Frequency-dependent dielectric constant prediction of polymers using machine learning

<jats:title>Abstract</jats:title><jats:p>The dielectric constant (<jats:italic>ϵ</jats:italic>) is a critical parameter utilized in the design of polymeric dielectrics for energy storage capacitors, microelectronic devices, and high-voltage insulations. However, agile discovery of polymer dielectrics with desirable <jats:italic>ϵ</jats:italic> remains a challenge, especially for high-energy, high-temperature applications. To aid accelerated polymer dielectrics discovery, we have developed a machine-learning (ML)-based model to instantly and accurately predict the frequency-dependent <jats:italic>ϵ</jats:italic> of polymers with the frequency range spanning 15 orders of magnitude. Our model is trained using a dataset of 1210 experimentally measured <jats:italic>ϵ</jats:italic> values at different frequencies, an advanced polymer fingerprinting scheme and the Gaussian process regression algorithm. The developed ML model is utilized to predict the <jats:italic>ϵ</jats:italic> of synthesizable 11,000 candidate polymers across the frequency range 60–10<jats:sup>15</jats:sup> Hz, with the correct inverse <jats:italic>ϵ</jats:italic> vs. frequency trend recovered throughout. Furthermore, using <jats:italic>ϵ</jats:italic> and another previously studied key design property (glass transition temperature, <jats:italic>T</jats:italic><jats:sub>g</jats:sub>) as screening criteria, we propose five representative polymers with desired <jats:italic>ϵ</jats:italic> and <jats:italic>T</jats:italic><jats:sub>g</jats:sub> for capacitors and microelectronic applications. This work demonstrates the use of surrogate ML models to successfully and rapidly discover polymers satisfying single or multiple property requirements for specific applications.</jats:p>