Adsorbents and adsorption models for capture of Kr and Xe gas mixtures in fixed-bed columns

• Published in: Chemical engineering journal (Lausanne, Switzerland : 1996) Vol. 375; no. C
• Main Authors: Ladshaw, Austin P., Wiechert, Alexander I., Welty, Amy K., Lyon, Kevin L., Law, Jack D., Jubin, Robert T., Tsouris, Costas, Yiacoumi, Sotira
• Format: Journal Article
• Language: English
• Published: United States Elsevier 25-06-2019
• Subjects: adsorption; discontinuous Galerkin; krypton; modeling; MOOSE; NUCLEAR FUEL CYCLE AND FUEL MATERIALS; OSPREY; xenon
• ISSN: 1385-8947; 1873-3212
• DOI: 10.1016/j.cej.2019.122073

Off-gases produced during the reprocessing of used nuclear fuel (UNF) include 129-I2, 3-HHO, 14-CO2, 85-Kr, and 135-Xe, which are volatilized out into the off-gas. In order to meet regulatory requirements for reprocessing plant emissions, these gases must be captured and removed from the o?-gas stream prior to o?-gas emission. Of particular interest are the noble gases, Kr and Xe, which can be fairly difficult to remove from the off-gas due to their low chemical reactivity. Thus, this work is focused on utilizing engineered adsorbents, AgZ-PAN and HZ-PAN, to capture Kr and Xe from a mixed-gas stream at relatively low temperatures (191–295 K) and various flow rates (50–2000 mL/min). Isothermal data for Kr and Xe on each adsorbent are analyzed to produce the Langmuir parameters needed to model the mixture adsorption capacities at relevant temperatures using the Extended Langmuir model. Those parameters are then incorporated into a fixed-bed adsorption model developed in this work using the Mulitphysics Object-Oriented Simulation Environment (MOOSE). That model is used to simulate breakthrough times for Kr and Xe in packed columns of AgZ-PAN and HZ-PAN, ranging in length from 6 to 20 in., at relevant temperatures and flow rates. Breakthrough times varied from nearly instantaneous for Kr in AgZ-PAN to 30 h for Xe in HZ-PAN. After the developed model was validated by comparisons with experimental break-through data, the model framework was used to simulate the performance of multiple fixed-bed columns connected in series.