Accelerating thermo-mechanical simulations using Neural Networks --- Applications to additive manufacturing and metal forming
In multiple industries, such as metal forming and additive manufacturing, the discrepancy between the desired shape and the shape really obtained is significant, which hinders the development of these manufacturing techniques. This is largely due to the complexity of the thermal and mechanical processes involved, resulting in a high computational simulation time.
The aim of this PhD is to significantly reduce this gap by accelerating thermo-mechanical finite element simulations, particularly through the design of a tailored neural network architecture, leveraging theoretical physical knowledge.
To achieve this, the thesis will benefit from a favorable ecosystem at both the LMS of École Polytechnique and CEA List: internally developed PlastiNN architecture (patent pending), existing mechanical databases, FactoryIA supercomputer, DGX systems, and 3D printing machines. The first step will be to extent the databases already generated from finite element simulations to the thermo-mechanical framework, then adapt the internally developed PlastiNN architecture to these simulations, and finally implement them.
The ultimate goal of the PhD is to demonstrate the acceleration of finite element simulations on real cases: firstly, through the implementation of feedback during metal printing via temperature field measurement to reduce the gap between the desired and manufactured geometry, and secondly, through the development of a forging control tool that achieves the desired geometry from an initial geometry. Both applications will rely on an optimization procedure made feasible by the acceleration of thermo-mechanical simulations.
Thermomechanical study of heterostructures according to bonding conditions
For many industrial applications, the assembly of several structures is one of the key stages in the manufacturing process. However, these steps are generally difficult to carry out, as they lead to significant increases in warpage. Controlling stresses and strains generated by heterostructures is however imperative. We proposes to address this subject using both experimental exploration and simulation through thermomechanical studies in order to predict and anticipate problems due to high deformations.
Thermally conductive yet electrically insulating polymer nanocomposite based on core-shell (nano)fillers oriented by magnetic field
Advances in power electronics, electric motors and batteries, for example, are leading to a significant increase in heat production during operation. This increase in power density combined with reduced heat exchange surfaces amplifies the challenges associated with heat dissipation. The absence of adequate dissipation leads to overheating of electronic components, impacting on their performance, durability and reliability. It is therefore essential to develop a new generation of heat dissipating materials incorporating a structure dedicated to this structure.
The objective and innovation of the PhD student's work will lie in the use of highly thermally conductive (nano)fillers that can be oriented in an epoxy resin in a magnetic field. The first area of work will therefore be to electrically isolate the thermally conductive (nano)charges with a high form factor (1D and 2D). The electrical insulation of these charges of interest will be achieved by a sol-gel process. The synthesis will be controlled and optimised with a view to correlating the homogeneity and thickness of the coating with the dielectric and thermal performance of the (nano)composite. The second part will focus on the grafting of magnetic nanoparticles (NPM) onto thermally conductive (nano)fillers. Commercial NPMs will be evaluated as well as grades synthesised in the laboratory. The (nano)composites must have a rheology compatible with the resin infusion process.
Optimisation of advanced mask design for sub-micrometer 3D lithography
With the advancement of opto-electronic technology, 3D patterns with sub micrometer dimensions are more and more integrated in the device, especially on imaging and AR/VR systems. To fabricate such 3D structures using standard lithography technique requires numerous process steps: multiple lithography and pattern transfer, which is time and resource consuming.
With optical grayscale lithography, such 3D structures can be fabricated in single lithography step, therefore reducing significantly the number of process steps required in standard lithography. For high volume manufacturing of such 3D patterns, optical grayscale lithography with Deep-UV (DUV), 248nm and 193nm are the most relevant, as it is compatible with industrial production line. This technique of 3D lithography is however more complex than it seems, which requires advance lithography model and data-preparation flow to design optical mask corresponding to the desired 3D pattern.
Quantum computing with nuclear spins
Nuclear spins in solids are amongst the quantum systems with the longest coherence times, up to minutes or even hours, and as such are attractive qubit candidates; however, controlling and reading out individual nuclear spins is highly challenging. In our laboratory, we have developed a new way to do so. The nuclear spin qubits are interfaced by an electron spin ancilla to which they are coupled by the hyperfine interaction. The electron spin is then measured by microwave photon counting at millikelvin temperatures [1,2]. Nuclear-spin single-shot readout is performed via the electron spin [3], and coherent control is achieved through the use of microwave Raman transitions [4]. The electron spins are Er3+ ions in a CaWO4 crystal, and the nuclear spins are 183W atoms in the matrix, which have a spin 1/2.
