Search

Physical reservoir computing (PRC) is a computing framework that harnesses the intrinsic dynamics of physical systems for computation. It offers a promising energy-efficient alternative to traditional von Neumann computing for certain tasks, particularly those demanding both memory and nonlinearity. As PRC is implemented across a broad variety of physical systems, the need increases for standardised tools for data processing and model training. In this manuscript, we introduce PRCpy, an open-source Python library designed to simplify the implementation and assessment of PRC for researchers. The package provides a high-level interface for data handling, preprocessing, model training, and evaluation. Key concepts are described and accompanied by experimental data on two benchmark problems: nonlinear transformation and future forecasting of chaotic signals. Throughout this manuscript, which will be updated as a rolling release, we aim to facilitate researchers from diverse disciplines to prioritise evaluating the computational benefits of the physical properties of their systems by simplifying data processing, model training and evaluation., Comment: 19 pages, 5 figures

Strongly-interacting nanomagnetic arrays are ideal systems for exploring reconfigurable magnonics. They provide huge microstate spaces and integrated solutions for storage and neuromorphic computing alongside GHz functionality. These systems may be broadly assessed by their range of reliably accessible states and the strength of magnon coupling phenomena and nonlinearities. Increasingly, nanomagnetic systems are expanding into three-dimensional architectures. This has enhanced the range of available magnetic microstates and functional behaviours, but engineering control over 3D states and dynamics remains challenging. Here, we introduce a 3D magnonic metamaterial composed from multilayered artificial spin ice nanoarrays. Comprising two magnetic layers separated by a non-magnetic spacer, each nanoisland may assume four macrospin or vortex states per magnetic layer. This creates a system with a rich $16^N$ microstate space and intense static and dynamic dipolar magnetic coupling. The system exhibits a broad range of emergent phenomena driven by the strong inter-layer dipolar interaction, including ultrastrong magnon-magnon coupling with normalised coupling rates of $\frac{\Delta \omega}{\gamma} = 0.57$, GHz mode shifts in zero applied field and chirality-selective magneto-toroidal microstate programming and corresponding magnonic spectral control.

Learning and pattern recognition inevitably requires memory of previous events, a feature that conventional CMOS hardware needs to artificially simulate. Dynamical systems naturally provide the memory, complexity, and nonlinearity needed for a plethora of different unconventional computing approaches. In this perspective article, we focus on the unconventional computing concept of reservoir computing and provide an overview of key physical reservoir works reported. We focus on the promising platform of magnetic structures and, in particular, skyrmions, which potentially allow for low-power applications. Moreover, we discuss skyrmion-based implementations of Brownian computing, which has recently been combined with reservoir computing. This computing paradigm leverages the thermal fluctuations present in many skyrmion systems. Finally, we provide an outlook on the most important challenges in this field., Comment: 19 pages and 3 figures

Two-dimensional (2D) van der Waals (vdW) magnets represent one of the most promising horizons for energy-efficient spintronic applications because their broad range of electronic, magnetic and topological properties. Of particular interest is the control of the magnetic properties of 2D materials by femtosecond laser pulses which can provide a real path for low-power consumption device platforms in data storage industries. However, little is known about the interplay between light and spin properties in vdW layers. Here, combining large-scale spin dynamics simulations including biquadratic exchange interactions and wide-field Kerr microscopy (WFKM), we show that ultrafast laser excitation can not only generate different type of spin textures in CrGeTe$_3$ vdW magnets but also induce a reversible transformation between them in a toggle-switch mechanism. Our calculations show that skyrmions, anti-skyrmions, skyrmioniums and stripe domains can be generated via high-intense laser pulses within the picosecond regime. The effect is tunable with the laser energy where different spin behaviours can be selected, such as fast demagnetisation process ($\sim$250 fs) important for information technologies. The phase transformation between the different topological spin textures is obtained as additional laser pulses are applied to the system where the polarisation and final state of the spins can be controlled by external magnetic fields. We experimentally confirmed the creation, manipulation and toggle switching phenomena in CrGeTe$_3$ due to the unique aspect of laser-induced heating of electrons. Our results indicate laser-driven spin textures on 2D magnets as a pathway towards ultrafast reconfigurable architecture at the atomistic level.

