Thermodynamic study of a distiller-electrochemical cell system for energy production from low temperature heat sources

Abstract

The “thermally regenerable batteries” have been proposed for exploiting the heat sources at very low temperature, below 150° (e.g. low-concentration solar and industrial waste heat) for production of electrical energy, in particular for small power applications. In this case, traditional techniques, such as organic Rankine cycle, Stirling engines and solid-state devices based on Seebeck effect, are not economically viable. In particular we consider an electrochemical method working with a closed cycle: a salinity-gradient-power device produces electrical current by consuming the concentration difference between two solutions; the mixed solutions it produces are sent to a distiller which restores the concentration difference, in turn exploiting the low-temperature heat source. Unfortunately, the efficiency of the devices proposed up to now is low. In this manuscript, we present for the first time a theoretical analysis of the whole cycle, in order to enable an educated choice of solutes and solvents with respect to the efficiency in the conversion of heat into electrical power. We find that the main requirement is a high boiling point elevation; minor advantages are obtained by solutions with a high latent heat of vaporisation and low specific heat capacity. The first two requirements could appear counter-intuitive, since they are detrimental in the case of distillation processes per se. While the above-mentioned requirements are connected to fundamental limitations of the energy and exergy efficiency of the device, the electrochemical parameters mainly affect the power density. Our results allow to devise solutions for single-effect processes that give a high exergy efficiency for very low temperature heat sources, i.e. to approach the Carnot cycle limit (e.g. 11% of energy efficiency for a temperature difference of 40 K), competitive with the more traditional techniques but much cheaper and easily down-scalable.