Papers

We study the real-frequency structure of the self-energy in the bilayer Hubbard model, using the dynamical cluster approximation. At half filling, the Mott insulator-band insulator (MI-BI) crossover involves a rearrangement of self-energy poles between the bonding and antibonding bands; these poles cross as the interlayer hopping $t_{\perp}$ is varied. Upon doping, this pole structure produces a band-selective pseudogap and enhances $s^{\pm}$-wave superconductivity. The order parameter is maximized near the MI-BI boundary, where low-energy anomalous self-energy poles develop simultaneously in both bands and cooperatively enhance the pairing. We further show that these self-energy poles can be interpreted as emergent fermionic excitations, offering an enhanced-pairing mechanism in common with the single-layer Hubbard model. The controllability of these poles through $t_{\perp}$ makes the bilayer system an unconventional platform for optimizing strongly correlated superconductivity.

Quasi-two-dimensional (2D) van der Waals (vdW) ferromagnets such as the series Fe$_{3-x}$GeTe$_2$, with a relatively high Curie temperature and robust metallicity, offer an ideal platform for investigating itinerant magnetism in reduced dimensions. Here, we present a comprehensive electron spin resonance (ESR) investigation of single-crystalline almost-stoichiometric Fe$_{3.03 \pm 0.03}$GeTe$_{2}$ across wide ranges of frequencies, temperatures, and magnetic fields to gain quantitative insights into its magnetic anisotropy and spin dynamics. Frequency-dependent ESR measurements establish Fe$_{3}$GeTe$_{2}$ as an easy-axis ferromagnet. Temperature-dependent high-field ESR reveals a large internal field that gradually decreases at higher temperatures. Remarkably, this internal field persists even above $T_\mathrm{C}$, evidencing short-range spin correlations in Fe$_{3}$GeTe$_{2}$. Analysis of spin-wave modes yields a strong uniaxial magnetocrystalline anisotropy $K_{\text{int}} \approx - 5 \times 10^6$ erg cm$^{-3}$ at 3 K and a large magnon gap $\Delta(3 \mathrm{K})$ $\approx$ 87.8 $\pm$ 13.7 GHz ($\approx$ 0.363 $\pm$ 0.057 meV). Our study highlights that small variations in Fe content in Fe$_{3}$GeTe$_{2}$ lead to substantial changes in the magnon gap at low temperatures, indicating the extreme sensitivity of spin dynamics to the chemical composition. These results establish Fe$_{3}$GeTe$_{2}$ as a model vdW ferromagnet for exploring tunable anisotropies and magnon excitations in metallic 2D magnets.

Bayesian calibration of material constitutive parameters from multimodal mechanical test data is often limited by the need to specify a joint likelihood across measurement modalities that differ in dimensionality, noise structure, and physical units. The resulting posteriors are often broad or strongly correlated, causing standard Markov Chain Monte Carlo (MCMC) samplers to mix poorly. Here, we present an amortized, likelihood-free framework that combines Gaussian process (GP) surrogates with Conditional Flow Matching (CFM) to learn conditional posteriors over constitutive parameters directly from synthetic multimodal parameter–observation pairs, avoiding hand-crafted likelihoods and repeated MCMC sampling. Once trained, the GP–CFM model generates posterior samples for each new specimen at negligible cost. The utility of this novel approach is demonstrated in this paper by estimating the values of Young's modulus and yield strength from the early portion of the force–displacement ($F$–$D$) curve and a Digital Image Correlation (DIC)-based displacement field measured in a Small Punch Test (SPT). It is observed that the $F$–$D$ data alone produce broad posteriors, consistent with limited parameter discrimination in the global response. Adding the DIC-measured displacement field was seen to contract the posteriors and shift them towards the independently measured tensile reference values. This work establishes a robust likelihood-free framework for the inference of material constitutive parameters from multimodal data, demonstrated through SPT–DIC integration.

