Failure of the DFT Ladder for the Fulminic Acid Challenge

Ashley M. Allen,Laura N. Olive Dornshuld,Patricia A. Gonzalez Franco,Wesley D. Allen,Henry F. Schaefer III
DOI: https://doi.org/10.26434/chemrxiv-2024-9vzls-v2
2024-10-21
Abstract:Properties of the historically pivotal fulminic acid (HCNO) molecule have been computed with a panoply of 473 density functionals of all varieties, providing a snapshot of the performance of contemporary density functional theory (DFT) for a challenging chemical system. Exhaustive tabulations and statistical analyses have been carried out for geometric parameters, vibrational frequencies, barriers to linearity, and the HCN–O dissociation energy. As the DFT ladder is climbed, only befuddlement ensues as to whether HCNO has a linear or bent equilibrium structure. The distinctive, extremely flat H–C–N bending potential of fulminic acid is almost universally misrepresented by the DFT functionals, which are thus incapable of solving quasilinear/quasibent issues of this molecule. High-ranking DFT functionals produce catastrophic errors in the HCN + O(<sup>3</sup><i>P</i>) → HCNO reaction energy, and lower rungs emerge as the best performers for many of the bond distances and harmonic vibrational frequencies. This research shows that the current DFT zoo of approximations is incapable of yielding any consensus or convergence on the equilibrium structure, H–C–N bending frequencies, or enthalpy of formation of HCNO. Additional analyses are performed on the side effects of popular dispersion corrections on the covalently bonded properties and thermochemistry of HCNO.
Chemistry
What problem does this paper attempt to address?
The problem this paper attempts to address is the evaluation of the performance of current Density Functional Theory (DFT) methods in handling the challenging chemical system of fulminic acid (HCNO). Specifically, the paper calculates the geometric parameters, vibrational frequencies, linear barriers, and HCN-O dissociation energies of fulminic acid using 473 different types of density functionals to test whether these methods can converge or agree on key properties such as the equilibrium structure of fulminic acid, the H-C-N bending frequency, or the enthalpy of formation. The study finds that as the level of DFT increases, the results for whether fulminic acid has a linear or bent equilibrium structure become very chaotic across different density functionals, especially in the description of the H-C-N bending potential, which is almost universally inaccurate. Additionally, advanced DFT functionals show significant errors in the reaction energy for HCN + O(3P) → HCNO, while lower-level methods perform better in some bond lengths and harmonic vibrational frequencies. This indicates that existing DFT approximation methods are still unable to provide reliable and consistent results when dealing with complex molecules like fulminic acid.