Non-invasively exploring the cerebellum microstructure with magnetic resonance
To better diagnose and monitor brain diseases, we need “non-invasive biopsies” to access the tissue cell-type composition and state without opening the skull. Magnetic resonance imaging (MRI) research efforts attempt to tackle the challenge but often lack cellular specificity because of the ubiquitous nature of water. Diffusion-weighted magnetic resonance imaging (dMRS) measures diffusion of intracellular and partly cell specific molecules in a region of interest, and forms a solid basis for resolving cell-types non-invasively. Among challenges, resolving signal contributions from the different cerebellar neurons could help monitor and understand neurodevelopmental and ataxic disorders. The cerebellum is a brain region representing 10% of the brain volume but containing more than half of the brain neurons, with the very large and complex Purkinje cells and the very small and round granule cells, both having very different functions and metabolism. The PhD project aims to disentangle these cells with complementary strategies: a classical dMRS approach and a quantum dMRS approach confronted to the state-of-the-art microstructure MRI methods.
Distributed multimodal learning for cooperative acoustic source localization and classification
In many complex environments, such as industrial sites, disaster-stricken buildings, or public spaces, it is necessary to automatically detect and localize sound events (falls, alarms, voices, mechanical failures). Mobile platforms equipped with cameras and microphones represent a promising solution, but a single platform remains limited: its microphone array provides an approximate direction towards the source but not a precise position in space, and its camera may be obstructed. This thesis proposes to study how a network of mobile platform, each carrying a calibrated audio-visual unit, can collaborate to localize and classify such events in 3D. Each platform analyses its own audio-visual observations and shares an estimate of the source direction with its neighbours; the network then combines these estimates to reconstruct the position of the event and identify it. The expected outcomes are a cooperative localization system that is robust to occlusions and partial platform failures.
Topologically Isolated Mode Acoustic Resonators
Timing is a key function in electronic circuits. Beyond on-chip signals synchronization, it also allows the synchronization of wireless data transmissions. Accurate time references require stable frequency sources, which also benefit to sensor applications. The gold standard for time or frequency generation is still quartz resonators, which are however bulky and difficult to miniaturize. Research is therefore still ongoing to provide high quality factor (> 10,000) resonators, ideally capable of operating at frequencies of several GHz. A key to reach such high quality factors is to confine strongly the mechanical vibration of micro-size structures in order to make them insensitive to external perturbations. Recently, the field of topological acoustics has demonstrated the capability to confine elastic waves in very small volumes concentrated at the interface between periodic structure, and to provide extremely high quality factor resonances.
This PhD position focuses on exploiting topologically protected modes in piezoelectric microstructures to provide next generations of high quality factor resonators, which may be used in oscillators or even filter circuits. Leveraging the know-how of CEA Leti in the design and fabrication of such components, the PhD will be part of an international collaboration with well established academic laboratories (Politecnico di Milano, Imperial College FEMTO-ST Institute) and industrial partners.
The candidate will model and design structures supporting topologically protected modes, combinining finite element simulations with simplified numerical approaches which reduce computation times. He will follow the fabrication of demonstrators in collaboration with the process integration teams in the CEA Leti clean rooms, and carry on measurements of the proposed resonators.
Potential synergy between NH3 and NaBH4 for improved H2 density and enhanced safety
The thesis focuses on the study of the hybrid ammonia–sodium borohydride system (NH3–NaBH4) as an innovative chemical energy carrier. The objective is to investigate the combination of ammonia (NH3), recognised for its high hydrogen density and mature industrial infrastructure, with sodium borohydride (NaBH4), a high-capacity chemical hydrogen storage material, in order to overcome certain limitations associated with each vector when considered separately.