[1] E. Albertinale et al., Nature 600, 434 (2021)
[2] Z. Wang et al., Nature 619, 276 (2023)
[3] J. Travesedo et al., arxiv (2024)
[4] J. O'Sullivan et al., arxiv (2024)
Magnetic DIsks as Transducer of Angular Momentum
The proposed topic is a collaborative project to exploit suspended magnetic disks as novel microwave transducers of orbital angular momentum. Our goal is to develop ultra-high fidelity opto-mechanical modulators operating at GHz frequencies by integrating magnetic materials into optical components. This innovative concept arises from recent progress in the study of angular momentum conservation laws by magnon modes in axi-symmetric cavities, leading to new opportunities to develop a more frugal, agile, and sustainable communications technology. Our proposed design has the potential to achieve coherent interconversion between the microwave frequency range in which wireless networks or quantum computers operate and optical network frequencies, which is the optimal frequency range for long-distance communications. In this regard, our proposal not only proposes new applications of magnonics to the field of optics not previously envisioned, but also builds a bridge between the spintronics and the electronic and quantum communities.
In this proposal, the elastic deformations are generated by the magnetization dynamics through the magneto-elastic tensor and its contactless coupling to a microwave circuit. We have shown that coherent coupling between magnons and phonons can be achieved by precisely tuning the magnetic resonance degenerate with a selected elastic mode via the application of an external magnetic field. We expect to achieve ultra-high fidelity conversion by focusing our study on micron-sized single crystal magnetic garnet structures integrated with GaAs photonic waveguides or cavities. In addition, we propose the fabrication of suspended cavities as a means to minimize further energy leakage (elastic or optical) through the substrate.
The first challenge is to produce hybrid materials that integrate high quality garnet films with semiconductors. We propose a radically new approach based on micron-thick magnetic garnet films grown by liquid phase epitaxy (LPE) on a gadolinium-gallium-garnet (GGG) substrate. The originality is to bond the flipped film to a semiconductor wafer and then remove most of the the GGG substrate by mechanical polishing. The resulting multi-layer is then processed using standard lithography techniques, taking advantage of the relative robustness of garnet materials to chemical, thermal or milling processes.
The second challenge is to go beyond the excitation of uniform modes and target modes with orbital angular momentum as encoders of arbitrarily large quanta of nJ? for mode multiplexed communication channels or multi-level quantum state registers. The project will take advantage of recent advances in spin-orbit coupling between azimuthal spin waves as well as elastic scattering of magnons on anisotropic magneto-crystalline tensors. In this project, we also want to go beyond uniformly magnetized state and exploit the ability to continuously morph the equilibrium magnetic texture in the azimuthal direction as a means of engineering the selection rules and thus coherently access otherwise hidden mode symmetries.
Brines for metal recycling
Critical metals are essential for a range of technologies that are vital to reduce our carbon dioxide emissions. However, the global recycling rate for metals contained in electronic waste is below 20%, indicating that electronic waste is a relatively untapped source of metals. Additionally, it is increasingly urgent to develop effective processes for recycling waste from products like solar panels, as the volume of waste solar pannels generated is set to rise significantly in the coming years. Currently, conventional hydrometallurgical methods often rely on toxic aqueous solutions to dissolve metals, which presents substantial environmental challenges.
This project proposes an innovative alternative by using concentrated brines (aqueous salt solutions) to oxidize and dissolve metals. This thesis will investigate the fundamental properties of brines and their ability to dissolve metals through various techniques, particularly electrochemical methods. Artificial intelligence methods developed within the lab will be employed to screen a wide range of brines that would allow metal dissolution. Subsequently, brine-based recycling processes will be developed to recover metals from printed circuit boards and solar panels. Finally, metal separation and the treatment of used brines will be explored using membrane and electrochemical processes.
Modelling spin shuttling in Si and Ge spin qubits
Silicon and Germanium spin qubits have made outstanding progress in the past few years. In these devices, the elementary information is stored as a coherent superposition of the spin states of an electron or hole confined in a quantum dot embedded in a Si/SiO2 or SiGe heterostructure. These spins can be manipulated electrically and are entangled through exchange interactions, allowing for a variety of one- and two-qubit gates required for quantum computing and simulation. Grenoble is promoting original spin qubit platforms based on Si and Ge, and holds various records in spin lifetimes and spin-photon interactions. At CEA/IRIG, we support the progress of these quantum technologies with state-of-the-art modelling. We are, in particular, developing the TB_Sim code, able to describe very realistic qubit structures down to the atomic scale if needed.
Spin shuttling has emerged recently as a resource for spin manipulation and transport. A carrier and its spin can indeed be moved (shuttled) coherently between quantum dots, allowing for the transport of quantum information on long ranges and for the coupling between distant spins. The shuttling dynamics is however complex owing to the spin-orbit interactions that couple the motion of the carrier to its spin. This calls for a comprehensive understanding of these interactions and of their effects on the evolution and coherence of the spin. The aim of this PhD is to model shuttling between Si/Ge spin qubits using a combination of analytical and numerical (TB_Sim) techniques. The project will address spin manipulation, transport and entanglement in arrays of spin qubits, as well as the response to noise and disorder (decoherence). The PhD candidate will have the opportunity to interact with a lively community of experimentalists working on spin qubits at CEA and CNRS.