Neuromorphic computing offers excellent processing capabilities compared to classical general purpose hardware based on von Neumann architectures for AI models and algorithms. Existing architectures with separate processing and memory blocks are not ideal for energy-efficient learning and inference. For edge applications, it is envisioned that classical von Neumann computing will be replaced by innovative architectures, where memory and processing will be performed in the same locations, the so-called Computing In Memory (CIM). NeuSPIN project pursues the ambition vision of transferring AI algorithms traditionally performed in the cloud to complex on-chip bio-inspired AI hardware with high accuracy and extremely low energy consumption. The project objectives are to develop and deploy cross-disciplinary hardware and software allowing Green AI for the edge computing. It will consist in a combination of new flavors of Non-Volatile Memory Spintronic (NVM) technologies, with novel neurons and synapses designs leading to CIM neural network architectures. Non Volatile Spintronic technologies are a very promising approach for in-memory computing systems due to their efficient implementations. However, implementing edge-AI adapted algorithms remains a serious challenge due to multiple non ideal properties, in particular stochasticity and variabilities. New specific training algorithms such as Bayesian machine learning models adapted to both technology imperfections and neuromorphic designs will be developed and used in the spintronic-based neuromorphic hardware. Finally, the project objectives are not only to deliver cross cutting research on edge AI implementations, but also to create a strong European partnership between two top level, internationally recognized research centers and universities.

This project will identify and exploit novel transport mechanisms in complex antiferromagnets (AFs). We will focus on crystalline, topological, and anomalous origins of the spin Hall effect, spontaneous Hall effect (HE) and their thermal counterparts (Nernst effects), arising in part by the spontaneous symmetry breaking in AFs. The project addresses fundamental questions in an emerging branch of spintronics based on transport phenomena governed by crystal symmetry, topology, and its interplay with AF order. This combination could prove essential for the development of robust large effects vital in new device concepts in the very active field of AF spintronics. In the field of spintronics, the inter-conversion between charge and spin currents has facilitated the progress of fundamental physics and fostered the emergence of new applications. In the past decade, it has emerged that the chief mechanism responsible for spin-charge conversion, the spin HE (SHE), reveals itself under various flavors. Indeed, HEs are either associated to the spin deflection upon extrinsic or intrinsic spin-orbit coupling (“anomalous” HE, AHE), to the non-trivial spin structures (“topological” HE, THE), or to specific atomic arrangements that break time- and spatial-reversal symmetry combinations (“crystal” HE, CHE). AFs are excellent platforms to investigate these different mechanisms, and their thermal (Nernst) counterparts, arising from the interplay of their band and spin structures. Note that the CHE has not yet been observed in experiment, and that crystal Nernst physics has not been addressed at all. Therefore, these effects are a key focus of our proposal. Another key feature of the project lies in the choice of the material (Mn5Si3) that will serve as a versatile platform to investigate the above mechanisms. The metamagnetic phase transition of Mn5Si3 (AF with a chiral spin structure below 65K and collinear above) will be used, e. g., to demonstrate and control the relative contributions of CHE and THE, and their thermal counterparts, respectively. This approach, based on the systematic study of a model system, is key to disentangle the origins of the Hall and Nernst responses. The proposed research program thus responds to the need for new knowledge and experimental designs that are essential to better understand, disentangle, and take advantage of these effects. The key aims of this program are to: - grow high-quality ordered Mn5Si3 thin films; - experimentally observe and theoretically model the CHE, THE and other Hall as well as Nernst effects in Mn5Si3 with different non-trivial spin structures, crystal orientations, and elastic strain; - test the assumed validity and universality of the Mott relation between charge and thermal transport in AFs with non-trivial topology; - investigate the impact on the inverse SHE by non-linear spin fluctuations near magnetic phase transitions, and exploit this effect to probe the magnetic order parameter variations.

