Ab initio instanton rate theory made efficient using Gaussian process regression.

Area of Science:

• Quantum Chemistry
• Chemical Kinetics
• Computational Methods

Background:

• Ab initio instanton rate theory accurately models tunnelling effects in chemical reactions.
• This method is computationally intensive, requiring numerous electronic-structure calculations.
• Extending conventional transition-state theory to deep-tunnelling regimes necessitates advanced computational approaches.

Purpose of the Study:

• To develop a more computationally feasible method for calculating chemical reaction rates.
• To accurately incorporate tunnelling effects using fewer ab initio calculations.
• To enable practical application of high-level electronic-structure methods for rate constant determination.

Main Methods:

• Gaussian process regression is employed to locally approximate the potential-energy surface.
• The method focuses on fitting the surface around the dominant tunnelling pathway.
• This approach is validated against on-the-fly ab initio instanton calculations.

Main Results:

• The proposed method achieves convergence to results comparable to on-the-fly calculations.
• It significantly reduces the computational cost by requiring fewer electronic-structure computations.
• Benchmark calculations for H + CH4 show fast convergence.
• New low-temperature rates for H + C2H6 in full dimensionality were evaluated.

Conclusions:

• The Gaussian process regression approach offers a practical and accurate alternative for calculating reaction rates.
• This method makes high-level electronic-structure calculations more accessible for complex chemical systems.
• The study provides accurate, low-temperature rate data for H + C2H6.