The proposed work specifically addresses the safer storage and transport of ammonia through its coupling with sodium borohydride, enabling a reduction in vapour pressure (compared to 8.88 bar at 21 °C for liquid ammonia) and less restrictive operating conditions. In parallel, the thesis aims to improve the stability (relative to the H2O–NaBH4 system) and operability of sodium borohydride which, when combined with ammonia molecules (acting as inert species), forms stable liquid or viscous phases that are potentially pumpable, thereby facilitating integration into energy-related processes.
The fundamental goal of the thesis is to understand the physicochemical mechanisms governing this hybrid system, particularly the role of dihydrogen interactions between the N–H bonds of ammonia and the B–H bonds of borohydride, and their influence on stability, reactivity, transport properties, and hydrogen release pathways (thermal and/or hydrolytic).
Beyond its storage function, the thesis also explores the potential of the NH3–NaBH4 system as a novel hybrid material with high gravimetric and volumetric hydrogen capacity, while considering realistic operational constraints relevant to energy applications in a dual-use context. At this stage, exhaustive optimisation is not the primary objective.
Intelligent control and optimization of DC microgrids using digital twins in real-time simulation
This thesis addresses the challenge of decarbonizing industrial and territorial systems by proposing a transition to direct current (DC) microgrids controlled by a Digital Twin. Faced with the saturation of alternating current (AC) grids due to the growth of photovoltaics, energy storage, and electric mobility, DC allows for a reduction in conversion losses (5 to 15%), improved flexibility, and a simplification of the electrical architecture.
The project is based on the development of a high-fidelity Digital Twin synchronized in real-time simulation. More than just a monitoring tool, it acts as a proactive decision-making system integrating advanced optimization algorithms, such as artificial intelligence and predictive control. It anticipates voltage instabilities, which are particularly critical in low-inertia DC grids, and continuously optimizes power flows to maximize self-consumption while preserving battery life.
Experimental validation relies on a Hardware-in-the-Loop approach within the CEA-Liten/G2Elab ecosystem, integrating physical converters. This methodology guarantees robustness, security, and resilience before any real-world deployment.
The expected outcomes are scientific (stability and real-time modeling), operational (provision of technical guides and decision-making tools), and strategic (strengthening French technological sovereignty in Smart Grids and accelerating the 2050 carbon neutrality trajectory advocated by ADEME).
High-Endurance Chalcogenide Memories for Next-Generation AI
Discover a unique phd opportunity where you will dive into the heart of innovation in memory technologies. You will develop strong expertise in areas such as electrical characterization and the understanding of degradation phenomena in chalcogenide-based memories.
By joining our multidisciplinary teams, you will play a key role in studying and improving the endurance of Phase-Change Memory (PCM) and Threshold Change Memory (TCM) devices—two promising technologies for high-performance artificial intelligence applications. You will take part in innovative projects combining scientific rigor and applied research on nanoscale devices, working closely with another CEA PhD student who conducts advanced physico-chemical analyses (TEM) to investigate degradation mechanisms.
