Experimental verification of a common assumption in the use of concentration-dependent interdiffusion coefficient

•Interdiffusion coefficient, D(C), is used to predict and analyze diffusion.•A general view that D(C) at any temperature and time can predict diffusion.•D(C) at long time fails to predict concentration profiles at shorter times.•Diffusion-induced stress varies and causes the D(C) to isothermally change. There has generally been a view that concentration-dependent interdiffusion coefficient, D(C), obtained at any diffusion temperature and time is a material constant and can be reliably used to predict diffusion effects at all other isothermal diffusion times. This notion which is implicitly predicated on the assumption that diffusion-induced stress (DIS) rapidly and completely relaxes once formed and does not influence D(C) is experimentally verified in the present work. The results of the study show that reliably computed D(C) from concentration profile obtained at long diffusion time fails to correctly predict concentration profiles at shorter diffusion periods, due to isothermal variation of the D(C) with time. Accordingly, instead of continuing to assume that D(C) is time-independent and can be reliably used to predict diffusion effects at any isothermal diffusion time, without appropriate experimental support, proper experimental verification of this crucial concept in other alloy systems, as done in the present work, is imperative.