In the past years, photovoltaics (PV) became one of the cheapest sources of energy. 97% of commercial solar cells are made of silicon, and their lab-scale record efficiencies of 28.2 % are now close to the theoretical limit (29.4 %). Yet, expectations of both the society and the PV industry are still high, and most of the research efforts are now dedicated to pushing forward the efficiency. Combining silicon and new materials in tandem devices is the most-regarded solution for next-generation photovoltaics. Current options are polycrystalline semiconductors like perovskites and inorganic Cu(In,Ga)(S,Se)2 or CdTe thin films, but they are still limited by both efficiency and/or stability issues that are hardly explained by current models. Further developments require a better understanding of the properties of low-cost thin-film materials. The goal of this project is twofold. From the one side combining CathodoLuminescence (CL) and PhotoLuminescence (PL) techniques will provide a multi-scale (from <10 nm to cm) analysis tool for elementary processes and properties of bulk materials and surfaces. On the other side, we aim at pushing our analysis a step further by calibrating our tool in absolute terms, i.e. extracting the number of CL photons emitted by the sample. Another long-term objective is to use advanced computational methods for correlative data analysis, and simulation to build a realistic model of thin-film solar cells using the measured quantities.