Assessment of Performance of Density Functionals for Predicting Potential Energy Curves in Hydrogen Storage Applications

Abstract

The availability of accurate computational tools for modeling and simulation is vital to accelerate the discovery of materials capable of storing hydrogen (H2) under given parameters of pressure swing and temperature. Previously, we compiled the H2Bind275 data set consisting of equilibrium geometries and assessed the performance of 55 density functionals over this data set (J. Chem. Theory Comput.20201649634982). As it is crucial for computational tools to accurately model the entire potential energy curve (PEC), in addition to the equilibrium geometry, we extended this data set with 389 new data points to include two compressed and three elongated geometries along 78 PECs for H2 binding, forming the H2Bind78 × 7 data set. By assessing the performance of 55 density functionals on this significantly larger and more comprehensive H2Bind78 × 7 data set, we identified the best performing density functionals for H2 binding applications: PBE0-DH, ωB97X-V, ωB97M-V, and DSD-PBEPBE-D3­(BJ). The addition of Hartree–Fock exchange improves the performance of density functionals, albeit not uniformly throughout the PEC. We recommend the usage of ωB97X-V and ωB97M-V density functionals as they offer good performance for both geometries and energies. In addition, we also identified B97M-V and B97M-rV as the best semilocal density functionals for predicting H2 binding energy at its equilibrium geometry.

Quantum chemical modeling of hydrogen binding in metal–organic frameworks: validation, insight, predictions and challenges

“…oB97M-V is the top-performing hybrid density functional in several large assessments, including the MGC84 database, 76 the large and diverse GMTKN55 benchmark dataset, 78,81,82 and the TMC151 transition metal database. 83 These findings are buttressed by an extensive benchmark 52,53 of density functionals for hydrogen storage.…”

“…133 Past endeavours at quantifying the binding of these H 2 adducts in Ca-porphyrin have used the local density approximation (LDA), and the generalized gradient approximation (GGA) with periodic DFT, giving binding estimates of À24.1 kJ mol À1 , and À9.6 kJ mol À1 per hydrogen respectively. In light of the significant difference in predicted binding energies between the LDA and GGA approximations, their limited treatment of dispersion effects, and the improvements that are possible with high-level functionals, 52,53 we employ one such functional here. This oB97M-V functional 79 employs the non-local and highly effective VV10 form 80 for capturing dispersion interactions and is well-suited for quantifying binding at these sites.…”

“…The exploration of hydrogen-binding interactions in MOFs that have been prepared or are possible candidates for preparation is possible with accurate modern density functional theory (DFT). 52,53 As we discuss in the second section, such calculations yield good numerical predictions on the one hand, and, on the other hand, insights into the origin of binding strength via the tools of energy decomposition analysis (EDA). 54,55 The fidelity of the calculations depends upon the model of the MOF: we use cluster calculations that aim to capture all relevant structural and electronic features close to a hydrogen binding site; cluster modeling is also reviewed.…”

Prediction of Multiple Hydrogen Ligation at a Vanadium(II) Site in a Metal–Organic Framework

“…ωB97M-V was the top-performing hybrid density functional in several large assessments, including the MGC84 database, the large and diverse GMTKN55 benchmark data set, − and the TMC151 transition metal database . These findings are buttressed by a recent extensive benchmark , of density functionals for hydrogen storage. Employing ωB97M-V limits calculations to a cluster node, as exact exchange is not feasible for periodic calculations.…”

Prediction of Multiple Hydrogen Ligation at a Vanadium(II) Site in a Metal-Organic Framework

“…ωB97M-V was the top-performing hybrid density functional in several large assessments, including the MGC84 database, 29 the large and diverse GMTKN55 benchmark dataset, [30][31][32] and the TMC151 transition metal database. 33 These findings are buttressed by a recent extensive benchmark 34,35 of density functionals for hydrogen storage. Employing ωB97M-V limits calculations to a cluster node, as exact exchange is not feasible for periodic calculations.…”