Would encoding quantum dispersion forces into interatomic potentials preserve accuracy across different loading rates and temperatures?

# Encoding Quantum Dispersion Forces Across Conditions

Your question asks whether encoding van der Waals interactions into classical interatomic potentials would remain accurate as loading rate and temperature change. This is a subtle materials-modeling question, and the archive offers relevant but partial insight.

The closest work is the DFTB+MBD framework [AG-2024.04-1576], which does exactly this: it encodes quantum dispersion forces (via many-body dispersion, or MBD) into a computationally efficient model for materials mechanics. The key finding is that simplified classical approximations fail to capture quantum effects on mechanical properties — meaning that if you freeze dispersion forces into a single, temperature-independent potential, you risk losing accuracy as conditions change.

Here's why: van der Waals interactions are *not* truly static. At higher temperatures, thermal motion broadens the effective range and strength of dispersion forces. Under fast loading, dynamic effects (phonon excitation, temporal correlations in electron density fluctuations) can modify how these forces manifest. A single interatomic potential fitted at one condition will not account for these rate- and temperature-dependent screening effects.

The MBD model itself is designed to address this by computing dispersion forces from first principles at each atomic configuration [AG-2024.04-1576], so it naturally adapts to changing conditions. But if you *extract* those forces and hard-code them into a classical pair potential, you lose that adaptability — much like freezing electronic response into a rigid structure.

The archive does not contain experimental validation of potential accuracy across wide ranges of loading rate and temperature, so I cannot cite a direct test of your specific scenario.