How accurate are approximate quantum chemical methods at modelling solute-solvent interactions in solvated clusters?

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Molecular Modeling

Background:

• Accurate calculation of solute-water interaction energies is crucial for understanding solvation.
• Assessing the performance of various Density Functional Theory (DFT) methods is necessary for reliable predictions.

Purpose of the Study:

• To evaluate a wide range of DFT methods for calculating interaction energies of thermal clusters of glycine in water.
• To identify the most accurate and efficient DFT functionals for solvation studies.

Main Methods:

• Molecular dynamics simulations to sample glycine-water clusters (1-40 water molecules) in different charge states.
• Benchmarking DFT methods against high-level ab initio composite methods (W1X-2, DLPNO-CCSD(T)/CBS).
• Evaluation of 16 DFT functionals from various rungs of Jacob's Ladder.

Main Results:

• DLPNO-CCSD(T)/CBS and DSD-PBEP86 methods provided accurate interaction energies.
• ωB97M-V and ωB97X-V functionals showed good performance for large clusters (MAD ∼4 kJ mol⁻¹).
• B3LYP-D3(BJ) exhibited the best performance with the lowest mean absolute deviation (MAD ∼1.7 kJ mol⁻¹).

Conclusions:

• B3LYP-D3(BJ) is the recommended DFT functional for accurate glycine-water interaction energy calculations.
• Relative interaction energy calculations can mitigate systematic errors.
• The ONIOM approximation offers a viable strategy for accelerating accurate calculations.