A precise electrochemical characterization of active materials (AMs) is crucial for the improvement of well-established lithium-ion battery (LIB) chemistries as well as exploring potential materials for next-generation batteries. AMs are typically micron-sized polycrystalline particles with a complex inner structure. Experimental measurement procedures for material characterization are commonly performed on thin electrode sheets or with advanced techniques on single particles. The extraction of material characteristics from such experimental data typically makes several simplifying assumptions. For example, the AM is usually simplified to be perfectly homogeneous, isotropic, spherical particles. The resulting extracted information thus includes uncertainties, but there is only limited understanding of their significance and how they are connected to the real AM microstructure. 3D microstructure-resolved models can help explaining and quantifying these relations, but so far there has been only limited focus on the AM microstructure [1]. For our investigations, we extended a previously developed electrochemical transport model for LIBs, cf. [2,3]. Ion and charge transport in the electrolyte and solid phases, including AM and current collectors, are described in a coupled system. This simulation framework allows to conduct virtual battery experiments including information on the AM microstructure. For a precise extraction of AM properties from the virtual measurement data, we apply a Bayesian optimization approach [4]. This approach provides a distinct advantage from most conventional methods. It not only yields an estimated value for a material property, but also quantifies the uncertainty linked to this prediction for each individual measurement. By combining the above methods, we investigate the influence of experimental measurement conditions and AM properties on the extractability of material parameters with a focus on the AM diffusivity. Further, microstructural AM properties like particle cracks or grain structures can be investigated in virtual experiments and linked to material and battery performance.