Thermodynamic Modelling of Complex Oxides for Smart Sensors

The search for more efficient materials follows a pattern that has changed very little over the years, involving poorly automated phases of synthesis, characterization and measurement of functional properties. Although this pattern has proved its strength in creating knowledge bases, it remains ineffective because it is time-consuming and generally covers a reduced range of compositions. The project Hiway-2-mat ( seeks to use high-throughput combinatorial approaches and develop autonomous configurations to explore the compositional spaces of complex oxide materials, with the aim of accelerating the discovery of materials for smart sensors. In this context, CALPHAD method is a valuable tool for materials exploration, as it can provide a number of useful insights into the role of oxidation state or oxygen partial pressure on phase stability, and on the degree of substitution of doping elements in an oxide matrix. The aim is to calculate phase diagrams of complex oxides based on available databases, either to better prepare combinatorial experiments, or to drive the autonomous robot on the fly, providing additional information for on-line characterization.
Your role will be to:
1)Perform thermodynamic simulations using CALPHAD method and Thermo-Calc Software to predict the stability range of a set of complex oxides (Ba/Ca/Sr)(Ti/Zr/Sn/Hf)O3 at different temperatures and oxygen partial pressures. In this step, the candidate will also perform a critical review of the thermodynamic data available in the literature.
2)Include key elements in the available database.
3)Develop a rapid screening method to search for the most promising compositions.
The candidate will work closely with the experimental platform development team to guide future trials and adapt the method to better meet the needs of large-scale production.

Exploring microfluidic solutions for manufacturing targets for fusion power generation

As part of a call for projects on "innovative nuclear reactors", the TARANIS project involves studying the possibility of energy production by a power laser-initiated inertial confinement fusion power plant. The current context, which encourages the development of low-carbon energies, and the fusion experiments carried out by the NIF's American teams, make it very attractive to conduct high-level research aimed at eventually producing an economically attractive energy source based on inertial fusion.
Among the many technical hurdles to be overcome, the production of fusion targets with a suitable reaction scheme compatible with energy production is a major challenge. The CEA has the know-how to produce batches of capsules containing the fusible elements of the reaction. However, the current process is not suitable for mass production of hundreds of thousands of capsules per day at an acceptable cost.
One high-potential avenue lies in the use of microfluidic devices, for which the Microfluidic Systems and Bioengineering Laboratory (LSMB) of the Health Technologies and Innovation Department (DTIS) of CEA's DRT has recognized expertise.

Hybrid ion exchangers for the traetment of radioactive organic liquids: molecular dynamics design assistance

The ECCLOR project (labelled 'Investissement pour le future') focuses on the treatment of radioactive organic effluents by developing porous materials capable of selectively eliminating alpha emitting ions. Research carried out at CEA has led to the design of hybrid materials with variable performance in capturing alpha emitters present in organic liquids. Understanding this performance at the molecular level is essential, but complex.
To address this challenge, this post-doctoral fellowship focuses on the use of classical molecular dynamics to rationalize these performances. The work will be carried out at the Marcoule research center's LILA laboratory, drawing on the expertise of teams specializing in the modeling of solid/liquid systems using classical molecular dynamics.
To support these simulations, experimental data may be provided by laboratories such as the Laboratoire des Procédés Supercritiques et de Décontamination (LPSD) and the Laboratoire de Formulation et Caractérisation des Matériaux minéraux (LFCM). The results obtained will be discussed at progress meetings and will be the subject of scientific publications.
In summary, this post-doctoral contract aims to couple theoretical approaches with experiment. Understanding the interactions within these materials at the molecular scale is essential to provide insights and improve the processes currently under study.

Causal learning

As part of a project that concerns the creation of innovative materials, we wish to strengthen our platform in its ability to learn from little experimental data.

In particular, we wish to work firstly on the extraction of causal links between manufacturing parameters and properties. Causality extraction is a subject of great importance in AI today and we wish to adapt existing approaches to experimental data and their particularities in order to select the variables of interest. Secondly, we will focus on these causal links and their characterization (causal inference) using an approach based on fuzzy rules, that is to say we will create fuzzy rules adapted to their representation.

Generative AI for model driven engineering

Generative AI and large language models (LLMs), such as Copilot and ChatGPT can complete code based on initial fragments written by a developer. They are integrated in software development environments such as VS code. Many papers analyse the advantages and limitations of these approaches for code generation. Besides some deficiencies, the produced code is often correct and the results are improving.

However, a surprisingly small amount of work has been done in the context of software modeling. The paper from Cámara et al. concludes that while the performance of the current LLMs for software modeling is still limited (in contrast to code generation), there is a need that (in contrast to code generation) we should adapt our model-based engineering practices to these new assistants and integrate these into MBSE methods and tools.

The goal of this post-doc is to explore generative AI in the context of system modeling and associated tool support. For instance, AI assistance can support the completion, re-factoring and analysis (for instance identified design patterns or anti-patterns) at the model level. Propositions are discussed in the team and in a second step prototyped and evaluated the mechanism in the context of the open-source UML modeler Papyrus.

