Surface segregation in multicomponent high entropy alloys: Atomistic simulations versus a multilayer analytical model

Abstract This paper compares two approaches for investigating the near-surface composition profile that results from surface segregation in the so-called Cantor alloy, an equi-molar alloy of CoCrFeMnNi. One approach consists of atomistic computer simulations by a combination of Monte Carlo, molecular dynamics and molecular statics techniques, and the other is a nearest neighbor analytical calculation performed in the regular solution approximation with a multilayer model, developed here for the first time for a N-component system and tested for the 5-component Cantor alloy. This type of comparison is useful because a typical computer simulation requires the use of ~100 parallel processors for 2 to 3 h, whereas a similar calculation by means of the analytical model can be performed in a few seconds on a laptop machine. The results obtained show qualitatively good agreement between the two approaches. Thus, while the results of the computer simulations are presumably more reliable, and provide an atomic scale picture, if massive computations are required, for example, in order to optimize the composition of a multicomponent alloy, then an initial screening of the composition space by the analytical model could provide a highly useful means of narrowing the regions of interest, in the same way that the CALPHAD method allows rapid investigation of phase diagrams in complex multinary systems.