A proof-of-concept for automated AI-driven stellarator coil optimization with in-the-loop finite-element calculations

2026-03-16T13:12:41Z

Finding feasible coils for stellarator fusion devices is a critical challenge of realizing this concept for future power plants. Years of research work can be put into the design of even a single reactor-scale stellarator design. To rapidly speed up and automate the workflow of designing stellarator coils, we have designed an end-to-end ``runner'' for performing stellarator coil optimization. The entirety of pre and post-processing steps have been automated; the user specifies only a few basic input parameters, and final coil solutions are updated on an open-source leaderboard. Two policies are available for performing non-stop automated coil optimizations through a genetic algorithm or a context-aware LLM. Lastly, we construct a novel in-the-loop optimization of Von Mises stresses in the coils, opening up important future capabilities for in-the-loop finite-element calculations.

Measuring impurity-induced shifts in Coulomb crystallization

2026-03-16T08:40:55Z

We report a laboratory measurement of how impurities shift Coulomb crystallization in a strongly interacting ionic system. This is achieved by using laser cooled Ca$^+$ crystals doped with a controlled number of Xe$^{12+}$ highly charged ions. We find that the crystallization threshold is unchanged at low impurity concentration, but shows a clear crossover once the impurity content becomes sufficiently large, after which the shift grows approximately linearly. Complementary measurements reveal that this global effect originates from a local pinning of the crystal around the impurities. We further show how the measured shift could impact standard models of crystallization in white dwarfs and neutron stars. Our results provide an experimental route to incorporating impurity effects into models of multicomponent Coulomb matter, relevant to stellar crystallization and strongly coupled plasmas.

Nonlinear Gaussian process tomography with imposed non-negativity constraints on physical quantities for plasma diagnostics

2026-03-16T07:56:08Z

We propose a novel tomographic method, nonlinear Gaussian process tomography (nonlinear GPT), that uses the Laplace approximation to impose constraints on non-negative physical quantities, such as the emissivity in plasma optical diagnostics. While positive-valued posteriors have previously been introduced through sampling-based approaches in the original GPT method, our alternative approach implements a logarithmic Gaussian process (log-GP) for faster computation and more natural enforcement of non-negativity. The effectiveness of the proposed log-GP tomography is demonstrated through a case study using the Ring Trap 1 (RT-1) device, where log-GPT outperforms existing methods, standard GPT, and the Minimum Fisher Information (MFI) methods in terms of reconstruction accuracy. The results highlight the effectiveness of nonlinear GPT for imposing physical constraints in applications to an inverse problem.

Nonlinear Bayesian Doppler Tomography for Simultaneous Reconstruction of Flow and Temperature

2026-03-16T07:31:11Z

We present a nonlinear Bayesian tomographic framework for Doppler spectral imaging that enables simultaneous reconstruction of emissivity, ion temperature, and flow velocity from line-integrated spectra. The method employs nonlinear Gaussian process tomography (GPT) with a Laplace approximation while retaining the full Doppler forward model. A log-Gaussian process prior stabilizes the velocity reconstruction in low-emissivity regions where Doppler information becomes weak, preventing the unphysical divergence of velocity estimates commonly encountered in conventional spectral tomography. The reconstruction method is verified using synthetic phantom data and applied to coherence imaging spectroscopy (CIS) measurements in the RT-1 device, resolving spatial structures of ion temperature and toroidal ion flow characteristic of magnetospheric plasma in the RT-1 device. The framework extends existing CIS tomography to regimes with strong flows and large temperature variations and provides a general Bayesian approach for Doppler spectral tomography that can be integrated with complementary spectroscopic diagnostics.

Near-Wall Pathways of Anomalous Electron Transport in Hall Thrusters Revealed by 3D PIC Simulations

