A materials informatics approach to understand processing-structure-property relationships in boron carbide ceramics

Date

2025-4-10

Author

Tarman, Ömer Gökhan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

272
views

0
downloads

Cite This

The limited quantitative understanding of complex processing–structure–property relationships in ceramics, which hinders the development of new compositions and production routes, is particularly critical for boron carbide (B4C) ones due to their extremely difficult sintering behavior, requiring very high temperatures and prolonged processing times with pressureless routes, and hence substantially increasing production costs. The application of machine learning algorithms in materials science offers an efficient alternative by enabling the quantification of these relationships based on existing data, thus reducing the need for synthesis and characterization of each candidate material. In this study, several machine learning models have been developed with respectable accuracies to predict the structural descriptors, namely density and grain size, and mechanical properties, such as fracture toughness, hardness, and flexural strength of B4C ceramics. By using data extracted from scientific literature, these models input processing parameters including the raw materials and sintering variables. The decision mechanisms of the models are analyzed to quantify the relationships between processing, structure, and properties. The identified patterns among these relationships were not solely case-specific but generally applicable, enabling the study to offer a broader and more comprehensive understanding. It has been shown that ceramics informatics holds significant potential as a research field, despite the inherently diverse and scattered nature of ceramics data. By reducing reliance on time-consuming and costly trial-and-error experiments, the developed materials informatics models significantly accelerate the design, optimization, and deployment of B₄C ceramics. This enables a more systematic, data-driven approach through the quantitative understanding of processing–structure–property linkages.

Subject Keywords

Materials informatics, Data analysis, Boron carbide, Mechanical properties

URI

https://hdl.handle.net/11511/114589

Collections

Graduate School of Natural and Applied Sciences, Thesis