Thermal regeneration of ammonium bi-carbonate solutions for closed-loop reverse electrodialysis

Abstract

Reverse electrodialysis is a novel technology that exploits a salinity gradient to generate electrical energy. The salinity gradient can be available from natural waters such as seawater and river water or they can be artificially generated and used within closed-loop applications. This last option has been recently investigated leading to the development of the RED heat engine concept. In this case, the deployed salinity gradient exiting the RED unit is regenerated in a thermally-driven unit using low-temperature heat, thus being able to convert heat to power within an integrated system. Among the different regeneration alternatives, the use of thermolytic salts has been presented as a promising option, due to the possible use of very-low grade heat (40-60°C) to regenerate the solutions by means of degradation/stripping/re-absorption processes. In the present work, different regeneration strategies for ammonium bi-carbonate aqueous solutions have been investigated by means of process simulation (Aspen Plus) and experimental tests. Simulations have been performed looking at two different regeneration methods: i) stripping with air and ii) distillation (in practice, also a stripping process, but with vapour). A sensitivity analysis has been performed to study the effect of different operating variables (streams’ temperature, pressure, flow rate, inlet concentration) on the regeneration performance. The experimental campaign has been carried out mainly on the air stripping concept, which did turn to be the most promising within the expected range of operating conditions. Experiments were also aimed at identifying the main dependences (e.g. the effect of different packing materials and operating conditions on the performance indicators), technological limitations and relevant solutions. Eventually, a comparison between experimental information and model predictions has been performed in order to highlight the main discrepancies and validate model prediction capabilities.