arXiv subjects

Johan H. Mentink

Publications and source records attributed to Johan H. Mentink.

At least 19 recordsLinked to original sources

arXiv2026

Exciton-induced magnons carrying orbital angular momentum in CrI3

Magnons are collective spin excitations that contain and transport spin angular momentum in magnetic materials. It has been suggested that they can also carry orbital angular momentum in analogy to the electronic motion around the nucleus. We explore the real-space topology of magnon wave-packets emanating from atomic-like excitons in the ferromagnetic insulator CrI3 and demonstrate the existence of orbital angular momentum in such wave-packets. We reveal that orbital angular momentum of magnons is nearly equal to their spin angular momentum and compensates the latter. This illustrates the existence of an unexplored internal angular momentum balance and demonstrates that the magnetization can be quenched without the need of angular momentum exchange with the lattice.

cond-mat.mtrl-sci↗

arXiv2026

Double-pulse control of all optical magnetization reversal in Tb/Co multilayers

Recent experiments have shown that femtosecond laser pulse with a Gaussian intensity profile can induce magnetization reversal in Tb/Co multilayers with a ring-shaped switching pattern within the laser-irradiated area. Here, we investigate the ultrafast magnetization dynamics leading to such a ring-shaped switching by using double-pulse laser excitation. The laser pulses cause heat-induced quenching and subsequent recovery of the magnetic anisotropy in the multilayers and drive the precessional magnetization switching in the magnetic multilayers. By adjusting the delay between the two pump pulses, we demonstrate that the recovery process can be manipulated and show, experimentally and numerically, that this allows control over the final magnetic domain pattern.

cond-mat.mtrl-sci↗

arXiv2026

Probabilistic Computers for Neural Quantum States

Neural quantum states efficiently represent many-body wavefunctions with neural networks, but the cost of Monte Carlo sampling limits their scaling to large system sizes. Here we address this challenge by combining sparse Boltzmann machine architectures with probabilistic computing hardware. We implement a probabilistic computer on field-programmable gate arrays (FPGAs) and use it as a fast sampler for energy-based neural quantum states. For the two-dimensional transverse-field Ising model at criticality, we obtain accurate ground-state energies for lattices up to 80$\times$80 (6400 spins) using a custom multi-FPGA cluster. Furthermore, we introduce a dual-sampling algorithm to train deep Boltzmann machines, replacing intractable marginalization with conditional sampling over auxiliary layers. This enables the training of sparse deep models and improves parameter efficiency relative to shallow networks. We further implement this algorithm on a single FPGA, demonstrating the training of deep Boltzmann machines for systems as large as $30 \times 30$ (900 spins). Together, these results demonstrate that probabilistic hardware can overcome the sampling bottleneck in variational simulation of quantum many-body systems, opening a path to larger system sizes and deeper variational architectures.

quant-ph↗

arXiv2026

Predicting sampling advantage of stochastic Ising Machines for Quantum Simulations

Stochastic Ising machines, sIMs, are highly promising accelerators for optimization and sampling of computational problems that can be formulated as an Ising model. Here we investigate the computational advantage of sIM for simulations of quantum magnets with neural-network quantum states (NQS), in which the quantum many-body wave function is mapped onto an Ising model. We study the sampling performance of sIM for NQS by comparing sampling on a software-emulated sIM with standard Metropolis-Hastings sampling for NQS. We quantify the sampling efficiency by the number of computational steps required to reach iso-accurate stochastic estimation of the variational energy and show that this is entirely determined by the autocorrelation time of the sampling. This enables predictions of sampling advantage without direct deployment on hardware. Although sampling of the quantum Heisenberg models studied exhibits much longer autocorrelation times on sIMs, the massively parallel sampling of hardware sIMs leads to a projected speed-up of 100 to 10000, suggesting great opportunities for studying complex quantum systems at larger scales.