In CES 39, MISTRAL takes part of the "Sécurité globale et cybersécurité" topic (8.8) of the ANR AAP2019 work program. More specifically, this project aims to experimentally develop secured schemes to protect objects and embedded systems as listed as key points in "Cybersécurité: liberté et sécurité dans le cyberespace, sécurisation des systèmes d'information, lutte contre la cybercriminalité".This item detailed in the paragraph E.8, "Domaines transverses". Moreover, MISTRAL adresses the topic (5.6), "Modèles numériques, simulation, applications" by leveraging technological solutions of secured embedded systems with MRAM NVM memory and the energy consumption of countermeasures in LWC algorithms. So far connected objects have been designed and deployed with strong cost and power consumption constraints, postponing security to secondary requirements. Recent successful attacks have proved that the security of IoT will become a major and crucial issue. Technical solutions, like Light Weight Cryptographic (LWC) and countermeasures against physical attacks, have to be designed to bridge the gap between security needs and cost constraints. The implementation of such solutions is a key point for both academic and industrial actors. MISTRAL is addressing the security of the cryptography embedded in connected objects at its highest standards while keeping concern by the energy footprint. Consequently, the project aims at proposing innovative research about the MRAM and CMOS hybridization to secure LWC algorithms with a particular focus on the resistance against physical attacks at lowest energetic impact. The proposed methodology and estimated results rely on: - LWC algorithm benchmarking as reference point to compare future results: including the report overhead in terms of silicon and power consumption. - Specifications of countermeasures against fault attacks taking benefits from MRAM/CMOS hybridization properties: Attacks scenarii that can be faced with the help of permanent states stored in the logic will be fully documented. - LWC algorithm designs: CMOS-based circuit as reference, hybridized and embedding MRAM-based coutermeasures: Design these non-volatile strengthening up to `place and route' on 28 nm FDSOI process. This hybridization approach can be built using NV process design kit. This is fully relevant as regard to ecosystems in STT-MRAM that is announced these days. As a result, the proposed circuits will be simulated (electrical, logic) to determine effective robustness of our solution against fault attacks as well as energy footprint compared to a CMOS built-in reference. - Security characterization of the MRAM bitcells: It is mandatory to insure that innovation will not bring new vulnerabilities, or to mitigate these one. The side-channel robustness will be evaluated on identified use cases. The power consumption traces will be estimated by simulation, challenged power analysis-based attacks and compared to the CMOS built-in reference. Vulnerabilities versus fault attacks will be characterized on dedicated samples (STT-MRAM bitcells) manufactured for the purpose of the project. They will be electrically characterized prior and after to any physical attack as Laser or Electromagnetic pulses. Modelization of the effects will be done and included in the simulation flow. Then a hardened STT-MRAM will be fabricated and validated following the same characterization sequence. To further improve this MRAM study, the SOT-MRAM (Spin Orbit Torque) will also be considered for simulations, nanofabrication and characterization.

MNEMOSYN will develop and optimize techniques to grow large-scale two-dimensional (2D) magnetic materials, co-integrated with strong spin-orbit coupling materials, graphene and 2D ferroelectrics in high-quality multilayer van der Waals heterostructures. The main objective is to demonstrate a large power reduction for magnetization actuation by spin-orbit-torques (SOT) and by multiferroic proximity effects. The consortium will benefit from state-of-the-art molecular beam epitaxy equipment provided by three different partners to accelerate the study of various selected materials combinations in a concerted fashion. Experimental developments will be supported and guided by simulations combining first-principles calculations with large-scale simulations of SOT figures of merit on properly designed Hamiltonians. The outcomes of the project will consist in establishing the best growth and integration conditions, the corresponding upper limit of SOTs efficiency, and the demonstration of 2D multiferroicity, targeting room-temperature operation and scalable fabrication of memory building blocks. The feasibility of further fab developments will be also ascertained to achieve higher technology readiness levels (TRL).

The project OISO (OxIde-based SpinOrbitronics) explores the potential of transition metal oxide (TMO) perovskites for SpinOrbitronics. SpinOrbitronics exploits the spin-orbit coupling (SOC) to obtain spin currents without ferromagnets (FM), more efficient torques to switch magnetization and reduced heat dissipation for low power scalable devices. TMO constitute a material platform of structurally well-matched compounds including FMs with low damping coefficients for long-lasting magnetization coherence, materials with large SOC for spin-charge interconversion, compounds for low power data manipulation. To show the potential of TMO-SpinOrbitronics, we will define and test reconfigurable magnonic circuits combining low damping FMs and ferroelectrics or SOC materials to tune spin wave attenuation. They will be uses to design an active phase shifter and a spin wave interferometer for RF electronics and future magnonic logics.