We investigate high-dimensional, non-linear dynamical systems when exposed to incoherent periodic inputs and Hebbian-like synaptic plasticity. Our findings reveal a striking phenomenon: depending on the interplay between the strength of the periodic drive and synaptic plasticity, the system's phase diagram can give rise to a region where, once both inputs are removed, the collective dynamics spontaneously settles into a limit cycle. This suggests that periodic drives can imprint lasting rhythmic patterns into the network through plasticity, effectively teaching it to oscillate on its own. Numerical simulations on finite size systems show that the limit cycle phase can be easily detected on single-sample trajectories, while averaged curves are affected by strong finite size effects due to sample-to-sample fluctuations of the period of the limit cycles.

We introduce a data-efficient neural-vision assisted method to automate crystallographic alignment during molecular beam epitaxy (MBE) growth. Trained on reflection high-energy electron diffraction (RHEED) patterns from only 15 CdTe structures, our model - enabled by physics-aware postprocessing - reliably infers crystallographic directions, replacing manual frame-by-frame inspection. To this end, we design, test, and critically compare neural-network architectures based on 2D and 3D ResNet configurations, both with and without postprocessing that leverages the physical constraints of RHEED image acquisition. Our work delivers (i) a fully trained neural system ready for closed-loop deployment in future CdTe growth experiments and (ii) a generalizable pipeline for new materials where access to diverse RHEED datasets is limited. More broadly, this study represents a step toward AI-driven MBE growth and demonstrates the potential of machine-learning-assisted automation in thin-film synthesis.

Long-wavelength Bogoliubov phonons in an ordinary condensate realize the standard acoustic Lorentz metric. We show that a condensate formed in a hyperbolic polariton band realizes a different collective geometry with an indefinite Bogoliubov metric whose spatial signature is inherited from the opposite signs of the band curvatures. This metric converts the acoustic light cone into a hyperbolic stability wedge, separating propagating quasiparticles from dynamically unstable ones. In a driven-dissipative condensate, gain saturation turns this metric relation into a non-Hermitian Bogoliubov dispersion. The zero-discriminant surface becomes an exceptional cone in the parametric space $(q_x,q_y,\Delta_{\rm NH})$, appearing as an exceptional hyperbola at fixed gain saturation. Across this surface the Bogoliubov branches coalesce, the biorthogonal phase rigidity collapses, and the spectrum changes from propagation to amplified or overdamped dynamics. Our results identify hyperbolic polariton condensates as a controllable setting where non-Hermitian exceptional degeneracies are organized by an effective Bogoliubov metric.

We introduce a general framework to solve a class of combinatorial opti- mization problems, including the matching problem, the Traveling Salesman Problem, and also the minimum weight k-factor problem. By reformulating these problems as an arrangement model, we recast the optimization task into a grand-canonical ensemble, where chemical potentials are used to re- lax strict topological constraints. The analytical solution found can serve as a polynomial-time algorithm to compute an approximate minimum cost for arbitrary k and link-weight distributions. Our framework is complementary to existing approaches and reveals new connections between combinatorial optimization and the statistical physics of disordered systems.

An upstream neutral mode can influence quasiparticle tunneling even when electric charge propagates only downstream. We propose a three-path fractional quantum Hall interferometer that probes this directional structure through the flux dependence of average terminal currents. Interference between direct and two-step tunneling first contributes at cubic order in the tunneling amplitudes of three weak quantum point contacts. In the local theory, equal drain voltages separate the downstream and upstream responses: for a purely downstream tunneling excitation, the cubic fundamental Aharonov-Bohm harmonic vanishes at one drain, while the other remains bright. Averaging measurements with exchanged unequal drain voltages extends this separation to a configuration in which all contacts are biased. It requires independent calibration of the electrostatic phase shift and fixed or known tunneling magnitudes. For a selected Abelian excitation with two-point correlation exponent $0<\Delta_{\boldsymbol{\ell}}<1$, the common low-temperature bias power and directional amplitude ratio determine its scaling dimension and principal exchange angle. We also show that a weak static density interaction spanning two tunneling points can activate the dark harmonic without adding an upstream mode. In the high-bias static limit, the first-order Laughlin activated amplitude scales as $E^{2\nu-1}$, where $E$ is the source-drain bias energy and $\nu$ is the filling factor. In the weak localized single-mode infrared regime, the bright harmonic retains its leading scaling and can serve as a reference in the same device. The calculation assumes one coherent tunneling species and common-temperature edge correlations with unresolved propagation times.

