Insights into capacitance variance mechanisms via a machine learning-biased evolutionary approach

Abstract

Dielectric particles are often added to ink formulations to tailor the macro level permittivity of printed dielectric substrates and coatings. In these inks, the combined role of particle morphology, discrete spatial arrangement and material properties on variance is difficult to distinguish experimentally and hence poorly understood. This is primarily due to the large parameter space of processing variables as well as electrical sensitivity to local heterogeneities. We address this challenge by combining a finite element capacitor model with a neural network biased genetic algorithm (NBGA) to optimize the volume fraction, particle size, and permittivity distributions of dielectric particles to identify systems with high capacitance variance. Analysis of the database generated from the optimization process provided insights on effect of polydisperse particles on variance of the system. Design rules/strategies were also identified for achieving target variance. Unsupervised machine learning techniques were applied to the NBGA-created database to extract correlations between the spatial/material distributions of the dielectric particles and the capacitance variance. Collectively, this study provides a useful framework to correlate electrical performance with both macro- and microstructural variation sources, which is key to accelerating the development of 3-D printing materials.