Thèse Dégradation des Technologies Photovoltaïques Émergentes dans des Conditions Réelles en Extérieur Caractérisation en Extérieur Vieillissement Accéléré en Laboratoire et Validation Croisée H/F - Doctorat.Gouv.Fr

Établissement : Université Paris-Saclay GS Sciences de l'ingénierie et des systèmes École doctorale : Electrical, Optical, Bio-physics and Engineering Laboratoire de recherche : Laboratoire de Génie Electrique et Electronique de Paris Direction de la thèse : Anne MIGAN-DUBOIS ORCID 0000000171072341 Début de la thèse : 2026-10-01 Date limite de candidature : 2026-05-05T23:59:59 Les technologies photovoltaïques (PV) à base de pérovskite se sont imposées comme des candidates très prometteuses pour la conversion de l'énergie solaire de nouvelle génération, avec des rendements records dépassant désormais 26 % pour les cellules à jonction unique et 34 % pour les configurations en tandem. Malgré ces progrès remarquables, leur commercialisation reste freinée par une stabilité opérationnelle à long terme insuffisante dans des conditions réelles en extérieur. Contrairement aux cellules PV classiques en silicium, les dispositifs à pérovskite présentent un comportement complexe et dépendant de l'historique (notamment la métastabilité, la migration ionique et la récupération partielle) que les protocoles de qualification standard en intérieur ne parviennent pas à saisir de manière fiable.
Cette thèse de doctorat propose une étude systématique de la dégradation des cellules PV à pérovskite en combinant une surveillance continue en extérieur de dispositifs réels avec des essais de vieillissement accéléré en laboratoire soigneusement conçus. L'objectif scientifique principal est d'identifier les facteurs de contrainte dominants à l'origine des pertes de performance irréversibles, de les distinguer des effets transitoires réversibles et, enfin, d'établir des relations d'équivalence entre les conditions de contrainte en laboratoire et les historiques d'exposition en extérieur. Une des principales nouveautés du projet réside dans le déploiement conjoint de caractérisations électriques (IV sous lumière et dans l'obscurité) et de l'imagerie par luminescence (EL/PL) en conditions extérieures, ce qui permettra d'extraire des signatures de dégradation multidimensionnelles et résolues spatialement, au-delà des capacités des mesures scalaires conventionnelles.
Les travaux seront menés au sein du laboratoire GeePs (Université Paris-Saclay) et en collaboration avec l'IPVF, ce qui permettra d'accéder à des plateformes de vieillissement en intérieur à la pointe de la technologie et à une infrastructure de surveillance en extérieur bien établie depuis plusieurs années. Le doctorant développera des pipelines d'acquisition de données automatisés, un cadre commun de « descripteurs de contrainte » pour comparer les ensembles de données entre les différentes conditions d'essai, ainsi que des modèles de corrélation reliant les signatures optiques et électriques aux voies de dégradation. Les résultats contribueront directement à la conception de normes de qualification plus prédictives pour les technologies PV à pérovskite et en tandem, conformément à l'objectif plus large d'accélérer leur déploiement fiable à l'échelle industrielle. Perovskite PV modules are attracting strong interest owing to their high efficiency [1], low material consumption, and compatibility with low-cost, scalable manufacturing routes [2-5]. Yet their long-term operational stability under outdoor conditions remains a critical barrier to commercial deployment, as performance is strongly governed by real climate stressors and cumulative operating history.
Unlike mature silicon PV, perovskite devices can exhibit non-linear, history-dependent outdoor behaviour (including metastability, partial recovery, and hysteresis) typically associated with ion redistribution, interfacial charge accumulation, and defect evolution [6,7]. These complex phenomenological effects are difficult to capture under simplified laboratory test conditions, which is why continuous outdoor monitoring is increasingly central to perovskite research: it enables direct tracking of device response under realistic irradiance variability, diurnal and seasonal temperature cycling, humidity fluctuations, wind loading, and diverse electrical operating modes. To date, the longest published time-series outdoor dataset for perovskite modules spans four years [8,9]. However, most outdoor studies remain primarily descriptive and are not always directly connected to controlled indoor ageing experiments performed on comparable device architectures. This limits the ability to translate outdoor observations into predictive qualification protocols.
