic study shows that for classical measurement cells, the membrane resistance is inversely proportional to the square root of its area. Whereas it's currently assumed in the literature that the resistance and the area of a membrane are inversely proportional. Based on this assumption, the power density is believed to be independent of the membrane area (reason why it is used as a comparison criteria), which we prove to be wrong. This result has a big impact on the reliability of the power density announced in the literature for different devices: the smaller the area of the tested membrane the higher is the output power density: from 0,5 W/m² up to 150 W/m² for the same membrane, same salt ratio (100) different membrane areas. This leads to overly optimistic extrapolation of electrical performances for membranes engineered and tested at sub millimeter sizes that will be industrially used at meter sizes. To explain these results, we show that for small membranes (compared to the electrode size and distance) there is an access resistance at the scale of the membrane. It results that the area mismatch between the membrane and the electrodes add a new resistance in the system that does not scale with the membrane area and that is taken into account in membrane characterizations. Experimental data and theoretical modeling are corroborating without any adjustment parameters. Consequently, the power per unit area obtained at micro scale cannot be extrapolated a priori to larger membranes. So, the comparison is biased between large scale commercially available membranes, and nano-engineered microscale membranes. This raises the urgent need for new indicators or standard measurement protocols. Based on these conclusions, we formulate a few recommendations to perform scalable and meaningful measurements of membrane resistance and power density.