Proton Transport in Perfluorinated Ionomer Simulated by Machine-Learned Interatomic Potential.

Area of Science:

• Materials Science
• Computational Chemistry
• Polymer Science

Background:

• First-principles methods present significant challenges for accurately modeling polymer materials.
• Perfluorinated ionomers are crucial in various applications but are complex to simulate.
• Understanding their structural and dynamical properties is essential for material design.

Purpose of the Study:

• To apply machine-learned interatomic potentials for simulating perfluorinated ionomers.
• To develop an efficient and transferable model for amorphous polymers.
• To investigate proton and water diffusion in dry and hydrated states.

Main Methods:

• Utilized an improved active-learning algorithm with minimal descriptors to build a machine-learned potential.
• Performed molecular dynamics simulations accelerated by the developed potentials.
• Analyzed the formation of hydrophilic/hydrophobic domains and diffusion coefficients under various humidity conditions.

Main Results:

• Accurately reproduced heterogeneous hydrophilic and hydrophobic domains in the polymer.
• Successfully predicted proton and water diffusion coefficients across different humidity levels.
• Identified significant contributions of short Grotthuss chains (2-3 water molecules) to proton mobility.

Conclusions:

• Machine-learned potentials offer an efficient and accurate approach for simulating complex polymer systems like perfluorinated ionomers.
• The study elucidates the mechanisms of proton transport, highlighting the role of water molecule clusters.
• Findings provide valuable insights for designing advanced ionomer materials with tailored transport properties.