摘要
Multi-principal element alloys (MPEAs), including high entropy alloys, show exceptional mechanical properties and corrosion resistance. These unusual properties are mostly attributed to their particular diffusion behaviors, such as sluggish diffusion. However, this phenomenon is still controversial. Although state-of-the-art simulations report that the percolation effect plays a crucial role in triggering sluggish diffusion in binary alloys, whether this correlation could be generalized to a broad class of MPEAs is unknown. In this work, we combine machine learning (ML) with kinetic Monte Carlo (kMC) to accurately determine the diffusion coefficients in ternary CoCrNi MPEAs. Special care is taken to understand the effects of percolation on diffusion. We find that percolation does not always lead to sluggish diffusion. Instead, the diffusion properties are exclusively governed by the potential energy landscape (PEL) that the diffusing objects experience. By comparing the reduced systems of CoNi and FeNi, we reveal that the percolation effect on triggering the sluggish diffusion strongly depends on the diffusion properties of the slow-diffuser in these systems. These results shed light on the understanding of the complex diffusion behaviors in MPEAs, highlighting the importance of tunning species-dependent PEL in order to tailor their diffusion properties.