For PV terrestrial applications, it would be very interesting to replace with Si the costly Ge substrate of III-V/Ge multijunction solar cells already developed for space applications. Effectively, according to a realistic model, we have shown that in a first instance a GaAs/Si tandem cell could achieve an efficiency of 29.2% under AM1.5G spectrum [1] and this would also open the route for low cost high efficiencies triple junctions. The main challenge is the growth of a lattice mismatched material without dislocations nor antiphase domains. Using Chemical Beam Epitaxial Lateral Overgrowth (CBE ELO) on Tunnel Oxide from nano-seeds with the metalorganic precursors trimethylgallium (TMGa) and tertiary-butyl-arsine (TBAs), we were able to grow such good quality micrometer size GaAs crystals on silicon [2]. To further develop a GaAs/Si tandem cell, the doping of these GaAs crystals has to be accurately controlled. Due to their micrometric size, C(V) measurements cannot be used for characterizing the active dopant density. We have then developed a contactless technique based on the bandgap versus temperature evolution investigated with microphotoluminescence. For many years, the Varshni’s semi-empirical relation has been used to describe semiconductors bandgap dependence with temperature [3]. However the inadequate analytical structure of Varshni’s formula was explained by Pässler in 1999 who proposed an analytical four-parameters expression capable of providing better numerical fittings and estimations of physical parameters [4]. We used this last approach taking also into account the band gap narrowing in substantially doped materials. As PL measurements can also be affected by the different sizes of our GaAs crystals in the submicrometric range, we firstly assessed the method on full plate GaAs layers homoepitaxially grown on GaAs wafers, the crystal orientation of which was carefully chosen to mimic the heteroepitaxial growth on silicon. The application of our method will be illustrated with the investigation of the influence on the residual doping of the V/III precursors ratio in the vapour phase. [1] D. Mencaraglia et al., EUPVSEC 2019 Proceedings, doi: 10.4229/EUPVSEC20192019-3BV.2 [2] C. Renard et al, Sci. Rep. 6 (2016) 25328. doi:10.1038/srep25328. [3] Y. P. Varshni, Physica 34, 149 (1967) [4] R. Pässler, Phys. Stat. Sol. (b) 216, 975 (1999)