(539h) Modeling the Adsorption Behavior in Irmofs Using Monte Carlo Simulations

Metal-organic frameworks are porous materials built from inorganic nodes connected by organic linkers that are increasingly being used for applications in gas adsorption and storage owing to their high porosities and crystallinity. Isoreticular MOFs, a series of MOFs existing in the same framework topology but differing in the substituent groups on the linker installed during formation of the framework or through post-synthetic modifications, have shown promise to tailor the guest-host interactions, thereby altering the adsorption behavior of the guest molecules in the MOFs.¹ The large number of MOFs resulting from this tunability can be investigated computationally for evaluating their effectiveness for a desired application. The structural flexibility of these materials due to the relatively weaker coordination links between the inorganic and organic units, however, allows for structural deformations of the framework in the presence of guest molecules, thereby making it challenging to model the adsorption behavior of these materials.² Conventional grand canonical Monte Carlo (GCMC) adsorption simulations treat the framework as rigid, thereby limiting their ability to model the flexibility of MOFs in the presence of guest molecules. Recent efforts to address this challenge either employ complex hybrid (Monte Carlo + molecular dynamics) simulation protocols² or use GCMC simulations in the “osmotic” statistical ensemble which may require apriori knowledge of the different deformed configurations of the framework.^3,4

In this work, we have employed force-field-based Monte Carlo simulations in the N_guestN_hostPT-Gibbs ensemble to investigate the adsorption of N,N-dimethylformamide (DMF), para-xylene, and ortho-xylene in the classic IRMOF-1 (MOF-5) and its linker-functionalized isoreticular analogue, IRMOF-3, which are both modelled as flexible MOFs. In this ensemble, only guest molecules are allowed to transfer between reservoir phase and MOF phase, the volume of the reservoir phase can change only isotropically, whereas changes in individual elements of the H matrix allow for both volume and shape fluctuations for the MOF phase. MOF-5 is built from inorganic Zn₄O nodes connected by 1,4-benzenedicarboxylate (BDC) linkers while the inorganic nodes in IRMOF-3 are coordinated by 2-amino-1,4-benzenedicarboxylate (NH₂-BDC) linkers. We observe lower onset pressures for sorbate uptake in IRMOF-3 as compared to MOF-5 owing to favorable interactions of the sorbates with the linker amine groups in IRMOF-3 as evidenced from their respective radial and spatial distribution plots. The isotherms for both p-xylene and o-xylene adsorption in MOF-5 and IRMOF-3 at 295 K indicate a “pore-filling” adsorption mechanism in these MOFs with negligible framework deformation. Remarkably, these isoreticular MOFs exhibit different DMF adsorption characteristics. While the isotherm for DMF adsorption in MOF-5 at 295 K shows a steep increase in the DMF uptake corresponding to a “pore-filling” adsorption mechanism, the DMF adsorption isotherm in IRMOF-3 at 295 K shows a gradual increase in the DMF uptake which is significantly influenced by the framework deformation. Indeed, at P/P_o = 0.001, we observe an approximately 30% reduction in the unit cell volume and a change in the unit cell shape from cubic to triclinic in order to maximize the DMF uptake at this intermediate pressure. Moreover, our TraPPE-based “flexible” force field employed for the IRMOFs in this work along with a new TraPPE force field proposed for DMF is able to capture the previously reported dynamic binding of DMF to the Zn²⁺ ions of the node thereby leading to an increased coordination number of the Zn²⁺ ions in the MOF. Interestingly, we also observe a bimodal distribution of the sorbate loading in individual pores of the MOF, thereby pointing to the existence of two types of pores in the MOF. This behavior is attributed to preferred linker orientations identified in the MOF resulting in adjacent pores of differing pore volume and hence, different sorbate loadings.

References:

(1) Babarao, R.; Jiang, J. Molecular Screening of Metal−Organic Frameworks for CO₂ Storage. Langmuir 2008, 24 (12), 6270–6278. https://doi.org/10.1021/la800369s.

(2) Rogge, S. M. J.; Goeminne, R.; Demuynck, R.; Gutiérrez‐Sevillano, J. J.; Vandenbrande, S.; Vanduyfhuys, L.; Waroquier, M.; Verstraelen, T.; Van Speybroeck, V. Modeling Gas Adsorption in Flexible Metal–Organic Frameworks via Hybrid Monte Carlo/Molecular Dynamics Schemes. Adv. Theory Simul. 2019, 2 (4), 1800177. https://doi.org/10.1002/adts.201800177.

(3) Jeffroy, M.; Fuchs, A. H.; Boutin, A. Structural Changes in Nanoporous Solids Due to Fluid Adsorption: Thermodynamic Analysis and Monte Carlo Simulations. Chem. Commun. 2008, No. 28, 3275. https://doi.org/10.1039/b805117h.

(4) Zhao, X. S.; Siepmann, J. I.; Xu, W.; Kiang, Y.-H.; Sheth, A. R.; Karaborni, S. Exploring the Formation of Multiple Layer Hydrates for a Complex Pharmaceutical Compound. J. Phys. Chem. B 2009, 113 (17), 5929–5937. https://doi.org/10.1021/jp808164t.