Related Experiment Video
Updated: Jun 17, 2026
07:24
Quantitative Atomic-Site Analysis of Functional Dopants/Point Defects in Crystalline Materials by Electron-Channeling-Enhanced Microanalysis
Published on: May 10, 2021
Digital discovery of large-scale optoelectronic materials via MPNICE machine-learning force fields
Hadi Abroshan1, H Shaun Kwak1, David J Giesen1
1Materials Science, Schrödinger, Inc., 1540 Broadway, 24th Floor, New York, NY, 10036, USA. hadi.abroshan@schrodinger.com.
Physical Chemistry Chemical Physics : PCCP
|June 16, 2026
Summary
Machine-learning force fields (MLFFs) using MPNICE accurately predict OLED material properties. This accelerates the design and optimization of next-generation optoelectronic devices by enabling rapid simulations with near quantum-mechanical accuracy.
Area of Science:
- Materials Science
- Computational Chemistry
- Organic Electronics
Background:
- Accurate prediction of structure-property relationships in organic light-emitting diode (OLED) materials is crucial for designing efficient devices.
- Existing computational methods struggle to efficiently capture conformational flexibility, dynamic disorder, and diverse bonding environments.
Purpose of the Study:
- To demonstrate the application of machine-learning force fields (MLFFs) based on a message passing network with iterative charge equilibration (MPNICE) for predictive modeling of OLED materials.
- To enable rapid geometry optimization and molecular dynamics simulations with near quantum-mechanical accuracy for OLED materials.
Main Methods:
- Developed and applied MLFFs based on MPNICE, trained on high-level electronic-structure data.
- Incorporated iterative charge equilibration and long-range electrostatics into the MLFF framework.
- Validated the approach through geometry optimization, molecular dynamics simulations, isomerization and conformational analyses, and excited-state calculations.
Main Results:
- MPNICE-optimized geometries closely reproduced density functional theory (DFT) reference structures for diverse OLED materials, reducing computational cost by orders of magnitude.
- MPNICE-based sampling accurately captured relative energetics and conformational landscapes.
- MLFF-driven molecular dynamics simulations provided insights into the impact of dynamics on optoelectronic response, with calculated spectra agreeing well with experimental data.
Conclusions:
- MPNICE-based MLFFs provide a scalable and physically grounded platform for data-driven OLED materials design.
- This approach accelerates the optimization of next-generation optoelectronic systems by enabling high-throughput screening and rapid property evaluation.
