Thèse Technologies de Détection Hybrides pour l'Évaluation de la Qualité des Grains de Café Impédancemétrie et Spectrométrie de Masse en Phase Gazeuse H/F - Doctorat.Gouv.Fr

Établissement : Université de Lille École doctorale : ENGSYS Sciences de l'ingénierie et des systèmes Laboratoire de recherche : Institut d'Electronique, de Microélectronique et de Nanotechnologie Direction de la thèse : Bilel HAFSI ORCID 0000000206467512 Début de la thèse : 2026-10-01 Date limite de candidature : 2026-09-30T23:59:59 Coffee quality assessment is a major scientific, economic, and societal challenge, as coffee is one of the most widely traded agricultural commodities worldwide. The quality of green and roasted coffee has a direct influence on market value, consumer perception, and the overall sustainability of the coffee supply chain. Quality variability arises from a complex interplay of factors, including plant physiology, storage conditions, microbial contamination, and environmental stress conditions. The global coffee industry represents a market of several hundred billion euros annually, with quality grading playing a central role in price determination. Producing regions are increasingly exposed to climate variability, which affects bean composition and aroma profiles, while consumers demand higher and more consistent sensory quality. Consequently, there is a growing need for rapid, objective, and valuable intelligent tools capable of assessing coffee quality throughout the production chain, from farm to cup.
The main objective of the ORBIT project is the development of a multi-sensory platform capable of simultaneously discriminating multiple chemical markers indicative related to coffee quality, defect signatures, and authenticity, including the identification of blends and non-authentic samples. These markers are related to VOCs associated with aroma, freshness, processing methods, contamination, and aging. The project focuses on the design, fabrication, and
characterization of flexible impedimetric sensor arrays using low-cost methods and organic functional materials. A key challenge lies in achieving sensitive and reproducible detection of low concentrations of VOCs within complex chemical environments. Coffee VOCs originate from diverse biochemical pathways and include aldehydes, alcohols, esters, ketones, furans, phenolic compounds, and sulfur-containing species. Their relative concentration evolves dynamically depending on processing and storage conditions. The project therefore aims to develop a sensory system capable not only of detecting individual compounds, but also of capturing multivariate signatures that reflect overall quality states. Another major challenge concerns data interpretation. The large volume of data generated by multi-sensor arrays requires advanced signal processing and machine learning strategies to extract relevant features, cluster responses, and quantify quality-related patterns. The integration of embedded intelligence is essential to enable real-time decisionmaking and autonomous operation.
The study will focus in particular on wetprocessing deposition techniques for organic materials. Two multi-material cointegration approaches will be considered: a top-down approach (via drop casting) and a bottom-up approach (via electro-grafting). The materials of interest will mainly
involve conductive polymers formulated with additives designed to optimize the chemo-specificity of the material through the incorporation of complexing compounds. These compounds will rely on fast, reversible chemistry, enabling sensor self-regeneration under operating conditions. They will include electron-acceptor species (Lewis's acids). Particular attention will be paid to the
environmental footprint of the materials employed, guiding their selection for integration into a functional detection platform (e.g., porphyrins and metalloporphyrins will be considered as key functional materials due to their strong molecular recognition capabilities). The electrical characterization of the sensors under various conditions (laboratory and real-world environments), together with the implementation of machine-learning algorithms for data processing, clustering, and quantification of target molecules, will be investigated within the scope of this project. L'évaluation de la qualité du café constitue un enjeu scientifique, économique et sociétal majeur, le café étant l'une des matières premières agricoles les plus échangées au monde. La qualité du café vert et torréfié a une influence directe sur sa valeur marchande, la perception des consommateurs et la durabilité globale de la chaîne d'approvisionnement du café. La variabilité de la qualité résulte d'une interaction complexe entre plusieurs facteurs, notamment la physiologie de la plante, les conditions de stockage, la contamination microbienne et les stress environnementaux. L'industrie mondiale du café représente un marché de plusieurs centaines de milliards d'euros par an, dans lequel la classification de la qualité joue un rôle central dans la détermination des prix. Les régions productrices sont de plus en plus exposées à la variabilité climatique, qui affecte la composition des grains et les profils aromatiques, tandis que les consommateurs exigent une qualité sensorielle plus élevée et plus constante. Par conséquent, il existe un besoin croissant d'outils intelligents rapides, objectifs et à forte valeur ajoutée, capables d'évaluer la qualité du café tout au long de la chaîne de production, de la ferme à la tasse. Développement d'un système de détection autonome pour le suivie de la qualité des grains de café

Electronique Organique
Systèmes électroniques
Programmation Python
Réalisation de cartes électroniques PCB
Capteurs

• Python
• Machine learning