2026-03-16T05:48:18Z

Cross-field electron transport in Hall thrusters is widely attributed to high-frequency $E\times B$ instabilities, yet its net spatial pathway remains poorly resolved. Here we perform instability-resolving three-dimensional particle-in-cell simulations of a Hall thruster using a boundary-faithful and highly integrated framework. The model incorporates a realistic magnetic-field configuration, self-consistent dielectric wall charging, secondary electron emission, Monte Carlo ionization collisions, a self-consistent continuum neutral-gas evolution model, and an open near-plume outflow treatment. From the strongly oscillatory three-dimensional fields, we extract the net instability-driven transport by time and azimuthal averaging of the correlation term $\langle n_e E_y\rangle$ and the corresponding effective perpendicular mobility. The simulations reveal that anomalous electron transport is not distributed uniformly across the channel cross section. Instead, it self-organizes into persistent near-wall pathways connected to the near-exit region. By comparing conducting-wall, ceramic-wall-with-secondary-emission, and open-outflow closures, we show that the near-wall transport topology is robust, while the boundary treatment mainly redistributes the detailed strength of the pathway and its coupling to the exit and near-plume region. These results demonstrate a previously unresolved spatial organization of instability-driven anomalous transport in Hall thrusters and highlight the unique role of 3D PIC simulations in revealing it.

Long lasting plasma density structures utilizing tailored density profiles

2026-03-15T19:42:38Z

Using fully kinetic Particle In Cell simulations, we investigate the stability and performance of autoresonant plasma beat wave excitation in plasmas with tailored density profiles. We show that a prescribed spatial variation of the background density sustains continuous phase locking between the driving laser beat and the excited plasma mode, thereby enabling precise control of the plasma wave packet shape and group velocity and providing an alternative to frequency chirping of the drive lasers. The density gradient scale is found to govern the nonlinear autoresonant growth, and the attainable saturation amplitude can exceed the classical Rosenbluth Liu prediction and, for appropriate laser intensities, approach the nonrelativistic wave breaking limit. We show that a four laser configuration in a steep parabolic density profile can generate a specially confined two phase quasiperiodic plasma lattice. The generation of such structures may lead to novel applications in plasma photonics.

Reduced One-Fluid GENERIC Closure from Relativistic Moment Kinetics

2026-03-15T16:09:35Z

In this work we derive a reduced one-fluid plasma model from the relativistic Vlasov--Boltzmann--Maxwell system using a moment hierarchy reduction combined with strong-guide-field anisotropic ordering. The unresolved higher-moment sector of the hierarchy is projected onto its dominant slow thermodynamic mode, producing a scalar regulator variable that represents a coarse-grained combination of charge imbalance, pressure anisotropy, and irreversible kinetic production channels. The resulting reduced state vector admits a GENERIC (General Equation for Non-Equilibrium Reversible--Irreversible Coupling) representation in which the reversible sector reproduces reduced electromagnetic field-line dynamics while the irreversible sector governs slow thermodynamic relaxation. Linearization yields a pair of electromagnetic eigenmodes together with an additional real thermodynamic eigenvalue. The fast modes recover the standard gyrotropic cold-plasma response, including familiar limits such as whistler dispersion, while the slow mode drives gradual drift of the effective electromagnetic spectrum and provides a reduced mechanism for variability in relativistic magnetized plasmas with slowly evolving macroscopic equilibrium. The previous 3-field model captured only the thermodynamic slow-mode sector, whereas the fully closed extended GENERIC model also contains the explicit reversible nonneutral charge degree of freedom, whose frozen-thermodynamic limit recovers the nonneutral whistler-Alfven equations as a strict subset. The model is formulated within the GENERIC framework, ensuring consistency with first-principles nonequilibrium thermodynamics.

Inertial-Range Suppression and Ponderomotive Density Cavitation in Broadband Sub-Alfvénic Turbulence under Plasma Sheet Boundary Layer Conditions

2026-03-15T15:27:04Z

Kinetic Alfvén waves (KAWs) are among the most pervasive electromagnetic fluctuations in magnetized astrophysical plasmas, from Earth's magnetospheric boundary layers to the turbulent intracluster medium of galaxy clusters. Their ponderomotive coupling to compressive density fluctuations remains incompletely understood in the broadband turbulent regime. We present two-dimensional pseudospectral simulations of the modified nonlinear Schrödinger--magnetosonic (MNLS--MS) system governing KAW envelopes, initialized with a broadband power-law spectrum ($|ψ(\mathbf{k})|^2\propto k^{-5/6}$) spanning many interacting modes, at $β\sim 0.1$--$0.3$ representative of plasma sheet boundary layer (PSBL) conditions. A fourth-order Runge--Kutta scheme on a $256\times 256$ grid integrates the system to $t = 40$ (normalized), with total energy conserved to within $0.085\%$. The nonlinearity parameter $χ_\mathrm{NL} \approx 0.25$ confirms broadband sub-Alfvénic turbulence throughout. Magnetic field intensity and plasma density develop spatially intermittent, filamentary structures within the first few wave periods, consistent with ponderomotive density cavitation and plasma expulsion from wave-intense regions. The magnetic energy spectra show inertial-range suppression, with a rapid transition from injection ($k < 0.3$) to dissipation without an extended power-law cascade, in agreement with the moderate magnetic Reynolds number ($\mathrm{Re} \sim 250$--$370$) of the simulation and the observationally constrained range for PSBL turbulence. These results provide numerical evidence that broadband KAW turbulence self-organizes into coherent density structures at kinetic scales, and that the spectral character of such turbulence is governed primarily by moderate-Reynolds-number constraints rather than by the wave physics alone.