You will have the opportunity to contribute actively to tasks such as:
Electrical characterization of PCM and TCM devices to analyze cycling-induced degradation
Development and evaluation of innovative programming protocols to extend endurance limits
Proposing solutions to improve the reliability and performance of next-generation memories
Regular collaboration and discussion with the CEA PhD student to interpret TEM results and draw conclusions about degradation mechanisms
Ultra-low frequency wireless power transmission for sensor node charging
Wireless power transfer (WPT) technologies are rapidly expanding, particularly for wireless charging of everyday electronic devices and for powering wireless communicating sensor nodes. However, their transmission ranges remain limited, and the high operating frequencies typically used prevent energy transfer in the presence of, or through, conductive media (such as metallic barriers or seawater). This constraint significantly limits their adoption in complex environments (industrial, biomedical, etc.).The ultra-low-frequency technology investigated in our laboratory is based on an electromechanical receiver system comprising a coil and a magnet set into motion by a remotely generated magnetic field. The objective of this PhD project is to propose and develop novel ultra-low-frequency concepts to increase transmission range while maintaining sufficient power density for supplying sensor systems. The work will therefore involve studying, designing, optimizing, and experimentally validating the performance of new topologies (emitter field shaping, receiver geometries and materials, etc.). The candidate will develop analytical and numerical models to identify key system parameters and compare performance with the state of the art (range, power density, sensitivity to orientation). In addition, the candidate will propose, design, and experimentally evaluate innovative energy conversion electronics, on the transmitter and/or receiver side, to assess their impact on the overall system performance. A joint optimization of the electromechanical system and its associated power electronics will ultimately lead to the realization of a high-performance wireless power transfer system. A multidisciplinary profile with a strong orientation toward physics and mechatronics is sought for this PhD project. In addition to solid theoretical foundations, the PhD candidate must demonstrate the ability to work effectively in a team environment as well as a strong aptitude for experimental work. The PhD candidate will be integrated into the Systems Department of CEA-Leti, within a team of researchers with strong expertise in the development and optimization of electronic and mechatronic systems, combining innovative solutions for energy harvesting, wireless power transfer, low-power electronics, and sensor integration aimed at the development of autonomous systems.
Learning Mechanisms for Detecting Abnormal Behaviors in Embedded Systems
Embedded systems are increasingly used in critical infrastructures (e.g., energy production networks) and are therefore prime targets for malicious actors. The use of intrusion detection systems (IDS) that dynamically analyze the system's state is becoming necessary to detect an attack before its impacts become harmful.
The IDS that interest us are based on machine learning anomaly detection methods and allow learning the normal behavior of a system and raising an alert at the slightest deviation. However, the learning of normal behavior by the model is done only once beforehand on a static dataset, even though the embedded systems considered can evolve over time with updates affecting their nominal behavior or the addition of new behaviors deemed legitimate.
The subject of this thesis therefore focuses on studying re-learning mechanisms for anomaly detection models to update the model's knowledge of normal behavior without losing information about its prior knowledge. Other learning paradigms, such as reinforcement learning or federated learning, may also be studied to improve the performance of IDS and enable learning from the behavior of multiple systems.
SiGe HBT LNA for cryogenic applications: design, characterization and optimization
The global race to build a quantum computer is heating up! These cutting-edge systems operate at temperatures below 4 K to preserve the delicate quantum states essential for computation. To achieve efficient control and detection, conventional electronic circuits must perform reliably at cryogenic temperatures, in close proximity to the quantum processor, thereby reducing wiring complexity and boosting performance. Beyond quantum computing, other domains—such as space exploration, high-performance computing, or high-energy physics—also require transistors capable of operating below 100 K.
During this phD, you will perform radiofrequency (RF) electrical characterization and modeling of Silicon-Germanium Heterojunction Bipolar Transistors in cryogenic environment, contributing to a deeper understanding of their behavior and optimizing their potential for extreme-condition applications. The objectives are twofold:
1.RF Electrical Characterization and Modeling:
•Conduct RF electrical measurements of SiGe HBTs at cryogenic temperatures.
•Develop accurate models to describe their behavior in cryogenic environments.
2.Optimization of Low-Noise Amplifiers (LNAs):
•Study the low-temperature behavior of individual passive and active devices composing an LNA.
•Optimize the design of low-noise amplifiers (LNAs) for cryogenic applications.
Side-Channel based Reverse-Engineering
The characterization of the security of embedded systems in "black box" or "gray box" against Side-Channel attacks often requires a preparatory phase of Reverse-Engineering, which can be particularly time-consuming, especially on a complex System-on-Chip that can be found in smartphones or in the automotive industry. This phase can, for example, consist of detecting a cryptographic primitive within Side-Channel measurements for a future observation attack, or a target routine for a fault injection attack. The objective of this thesis is to develop a methodology and non-profiled tools that allow the automation of this detection phase, while enabling the exploitation of prior knowledge of a potential attacker.