quant-ph↗

arXiv2025

A fluctuation-free pathway for a topological magnetic phase transition

Topological magnetic textures are particle-like spin configurations stabilized by competing interactions. Their formation is commonly attributed to fluctuation-driven, first-order nucleation processes requiring activation over a topological energy barrier. Here, we demonstrate an alternative barrier- and fluctuation-free pathway for nucleating topological magnetic textures, triggered in our experiments by an excitation-induced spin reorientation transition. By combining x-ray imaging, scattering and micromagnetic simulations, we show that the system follows a deterministic cascade of symmetry-breaking phase transitions after excitation. First, the system undergoes a second-order phase transition from a homogeneous state to weak stripe domains, then a first-order transition to topologically trivial bubbles, and finally a topological switching event into skyrmionic textures. Through simulations, we generalize our findings and demonstrate that this pathway is active in a vast range of low-anisotropy materials. This previously unrecognized, spontaneous transition pathway suggests strategies for rapid, low-energy generation of topological spin textures and points to a general role of intrinsic modulational instabilities in phase transitions beyond magnetism.

cond-mat.mtrl-sci↗

arXiv2025

Instability of explicit time integration for strongly quenched dynamics with neural quantum states

Neural quantum states have recently demonstrated significant potential for simulating quantum dynamics beyond the capabilities of existing variational ansätze. However, studying strongly driven quantum dynamics with neural networks has proven challenging so far. Here, we focus on assessing several sources of numerical instabilities that can appear in the simulation of quantum dynamics based on the time-dependent variational principle (TDVP) with the computationally efficient explicit time integration scheme. Focusing on the restricted Boltzmann machine architecture, we compare solutions obtained by TDVP with analytical solutions and implicit methods as a function of the quench strength. Interestingly, we uncover a quenching strength that leads to a numerical breakdown in the absence of Monte Carlo noise, despite the fact that physical observables don't exhibit irregularities. This breakdown phenomenon appears consistently across several different TDVP formulations, even those that eliminate small eigenvalues of the Fisher matrix or use geometric properties to recast the equation of motion. We provide evidence that the nature of the instability stems from stiffness of the dynamics of the variational parameters, despite the absence of stiffness in the exact quantum dynamics. We conclude that alternative methods need to be developed to leverage the computational efficiency of explicit time integration of the TDVP equations for simulating strongly nonequilibrium quantum dynamics with neural-network quantum states.

quant-ph↗

arXiv2025

Picosecond localization dynamics following ultrafast nanoscale magnetic switching

Ultrashort laser pulses provide the fastest known way to switch magnetic order. Such excitation commonly creates nanometer-scale domains, even after homogeneous illumination when the position of nucleated domains is not externally defined. However, the physics of domain localization during such ultrafast phase transitions remains unresolved. Here, we use shot-resolved pump-probe resonant x-ray scattering together with a material featuring a periodically modulated magnetic anisotropy landscape to track, in real time, the laser-driven nucleation and localization of nanometer-scale spin textures. We find that nucleation and localization are two distinct processes. Nucleation occurs homogeneously via fluctuations at early times, whereas spatially periodic structures emerge only later and, under suitable conditions, localize in less than one nanosecond. Real-space simulations show that this localization is governed by strong lateral variations in spin-texture lifetimes. Our results demonstrate that ultrafast phase-transition dynamics fundamentally differ from conventional transitions, yet still can be controlled through moderate nanometer-scale tailoring of the energy landscape.

cond-mat.mes-hall↗

arXiv2025

Spin-correlation dynamics: A semiclassical framework for nonlinear quantum magnetism

Classical nonlinear theories are highly successful in describing far-from-equilibrium dynamics of magnets, encompassing phenomena such as parametric resonance, ultrafast switching, and even chaos. However, at ultrashort length and time scales, where quantum correlations become significant, these models inevitably break down. While numerous methods exist to simulate quantum many-body spin systems, they are often limited to near-equilibrium conditions, capture only short-time dynamics, or obscure the intuitive connection between nonlinear behavior and its geometric origin in the su(2) spin algebra. To advance nonlinear magnetism into the quantum regime, we develop a theory in which semiclassical spin correlations, rather than individual spins, serve as the fundamental dynamical variables. Defined on the bonds of a bipartite lattice, these correlations are inherently nonlocal, with dynamics following through a semiclassical mapping that preserves the original spin algebra. The resulting semiclassical theory captures nonlinear dynamics that are entirely nonclassical and naturally accommodates phenomenological damping at the level of correlations, which is typically challenging to include in quantum methods. As an application, we focus on Heisenberg antiferromagnets, which feature significant quantum effects. We predict nonlinear scaling of the mean frequency of quantum oscillations in the Néel state with the spin quantum number S. These have no classical analog and exhibit features reminiscent of nonlinear parametric resonance, fully confirmed by exact diagonalization. The predicted dynamical features are embedded in the geometric structure of the semiclassical phase space of spin correlations, making their physical origin much more transparent than in full quantum methods. With this, semiclassical spin-correlation dynamics provide a foundation for exploring nonlinear quantum magnetism.

