Future applications for brazing will demand new filler metals that perform at higher temperatures, improve on the current mechanical properties, have better oxidation resistance and have the capability to easily braze materials with different chemical natures (e.g., metal to ceramics). As a result of this complex mix of requirements, new alloys such as Eutectic High Entropy Alloys (EHEAs), have been sought as a possible solution. This is because of their good behaviour at high temperatures, as has been reported for some High Entropy Alloys (HEAs), but also because they can be expected to show lower melting points than might commonly be found for combinations of refractory elements. Moreover, their isothermal solidification behaviour implies no segregation, and their dual-phase microstructure could provide good mechanical properties. However, the design of new EHEAs represents a challenge, due to two factors; the millions of possible combinations of elements and stoichiometry that can be formed, and the lack of reliable design tools to help defining the eutectic composition. Machine learning has shown promising results distinguishing between HEAs and EHEAs, by the use of thermodynamic, electronic and atomic size features. This has allowed us to make and test new EHEAs, and here we introduce the initial characterization and performance of a high-temperature metal-ceramic braze using Hf_(0.4)FeCoMnNi_2, an example of an alloy designed with the use of machine learning algorithms.
展开▼