We study the metrological properties of a continuously monitored Kitaev chain in the presence of imperfect detection. The system is conditioned on a no-click record, while each emitted fermion is registered only with probability $0\leq q\leq 1$. Because the conditional dynamics remains Gaussian, the steady state is fully characterized by the fermionic correlation matrix. This allows a direct evaluation of the quantum Fisher information and of the mean Uhlmann curvature. For perfect detection, the monitored steady state retains a singular critical structure and the quantum Fisher information with respect to the chemical potential becomes super-extensive. For any $q<1$, imperfect detection introduces a finite smoothing length that rounds the singularity and restores extensive scaling. The detector efficiency behaves instead as a compatible mixed-state estimation parameter, as signaled by the vanishing mean Uhlmann curvature. These results show that incomplete trajectory information destroys the metrological enhancement associated with monitored criticality through a mechanism that differs from ordinary thermal smearing.

Finite-time driving of stochastic systems generates excess dissipation, causing the evolving probability distribution to lag behind the instantaneous equilibrium, and consequently degrading the convergence of nonequilibrium free energy estimators based on the Jarzynski equality. Escorted free energy simulations address the non-adiabatic lag by engineering control fields $\mathbf{u}$ that eliminate the lag, enforcing the trajectory-wise equality $\mathcal{W}_\mathbf{u} = \Delta \mathcal{F}$, and yielding zero-variance estimators. However, constructing the escorting field in closed form remains a challenge, approached variously through flow-field methods, targeted free energy perturbation, or learned diffeomorphisms. In this work, we construct a complementary numerical framework based on gauge-type transforms instead of generalized coordinate transforms for perfect escorting based on the exact spectral decomposition of the time-dependent Fokker-Planck generator. The biorthogonal decomposition of the Liouville operator directly yields a counterdiabatic correction whose action on the instantaneous equilibrium distribution exactly cancels the non-adiabatic lag at arbitrary driving speed in formal analogy with shortcuts-to-adiabaticity techniques such as Berry's transitionless driving for quantum systems. Numerical verification for simulations of an overdamped particle in a time-varying double-well potential and harmonic traps confirms that the counterdiabatic condition is satisfied to machine precision, with the non-adiabatic lag suppressed by roughly twelve orders of magnitude in total variation distance and sixteen orders in KL divergence relative to the unescorted dynamics. As a diagnostic, we demonstrate vanishing dissipated work $\mathcal{W}_{\text{diss}}(t) \approx 0$ for the deterministically propagated Fokker-Planck density across all protocol speeds.

The study of electron hydrodynamics provides a powerful framework for understanding transport in ultraclean conductors, yet experimental evidence has thus far been largely restricted to the linear-response regime. Here, we report the direct observation of a strongly nonlinear transport regime in a high-mobility two-dimensional electron system. By engineering devices with a Venturi-shaped wedge geometry specifically designed to enhance convective nonlinearities, we uncover a pronounced nonlinear voltage response and large diodicity in the current-voltage characteristics. Our experimental findings show quantitative agreement with a theoretical model that attributes the observed nonlinearity to the convective acceleration of the electron fluid, analogous to the Bernoulli effect. These results provide compelling evidence for the applicability of the hydrodynamic framework to two-dimensional electron transport and open new avenues for exploring nonlinear and preturbulent phenomena in solid-state systems.

As an homage to the memory of Hans Willem Capel, I present some personal reminiscences and thoughts that come to my mind in this special occasion.