The discovery of atomic monolayer magnetic materials has stimulated intense research activities in the two-dimensional (2D) van der Waals (vdW) materials community. The field is growing rapidly and there has been a large class of 2D vdW magnetic compounds with unique properties, which provides an ideal platform to study magnetism in the atomically thin limit. In parallel, based on tunneling magnetoresistance and magneto-optical effect in 2D vdW magnets and their heterostructures, emerging concepts of spintronic and optoelectronic applications such as spin tunnel field-effect transistors and spin-filtering devices are explored. While the magnetic ground state has been extensively investigated, reliable characterization and control of spin dynamics play a crucial role in designing ultrafast spintronic devices. Ferromagnetic resonance (FMR) allows direct measurements of magnetic excitations, which provides insight into the key parameters of magnetic properties such as exchange interaction, magnetic anisotropy, gyromagnetic ratio, spin-orbit coupling, damping rate, and domain structure. In this review article, we present an overview of the essential progress in probing spin dynamics of 2D vdW magnets using FMR techniques. Given the dynamic nature of this field, we focus mainly on broadband FMR, optical FMR, and spin-torque FMR, and their applications in studying prototypical 2D vdW magnets. We conclude with the recent advances in laboratory- and synchrotron-based FMR techniques and their opportunities to broaden the horizon of research pathways into atomically thin magnets., Comment: 42 pages, 16 figures, 3 tables

Physical neuromorphic computing, exploiting the complex dynamics of physical systems, has seen rapid advancements in sophistication and performance. Physical reservoir computing, a subset of neuromorphic computing, faces limitations due to its reliance on single systems. This constrains output dimensionality and dynamic range, limiting performance to a narrow range of tasks. Here, we engineer a suite of nanomagnetic array physical reservoirs and interconnect them in parallel and series to create a multilayer neural network architecture. The output of one reservoir is recorded, scaled and virtually fed as input to the next reservoir. This networked approach increases output dimensionality, internal dynamics and computational performance. We demonstrate that a physical neuromorphic system can achieve an overparameterised state, facilitating meta-learning on small training sets and yielding strong performance across a wide range of tasks. Our approach's efficacy is further demonstrated through few-shot learning, where the system rapidly adapts to new tasks.

Reservoir computing is a neuromorphic architecture that potentially offers viable solutions to the growing energy costs of machine learning. In software-based machine learning, neural network properties and performance can be readily reconfigured to suit different computational tasks by changing hyperparameters. This critical functionality is missing in ``physical" reservoir computing schemes that exploit nonlinear and history-dependent memory responses of physical systems for data processing. Here, we experimentally present a `task-adaptive' approach to physical reservoir computing, capable of reconfiguring key reservoir properties (nonlinearity, memory-capacity and complexity) to optimise computational performance across a broad range of tasks. As a model case of this, we use the temperature and magnetic-field controlled spin-wave response of Cu$_2$OSeO$_3$ that hosts skyrmion, conical and helical magnetic phases, providing on-demand access to a host of different physical reservoir responses. We quantify phase-tunable reservoir performance, characterise their properties and discuss the correlation between these in physical reservoirs. This task-adaptive approach overcomes key prior limitations of physical reservoirs, opening opportunities to apply thermodynamically stable and metastable phase control across a wide variety of physical reservoir systems, as we show its transferable nature using above(near)-room-temperature demonstration with Co$_{8.5}$Zn$_{8.5}$Mn$_{3}$ (FeGe)., Comment: Main manuscript: 14 pages, 5 figures. Supplementary materials: 13 pages, 10 figures