In this context, understanding degradation mechanisms and defining robust performance indicators are essential to (i) separate reversible transient effects from irreversible degradation, (ii) link performance losses to specific stressors and operating histories, and (iii) guide both materials and stack optimisation as well as qualification protocols that are genuinely predictive of field reliability.
Although accelerated indoor ageing tests are widely used to evaluate device stability, the diversity of testing protocols [10] frequently leads to discrepancies in reported degradation pathways and lifetime estimates. Recent studies emphasise the need to carefully design and standardise stress protocols in order to decouple the contributions of individual stressors from their combined effects. Furthermore, accumulating evidence suggests that the specific conditions of indoor ageing can substantially influence the resulting degradation behaviour. Notably, similar burn-in phases have been observed in both indoor- and outdoor-aged p-i-n IPVF devices, even under different ageing pathways, pointing to the potential existence of equivalence relationships between specific indoor stress conditions and real outdoor exposure, which could ultimately improve the reliability of lifetime predictions for perovskite technologies. The objectives of this thesis are to characterise and understand the degradation mechanisms of emerging photovoltaic devices, from small-area laboratory cells to full-scale modules, through the combined analysis of long-term outdoor field monitoring and targeted indoor accelerated ageing. The work aims to identify and rank the dominant environmental and operational stress factors driving performance loss under real field conditions, to establish physics-grounded equivalence relationships between indoor test conditions and outdoor exposure histories, and to develop robust, standardised characterisation protocols that are genuinely predictive of field reliability. The research programme is structured around three complementary work packages:
Platform upgrade and measurement automation: The outdoor monitoring infrastructure will be upgraded to ensure robust and reliable long-term operation, encompassing improved instrumentation, tighter system integration, and enhanced data continuity. Automated measurement sequences will be implemented for outdoor Dark IV acquisitions. Electrical measurements will be synchronised with EL/PL acquisition via shared hardware triggers and automated data pipelines, enabling consistent, time-stamped multi-modal datasets.
Outdoor behaviour assessment and indoor benchmarking: This work package will define a common set of key performance indicators (KPIs) to track device performance and stability over time, including stabilised power metrics, recovery indicators, and drift under different operating modes. Time-series data will be analysed to separate reversible from irreversible losses and to detect performance regime changes driven by environmental conditions and electrical history. In parallel, a consistent indoor/outdoor comparison framework will be built by translating both datasets into shared stress descriptors (cumulative irradiance dose, thermal cycling intensity, relative humidity exposure, and bias/operating history). Analysis tools and predictive models will be developed to quantify behavioural responses under real operating conditions, and to enable direct comparison between tandem PK/Si devices and single-junction perovskite and silicon references exposed to identical stress histories.
Outdoor EL/PL deployment and correlation to electrical performance: This work package will focus on deploying outdoor electroluminescence and photoluminescence measurements under calibrated, reproducible protocols, with robust signal normalisation, careful handling of temperature dependence, and systematic repeatability checks. Luminescence imaging provides considerably richer information than conventional scalar indicators such as Voc, Jsc and FF, enabling the identification of spatially resolved degradation causes. Key luminescence features, including emission intensity, spatial uniformity, degradation patterns, and spectral proxies when available, will be extracted, monitored, and tracked as a function of operating and electrical history. These EL/PL signatures will then be correlated with electrical parameters (Voc/FF trends, Rs/Rsh evolution, Dark IV indicators, and recombination-related proxies) to distinguish reversible effects from permanent degradation and to identify the most likely failure modes.

Nous recherchons un étudiant qui aime réaliser des caractérisations, monter des manips et traiter de gros jeux de données par diverses méthodes statistiques ou d'apprentissage automatique. L'étudiant devra être motivé par le développement de sources d'énergie bas carbone, l'envie de transmettre ses résultats à la communauté scientifique internationale, curieux et travailleur.