Strong gradient neoclassical transport in the plateau regime

2026-03-15T15:17:36Z

Strong gradient regions in tokamaks such as the pedestal or internal transport barriers are regions of reduced turbulence where neoclassical transport can play a dominant role. In pedestals, gradient lengths comparable to the ion poloidal gyroradius have been measured. Standard neoclassical theory can miss important strong gradient effects in these regions because it assumes that the gradient length scales of density, temperature and potential are larger than the ion poloidal gyroradius. We extend plateau regime neoclassical theory into regions of gradients of the order of the ion poloidal gyroradius to capture strong gradient effects on transport processes in the pedestal and internal transport barriers. The fundamental idea behind our new framework is to keep a scale separation between the orbit widths and the gradient length scales by performing a large aspect ratio expansion. In the plateau regime, strong gradients cause poloidal variation that is in-out as well as up-down asymmetric. We study two different test cases assuming either radial force balance or the absence of turbulence and show that strong gradient effects can enhance or reduce standard neoclassical theory predictions in the plateau regime in strong gradient regions.

Formation and rising phase of a flux rope through data-constrained simulations

2026-03-14T12:32:28Z

Context. Data-constrained models incorporate observed photospheric magnetic fields. However, due to the lack of magnetic field information in the rest of the solar atmosphere, models rely on extrapolations that, in most cases, neglect the Lorentz force. Nevertheless, this force is present in the lower atmosphere and may play a key role in destabilising the equilibrium configuration and triggering eruptions. Aims. This study seeks to understand and reproduce a solar eruption SOL2014-12-18T21:41 that occurred in active region NOAA 12241, preceded by an M6.9 flare, and to investigate the impact of relaxing the initial force-free assumption. Methods. The resistive and compressible magnetohydrodynamic simulation is initiated using a non-force-free magnetic field extrapolated from a photospheric vector magnetogram taken minutes before the flare. The simulation includes a stratified atmosphere and non-ideal effects such as thermal conduction and radiative cooling. Results. A flux rope forms and rises in the simulation, carrying away dense material from the lower solar atmosphere. Its formation results from the non-zero Lorentz force acting on the initial sheared arcade, without assuming pre-existing flux ropes or photospheric driving motions. The flux rope is then deflected toward regions of low magnetic pressure, escaping the domain at 350 km/s with approximately constant acceleration. Conclusions. A robust numerical framework for modelling flaring active regions was applied to the eruption of NOAA AR12241 as a case study, assuming a realistic non-force-free magnetic field near the flare onset. It exemplifies how an initial Lorentz force imbalance can successfully trigger a flux rope formation that later escapes the simulation domain. It also enables comparison with real observations through the addition of a stratified atmosphere spanning from the photosphere to the corona.

Enhancement of Proton Acceleration via Geometric Confinement in Near Critical Density-filled Targets

