New Permutationally Invariant Polynomial Potential Energy Surfaces for H₅O₂⁺ with Fast Analytical Gradients...

Area of Science:

• Computational chemistry
• Quantum chemistry
• Molecular dynamics

Background:

• The protonated water dimer is crucial for understanding the hydrated proton.
• Previous computational models had limitations in precision and gradient calculation speed.

Purpose of the Study:

• To develop improved potential energy surfaces (PESs) for the protonated water dimer.
• To enable faster and more accurate gradient calculations for simulations.

Main Methods:

• Linear regression with permutationally invariant polynomials (PIPs).
• Reverse differentiation for efficient gradient computation.
• Fitting to high-level CCSD(T) data up to 110,000 cm⁻¹.

Main Results:

• New PESs provide more precise fits to CCSD(T) data compared to previous models.
• Fast gradients are achieved via reverse differentiation, significantly outperforming numerical gradients.
• The new surfaces show good agreement with CCSD(T) benchmarks and Diffusion Monte Carlo zero-point energies.

Conclusions:

• The developed PESs offer a significant advancement for studying the hydrated proton.
• Enhanced computational efficiency through fast gradients will facilitate more extensive molecular simulations.
• These improved models contribute to a deeper understanding of proton transfer in aqueous systems.