My PhD research sits at the intersection of nanomagnetism, spintronics, and unconventional computing. The central object of study is the spin-torque vortex oscillator (STVO) — a nanoscale magnetic device whose rich nonlinear dynamics make it a promising candidate as a physical neuron in brain-inspired hardware.
I work within the Neuromorphic Engineering Group (NEnG) at UCLouvain’s IMCN institute, under the supervision of Prof. Flavio Abreu Araujo and Dr. Tristan da Câmara Santa Clara Gomes.
Spin-Torque Vortex Oscillators
A spin-torque vortex oscillator is a nanopillar device in which a spin-polarized current drives the steady precession of a magnetic vortex core. This produces a microwave-frequency electrical signal whose frequency, amplitude, and phase are highly sensitive to the applied current and external conditions — making STVOs inherently nonlinear and tuneable oscillators.
Their key properties for neuromorphic applications include:
- Nonlinearity — the oscillator responds non-trivially to inputs, enabling complex transformations
- Memory — the vortex core retains information about recent inputs through its phase dynamics
- Tunability — frequency and amplitude can be controlled via dc current or applied field
Neuromorphic Computing with STVOs
The broader goal of my PhD is to harness these properties for spintronic neuromorphic hardware. In reservoir computing, a network of nonlinear dynamical nodes (the “reservoir”) transforms input signals into a high-dimensional representation, which is then read out by a simple linear layer. STVOs serve as ideal physical reservoir nodes.
My work builds on the framework developed in the group, notably the hybrid Thiele Equation Approach (hybrid TEA), which combines the accuracy of full micromagnetic simulations with the computational speed of analytical models. This allows efficient simulation of STVO dynamics for neuromorphic tasks — a tool I use and aim to extend during my PhD.
PhD Work Packages
My thesis is structured around five work packages over four years:
| WP | Title | Period |
|---|---|---|
| WP1 | Analytical modelling and experimental preparation | Year 1–2 |
| WP2 | Experimental measurements and micromagnetic simulations | Year 1–2 |
| WP3 | Hybrid Thiele calibration and experimental validation | Year 2–3 |
| WP4 | Coupling and neuromorphic functionality | Year 3–4 |
| WP5 | Dissemination, open-source tools, and thesis writing | Year 1–4 |
WP1 establishes the analytical framework linking STVO device parameters to measurable observables, including a sensitivity analysis and implementation of a numerical solver.
WP2 covers the experimental characterisation of individual STVOs and the use of MuMax3 micromagnetic simulations to benchmark models against measurements.
WP3 uses the hybrid TEA to bridge the gap between model and experiment, building parametric corrections from multi-device validation data.
WP4 investigates dipolar coupling between STVOs and explores STVO networks as prototypes for neuromorphic circuits.
Methods & Tools
Simulation: MuMax3 (micromagnetic), Python, COMSOL, custom STVO solvers
Circuits & modelling: LTSpice, Eldo, Cadence, Verilog-AMS
Experimental (in training): RF/dc electrical characterisation, magnetotransport measurements
Group & Collaborations
| Partner | Institution | Topic |
|---|---|---|
| Prof. Flavio Abreu Araujo | UCLouvain (NEnG) | PhD supervisor — STVO neuromorphic hardware |
| Dr. Tristan da Câmara Santa Clara Gomes | INESC-MN, Portugal / UCLouvain | PhD co-supervisor — spintronic neuromorphic computing |
The NEnG group collaborates broadly with SPINTEC (CEA/CNRS, Grenoble), INL (Braga), C2N (CNRS, Palaiseau), Bar-Ilan University, and UT Dallas.
Key References
- J. Grollier et al., “Neuromorphic spintronics,” Nature Electronics 3, 360–370 (2020).
- D. Marković et al., “Physics for neuromorphic computing,” Nature Rev. Physics 2, 499–510 (2020).
- J. Torrejon et al., “Neuromorphic computing with nanoscale spintronic oscillators,” Nature 547, 428–431 (2017).
- V. S. Pribiag et al., “Magnetic vortex oscillator driven by dc spin-polarized current,” Nature Phys. 3, 498–503 (2007).
- A. Moureaux et al., “Neuromorphic spintronics accelerated by an unconventional data-driven Thiele equation approach,” arXiv:2301.11025 (2023).
- F. A. Araujo, C. Chopin & S. De Wergifosse, “Ampere–Oersted field splitting of the nonlinear STVO dynamics,” Sci. Rep. 12, 10605 (2022).