Chaos-based information processing using spintronic nano-oscillators

Informations projet

CHAOSPIN

N° de convention de subvention: 751344

DOI

10.3030/751344

Projet clôturé

Date de signature de la CE 27 Mars 2017

Date de début 1 Février 2018

Date de fin 31 Janvier 2020

Financé au titre de

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Coût total

€ 185 076,00

Contribution de l’UE

€ 185 076,00

Coordonné par

UNIVERSITE PARIS-SACLAY
France

Periodic Reporting for period 1 - CHAOSPIN (Chaos-based information processing using spintronic nano-oscillators)

Période du rapport: 2018-02-01 au 2020-01-31

Résumé du contexte et des objectifs généraux du projet

ChaoSpin was a frontier research project at the interface between nanomagnetism, spintronics, and nonlinear dynamics. It was motivated by the premise that the rich behaviour of nonlinear systems, in particular chaos, can be leveraged for alternative computing paradigms. The primary objective was to establish the utility and feasibility of the nanocontact vortex oscillator, a nanoscale spintronic device, as a universal building block for chaos-based information processing by demonstrating key technological functionalities, such as random number generation and communication using symbolic dynamics. The underlying idea is that the complexity required for computation and possible cognitive functions can be generated within a single system, without the need of a complex array of interconnected subsystems.

The project addresses the problems arising from the end of “Moore’s law”, a projection that defines the roadmap followed by the semiconductor industry which results in computer processing power being doubled roughly every 18 months. The future of this trend is in question since further miniaturisation of microelectronic components like transistors will not result in commensurate growth in performance, largely due to issues related to energy consumption and device-to-device variability when their dimensions reach the nanometre scale. As such, alternative computing paradigms are actively being explored, such as neuro-inspired and quantum computing. The aim of ChaoSpin is to show that paradigms based on chaotic phenomena could be useful to address these issues.

The project objectives were defined by three scientific questions related to magnetic vortex dynamics on the nanoscale, namely:
• Can the chaotic state be detected experimentally?
• How does the chaotic state respond to external forcing?
• Can the chaotic state be exploited for information processing?
These questions were addressed by combining the use of high-performance simulation tools and quantitative theories with state-of- the-art experiments involving high-frequency electrical characterisation. We succeeded in determining the chaotic nature of the vortex dynamics experimentally. We also showed that the nanocontact vortex oscillator can indeed produce aperiodic patterns, which contain sufficient entropy to be used for random number generation.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

We have performed an extensive experimental study of the nanocontact vortex oscillator dynamics under liquid nitrogen temperatures. The study was conducted primarily in the time domain, where the oscillator signal was acquired for different applied currents and analysed with a novel pattern filtering technique developed. This technique allows for recurring motifs to be identified and reconstructed in a time series, which removes spurious contributions from thermal noise. We showed that commensurate phases of the oscillator, resulting from periodic reversals of the vortex core, result accordingly in periodic repetitions of two basic patterns, while chaos appearing under certain ranges of current manifests itself as an aperiodic sequence of the same patterns. Based on this observation, we devised a scheme to encode binary information in these pattern sequences, which allowed us to quantify the complexity and entropy content of the oscillator signal. We showed that chaotic regime allows true random numbers to be generated at rates of >100 MHz. This behaviour was reproduced in large-scale micromagnetics simulations were the time evolution of the magnetization state in the oscillator was computed under similar conditions to experiment. These results have been disseminated at five international conferences and reported in three journal articles (publications in Physical Review Letters, Physical Review B, and Nature Communications).

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

The experimental results obtained represent progress beyond the state-of-the-art. Previous studies on similar spintronic oscillators have focused on properties such as the spectral characteristics of the oscillations and their capacity for external injection locking or mutual synchronization. Other magnetic systems have also been shown to exhibit chaos, but none exhibit features as simple as well-defined aperiodic patterns that can be exploited for information processing applications. The pattern filtering technique developed is also novel and will allow for more detailed studies in the future. From a broader perspective, these results show that well-controlled chaotic phases in nanoscale spintronic devices could be used for new kinds of applications beyond binary storage and logic. This opens up a new avenue of research for nanomagnetism and spintronics.

Vortex trajectories in the phase space of the dynamical system showing periodic and chaotic states.