P-13 Machine Learning Analysis of Microwave Dielectric Properties for Seven Structure Types: The Role of the Processing and Composition

Topic: Computation and theoretical aspects

Dr. Natalia KIREEVA
Frumkin Institute of Physical Chemistry and Electrochemistry RAS, Russia
Laboratory of new physical chemical problems

P-13_N. Kireeva_SSC2021_Machine learning analysis of microwave dielectric properties for seven structure types

N. Kireeva^1,2*, V. P. Solov’ev¹

¹ Institute of Physical Chemistry and Electrochemistry RAS, 31 Leninsky prosp, Moscow, 119071, Russian Federation; ² Moscow Institute of Physics and Technology, 9 Institutskiy per., Dolgoprudny, Moscow Region, 141700, Russian Federation

^* E-mail: nkireeva@gmail.com

Keywords: microwave dielectric characteristics, materials informatics, descriptors, grain boundaries, machine learning

During the last decades the area of wireless communications experienced the drastic growth demanding the search for new materials with tailored microwave dielectric characteristics. Microwave dielectric properties of ceramic materials have been extensively studied, the comprehensive database of microwave dielectric properties is now available [1, 2]. Several studies involving machine learning approaches have been performed [3, 4].

In this study, machine learning approaches have been used for synthesis-structure-property relationships assessment. The analysis has been performed for several structure types. The new descriptors encapsulating the information on the constituent building blocks, their relative number and connectivity have been introduced. These structure-characterizing parameters, the hybrid of descriptors recently proposed in [5] and [6], can be regarded as a flexible way to describe the inorganic crystal structures using the substructural bottom-up approach comprising both, the symmetry and the connectivity information.

The role of the processes observed at the grain boundaries as well as the re-crystallization processes for the microwave dielectric characteristics is discussed. The impact of the heat-treatment and the composition for the considered structure types is discussed.

References

[1] M. Sebastian (2008) Dielectric materials for wireless communication, Elsevier Science.

[2] M. T. Sebastian, R. Ubic, H. Jantunen (2015). Low-loss dielectric ceramic materials and their properties, International Materials Reviews, 60 (7), 392–412.

[3] D. Scott, P. Coveney, J. Kilner, J. Rossiny, N. N. Alford (2007). Prediction of the functional properties of ceramic materials from composition using artificial neural networks, Journal of the European Ceramic Society, 27 (16), 4425 – 4435.

[4] D. J. Scott, S. Manos, P. V. Coveney (2008). Design of electroceramic materials using artificial neural networks and multiobjective evolutionary algorithms, Journal of Chemical Information and Modeling, 48 (2), 262–273.

[5] A. R. Overy, A. B. Cairns, M. J. Cliffe, A. Simonov, M. G. Tucker, A. L. Goodwin (2016). Design of crystal-like aperiodic solids with selective disorder–phonon coupling, Nature Communications, 7, 10445.

[6] W. F. Reinhart, A. W. Long, M. P. Howard, A. L. Ferguson, A. Z. Panagiotopoulos (2017). Machine learning for autonomous crystal structure identification, Soft Matter 13 (27), 4733–4745.

Acknowledgments

Authors acknowledge Russian Foundation for Basic Research (Project No. 15-29- 09075) for the support. NK thanks Vladislav S. Pervov for fruitful discussions of this work, Denis Ostroumov, Alexander Petrov and Alexei Averin for their help at the beginning of this study.