Comparing the Expense and Accuracy of Methods to Simulate Atomic Vibrations in Rubrene

Area of Science:

• Materials Science
• Computational Chemistry
• Condensed Matter Physics

Background:

• Atomic vibrations are crucial for understanding material properties, influencing phenomena like charge transport and structural disorder.
• Predicting atomic vibrations via simulations is essential, but existing methods present challenges in balancing accuracy and computational cost.

Purpose of the Study:

• To evaluate and compare the accuracy-efficiency trade-offs of six distinct simulation methods for predicting atomic vibrational properties.
• To identify the most effective simulation approach for analyzing inelastic neutron scattering (INS) spectra using rubrene as a model system.

Main Methods:

• Six simulation techniques were employed: Density Functional Theory (DFT), Density Functional Tight Binding (DFTB), Chebyshev-corrected DFTB, a trained machine learning (ML) model, a pre-trained ML model (ANI-1), and a classical force field.
• The accuracy of each method was rigorously assessed by comparing its predictions against experimental inelastic neutron scattering (INS) data.

Main Results:

• All investigated methods demonstrated a degree of accuracy in predicting atomic vibrational properties across a broad energy range.
• The Chebyshev-corrected tight-binding method emerged as the optimal approach, delivering a superior combination of high accuracy and computational efficiency.
• The study provides practical guidelines for selecting simulation methods to achieve efficient and accurate INS spectrum predictions.

Conclusions:

• The Chebyshev-corrected DFTB method represents a highly effective tool for accurate and efficient prediction of atomic vibrational spectra.
• The findings offer valuable insights for researchers seeking to optimize computational strategies for materials characterization through vibrational spectroscopy.