Modeling the vacancy concentration in bi-metallic systems is a key step to understand thermal interdiffusion phenomena. For instance in the kinetics of precipitation, a strong difference between vacancy concentration in the precipitates and in the matrix, may change the precipitate mobility and the coarsening mechanism, the precipitate composition and morphology [1]. The origin of this inhomogeneous vacancy concentration is mainly due to the differences in migration energies ∆E and in formation energies ∆E of the vacancy in the two phases. The distribution of vacancy sinks and sources or the local lattice deformation should also play an important role. In this work, we focus on the influence of both ∆E and ∆E terms on the vacancy concentration profile in a phase separating a multilayer with planar (111) interfaces. The atomic diffusion model is a kinetic Ising-like model with a vacancy exchange mechanism. We perform both kinetic Mean-field and kinetic Monte-Carlo simulations in order to discuss how the vacancy concentration profile (VCP) reaches its steady-state. In the figure, we report the quasistationary VCP found at the vicinity of an (111) interface for ∆E = 0eV. As expected the vacancy strongly segregates into the two interface planes (p=-1,0) because of the tendency of the system to phase separate (the vacancy tends to reduce the number of AB pairs). More surprising is the decrease of the concentration in adjacent planes (p=-3,2,1,2). This marked decrease, that becomes asymmetric for ∆E = -0.1eV (see figure), is rationalized by invoking local correlation effects.