Ab Initio Flexible Force Field for Metal-Organic Frameworks Using Dummy Model Coordination Bonds

Nucleation of zeolitic imidazolate frameworks: from molecules to nanoparticles

We have studied the clusters involved in the initial stages of nucleation of Zeolitic Imidazolate Frameworks, employing a wide range of computational techniques. In the pre-nucleating solution, the prevalent cluster is the ZnIm₄ cluster (formed by a zinc cation, Zn²⁺, and four imidazolate anions, Im^-), although clusters such as ZnIm₃, Zn₂Im₇, Zn₂Im₇, Zn₃Im₉, Zn₃Im₁₀, or Zn₄Im₁₂ have energies that are not much higher, so they would also be present in solution at appreciable quantities. All these species, except ZnIm₃, have a tetrahedrally coordinated Zn²⁺ cation. Small Zn_xIm_y clusters are less stable than the ZnIm₄ cluster. The first cluster that is found to be more stable than ZnIm₄ is the Zn₄₁Im₈₈ cluster, which is a disordered cluster with glassy structure. Bulk-like clusters do not begin to be more stable than glassy clusters until much larger sizes, since the larger cluster we have studied (Zn₁₄₄Im₂₈₈) is still less stable than the glassy Zn₄₁Im₈₈ cluster, suggesting that Ostwald's rule (the less stable polymorph crystallizes first) could be fulfilled, not for kinetic, but for thermodynamic reasons. Our results suggest that the first clusters formed in the nucleation process would be glassy clusters, which then undergo transformation to any of the various crystal structures possible, depending on the kinetic routes provided by the synthesis conditions. Our study helps elucidate the way in which the various species present in solution interact, leading to nucleation and crystal growth.

Computer simulation of the early stages of self-assembly and thermal decomposition of ZIF-8

We employ all-atom well-tempered metadynamics simulations to study the mechanistic details of both the early stages of nucleation and crystal decomposition for the benchmark metal–organic framework (MOF) ZIF-8. To do so, we developed and validated a force field that reliably models the modes of coordination bonds via a Morse potential functional form and employs cationic and anionic dummy atoms to capture coordination symmetry. We also explored a set of physically relevant collective variables and carefully selected an appropriate subset for our problem at hand. After a rapid increase of the Zn–N connectivity, we observe the evaporation of small clusters in favor of a few large clusters, which leads to the formation of an amorphous highly connected aggregate. [Formula: see text] and [Formula: see text] complexes are observed with lifetimes in the order of a few picoseconds, while larger structures, such as four-, five-, and six-membered rings, have substantially longer lifetimes of a few nanoseconds. The free ligands act as “templating agents” for the formation of sodalite cages. ZIF-8 crystal decomposition results in the formation of a vitreous phase. Our findings contribute to a fundamental understanding of MOF’s synthesis that paves the way to controlling synthesis products. Furthermore, our developed force field and methodology can be applied to model solution processes that require coordination bond reactivity for other ZIFs besides ZIF-8.

Parametrization of Nonbonded Force Field Terms for Metal-Organic Frameworks Using Machine Learning Approach

The enormous structural and chemical diversity of metal-organic frameworks (MOFs) forces researchers to actively use simulation techniques as often as experiments. MOFs are widely known for their outstanding adsorption properties, so a precise description of the host-guest interactions is essential for high-throughput screening aimed at ranking the most promising candidates. However, highly accurate ab initio calculations cannot be routinely applied to model thousands of structures due to the demanding computational costs. Furthermore, methods based on force field (FF) parametrization suffer from low transferability. To resolve this accuracy-efficiency dilemma, we applied a machine learning (ML) approach: extreme gradient boosting. The trained models reproduced the atom-in-material quantities, including partial charges, polarizabilities, dispersion coefficients, quantum Drude oscillator, and electron cloud parameters, with accuracy similar to the reference data set. The aforementioned FF precursors make it possible to thoroughly describe noncovalent interactions typical for MOF-adsorbate systems: electrostatic, dispersion, polarization, and short-range repulsion. The presented approach can also readily facilitate hybrid atomistic simulation/ML workflows.

Atomistic Insight Into the Host-Guest Interaction of a Photoresponsive Metal-Organic Framework

Photoresponsive functional materials have gained increasing attention due to their externally tunable properties. Molecular switches embedded in these materials enable the control of phenomena at the atomic level by light. Metal-organic frameworks (MOFs) provide a versatile platform to immobilize these photoresponsive units within defined molecular environments to optimize the intended functionality. For the application of these photoresponsive MOFs (pho-MOFs), it is crucial to understand the influence of the switching state on the host-guest interaction. Therefore, we present a detailed insight into the impact of molecular switching on the intermolecular interactions. By performing atomistic simulations, we revealed that due to different interactions of the guest molecules with the two isomeric states of an azobenzene-functionalized MOF, both the adsorption sites and the orientation of the molecules within the pores are modulated. By shedding light on the host-guest interaction, our study highlights the unique potential of pho-MOFs to tailor molecular interaction by light.