Actuality from the project

At this stage, the project is proceeding as planned. Following the design phase of the COSMOSAT prototype, there was a lengthy process of drafting, publishing, and analyzing bids, as well as receiving the equipment necessary for building the prototype. The prototype was validated and tested throughout 2025, with an initial electrical mode analysis conducted in December 2025, which validated the first stage of development (Photo 1). The performance (reflectron, mass spectrometer, vacuum chamber) is very satisfactory (Photo 2), and initial results on meteorite samples are currently being analyzed (April 2026). A presentation of the instrument is scheduled for August 2026 at the M&M2026 international conference in Milwaukee. Meanwhile, Théo Mauger began his doctoral thesis in 2024 at Gipsa-lab and has carried out substantial work on the numerical modeling of spectra using two different approaches. One approach uses a dictionary (completed; ⟨hal-05233396⟩), and the other uses deep learning (in progress), to enable the automatic identification and quantification of elements. The very promising results are currently being evaluated, and a publication is in preparation.

Focus

We hope that COSMOSAT will be completed (with the laser mode tested) and operational by November 2026 so that two postdocs (two years each) can begin applied research on 1) determining the mechanisms of amorphous silicate formation in primitive extraterrestrial rocks, and 2) understanding the processes and timescales in extraterrestrial and ancient terrestrial materials (phosphates).

Before considering the integration and setup of a Europe-wide platform, we have taken steps to ensure that COSMOSAT is integrated into a network (Electron Microscopy and Microanalysis; https://www.regef.fr/microscopie-microanalyse-electroniques) at IR RéGEF (late 2026–early 2027). We will then seek to have it designated as a national instrument by the CNRS INSU (in 2030), so that it can benefit the entire Earth and Universe Sciences community.

Relation with others WP

Link to Research Area 5, “Simulations and Data Analysis Using High-Performance Computing/AI”—WP 5.1 and 5.3 (teams from CRAL, GIPSALab, and LabHC) focusing on AI-based signal processing.

Relation with industries

Collaboration with Cameca

Publication links to the project

– Brevet EP n°25306099.0 du 09/07/2025 pour « ChaSe DLD system: Towards a fast zero dead-time position sensitive detector for the Atom Probe Tomography » aux noms de CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE, UNIVERSITE DE ROUEN NORMANDIE, Institut National Des Sciences Appliquées De Rouen ; Bacchi C, G. Da Costa, F. Vurpillot

– Mauger T., Amblard P.-O., and Michel O.J.J. (2025). Approche dictionnaire pour l’identification et la quantification des ions atomiques d’un spectre de masse issu d’une sonde atomique. GRETSI 2025 – XXXème Colloque Francophone de Traitement du Signal et des Images, Strasbourg, France. ⟨hal-05233396⟩

– Veret T., Delaroche F., Blum Y., Houard J., Klaes B., Mouton I., De-Geuser F., Seydoux-Guillaume A.-M., and Vurpillot F. (2026). Exploring mechanisms leading to composition errors in monazite (CePO₄) analysed with atom probe tomography. Microscopy and Microanalysis. ⟨hal-05546893⟩