The MULTISOLSI project consists in studying, designing and optimizing the electrical junction between two relaxed lattice mismatched materials with zero or negligible strain relaxation defects, namely Ga1-xInxAs and Si in the first instance, in order to optimize a final GaAs/Si/ Ga0.24In0.76As triple junction multispectral solar cell demonstrator achieving a very high efficiency (33%) under the global AM1.5G spectrum. This goal is achievable due to previous know-how acquired by IEF in Ge lateral epitaxial overgrowth on ultra-thin Si oxide. The main objective is to study the resulting Ga1-xInxAs/Si heterojunction and design optimized structures in particular for the novel three dimensional tunnel junction design. The electrical characterization and the design of structures will be performed at the LGEP in collaboration with the IEF partner. For the IEF partner, the goal will consist in adapting, in close collaboration with CEMES, the growth process to the realisation of the designed junctions between Si substrate and Ga1-xInxAs epitaxial layer. This study involves the following complementary investigations : (1) identify electrically active defects at the interface and determine their density, evaluate the extent of doped areas at both sides of the SiO2 layer with I(V), C(V) and admittance spectroscopy measurements (LGEP), EBIC (Electron beam induced current) and AFM measurements (LGEP, IEF). 2) TEM imaging for the identification of structural defects (CEMES). Three main steps can be listed in this project: (i) The first one consists in studying the Ga1-xInxAs/oxidized Si interface. Structures will be processed and grown at IEF and then characterised at LGEP and CEMES for achieving the extraction of electrical and structural properties. The method envisaged for the GaxIn1-xAs growth integration process is the GaxIn1-xAs ELO (Epitaxial Lateral Overgrowth). For this purpose, the parameters to be optimized are the size, spacing distance and shape of the seeds, as well as the thickness of the SiO2 layer. The main output of this first step should be a thorough knowledge of the respective effects of these parameters on the carrier transport, in terms of density and identification of interfacial defects. (ii) The second step will consist in the adaptation of the above approach to the realisation of tunnel junctions between Si substrate and GaxIn1-xAs layer in agreement with the design proposed by the LGEP from the knowledge accumulated in the first step of the study. This part is the most challenging of this task and consists in the realisation of abrupt and thin doped GaxIn1-xAs layer (about 10 nm) on the SiO2 layer. (iii) The third step is the ultimate goal of this project: to develop a GaAs/Si/GaInAs triple junction having an efficiency for AM1.5G terrestrial solar spectrum conversion as close as possible to its theoretical limit (32.9%). We plan to limit the present proposal to the GaInAs family, but a further extension to the GaAlAs or InGaAsP branches can be conveniently envisaged as it exhibits the same mismatch with Si than GaAs. The methodology outlined above exploiting the GaInAs ternary system allows us to quantitatively project efficiencies of 32.9% for our triple junction demonstrator, an efficiency competitive with the triple junction terrestrial world record in these conditions (but on a Ge substrate), and significantly greater than the 26% single junction world record to date. Once the bluilding blocks of tunnel junction and heterojunction growth have been put in place, there is a clear potential to develop the phosphide based route of GaInP/Si/GaInAs structures. Our modeling then projects an efficiency potential of 45.5% under AM1.5G for available materials, presenting a major step towards terrestrial silicon based photovoltaics with very high efficiency, as this efficiency is very close to the theoretical limit (47.3 %) for an ideal triple junction under AM1.5G.