Experimental studies on optimal process of the iodine-sulfur cycle for nuclear hydrogen production

Abstract

For nuclear hydrogen production, we selected Iodine-Sulfur (I-S) cycle as the most promising one by screening process among 115 thermo-chemical water splitting technologies. We developed a thermo-physical model for the hydrogen-iodide (HI) VLE and decomposition behavior in the iodine-sulfur (IS) cycle to improve the conventional I-S cycle suggested by GA. Neumann’s modified NRTL model was improved by correcting an unphysical assumption for the non-randomness parameter, and using the two-step equilibrium approach for the HI decomposition modeling. However, the parameters of the model were decided through regression with the 271 sets of existing experimental data; the accuracy of the model should be improved by more experimental data over all operating ranges, especially, in the high temperature and high pressure regions. To obtain the data of those regions, an autoclave for high temperature and high pressure was designed and manufactured. Various materials and surface coating technologies were investigated for preventing corrosion from acids. However, we have currently failed to overcome the corrosion problems with highly corrosive acids at a high temperature and high pressure. We experimentally validated that azeotropic constraint between acid and $H_2O$ undermined the total efficiency of the I-S cycle. As mentioned above, the conventional I-S cycle suffers from low thermal efficiency and highly corrosive streams. To alleviate these problems, we have proposed the optimal operating conditions for the Bunsen reaction and devised a new KAIST flowsheet that produces highly enriched HI through spontaneous L-L phase separation and simple flash processes under low pressure. A series of phase separation experiments were performed to validate the new flowsheet and extend its feasibility. When the molar ratio of $I_2/H_2SO_4$ in the feed increased from 2 to 4, the molar ratio of $HI/(HI+H_2O)$ in the $HI_x$ phase improved from 0.157 to 0.22, which is high enough to gen...