Deep Neural Network Enhanced Mesoscopic Thermodynamic Model for Unlocking the Electrode/Electrolyte Interface

Area of Science:

• Electrochemistry
• Computational Materials Science
• Chemical Engineering

Background:

• The electrode/electrolyte interface is crucial for energy storage and conversion devices.
• Accurately modeling this interface, considering both molecular and external field effects, remains a challenge.

Purpose of the Study:

• To develop a computational model for accurately describing electrolyte interface properties.
• To enhance the efficiency of thermodynamic calculations for electrolytes.

Main Methods:

• Development of a mesoscopic thermodynamic model based on chemical potential.
• Integration of a deep neural network to enhance computational efficiency (DeepMT model).
• Statistical analysis of ion density distributions under complex conditions.

Main Results:

• The DeepMT model bridges micro-level ion characteristics and macro-level external field effects.
• Achieved a computational efficiency improvement of approximately four orders of magnitude compared to direct theoretical calculations.
• Accurately predicted key interface properties: ion adsorption, surface charge, and differential capacitance.

Conclusions:

• The DeepMT model offers a computationally efficient and accurate approach for studying electrode/electrolyte interfaces.
• This model enables precise prediction of ion behavior and interface properties in energy storage and conversion systems.
• Facilitates advancements in the design and optimization of electrochemical devices.