Adsorption of CO₂, CH₄, and N₂ on Zeolitic Imidazolate Frameworks: Experiments and Simulations

Abstract

Experimental measurements and molecular simulations were conducted for two zeolitic imidazolate frameworks, ZIF-8 and ZIF-76. The transferability of the force field was tested by comparing molecular simulation results of gas adsorption with experimental data available in the literature for other ZIF materials (ZIF-69). Owing to the good agreement observed between simulation and experimental data, the simulation results can be used to identify preferential adsorption sites, which are located close to the organic linkers. Topological mapping of the potential-energy surfaces makes it possible to relate the preferential adsorption sites, Henry constant, and isosteric heats of adsorption at zero coverage to the nature of the host-guest interactions and the chemical nature of the organic linker. The role played by the topology of the solid and the organic linkers, instead of the metal sites, upon gas adsorption on zeolite-like metal-organic frameworks is discussed.

The right isotherms for the right reasons? Validation of generic force fields for prediction of methane adsorption in metal-organic frameworks

“…Yang and Zhong [50] reported an over-estimation in the cases of CH4 on IRMOF-1 and Cu-BTC using OPLS-AA and suggested a re-parameterisation of the C and O atoms of the carboxylate groups, reducing εii by as much as 30%. A similar reduction in εii across all framework atoms was suggested for UFF by Fairen-Jimenez et al [51] and Pérez-Peritello et al [23] (who scaled εUFF by 0.59 and 0.69 respectively) based on simulations of methane adsorption in ZIF-8 and ZIF-69, while a similarly large over-estimation (~50%) has been reported in UiO-66(Zr) using UFF [22]. We would suggest that while UFF and OPLS-AA are suitable for qualitative prediction of adsorption isotherms in MOFs, both are likely to over-estimate the amount adsorbed, as well as low-coverage properties such as heats of adsorption or Henry's constants, by a significant amount.…”

“… While OPLS-AA performed well in predictions of gas-ligand binding, it is likely to over-estimate gas adsorption in MOFs which use carboxylic acid as a coordinating group and did not offer a significant improvement over UFF or DREIDING in the prediction of adsorption isotherms.  DREIDING offered the best performance of the three tested force fields in the prediction of adsorption isotherms in all the systems considered in this study, and in literature studies in which more than one force field was evaluated [23,48,57]. It also performed very well in the prediction of gas-ligand interactions and, as it considers each element to have only one set of LJ parameters, has the advantage of being easily implemented in simulations.…”

“…Since then, GCMC has been employed to predict CH4 adsorption isotherms in a wide range of MOFs and related structures [14,15,16,17,18,19,20,21,22,23], with reasonable agreement with experiment reported in many cases.…”

“…It is possible that the scaling factors suggested by Fairen-Jimenez [51] and Pérez-Peritello [23] may satisfactorily address this shortcoming, although this was not explored in the present work.  While OPLS-AA performed well in predictions of gas-ligand binding, it is likely to over-estimate gas adsorption in MOFs which use carboxylic acid as a coordinating group and did not offer a significant improvement over UFF or DREIDING in the prediction of adsorption isotherms.…”

“…Although DREIDING has been shown to significantly over-estimate methane adsorption when compared to experiment in ZIF-8 [23], very good agreement was observed in the case of UiO-66(Zr) [22] (~9-20% over-estimation for θ < 0.5) and in the present work, while simulations in IRMOF-1 and IRMOF-6 [12] and in IRMOF-8 [48] found excellent quantitative agreement with experiment (5-10% difference when averaged over the full isotherm). While there are too many variables to fully rationalise these apparent differences in quantitative agreement (e.g.…”