This work focuses on developing afully transferable model to predict adsorption in Metal-Organic Frameworks (MOFs) with coordinatively unsaturated metal sites (CUS). MOFs are of great interest to the scientific community due to theirlarge porosity, surface area and high degree of tailorability. In particular, MOFs containing CUS have demonstrated highly selective adsorption by forming strong coordination bonds with specific adsorbates, which gives them great potential for challenging gas separations. The caveat of MOF variability is that this introduces an overly deep pool of potential MOFs to be assessed (currently there are ~50,000 MOFs identified in the Cambridge database) through purely experimental means. Furthermore, experimentally measuring competitive binary adsorption is very difficult, and it is therefore often estimated on the basis of single-component adsorption through Ideal Adsorbed Solution Theory (IAST). However, IAST has been shown to be inaccurate for systems involving specific gas-solid interactions, as is the case in MOFs with CUS [0].
As such, there is potential for computational modelling to play a pivotal role in adsorbent material design, e.g., through high-throughput screening of MOFs. However, accurate predictions of adsorption in MOFs by computer simulation require the development of realistic molecular models, which is especially difficult for CUS-containing MOFs. It has recently been demonstrated that conventional “off-the-shelf” molecular models are unable to correctly describe adsorption in CUS-containing MOFs [1].Here we report the latest developments of a new approach that combines accurate quantum mechanical (QM) calculations with classical Monte Carlo simulations to enable accurate predictions of adsorption for these complex materials [2,3]. The new model was first validated against ethylene adsorption in Cu-BTC and showed good agreement with experiment. Crucially, we demonstrate that the QM-derived potential parameters are transferable to different adsorbates of the same type (e.g., ethylene to propylene) as well as between different MOFs with the same type of open metal site (e.g.,Cu-paddlewheel MOFs), as shown for NOTT-101. This has allowed us to predict multi-component adsorption isotherms, and corresponding selectivities, in the context of several challenging gas separations(e.g., ethylene/ethane; propylene/propane; CO/nitrogen) using MOFs with open metal sites. Overall, our new approach offers further insight into the molecular level adsorption mechanisms, from which new promising materials can be designed.
[0] Naomi F. Cessford, Nigel A. Seaton, Tina Düren, Ind. Eng. Chem. Res., 51, 4911-4921, 2012
[1] M. Fischer, J.R. B. Gomes, M.Jorge, Mol. Simul., 40, 537-556, 2014.
[2] M. Fischer, J. R. B. Gomes, M.Froba, M. Jorge, Langmuir. 28, 8537-8549, 2012
[3] Christopher Campbell, Carlos A. Ferreiro-Rangel, Michael Fischer, José R. B.Gomes, Miguel Jorge, J. Phys. Chem. C, 121, 441-458, 2017
Advances in molecular simulation , Carbon capture and other industrial applications
