(408h) Modeling Hysteretic Isotherms and Parameter Estimation for Pressure Swing Adsorption Process Using Gate-Opening Metal-Organic Frameworks

Improvement in the performance of adsorbent may reduce the cost of CO₂ capture substantially. Metal-Organic Frameworks (MOFs) have been gaining attention in recent years, which are hybrid crystalline porous materials that consist of a regular array of positively charged metal ions surrounded by organic molecules. Among various MOF structures reported in literature, we investigate ELM-11, which has a sigmoidal isotherm shape. This material can captures and releases CO₂ with a small change of pressure^[1], which would be a substantial advantage for applications in Pressure Swing Adsorption (PSA).

PSA process using ELM-11 has not been investigated in the past partly because of the difficulty of modeling the complex hysteresis of the isotherms due to the “gate adsorption” mechanism; the structure of ELM-11 changes as CO₂ molecules are captured in the framework. Figure 1 shows isotherms of ELM-11. By increasing the partial pressure of CO₂ from the state without CO₂ molecules in the framework, the space between layers of ELM-11 opens (gate opening) and the CO₂ molecules are captured in the space. This mechanism is reversed in desorption; when the partial pressure of CO₂ around ELM-11 is decreased, the captured CO₂ molecules are released, which leads to the closure of the space between the layers (gate closing). Note that there is a gap between the two trajectories of the adsorption and desorption isotherms.

In this research, we model the isotherms and estimate the parameters in the model from experimental data. We employ the Markov chain Monte Carlo (MCMC) method, a sampling algorithm for solving Bayesian estimation. This approach provides not only the estimate of parameter values but also the uncertainty in the estimates as probability distributions. The isotherm with the obtained parameters, which is a hybrid model, is incorporated into a PSA model to estimate the process performance.

Reference

[1] Tanaka, H. et al, 2015. J. Phys. Chem. C 119, 11533–11543.