Intelligent data analysis of electrophysical characteristics of nanocomposites for microelectronics and biosensors
Financial University under the Government of the Russian Federation, Moscow, Russia
Abstract
Intelligent analysis of multimodal electrophysical data from nanocomposites opens new pathways for developing high-performance microelectronic and biophotonic sensor systems. This work presents a comprehensive approach that integrates experimental measurements (impedance spectroscopy, RF characterization, dielectric permittivity profiling, and optical spectra) with machine learning and physics‑informed modeling. The proposed pipeline comprises signal preprocessing, feature extraction using multi-domain transforms and time–frequency analysis, selection of informative features, and training of hybrid models (ensembles, graph neural networks, and PINN‑style architectures) to recover material parameters and predict functional performance. We show that incorporating physics-based constraints and priors improves estimate robustness and interpretability under limited training data. Using polymer‑filled nanocomposites with plasmonic and ferroelectric inclusions as a case study, the approach yields superior accuracy in estimating conductivity, dielectric permittivity, and resonance spectra compared to conventional processing methods. Applications in microelectronic components (dynamic filtering, embedded diagnostics) and biosensors (enhanced sensitivity and selectivity via nanoparticle optimization and signal fusion) are discussed. The study highlights the role of intelligent methods in accelerating material development and integrating nanocomposites into biophotonic microsystems, and it outlines prospects for automated quality control and adaptive materials design.
Speaker
Sergey Korchagin
Financial University under the Government of the Russian Federation
Russia
Discussion
Ask question
