Treating London-Dispersion Effects with the Latest Minnesota Density Functionals: Problems and Possible Solutions

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Accurate description of London-dispersion interactions is crucial for predicting molecular properties.
• Many modern density functionals struggle to capture these long-range van der Waals forces.
• Existing empirical corrections may introduce artifacts like double-counting.

Purpose of the Study:

• To evaluate the performance of recent Minnesota density functionals regarding London-dispersion interactions.
• To assess the effectiveness of common correction methods like DFT-D3 and VV10.
• To identify inherent limitations in current density functional approximations for dispersion forces.

Main Methods:

• Systematic testing of Minnesota density functionals (SOGGA11, M11-L, N12, MN12-L, SOGGA11-X, M11, N12-SX, MN12-SX).
• Application of Grimme's DFT-D3 correction and the VV10 van der Waals kernel.
• Analysis of dissociation curves, particularly for the argon dimer.

Main Results:

• Minnesota functionals exhibit deficiencies in describing London-dispersion interactions.
• DFT-D3 and VV10 corrections offer partial improvements but can lead to double-counting of electron correlation.
• Unphysical dissociation curves for the argon dimer highlight fundamental issues with these functionals.
• Empirical parameter optimization alone is insufficient without addressing long-range behavior.

Conclusions:

• Current density functional approximations, even recent ones, require careful consideration for studies involving London-dispersion.
• The problem of London-dispersion is unlikely to be solved solely by empirical parameterization.
• Future development should focus on incorporating correct asymptotic long-range behavior into functionals.
• Chemists must be aware of these limitations when selecting computational methods.