Development of noise-based artifical intellgence approaches

Current approaches to AI are largely based on extensive vector-matrix multiplication. In this postdoctoral project we would like to pose the question, what comes next? Specifically we would like to study whether (stochastic) noise could be the computational primitive that the a new generation of AI is built upon. This question will be answered in two steps. First, we will explore theories regarding the computational role of microscopic and system-level noise in neuroscience as well as how noise is increasingly leveraged in machine leaning and artificial intelligence. We aim to establish concrete links between these two fields and, in particular, we will explore the relationship between noise and uncertainty quantification.
Building on this, the postdoctoral researcher will then develop new models that leverage noise to carry out cognitive tasks, of which uncertainty is an intrinsic component. This will not only serve as an AI approach, but should also serve as a computational tool to study cognition in humans and also as a model for specific brain areas known to participate in different aspects of cognition, from perception to learning to decision making and uncertainty quantification.
Perspectives of the postdoctoral project should inform how future fMRI imaging and invasive and non-invasive electrophysiological recordings may be used to test theories of this model. Additionally, the candidate will be expected to interact with other activates in the CEA related to the development of noise-based analogue AI accelerators.

Eco-innovation of insulating materials by AI, for the design of a future cable that is long-lasting, resilient, bio-sourced and recyclable.

This topic is part of a larger upcoming project for the AI-powered creation of a new electrical cable for future nuclear power plants. The goal is to design cables with a much longer lifetime than existing cables in an eco-innovative approach.
The focus is on the cable insulation because it is the most critical component for the application and the most sensitive to aging. The current solution is based on adding additives (anti-rad and antioxidants) to the insulation to limit the effects of irradiation and delay aging as much as possible. However, there is another solution that has never been tested before: self-repairing materials.
The project to which this topic is attached aims to design and manufacture several test model of insulation specimens. With several test characterization protocols, in order to verify the gain in terms of reliability and resilience. The results obtained will begin to fill a future database for the AI platform Expressif, developed at CEA List, which will be used to design the future cable.

Co-design strategy (SW/HW) to enable a structured spatio-temporal sparsity for NN inference/learning

The goal of the project is to identify, analyze and evaluate mechanisms for modulating the spatio-temporal sparsity of activation functions in order to minimize the computational load of transformer NN model (learning/inference). A combined approach with extreme quantization will also be considered.
The aim is to jointly refine an innovative strategy to assess the impacts and potential gains of these mechanisms on the model execution under hardware constraints. In particular, this co-design should also enable to qualify and to exploit a bidirectional feedback loop between a targeted neural network and a hardware instantiation to achieve the best tradeoff (compactness/latency).

LLMs hybridation for requirements engineering

Developing physical or digital systems is a complex process involving both technical and human challenges. The first step is to give shape to ideas by drafting specifications for the system to be. Usually written in natural language by business analysts, these documents are the cornerstones that bind all stakeholders together for the duration of the project, making it easier to share and understand what needs to be done. Requirements engineering proposes various techniques (reviews, modeling, formalization, etc.) to regulate this process and improve the quality (consistency, completeness, etc.) of the produced requirements, with the aim of detecting and correcting defects even before the system is implemented.
In the field of requirements engineering, the recent arrival of very large model neural networks (LLMs) has the potential to be a "game changer" [4]. We propose to support the work of the functional analyst with a tool that facilitates and makes reliable the writing of the requirements corpus. The tool will make use of a conversational agent of the transformer/LLM type (such as ChatGPT or Lama) combined with rigorous analysis and assistance methods. It will propose options for rewriting requirements in a format compatible with INCOSE or EARS standards, analyze the results produced by the LLM, and provide a requirements quality audit.

Thermodynamic modelling of protective coating for solid oxide electrolysis cells

In the pursuit of a sustainable energy future, solid oxide electrolysis cells (SOECs) are a highly promising technology for producing clean hydrogen by electrolysis of water at high temperature (between 500 and 850°C). Although high operating temperature offers many benefits (high efficiency and low power consumption), it can lead to degradation of the interconnectors. Coatings are proposed to improve the long-term performance of interconnectors and reduce corrosion problems. The aim is to find the best coating candidates with high thermodynamic stability, high electrical conductivity and low cation diffusivity. In this context, you will join the LM2T team within the DIADEM Project ( for innovative materials.
Your role will be to:
1)Perform thermodynamic simulations using CALPHAD method and Thermo-Calc Software to predict the stability range of a set of coating candidates (e.g. spinel oxides and perovskites) and the possible decomposition reactions in different atmosphere conditions (temperature and oxygen partial pressure). In this step, the candidate will also perform a critical review of the thermodynamic data available in the literature.
2)To couple information obtained from CALPHAD calculations and the thermodynamic database to estimate the thermal expansion and electrical conductivity of the most promising compositions.
The candidate will work closely with the experimental team (ISAS/LECNA and UMR-IPV) producing the coatings to guide future trials and adapt the method to better meet large-scale production needs.