cond-mat.mes-hall↗

arXiv2025

Multi-value Probabilistic Computing with current-controlled Skyrmion Diffusion

Magnetic systems are highly promising for implementing probabilistic computing paradigms because of the fitting energy scales and conspicuous non-linearities. While conventional binary probabilistic computing has been realized, implementing more advantageous multi-value probabilistic computing (MPC) remains a challenge. Here, we report the realization of MPC by leveraging the thermally activated diffusion of magnetic skyrmions through an effectively non-flat energy landscape defined by a discrete number of pinning sites. The time-averaged spatial distribution of the diffusing skyrmions directly realizes a discrete probability distribution, which is tunable by current-generated spin-orbit torques, and can be quantified by non-perturbative electrical measurements. Even a very straightforward implementation with global tuning, already allows us to demonstrate the softmax computation - a core function in artificial intelligence. As a key advance, we demonstrate invertible logic without the need to create a network of probabilistic devices, offering major scalability advantages. Our proof of concept can be generalized to multiple skyrmions and can accommodate multiple locally tunable inputs and outputs using magnetic tunnel junctions, potentially enabling the representation of highly complex distribution functions.

cond-mat.mtrl-sci↗

arXiv2025

Pushing the Boundary of Quantum Advantage in Hard Combinatorial Optimization with Probabilistic Computers

Recent demonstrations on specialized benchmarks have reignited excitement for quantum computers, yet whether they can deliver an advantage for practical real-world problems remains an open question. Here, we show that probabilistic computers (p-computers), when co-designed with hardware to implement powerful Monte Carlo algorithms, provide a compelling and scalable classical pathway for solving hard optimization problems. We focus on two key algorithms applied to 3D spin glasses: discrete-time simulated quantum annealing (DT-SQA) and adaptive parallel tempering (APT). We benchmark these methods against the performance of a leading quantum annealer on the same problem instances. For DT-SQA, we find that increasing the number of replicas improves residual energy scaling, in line with expectations from extreme value theory. We then show that APT, when supported by non-local isoenergetic cluster moves, exhibits a more favorable scaling and ultimately outperforms DT-SQA. We demonstrate these algorithms are readily implementable in modern hardware, projecting that custom Field Programmable Gate Arrays (FPGA) or specialized chips can leverage massive parallelism to accelerate these algorithms by orders of magnitude while drastically improving energy efficiency. Our results establish a new, rigorous classical baseline, clarifying the landscape for assessing a practical quantum advantage and presenting p-computers as a scalable platform for real-world optimization challenges.

quant-ph↗

arXiv2025

Ultrafast coherent magnon spin currents in antiferromagnets

Generating coherent magnon spin currents with the highest frequencies and shortest wavelengths is a key challenge in ultrafast spintronics and magnonics. A promising route is to excite counter-propagating magnon pairs. In antiferromagnets, such pairs can be accessed in the ultrafast regime, where coherent dynamics are dominated by magnons at the edge of the Brillouin zone. However, it has seemed impossible to generate a net spin current from coherent magnon pairs. Here we show that a coherent superposition of multiple magnon-pair modes can produce such a current in parity-time symmetric antiferromagnets. The ultrafast coherent spin currents are excited with linearly polarized light, with the light polarization steering the current direction. Finally, by superposing two orthogonal spin currents, circular spin currents can be generated, which have not been discussed for steady-state currents.