Two-dimensional topological insulators (2DTIs) harbor spin-polarized edge states that are topologically protected by time-reversal symmetry against non-magnetic structural disorder. However, coupling to magnetic impurities breaks this symmetry, inducing backscattering and destroying perfect quantization. While the impact of isolated dilute magnetic impurities is well understood, the transport properties in the presence of dense, disordered ensembles of magnetic moments remain poorly understood. In this work, we develop an analytical framework, supported by extensive numerical simulations, that captures the behavior of edge transport in two-dimensional topological insulators (2DTIs) with a finite concentration of magnetic impurities. We predict the onset of Anderson localization and uncover an anomalous localization regime characterized by a sub-exponential decay of the conductance, scaling as $\ln {\cal G} \propto -\sqrt{L}$, where $L$ is the system length. Furthermore, we demonstrate that the transport exhibits a universal scaling behavior governed solely by the effective impurity concentration. Applying our model to Mn-doped HgTe quantum wells, we find excellent agreement with experimental data. These findings provide a theoretical foundation for understanding anomalous localization phenomena in magnetically doped topological phases.

Our work explores large-area CVD graphene as a scalable platform for photodetectors. We show that substrate-induced photogating in graphene/SiO$_2$/Si structures enables a photoconductivity signal in mono-, bi-, and trilayer CVD graphene. The devices respond to illumination with photon energies above the Si band gap and do not respond to telecom frequency, suggesting the photoconductivity mechanism through the photogating. Large sample area and gate voltage-dependent resistivity measurements allow us to prove the dominant role of the photogating directly, demonstrating similarity of the capacitive recharging current and the photoresponse. The sensitivity is the highest (995 A/W) for the monolayer graphene. Increasing the number of graphene layers reduces the sensitivity due to screening and parallel conduction, but can improve detectivity by lowering the noise level. The sensitivity of CVD graphene device depends on fabrication route of photolitography: device with metal contacts fabricated before graphene transfer show higher photoresponce. Thus, our work demonstrate the potential of industrially relevant CVD graphene for sensitive photogating-based photodetectors.

Among the other non-noble metals, lead (Pb) is a material of particular interest for plasmonic applications in the deep ultraviolet spectral region. However, experimental studies on its plasmonic performance have not yet been conducted. In this work, the dependence of the optical properties of spherical lead nanoparticles on their diameter is demonstrated. The plasmonic performance of chemically synthesized lead nanoparticles is evaluated at the single-particle level by means of a combination of scanning transmission electron microscopy and electron energy loss spectroscopy. Our findings demonstrate that these nanoparticles support localized surface plasmon resonances across the entire spectrum from near-infrared to deep-ultraviolet. This range was identified as the most extensive among all plasmonic elemental metals, extending to wavelengths below 200 nanometers. Consequently, lead nanoparticles exhibit stable plasmonic performance over a remarkably broad wavelength range, thereby substantiating their potential as a multispectral plasmonic platform.

In this work, we introduce a class of hybrid-parity sliding multiferroics in which the spontaneous ferroelectric polarization is coupled to certain nonrelativistic spin splitting components through interlayer sliding, allowing these components to be reversibly switched in an electrical way. Symmetry analysis identifies coplanar magnets as natural platforms for realizing this form of sliding multiferroicity. First-principles calculations establish bilayer VBr$_2$ as a representative example, demonstrating the coupled reversal of the out-of-plane ferroelectric polarization and the signs of both even- and odd-parity nonrelativistic spin splitting components via an interlayer-sliding pathway. The signs of these nonrelativistic spin splitting components are locked to the sliding-switchable ferroelectric polarization and encoded in the spin-current responses, providing a signature of the coupled ferroic switching. Our findings expand the scope of sliding multiferroics and the functionality of sliding ferroelectrics for low-energy, nonvolatile logic devices.

Large ecosystems balance competition and cooperation, yet standard generalized Lotka–Volterra models make mutualism destabilizing by amplifying disorder and driving unbounded growth. We show that Monod-like saturation resolves this paradox: dynamical mean-field theory and random-matrix analysis reveal a broader stable phase and enhanced survival. Network architecture provides a second control mechanism, but nestedness offers no intrinsic stability advantage. Instead, it is a byproduct of degree distributions with high connectivity necessary for stability.

