Subsurface modelling using geoscientific data is essential to understand the Earth and to sustainably manage natural resources. Geology and geophysics are two critical aspects of such modelling. Geological and geophysical models have different resolutions and are sensitive to different features. Considering only geological or geophysical aspects often leads to contradictions as creating an Earth model is a highly non-unique problem. In addition, the sensitivity of the data is limited and many objects cannot be differentiated by a single discipline. The only way to address this is solving the longstanding challenge of integrating of geological data and knowledge (orientation data, contacts and ontologies) and geophysical methods (physical fields). Recent techniques usually focus on features the data is sensitive to and merely use one discipline to falsify hypotheses from the other. Such approach prevents considering the full range of potential outcomes, and fails to exploit the sensitivity of both approaches. This project proposes a different philosophy to solve the challenge of connecting geological and geophysical modelling. It first involves the development of a novel method integrating the two model types in a single framework giving them equal importance. Geological and geophysical data will be modelled simultaneously through an implicit functional mapping one domain into the other by linking their respective models. This will allow the simultaneous recovery of compatible geological and geophysical models. Secondly, this project will use a new hybrid deterministic-stochastic optimisation technique to explore the range of subsurface scenarios to estimate the diversity of features that cannot be differentiated based on the available data. Thirdly, after proof-of-concept, the method will be applied to two cases: imaging of a mantle uplift in the Pyrenees Mountains (France/Spain), and study of potential new subsurface scenarios around the Kevitsa mine (Finland).

Light elements such as hydrogen and nitrogen present large isotope variations among solar system objects and reservoirs (including planetary atmospheres) that remain unexplained at present. Works based on theoretical approaches are model-dependent and do not reach a consensus. Laboratory experiments are required in order to develop the underlying physical mechanisms. The aim of the project is to investigate the origins of and processes responsible for isotope variations of the light elements and noble gases in the Solar System through an experimental approach involving ionization of gaseous species. We will also investigate mechanisms and processes of isotope fractionation of atmophile elements in planetary atmospheres that have been irradiated by solar UV photons, with particular reference to Mars and the early Earth. Three pathways will be considered: (i) plasma ionisation of gas mixtures (H2-CO-N2-noble gases) in a custom-built reactor; (ii) photo-ionisation and photo-dissociation of the relevant gas species and mixtures using synchrotron light; and (iii) UV irradiation of ices containing the species of interest. The results of this study will shed light on the early Solar System evolution and on processes of planetary formation.

The PV sector is experiencing a fast growth and is becoming one of the main pillars for the energy transition. The goal of 30% efficiency (challenging roadmap issued during COP21, initiated by IPVF and named “30-30-30”) will be extremely difficult to reach with classical solar cells based on single junctions. On contrary, tandem cells could reach 43% theoretically by combining a VIS/NIR bottom cell, Si-made, with a top cell collecting the blue/UV range of the solar spectrum, which could be a CIGS top cell specially optimized with a bandgap about 1.7 eV. We propose a new and disruptive approach based on using wideband gap III-V (GaP-based) buffer layers between Si and CIGS. We expect that epitaxial (CIGS/III-V/Si substrate) developed in this project will reach 18% at 1.7 eV. Original designs of (CIGS/III-V/Si/Si substrate) tandem cells with two terminals and a efficiency of 25% are proposed at the end of the project.