cond-mat.mes-hall↗

arXiv2025

Revealing domain wall stability during ultrafast demagnetization

The ultrafast control of nanoscale spin textures such as magnetic domain walls or skyrmions is essential for advancing high-speed, high-density spintronics. However, imaging their dynamics will require a technique that combines nanometer spatial and femtosecond temporal resolution. Introducing ultrafast sub-wavelength imaging in the extreme ultraviolet, we track domain wall properties during ultrafast demagnetization in ferro- and ferrimagnetic thin films. We reveal that domain walls remain invariant in position, shape, and width, down to a demonstrated sub-nanometer precision, for up to 50% demagnetization. Stronger excitation causes stochastic nanoscale domain switching. This previously unobservable robustness of laser-excited domain walls highlights the localized nature of photoinduced demagnetization and presents both challenges and opportunities for all-optical magnetic control. The presented technique can be generalized to directly probe nanoscale dynamics in spintronic materials and devices.

cond-mat.mtrl-sci↗

arXiv2025

Effective Theory of Ultrafast Skyrmion Nucleation

Laser-induced ultrafast skyrmion nucleation has been experimentally demonstrated in several materials. So far, atomistic models have been used to corroborate experimental results. However, such simulations do not provide a simple intuitive understanding of the underlying physics. Here, we propose a coarse-grained effective theory where skyrmions can be nucleated or annihilated by thermal activation over energy barriers. Evaluating these two processes during a heat pulse shows good agreement with atomistic spin dynamics simulations and experiments while drastically reducing computational complexity. Furthermore, the effective theory provides a direct guide for experimentally optimizing the number of nucleated skyrmions. Interestingly, the model also predicts a novel pathway for ultrafast annihilation of skyrmions. Our results pave the way for a deeper understanding of ultrafast nanomagnetism and the role of non-equilibrium physics.

cond-mat.mes-hall↗

arXiv2024

Laser-induced helicity and texture-dependent switching of nanoscale stochastic domains in a ferromagnetic film

Controlling magnetic textures at ever smaller length and time scales is of key fundamental and technological interest. Achieving nanoscale control often relies on finding an external stimulus that is able to act on that small length scales, which is highly challenging. A promising alternative is to achieve nanoscale control using the inhomogeneity of the magnetic texture itself. Using a multilayered ferromagnetic Pt/Co/Pt thin-film structure as a model system, we employ a magnetic force microscope to investigate the change in magnetic nanotextures induced by circularly polarized picosecond laser pulses. Starting from a saturated magnetic state, we find stochastic nucleation of complex nanotextured domain networks. In particular, the growth of these domains depends not only on their macroscopic magnetization but also on the complexity of the domain texture. This helicity and texture-dependent effect contrasts with the typical homogeneous growth of magnetic domains initiated by an effective magnetic field of a circularly polarized laser pulse. We corroborate our findings with a stochastic model for the nucleation of magnetic domains, in which the nucleation and annihilation probability not only depends on the helicity of light but also on the relative magnetization orientation of neighboring domains. Our results establish a new approach to investigate ultrafast nanoscale magnetism and photo-excitation across first-order phase transitions.

cond-mat.mtrl-sci↗

arXiv2024

Spontaneous and impulsive stimulated Raman scattering from two-magnon modes in a cubic antiferromagnet

Exchange interactions govern the ordering between microscopic spins and the highest-frequency spin excitations - magnons at the edge of the Brillouin zone. As well known from spontaneous Raman scattering (RS) experiments in antiferromagnets, such magnons couple to light in the form of two-magnon modes - pairs of magnon with opposite wavevectors. Experimental works on two-magnon modes driven by exchange perturbation in impulsive stimulated Raman scattering (ISRS) experiment posed a question about consistency between spin dynamics measured in the ISRS and RS experiments. Here, based on an extended spin correlation pseudovector formalism, we derive the analytical expression for observables in both types of experiments to determine a possibly fundamental differences between the detected two-magnon spectra. We find that in both cases the magnons from the edge of the Brillouin zone give the largest contribution to the measured spectra. However, there is the difference in the spectra which stems from the fact that RS probes population of a continuum of incoherent modes, while in the case of impulsively driven modes, they are coherent and their phase and amplitudes are detected. We show that for the continuum of modes, the sensitivity to the phase results in a relative shift of the main peaks in the two spectra, and the spectrum of the ISRS is significantly broadened and extends to the range above the maximum two-magnon mode frequency. Formally, this is manifested in the fact, that the RS is described by an imaginary part of the Green function only, while the ISRS is described by the absolute value and hence additionally carries information about the real part of the Green function. We further derive two-magnon Raman tensor dispersion and the weighting factors, which define the features of the coupling to light of the modes from the different domains in the Brillouin zone.