Heterogeneous density functional theory (DFT) calculations, particularly plane-wave implementations, introduce systematic formation energy errors ranging from tens to hundreds of meV/atom, depending on the selection of exchange-correlation functionals, kinetic energy cutoffs, pseudopotentials, and dispersion corrections. As demonstrated by the MatPES dataset, identical structures can exhibit an average energy discrepancy of 107 meV/atom between PBE and r2SCAN calculations. Such method-dependent discrepancies hinder the integration of multi-source DFT data, greatly limiting the scale and quality of datasets for training robust materials AI models. Here, we resolve this fundamental data silo barrier via graph-based transfer learning. Leveraging 380,190 structurally paired PBE-r2SCAN entries from the MatPES database, we train a structure-aware graph neural network to predict cross-functional energy residuals and align inconsistent DFT energy scales. By adopting GPTFF model architecture, the model converts conventional PBE energies to r2SCAN-level accuracy with a mean absolute error of 14.3 meV/atom, compared with 18.2 meV/atom achieved by CHGNet. This versatile approach effectively upgrades massive legacy PBE datasets to high-precision r2SCAN standards. It enables reliable predictions of phase stability, battery voltage profiles, and reaction thermodynamics, while allowing the integration of multi-source DFT data to advance the development of high-performance materials foundation models.

We develop from first principles the hydrodynamics of a two-dimensional chiral fluid, i.e. one carrying a net microscopic angular-momentum (spin) field. Enforcing angular-momentum conservation without imposing stress-tensor symmetry, we derive the full form of the stress tensor and of the spin flux, and we show that the entire chiral response is generated from the classical Newtonian one by a single operation of direct physical origin — the $90^{\circ}$ rotation through which chirality acts, mirrored at the particle level by transverse forces (Caprini & Marini Bettolo Marconi 2025). Applied to the irreducible (deviatoric) decomposition, the rotation assigns to each classical channel a chiral partner — pressure to chiral pressure, bulk and shear viscosities to their odd counterparts, the spin-flux gradient to its rotated image — one coefficient and one mechanical action per channel, with no further cross-couplings. In this representation the steady base states become elementary. Quiescent states are organised by a holomorphic chiral complex potential, the mechanical and chiral pressures forming a conjugate harmonic pair subject to a topological existence condition; inhomogeneous activity forces azimuthal flows; and a boundary-driven confined flow, the chiral Stokes cavity, obeys a modified Helmholtz–Poisson system, solved in closed form in a circular domain and numerically in a square one. A single dimensionless group controls both geometries and sets the crossover from a screened, single-vortex regime to a sequence of sign-reversing vortical structures. The theory yields quantitative predictions, amenable to direct comparison with experiments on air-fluidised chiral disks (López-Castaño et al. 2022), and recovers the phenomenological frameworks of the chiral-fluid literature as particular cases.

The vast chemical design space and complex, interdependent design variables make catalyst discovery for targeted properties highly labor- and resource-intensive. Although generative models have emerged as a promising solution, existing approaches are generally limited to single-property conditioning or narrow chemical spaces. Here, we present Catalyst Diffusion Transformer (CatDiT), a unified framework for inverse catalyst design that generates valid and novel structures ranging from intermetallic alloys to oxide surfaces. By learning compressed latent representations, CatDiT enables efficient training and rapid sampling while supporting simultaneous conditioning on adsorbate type, binding energy, and catalyst class. The model provides reliable control of discrete properties and directional control of continuous properties, enriching candidate pools for reaction-specific catalyst discovery. As a representative application, multi-conditional generation for the nitrogen reduction reaction (NRR) yields 28 density functional theory (DFT)-relaxed alloy candidates that satisfy the target activity window and lie above the pure-metal *N-*H scaling line, corresponding to a ~1.5-fold enrichment over the source distribution. These results establish CatDiT as a practical and scalable approach for property-directed catalyst inverse design and targeted catalyst generation.