Layered van der Waals (vdW) magnets can maintain a magnetic order even down to the single-layer regime and hold promise for integrated spintronic devices. While the magnetic ground state of vdW magnets was extensively studied, key parameters of spin dynamics, like the Gilbert damping, crucial for designing ultra-fast spintronic devices, remains largely unexplored. Despite recent studies by optical excitation and detection, achieving spin wave control with microwaves is highly desirable, as modern integrated information technologies predominantly are operated with these. The intrinsically small numbers of spins, however, poses a major challenge to this. Here, we present a hybrid approach to detect spin dynamics mediated by photon-magnon coupling between high-Q superconducting resonators and ultra-thin flakes of Cr$_2$Ge$_2$Te$_6$ (CGT) as thin as 11\,nm. We test and benchmark our technique with 23 individual CGT flakes and extract an upper limit for the Gilbert damping parameter. These results are crucial in designing on-chip integrated circuits using vdW magnets and offer prospects for probing spin dynamics of monolayer vdW magnets.

In a data-driven economy, virtually all industries benefit from advances in information technology -- powerful computing systems are critically important for rapid technological progress. However, this progress might be at risk of slowing down if we do not address the discrepancy between our current computing power demands and what the existing technologies can offer. Key limitations to improving energy efficiency are the excessive growth of data transfer costs associated with the von Neumann architecture and the fundamental limits of complementary metal-oxide-semiconductor (CMOS) technologies, such as transistors. In this perspective article, we discuss three technologies that will likely play an essential role in future computing systems: memristive electronics, spintronics, and electronics based on 2D materials. We present how these may transform conventional digital computers and contribute to the adoption of new paradigms, like neuromorphic computing., Comment: 28 pages, 7 figures

Proximity-induced long-range spin-triplet supercurrents, important for the field of superconducting spintronics, are generated in superconducting/ferromagnetic heterostructures when interfacial magnetic inhomogeneities responsible for spin mixing and spin flip scattering are present. The multilayer stack Nb/Cr/Fe/Cr/Nb has been shown to support such exotic currents when fabricated into Josephson junction devices. However, creating pure spin currents controllably in superconductors outside of the Josephson junction architecture is a bottleneck to progress. Recently, ferromagnetic resonance was proposed as a possible direction, the signature of pure supercurrent creation being an enhancement of the Gilbert damping below the superconducting critical temperature, but the necessary conditions are still poorly established. Consistent with theoretical prediction, we demonstrate conclusively that pumping pure spin currents into a superconductor is only possible when conditions supporting proximity-induced spin-triplet effects are satisfied. Our study is an important step forward for superconducting pure spin current creation and manipulation, considerably advancing the field of superconducting spintronics., Comment: Main Text pages 1-21, 3 figures. Supplementary 23-34, 11 figures

We experimentally and theoretically demonstrate that nonlinear spin-wave interactions suppress the hybrid magnon-photon quasiparticle or "magnon polariton" in microwave spectra of an yttrium iron garnet film detected by an on-chip split-ring resonator. We observe a strong coupling between the Kittel and microwave cavity modes in terms of an avoided crossing as a function of magnetic fields at low microwave input powers, but a complete closing of the gap at high powers. The experimental results are well explained by a theoretical model including the three-magnon decay of the Kittel magnon into spin waves. The gap closure originates from the saturation of the ferromagnetic resonance above the Suhl instability threshold by a coherent back reaction from the spin waves., Comment: 6 pages