cond-mat.str-el↗

arXiv2024

Self-induced Floquet magnons in magnetic vortices

Driving condensed matter systems with periodic electromagnetic fields can result in exotic states not found in equilibrium. Termed Floquet engineering, such periodic driving applied to electronic systems can tailor quantum effects to induce topological band structures and control spin interactions. However, Floquet engineering of magnon band structures in magnetic systems has proven challenging so far. Here, we present a class of Floquet states in a magnetic vortex that arise from nonlinear interactions between the vortex core and microwave magnons. Floquet bands emerge through the periodic oscillation of the core, which can be initiated by either driving the core directly or pumping azimuthal magnon modes. For the latter, the azimuthal modes induce core gyration through nonlinear interactions, which in turn renormalizes the magnon band structure. This represents a self-induced mechanism for Floquet band engineering and offers new avenues to study and control nonlinear magnon dynamics.

cond-mat.mes-hall↗

arXiv2023

Selection rules for ultrafast laser excitation and detection of spin correlations dynamics in a cubic antiferromagnet

Exchange interactions determine the correlations between microscopic spins in magnetic materials. Probing the dynamics of these spin correlations on ultrashort length and time scales is, however rather challenging, since it requires simultaneously high spatial and high temporal resolution. Recent experimental demonstrations of laser-driven two-magnon modes - zone-edge excitations in antiferromagnets governed by exchange coupling - posed questions about the microscopic nature of the observed spin dynamics, the mechanism underlying its excitation, and their macroscopic manifestation enabling detection. Here, on the basis of a simple microscopic model, we derive the selection rules for cubic systems that describe the polarization of pump and probe pulses required to excite and detect dynamics of nearest-neighbor spin correlations, and can be employed to isolate such dynamics from other magnetic excitations and magneto-optical effects. We show that laser-driven spin correlations contribute to optical anisotropy of the antiferromagnet even in the absence of spin-orbit coupling. In addition, we highlight the role of subleading anisotropy in the spin system and demonstrate that the dynamics of the antiferromagnetic order parameter occurs only in next-to-leading order, determined by the smallness of the magnetic anisotropy as compared to the isotropic exchange interactions in the system. We expect that our results will stimulate and support further studies of magnetic correlations on the shortest length and time scale.

cond-mat.str-el↗

arXiv2023

Benchmarking energy consumption and latency for neuromorphic computing in condensed matter and particle physics

The massive use of artificial neural networks (ANNs), increasingly popular in many areas of scientific computing, rapidly increases the energy consumption of modern high-performance computing systems. An appealing and possibly more sustainable alternative is provided by novel neuromorphic paradigms, which directly implement ANNs in hardware. However, little is known about the actual benefits of running ANNs on neuromorphic hardware for use cases in scientific computing. Here we present a methodology for measuring the energy cost and compute time for inference tasks with ANNs on conventional hardware. In addition, we have designed an architecture for these tasks and estimate the same metrics based on a state-of-the-art analog in-memory computing (AIMC) platform, one of the key paradigms in neuromorphic computing. Both methodologies are compared for a use case in quantum many-body physics in two dimensional condensed matter systems and for anomaly detection at 40 MHz rates at the Large Hadron Collider in particle physics. We find that AIMC can achieve up to one order of magnitude shorter computation times than conventional hardware, at an energy cost that is up to three orders of magnitude smaller. This suggests great potential for faster and more sustainable scientific computing with neuromorphic hardware.

cs.ET↗