Non-reciprocal interactions find broad applicability in non-equilibrium and living systems. Their canonical implementation involves asymmetric couplings between two entities, which generally induce spatio-temporal patterns and time-dependent steady states that break time-translational invariance, representing a clear deviation from equilibrium physics. Although their phenomenology is well understood, whether non-reciprocal interactions induce new universality classes, and if so, under what conditions, remains an open question. In the present work, we perform a field-theoretic renormalization group (RG) analysis of the dynamics of two non-reciprocally coupled $n$-vector order parameters possessing a $U(n)$ symmetry, generalizing previous results to order parameters with multiple components, a feature that has been shown to generate novel non-equilibrium critical behavior in certain non-reciprocal systems. To lowest order in $\epsilon = 4-d$, we find that the non-reciprocal coupling is RG-irrelevant, and the critical behavior is governed by the equilibrium fixed point of the Model A universality class of Hohenberg and Halperin with $2n$ vector components, even though the transition is into a non-equilibrium state. Our results demonstrate that non-reciprocity alone may not be sufficient to induce novel universality classes.

Square-net topological materials exhibit various interesting properties arising from the interplay of their structural diversity, topologically protected bands and different possible electronic features, including intrinsic magnetic ordering and superconductivity. However, little is known about the low-dimensional structures of these materials and their properties. Realization of low-dimensional topological square-net materials has the potential of enabling enhancement of their quantum behavior arising from quasi-compact 1D geometry and high surface-to-volume ratio, and their integration into functional devices. This work reports on the synthesis and properties of high-quality, single-crystal nanowires of Si square-net material TaSiAs. The chemical-vapor-transport produces TaSiAs nanowires that are chemically encapsulated with a thin dielectric shell of SiO2, enabling remarkable ambient stability and a pristine surface, which are critical for observing robust topologically protected surface states. Atomic-resolution structural analysis reveals a sharp core-shell interface, and a Si square-net lattice extending along the nanowire axis. These chemically protected nanowires allowed us to observe rich electrical and magnetotransport features, including 4 to 11 times lower room-temperature resistivity than the closest bulk analogues and non-saturating linear magnetoresistance, revealing topologically protected coherent surface transport. First-principles calculations show a wide-range (4 eV) linear band dispersion, alongside different Dirac cones that are protected by either symmorphic (C4) or non-symmorphic symmetries, predicting transport features consistent with our results. The findings demonstrate the unique properties of low-dimensional square-net topological materials and their potential applications, including next-generation interconnects, spintronic devices and quantum computing.

We show that retarded correlators in conformal field theories of scalar primary operators evaluated at finite temperature using $AdS/CFT$ scale anomalously at large light like momenta. The anomalous scaling depends on the curvature of the horizon. Setting the momenta equal to the frequency $\omega$, the retarded correlator for black holes with planar horizons scales as $\omega^{\frac{2}{d+2} (2\Delta - d) }$ as opposed to the scaling behaviour of $\omega^{2\Delta - d}$ expected by dimensional analysis at generic fixed momenta and large frequencies. $\Delta$ is the dimension of the primary and $d$, the number of space-time dimensions. For black holes with spherical and hyperbolic horizons when the frequency squared equals the Casimir along the horizon, the retarded correlator scales as $\omega^{\frac{2\Delta -d}{2} }$. We establish this using exact results in $d=2$, and a numerical analysis of the Heun equation for the $AdS_5$ planar black hole and finally using the WKB approximation in general $d$. The exact result for black holes with hyperbolic horizon as well as the analysis at large $d$ provides additional checks for the anomalous scaling behaviour. Finally we evaluate the anomalous scaling exponent for the stress tensor correlator in all its 3 channels.

