The robust design of microgrids based on optimization methods is a challenging process which usually requires multiple system simulations and implies the use of suitable models ensuring a good compromise between complexity and accuracy. These models also have to include the main couplings within systems, which have a major impact on design criteria and constraints. In this paper, we particularly illustrate this context with regard to the choice of battery models integrating energy efficiency and aging for the design of microgrids. Using a simple case study, we demonstrate the importance of taking into account battery capacity loss due to aging to accurately assess the microgrid's self-sufficiency and cost over its lifetime.