Accelerated design of eutectic high-entropy alloys using high-throughput phase diagram calculations and machine learning

Designing eutectic high-entropy alloys (HEAs) with a stable dual-phase structure is prospective for high-temperature structural materials, but remains a challenge due to the complexity of multicomponent compositional space. This study introduces a machine learning (ML) based classification model to assist the high-throughput thermodynamic calculations (HTCs), improving the efficiency by identifying eutectic, near-eutectic and non-eutectic compositions in the VNbTiTaSi system using the parameters that determine the eutectic forming ability. These parameters include the volume fraction of silicides, the liquidus temperature, the eutectic reaction temperature, the melting point, and the temperature interval, obtained from HTCs. The design of experiments (DOE) method was applied to verify the eutectic-like duplex microstructure and to narrow down the range of the abovementioned parameters. Ultimately, three eutectic alloys, V 36 Nb 31 Ti 13 Ta 5 Si 15 , V 32 Nb 29 Ti 17 Mo 7 Si 15 , and V 32 Nb 28 Ti 13 Ta 5 Mo 7 Si 15 , were selected from the eutectic zone identified by the ML models and demonstrated exceptional compression properties at elevated temperatures. This ML combining HTCs eutectic design method is expected to be extendable to multi-component alloys with even more elements and other alloy systems.