Do two Hamiltonians related by a non-invertible transformation necessarily share the same finite-temperature phase diagram? For a cluster-model interpolation $H(s)$ and its Kennedy–Tasaki dual $\tilde H(s)$, the map is a partial isometry and equates only their all-plus-sector partition functions. In one dimension, thermal order is forbidden on both sides, leaving only the common zero-temperature transition at $s=\tfrac12$. In three dimensions, however, the inequivalence is exact already at $s=0$: the cluster model has an analytic paramagnetic free energy, whereas its dual $\mathbb Z_2$ gauge theory has a deconfinement transition at $T_c\approx1.31$. Quantum Monte Carlo shows that the mismatch occupies a finite window of the interpolation, marked by a self-dual frozen wedge on the cluster side and a deconfined dome on the gauge side; once both close the two phase diagrams agree again, coinciding over the entire remaining interval up to the shared trivial endpoint $s=1$.

The persistence of ferromagnetic long-range order in monolayers of the van der Waals semiconductor CrI$_3$ opens new routes for spintronic applications based on two-dimensional quantum magnets. In the fabrication of such devices, the constituent materials inevitably experience anisotropic strain, which modifies their intrinsic electronic properties. At the same time, strain can serve as a powerful tuning parameter, driving the material to desired regimes. While several theoretical studies have investigated the effect of biaxial in-plane strain on CrI$_3$ numerically, experiments are widely limited to the application of hydrostatic pressure. Here, we perform high-resolution magnetostriction experiments on bulk CrI$_3$ samples, and ab-initio-based magnetoelastic calculations, to elucidate the role of uniaxial lattice strain on the magnetic properties. Our data show that magnetostriction in CrI$_3$ is unexpectedly sensitive to surface effects, which enables us to investigate the influence of in-plane and out-of-plane strain separately, in both the bulk ferromagnetic (BFM) phase emerging at $T_{\rm C}=61\,\mathrm{K}$ and the surface antiferromagnetic (SAFM) phase below $T^* \simeq 50\,\mathrm{K}$. In particular, we quantify the uniaxial strain dependence of the surface interlayer coupling $J^{\rm SAFM}_{\perp}$ and the surface spin-flip field $B^*$, which drastically exceed the strain effects in the BFM phase by a factor of $\sim 30$. The large magnetostrictive response allows us to study the magnetoelastic coupling in few-layer CrI$_3$ through experiments on bulk single crystals, without requiring exfoliation.

Understanding how the breaking of reciprocity influences thermal flows and microscopic thermodynamic processes such as refrigeration and energy conversion is of broad current interest. We report experimental measurements of thermal flows and refrigeration in nanomechanical resonator networks in which nonreciprocity is controlled through an optically-induced synthetic magnetic flux. Time-modulated optomechanical interactions allow controlled coupling, gauge fields, and refrigeration processes through nanomechanical frequency conversion. We quantify nonequilibrium heat flows between resonators coupled to dissipative baths of different occupation and image heat and effective temperature in networks through measuring correlations of fluctuations. Synthetic magnetism is shown to imprint chirality on thermal fluctuations in a loop of resonators, leading to pronounced chiral flows with different handedness in distinct frequency bands. We find that the heat flows in a non-equilibrium system are tuned by the synthetic magnetic flux, which redistributes energy in the thermal steady-state. Specifically, we illustrate how nonreciprocity enhances the refrigeration of a resonator in the strong coupling regime, reducing its temperature below the bound that applies to time-reversal symmetric networks. These results experimentally demonstrate the impact of nonreciprocity on thermodynamic machines, and provide new methods to characterize them at the microscopic level.

The quantum metric integral (QMI) in two-dimensional (2D) systems is conventionally bounded from below by the Chern number. For systems with zero Chern number or identically vanishing Berry curvature, however, this bound becomes trivial and provides no useful geometric constraints. Here, we develop a dimension-reduction framework that decomposes the 2D QMI into lower-dimensional components in a nested-loop way. With this method, we establish a nonzero lower bound on the QMI arising from one-dimensional topological obstructions even when the conventional 2D topology is trivial. We explicitly demonstrate this mechanism in a tilted 2D Su-Schrieffer-Heeger model and an anisotropic Wilson-Dirac model with chiral symmetry. The resulting lower bounds of QMI are determined by the quantized Wannier bands along two different directions. We further investigate the quantum geometry in higher-order topological phases following the same strategy. By introducing Wannier-band basis obtained from the nested Wilson loop, we demonstrate that the Wannier-band QMI is bounded from below by the higher-order topological invariant, e.g. the quadrupole moment in Benalcazar-Bernevig-Hughes model. Our results establish nonzero lower bounds on QMI from a dimension-reduction framework, thereby generalizing the fundamental relation between quantum geometry and topology.

