Ubiquitous Negative Electron Densities Discredit Smooth Energy Interpolation in Density Functional Theory

Area of Science:

• Quantum Chemistry
• Theoretical Chemistry
• Computational Chemistry

Background:

• Accurate prediction of molecular properties is crucial in chemistry.
• Interpolation methods are used to estimate properties between known points.
• Understanding electron density behavior is key to molecular property prediction.

Purpose of the Study:

• To develop a unified theoretical framework for interpolating molecular properties.
• To investigate the validity of different interpolation models for charge transfer.
• To identify the most physically realistic interpolation method for molecular systems.

Main Methods:

• Developed a unified theoretical framework for property interpolation.
• Analyzed electron density behavior under various interpolation models.
• Examined the implications of parabolic and linear interpolation models.

Main Results:

• Smooth interpolatory models, including quadratic ones, predict negative electron densities in certain spatial regions and electron counts.
• Parabolic interpolation leads to unphysical negative electron populations in specific cases (e.g., HF^+0.5).
• Linear (zero-temperature grand canonical) interpolation is identified as the most physically sensible model for charge transfer in most scenarios.

Conclusions:

• Non-linear interpolation models for charge transfer can produce physically unrealistic electron density distributions.
• Linear interpolation provides a more robust and physically grounded approach for estimating molecular properties.
• The choice of interpolation model significantly impacts the accuracy and interpretability of predicted molecular properties.