Recently, it has been shown that the chiral magnetic insulator Cu$_2$OSeO$_3$ hosts skyrmions in two separated pockets in temperature and magnetic field phase space. It has also been shown that the predominant stabilization mechanism for the low-temperature skyrmion (LTS) phase is via the crystalline anisotropy, opposed to temperature fluctuations that stabilize the well-established high-temperature skyrmion (HTS) phase. Here, we report on a detailed study of LTS generation by field cycling, probed by GHz spin dynamics in Cu$_2$OSeO$_3$. LTSs are populated via a field cycling protocol with the static magnetic field applied parallel to the $\langle{100}\rangle$ crystalline direction of plate and cuboid-shaped bulk crystals. By analyzing temperature-dependent broadband spectroscopy data, clear evidence of low-temperature skyrmion excitations with clockwise (CW), counterclockwise (CCW), and breathing mode (BR) character at temperatures below $T$ = 40 K are shown. We find that the mode intensities can be tuned with the number of field-cycles below the saturation field. By tracking the resonance frequencies, we are able to map out the field-cycle-generated LTS phase diagram, from which we conclude that the LTS phase is distinctly separated from the high-temperature counterpart. We also study the mode hybridization between the dark CW and the BR modes as a function of temperature. By using two Cu$_2$OSeO$_3$ crystals with different shapes and therefore different demagnetization factors, together with numerical calculations, we unambiguously show that the magnetocrystalline anisotropy plays a central role for the mode hybridization., Comment: 29 pages; main text: pages 1-29 with 7 figures, supplementary materials: pages 23-29 with 3 figures

We present the discovery of a charge density wave (CDW) ground state in heavily electron-doped molybdenum disulfide (MoS$_2$). This is the first observation of a CDW in any $d^2$ (column 6) transition metal dichalcogenide (TMD). The band structure of MoS$_2$ is distinct from the $d^0$ and $d^1$ TMDs in which CDWs have been previously observed, facilitating new insight into CDW formation. We demonstrate a metal-insulator transition at 85 K, a 25 meV gap at the Fermi level, and two distinct CDW modulations, $(2\sqrt{3}\times2\sqrt{3})$R$30^\circ$ and $2\times2$, attributable to Fermi surface nesting (FSN) and electron-phonon coupling (EPC), respectively. This simultaneous exhibition of FSN and EPC CDW modulations is unique among observations of CDW ground states, and we discuss this in the context of band folding. Our observations provide a route toward the resolution of controversies surrounding the origin of CDW modulations in TMDs., Comment: 19 pages, 4 figures, supplementary information

Strongly-interacting artificial spin systems are moving beyond mimicking naturally-occurring materials to emerge as versatile functional platforms, from reconfigurable magnonics to neuromorphic computing. Typically artificial spin systems comprise nanomagnets with a single magnetisation texture: collinear macrospins or chiral vortices. By tuning nanoarray dimensions we achieve macrospin/vortex bistability and demonstrate a four-state metamaterial spin-system 'Artificial Spin-Vortex Ice' (ASVI). ASVI can host Ising-like macrospins with strong ice-like vertex interactions, and weakly-coupled vortices with low stray dipolar-field. Vortices and macrospins exhibit starkly-differing spin-wave spectra with analogue-style mode-amplitude control and mode-frequency shifts of df = 3.8 GHz. The enhanced bi-textural microstate space gives rise to emergent physical memory phenomena, with ratchet-like vortex training and history-dependent nonlinear fading memory when driven through global field cycles. We employ spin-wave microstate fingerprinting for rapid, scaleable readout of vortex and macrospin populations and leverage this for spin-wave reservoir computation. ASVI performs linear and non-linear mapping transformations of diverse input signals as well as chaotic time-series forecasting. Energy costs of machine learning are spiralling unsustainably, developing low-energy neuromorphic computation hardware such as ASVI is crucial to achieving a zero-carbon computational future.

The discovery of an ever increasing family of atomic layered magnetic materials, together with the already established vast catalogue of strong spin-orbit coupling (SOC) and topological systems, calls for some guiding principles to tailor and optimize novel spin transport and optical properties at their interfaces. Here we focus on the latest developments in both fields that have brought them closer together and make them ripe for future fruitful synergy. After outlining fundamentals on van der Waals (vdW) magnetism and SOC effects, we discuss how their coexistence, manipulation and competition could ultimately establish new ways to engineer robust spin textures and drive the generation and dynamics of spin current and magnetization switching in 2D materials-based vdW heterostructures. Grounding our analysis on existing experimental results and theoretical considerations, we draw a prospective analysis about how intertwined magnetism and spin-orbit torque (SOT) phenomena combine at interfaces with well-defined symmetries, and how this dictates the nature and figures-of-merit of SOT and angular momentum transfer. This will serve as a guiding role in designing future non-volatile memory devices that utilize the unique properties of 2D materials with the spin degree of freedom., Comment: 26 pages, 5 figures, 1 table and 1 textbox