We propose a four-terminal, X-shaped chiral Majorana interferometer with a central floating superconducting island, enabling the direct detection of the non-Abelian statistics of Ising anyons via the linear-response DC conductance tensor in charge transport experiments. Here, Ising anyons are realizable as edge vortices nucleated at Josephson line junctions defining the superconducting island, where both edge-vortex and Majorana-fermion tunneling processes can occur. We show that in such a multi-terminal Majorana interferometer, both the vacuum and the fermionic fusion channel for Ising anyons are possible. This is in contrast to two-arm interferometers, where only the vacuum fusion channel is accessible and the DC conductance contribution from edge vortices always vanishes. Using a low-energy effective theory derived via chiral bosonization, we find that in the X-shaped interferometer, the DC conductance tensor is completely isotropic, yielding a non-zero conductance when simultaneous edge-vortex and Majorana tunneling activates the fermionic fusion channel. Apart from conductance oscillations in a gate-tunable charge parameter, which display an offset related to the anyon topological spin, measuring a finite conductance can already provide direct evidence for non-Abelian statistics in this geometry.

Here we show that an autocatalytic reaction network can exhibit dynamic hysteresis as a result of the competition between its intrinsic relaxation time scale and that of the periodic drive. The autocatalytic reaction steps generate bistability in the concentration of the autocatalytic species, and periodic pumping of the product species provides the external drive. Hysteresis arises from the lag between the concentration response of the autocatalytic species and the external periodic drive. The resulting hysteresis-loop area quantifies the extent of the system's hysteretic response. Using the periodically driven Schlögl model as a representative bistable chemical system, we use this loop area to determine how the magnitude of the hysteretic response is controlled by the driving protocol, intrinsic fluctuations, and the size of the system. Varying the driving frequency and the strength of the fluctuations causes a turnover of the hysteresis loop area, whereas the area changes monotonically with driving amplitude and system size. These trends identify the driving protocol and fluctuation strength as primary controls on the magnitude of dynamic hysteresis. In contrast to stochastic resonance, which is typically associated with weak periodic forcing and noise-assisted amplification, dynamic hysteresis can characterize the reaction-system response over a wider range of external control conditions. We further show that the delayed concentration response is mirrored in Shannon entropy and in the total entropy production rate, connecting dynamic hysteresis to information-theoretic and stochastic-thermodynamic measures of irreversibility. Overall, we identify and interpret the role of the controlling factors in chemical dynamic hysteresis and suggest implications for efficient chemical logic gates and eventually, chemical computers.

In their paper published in Science 392, 221 (2026), Widmann and collaborators reported interferometry experiments to explore the emission coherence decay of a two-dimensional polariton condensate generated in an array of coupled resonators. The authors claim evidence of Kardar–Parisi–Zhang (KPZ) universal scaling in the measured spatio-temporal coherence decay. We argue in the following that the data were not properly analyzed. We re-analyze the experimental data acquired both with the square and the triangular lattices for various values of the excitation power. Instead of stretched exponential decays, in the space (time) windows considered in the paper we find that the measured $|g^{(1)}(\delta {\bf r}, \delta t)|$ at $\delta t=0$ ($|\delta {\bf r}| $ close to $0$) rather show Gaussian (exponential) decay for all excitation powers. As a result, using as temporal and spatial exponents $\beta=0.5$ and $\chi=1$, the data for all pump powers are found to collapse onto a single curve, which is not the KPZ scaling function. In particular, we show that the data collapse onto the KPZ scaling function presented in the paper is an artifact stemming from incorrect data normalization. We thus conclude that the main claim of the paper is not justified as the spatio-temporal decays of the coherence over the space-time windows analyzed in the paper are not well described by the KPZ universal behavior.