High accuracy barrier heights, enthalpies, and rate coefficients for chemical reactions

Abstract

Quantitative chemical reaction data, including activation energies and reaction rates, are crucial for developing detailed kinetic mechanisms and accurately predicting reaction outcomes. However, such data are often difficult to find, and high-quality datasets are especially rare. Here, we use CCSD(T)-F12a/cc-pVDZ-F12//ωB97X-D3/def2-TZVP to obtain high-quality single point calculations for nearly 22,000 unique stable species and transition states. We report the results from these quantum chemistry calculations and extract the barrier heights and reaction enthalpies to create a kinetics dataset of nearly 12,000 gas-phase reactions. These reactions involve H, C, N, and O, contain up to seven heavy atoms, and have cleaned atom-mapped SMILES. Our higher-accuracy coupled-cluster barrier heights differ significantly (RMSE of ∼5 kcal mol−1) relative to those calculated at ωB97X-D3/def2-TZVP. We also report accurate transition state theory rate coefficients $${k}_{\infty }(T)$$ k ∞ ( T ) between 300 K and 2000 K and the corresponding Arrhenius parameters for a subset of rigid reactions. We believe this data will accelerate development of automated and reliable methods for quantitative reaction prediction.

Fast Predictions of Reaction Barrier Heights: Toward Coupled-Cluster Accuracy

“…To train our model, we leverage a recently refined gas-phase data set of elementary reactions with atom-mapped SMILES. , These data span a diverse set of reactions, whose neutral molecules involve carbon, hydrogen, nitrogen, and oxygen and contain up to seven heavy atoms. Reactions are available at three levels of theory: B97-D3/def2-mSVP, ωB97X-D3/def2-TZVP, and CCSD­(T)-F12a/cc-pVDZ-F12//ωB97X-D3/def2-TZVP.…”

“…These energies used scaled vibrational frequencies to account for anharmonic effects. , The barrier heights Δ E 0 were calculated by subtracting the resulting TS and reactant energies. Reaction enthalpies were calculated by subtracting the resulting product and reactant energies, which include bond additivity corrections as described in ref . These reactions largely overlap with those used to train the model in ref .…”

“…Despite all of these advances, calculations of rate coefficients are still limited by the accurate yet expensive quantum methods used to find the saddle point geometry and the poor scaling of the even more expensive methods used to compute its energy. For example, coupled-cluster CCSD­(T)-F12 calculations are commonly considered the gold-standard in quantum chemistry due to their reliable single point energies and reaction energies, − yet they scale as O ( N 7 ), where N is the number of orbitals . For some reactions, even CCSD­(T) is not accurate enough, and more expensive calculations are required …”

“…However, the great majority of the species in their data set have rotatable bonds and a large number of low-energy conformers, and many of the “product” structures and vibrational frequencies they used were actually properties of van der Waals complexes of two or three product molecules. Partition functions are drastically affected by conformers and rotors, − which is why the refined data set from ref published high-pressure limit TST rate coefficients for only a small subset of the reactionsthose where the TS and the reactant(s) were rigid so there was only one low-lying conformer. Furthermore, in the literature, no analysis was provided for how sensitive the model predictions are to the input 3D geometry; as shown in our results section below, this could be a significant problem.…”

Concluding remarks: Faraday Discussion on unimolecular reactions

“…Please do not adjust margins Please do not adjust margins But fortunately many important reactions can be addressed using methods based on CCSD(T). With modern supercomputers, it is now possible to perform CCSD(T) calculations on thousands of reactions [12], so many that one can use them as training data for machine learning [13], to construct fast estimators for barrier heights. These fast estimators are not 100% reliable, for an histogram of the deviations between our fast estimator and the actual CCSD(T) barrier heights see Fig.…”