Machine learning-guided design and development of multifunctional flexible Ag/poly (amic acid) composites using the...

Area of Science:

• Materials Science
• Nanotechnology
• Artificial Intelligence in Materials Science

Background:

• Flexible composites are crucial for advancing flexible electronics.
• Integrating artificial intelligence (AI) with materials research can significantly enhance efficiency in design, synthesis, characterization, and application.
• Optimizing fabrication processes for flexible materials often requires extensive experimentation.

Purpose of the Study:

• To develop a machine learning model for predicting the electrical properties of flexible Ag/poly(amic acid) (PAA) composites.
• To optimize the fabrication conditions for Ag/PAA composites using AI.
• To demonstrate the applicability of the optimized materials in flexible electronic devices.

Main Methods:

• A backpropagation (BP) neural network optimized by the differential evolution (DE) algorithm was employed.
• Input parameters included PAA concentration, AgNO3 ion exchange time, NaBH4 concentration, and reduction time.
• Output was the product of sheet resistance and processing time; DE algorithm optimized BP network's initial threshold, weight, and data import model.

Main Results:

• A highly accurate machine learning model was established using 1077 learning and 49 predictive samples.
• The model achieved prediction errors of less than 1.96%.
• Optimized fabrication conditions for Ag/PAA composites suitable for strain sensors and electrodes were successfully predicted.

Conclusions:

• The developed machine learning approach effectively optimizes material fabrication processes for flexible electronics.
• Successfully fabricated strain gauge sensors, triboelectric nanogenerators (TENGs), and capacitive pressure sensor arrays using optimized parameters.
• This work highlights the significance of AI in guiding material and process design for rapid development of flexible materials and devices.