Related Experiment Video
Updated: Jun 20, 2026
07:31
Realistic Membrane Modeling Using Complex Lipid Mixtures in Simulation Studies
Published on: September 1, 2023
How Long Is Long Enough? Extrapolation of Machine-Learning Interatomic Potentials for Oligomeric and Polymeric
Natalie E Hooven1, Arthur Y Lin1, Charles H Carroll1
1Department of Chemical and Biological Engineering, University of Wisconsin-Madison, Madison, Wisconsin 53706, United States.
Journal of Chemical Theory and Computation
|June 19, 2026
Summary
Machine-learning interatomic potentials (MLIPs) can accurately simulate larger molecules if trained on smaller systems where local chemical environments match. Careful neighbor list construction is key for predicting polymer behavior using these transferable MLIPs.
Area of Science:
- Computational Chemistry
- Materials Science
- Machine Learning
Background:
- Machine-learning interatomic potentials (MLIPs) offer high-fidelity molecular simulations at larger scales.
- Training MLIPs for large molecules (polymers, biomolecules) is challenging due to the difficulty of obtaining ab initio data.
- Smaller, analogous chemical systems are often used to train MLIPs for complex molecular architectures.
Purpose of the Study:
- To investigate the conditions under which MLIPs trained on small molecules can accurately extrapolate to longer molecular chains and complex architectures.
- To establish criteria for developing transferable MLIPs for macromolecular materials.
- To provide a data-driven approach for designing effective MLIPs.
Main Methods:
- A control study using linear alkanes (n=1-8) to assess MLIP extrapolation capabilities.
- Analysis of MLIP performance combined with environment-resolved Smooth Overlap of Atomic Positions (SOAP) descriptors.
- Application of Principal Covariates Classification (PCC) to identify key chemical environments.
- Investigation of neighbor list construction's impact on intermolecular energetics.
Main Results:
- Reliable extrapolation of MLIPs to longer chains is achieved when local chemical environments in training and target systems converge.
- The accuracy of MLIPs for polymeric behavior is critically dependent on the accurate learning of intermolecular energetics.
- Optimized neighbor list construction significantly enhances the learnability of intermolecular interactions.
Conclusions:
- Transferable MLIPs for macromolecular materials can be designed by ensuring convergence of local chemical environments between training and target systems.
- Neighbor list optimization is crucial for accurately capturing intermolecular energetics, essential for simulating polymer properties.
- This study provides a practical framework for developing robust MLIPs for large-scale molecular simulations.
Related Concept Videos
01:27¹H NMR: Long-Range Coupling
The coupling interactions of nuclei across four or more bonds are usually weak, with J values less than 1 Hz. While these are usually not observed in spectra, the presence of multiple bonds along the coupling pathway can result in observable long-range coupling.
In alkenes, spin information is communicated via σ–π overlap, as seen in allylic (four-bond) and homoallylic (five-bond) couplings. These coupling interactions are stronger when the σ bond is parallel to the alkene π orbitals.
In alkenes, spin information is communicated via σ–π overlap, as seen in allylic (four-bond) and homoallylic (five-bond) couplings. These coupling interactions are stronger when the σ bond is parallel to the alkene π orbitals.
01:10Polymers: Molecular Weight Distribution
For any given polymer, the weight average molecular weight (Mw) is higher than, if not equal to, the number average molecular weight (Mn). The only situation in which the weight average molecular weight and the number average molecular weight are equal is when a polymer consists only of chains with equal molecular weight. However, this never happens in a synthetic polymer, since it is difficult to control the polymerization process up to a molecular level with accuracy to a hundred percent.
02:56Intermolecular Forces and Physical Properties
02:00Molecular Models
Physical models representing molecular architectures of chemical compounds play essential roles in understanding chemistry. The use of molecular models makes it easier to visualize the structures and shapes of atoms and molecules.
03:51Molecular Orbital Theory II
Molecular Orbital Energy Diagrams
02:40MO Theory and Covalent Bonding
The molecular orbital theory describes the distribution of electrons in molecules in a manner similar to the distribution of electrons in atomic orbitals. The region of space in which a valence electron in a molecule is likely to be found is called a molecular orbital. Mathematically, the linear combination of atomic orbitals (LCAO) generates molecular orbitals. Combinations of in-phase atomic orbital wave functions result in regions with a high probability of electron density, while...