Topologically protected nanoscale spin textures, known as magnetic skyrmions, possess particle-like properties and feature emergent magnetism effects. In bulk cubic heli-magnets, distinct skyrmion resonant modes are already identified using a technique like ferromagnetic resonance in spintronics. However, direct light-matter coupling between microwave photons and skyrmion resonance modes has not been demonstrated yet. Utilising two distinct cavity systems, we realise to observe a direct interaction between the cavity resonant mode and two resonant skyrmion modes, the counter-clockwise gyration and breathing modes, in bulk Cu$_2$OSeO$_3$. For both resonant modes, we find the largest coupling strength at 57 K indicated by an enhancement of the cavity linewidth at the degeneracy point. We study the effective coupling strength as a function of temperature within the expected skyrmion phase. We attribute the maximum in effective coupling strength to the presence of a large number of skyrmions, and correspondingly to a completely stable skyrmion lattice. Our experimental findings indicate that the coupling between photons and resonant modes of magnetic skyrmions depends on the relative density of these topological particles instead of the pure spin number in the system., Comment: 11 pages, main text: page 1-8 with 4 figures, supplementary material: page 9-11 with 6 figures

An inhomogeneous magnetic exchange field at a superconductor/ferromagnet interface converts spin-singlet Cooper pairs to a spin-aligned (i.e. spin-polarized) triplet state. Although the decay envelope of such triplet pairs within ferromagnetic materials is well studied, little is known about their decay in non-magnetic metals and superconductors, and in particular in the presence of spin-orbit coupling (SOC). Here we investigate devices in which triplet supercurrents are injected into the s-wave superconductor Nb. In the normal state of Nb, triplet supercurrents decay over a distance of 5 nm, which is an order of magnitude smaller than the decay of spin singlet pairs due to the SOC interacting with the spin associated with triplet pairs. In the superconducting state of Nb, triplet supercurrents are not able to couple with the singlet wavefunction and thus blocked by the absence of available equilibrium states in the singlet gap. The results offer new insight into the dynamics between s-wave singlet and s-wave triplet states., Comment: 19 pages, 5 figures

Electrical control of magnetism of a ferromagnetic semiconductor offers exciting prospects for future spintronic devices for processing and storing information. Here, we report observation of electrically modulated magnetic phase transition and magnetic anisotropy in thin crystal of Cr$_2$Ge$_2$Te$_6$ (CGT), a layered ferromagnetic semiconductor. We show that heavily electron-doped ($\sim$ $10^{14}$ cm$^{-2}$) CGT in an electric double-layer transistor device is found to exhibit hysteresis in magnetoresistance (MR), a clear signature of ferromagnetism, at temperatures up to above 200 K, which is significantly higher than the known Curie temperature of 61 K for an undoped material. Additionally, angle-dependent MR measurements reveal that the magnetic easy axis of this new ground state lies within the layer plane in stark contrast to the case of undoped CGT, whose easy axis points in the out-of-plane direction. We propose that significant doping promotes double-exchange mechanism mediated by free carriers, prevailing over the superexchange mechanism in the insulating state. Our findings highlight that electrostatic gating of this class of materials allows not only charge flow switching but also magnetic phase switching, evidencing their potential for spintronics applications.

