AI-Assisted Virtual Testing Method for Mechanical Performance of Stretchable Silicon Arrays

This paper reports a novel approach for virtual mechanical performance testing of stretchable silicon arrays using artificial neural network (ANN). By inputting key geometric feature sizes of stretchable silicon arrays measured by automatic optical inspection (AOI), the proposed ANN model enabled to predict the equivalent Young's modulus and elastic stretchability of the stretchable silicon array with accuracy up to 95%, without the requirements of destructive uniaxial tensile tests. Consequently, this method allows for the quantitative analysis of inevitable parameter deviations in the wafer-level manufacturing process, facilitating high-consistency mass manufacturing of high-performance stretchable silicon arrays.