Recent ferromagnetic resonance experiments and theory of Pt/Nb/Ni8Fe2 proximity-coupled structures strongly suggest that spin-orbit coupling (SOC) in Pt in conjunction with a magnetic exchange field in Ni8Fe2 are the essential ingredients to generate a pure spin supercurrent channel in Nb. Here, by substituting Pt for a perpendicularly magnetized Pt/Co/Pt spin-sink, we are able to demonstrate the role of SOC, and show that pure spin supercurrent pumping efficiency across Nb is tunable by controlling the magnetization direction of Co. By inserting a Cu spacer with weak SOC between Nb and Pt/(Co/Pt) spin-sink, we also prove that Rashba type SOC is key for forming and transmitting pure spin supercurrents across Nb. Finally, by engineering these properties within a single multilayer structure, we demonstrate a prototype superconducting spin-wave (SW) device in which lateral SW propagation is gateable via the opening or closing of a vertical pure spin supercurrent channel in Nb., Comment: 28 pages, 4 figures

We report Abrikosov vortex nucleation in Pt/Nb/Ni80Fe20/Nb/Pt proximity-coupled structures under oblique ferromagnetic resonance (FMR) that turns out to be detrimental to superconducting spin pumping. By measuring an out-of-plane field-angle {\theta}H dependence and comparison with Pt-absent control samples, we show that as {\theta}H increases, the degree of enhancement (suppression) of spin pumping efficiency in the superconducting state for the Pt-present (Pt-absent) sample diminishes and it reverts to the normal state value at {\theta}H = 90{\deg}. This can be explained in terms of a substantial out-of-plane component of the resonance field for the Ni80Fe20 layer (with in-plane magnetization anisotropy and high aspect ratio) that approaches the upper critical field of the Nb, turning a large fraction of the singlet superconductor volume into the normal state., Comment: 21 pages, 5 figures, 1 table

We investigate the influence of Meissner screening and trapped magnetic flux on magnetization dynamics for a Ni80Fe20 film sandwiched between two thick Nb layers (100 nm) using broadband (5-20 GHz) ferromagnetic resonance (FMR) spectroscopy. Below the superconducting transition Tc of Nb, significant zero-frequency line broadening (5-6 mT) and DC resonance field shift (50 mT) to a low field are both observed if the Nb thickness is comparable to the London penetration depth of Nb films (>= 100 nm). We attribute the observed peculiar behaviors to the increased incoherent precession near the Ni80Fe20/Nb interface and the effectively focused magnetic flux in the middle Ni80Fe20 caused by strong Meissner screening and (defect-)trapped flux of the thick adjacent Nb layers. This explanation is supported by static magnetic properties of the samples and comparison with FMR data on thick Nb/Ni80Fe20 bilayers. Great care should therefore be taken in the analysis of FMR response in ferromagnetic Josephson structures with thick superconductors, a fundamental property for high-frequency device applications of spin-polarized supercurrents., Comment: 23 pages, 5 figures

The interplay between spin, charge, and orbital degrees of freedom has led to the development of spintronic devices like spin-torque oscillators, spin-logic devices, and spin-transfer torque magnetic random-access memories. In this development spin pumping, the process where pure spin-currents are generated from magnetisation precession, has proved to be a powerful method for probing spin physics and magnetisation dynamics. The effect originates from direct conversion of low energy quantised spin-waves in the magnet, known as magnons, into a flow of spins from the precessing magnet to adjacent normal metal leads. The spin-pumping phenomenon represents a convenient way to electrically detect magnetisation dynamics, however, precessing magnets have been limited so far to pump pure spin currents, which require a secondary spin-charge conversion element such as heavy metals with large spin Hall angle or multi-layer layouts to be detectable. Here, we report the experimental observation of charge pumping in which a precessing ferromagnet pumps a charge current, demonstrating direct conversion of magnons into high-frequency currents via the relativistic spin-orbit interaction. The generated electric current, differently from spin currents generated by spin-pumping, can be directly detected without the need of any additional spin to charge conversion mechanism and amplitude and phase information about the relativistic current-driven magnetisation dynamics. The charge-pumping phenomenon is generic and gives a deeper understanding of the recently observed spin-orbit torques, of which it is the reciprocal effect and which currently attract interest for their potential in manipulating magnetic information. Furthermore, charge pumping provides a novel link between magnetism and electricity and may find application in sourcing alternating electric currents., Comment: 3 figures