2026-03-14T08:45:52Z

High-quality proton beams generated by laser-plasma interactions are of significant interest for applications ranging from tumor therapy to fast ignition in inertial confinement fusion. However, simultaneously achieving high energy coupling efficiency and beam collimation remains a challenge. In this work, we investigate the enhancement of proton acceleration via geometric confinement in Near-Critical Density (NCD) plasma-filled micro-structured targets using two-dimensional particle-in-cell (PIC) simulations. To optimize laser-to-particle energy transfer, we systematically compared various target configurations, such as rectangular tubes, hybrid funnels, and straight cones. Our results reveals that increasing geometric complexity does not necessarily translate to superior acceleration performance. Instead, the relatively simple NCD-filled straight-cone target outperforms more complex hybrid geometries, achieving a maximum proton cutoff energy of 181.7 MeV and a reduced divergence of approximately $12^{\circ}$ at a laser intensity of $5.5 \times 10^{20}$ W/cm$^2$. This enhancement is attributed to the synergistic effect of relativistic laser self-focusing within the NCD channel and the strong spatial confinement of hot electrons by the conical walls. Furthermore, we identify a unique double-peak structure in the temporal evolution of the electron energy, which serves as a signature of sustained electron refluxing. This refluxing mechanism maintains a robust sheath field over an extended duration, driving the superior acceleration. The proposed target design offers a robust pathway for generating high-flux, high-energy proton beams suitable for next-generation high-repetition-rate laser facilities.

An analytical optimization of plasma density profiles for downramp injection in laser wake-field acceleration

2026-03-14T07:17:38Z

We propose and detail a multi-step analytical procedure, based on an improved fully relativistic plane model for Laser Wake Field Acceleration, to tailor the initial density of a cold diluted plasma to the laser pulse profile, so as to control and optimize the partial wave-breaking of the plasma wave and maximize the acceleration of small bunches of electrons self-injected by the first wave-breaking at the density down-ramp, at least in the first stages of their acceleration phase. We find an excellent agreement with the results of Particle In Cell simulations obtained with the same input data.

Experimental Determination of Gamma-Ray Polarization in Strong-Field Nonlinear Compton Scattering

2026-03-13T14:09:57Z

The polarization of gamma rays produced in strong-field quantum electrodynamics (SFQED) is a fundamental and long-standing prediction, the verification of which has remained elusive, limiting both foundational tests and applications. Here, we report the first experimental measurement of gamma-ray polarization generated via all-optical nonlinear Compton scattering. Colliding a laser-wakefield-accelerated electron beam with an intense counter-propagating laser pulse reflected from a plasma mirror, we produce bright gamma rays in the strong-field regime ($a_0 > 1$). For gamma rays with $a_0 \approx 3$, a linear polarization degree of $\sim 50\%$ is measured via the azimuthal asymmetry of photoneutrons from a deuterium target, and independently verified by a Compton polarimeter.The results show excellent agreement with SFQED calculations employing the locally monochromatic approximation, while diverging from predictions based on the locally constant field approximation, highlighting the importance of quantum interference effects in this regime. Our work provides experimental evidence for polarization dynamics in SFQED, supports a key prediction of nonperturbative QED, and paves the way for compact, laser-driven sources of polarized gamma rays.}

On the timescales of controlled termination of tokamak plasmas

2026-03-13T13:16:54Z

The RAPTOR code is used to model how the time required for controlled termination of Ohmic plasmas scales from present tokamaks like TCV and JET, to reactor-grade tokamaks like ITER and DEMO. We show that ramping the plasma current $I_p$ down to 20% of the flat-top value over a time $Δt_{ramp-down}=τ_{LR}=L_i/R$, with internal inductance $L_i$ and resistance $R$ evaluated at flat-top conditions, results in an approximately self-similar peaking of the current density for these four tokamaks, indicating the adequacy of $τ_{LR}$ as a relevant timescale for cross-machine comparison, yielding $τ_{LR} =$ 0.033s (TCV), 2.87s (JET), 63.2s (ITER) and 166.9s (DEMO). Note that $τ_{LR}$ is easy to evaluate, both in systems codes and on a real-time control system. For the simulated ramp-downs with $Δt_{ramp-down}=τ_{LR}$, the end-of-ramp-down normalized internal inductance $\ell_{i3}$ is limited below 2. An $I_p$ ramp-down faster than $τ_{LR}=L_i/R$ requires a reversal of the boundary loop voltage and leads to the formation of a broad plasma layer carrying current in the direction opposite to the total plasma current, concomitant with $\ell_{i3}>2$, a central region with low magnetic shear and strongly peaked pressure profiles. Significant reduction of plasma volume and elongation, as foreseen for ITER and DEMO, is shown to counteract the reversal of current density and the $\ell_{i3}$ increase, while easing vertical stability control, potentially enabling faster $I_p$ ramp-down scenarios. Experimental and theoretical studies should be performed to test the feasibility of such fast termination scenarios, notably with respect to vertical position control, shape control and (resistive) beta limits. A simple analytical model is proposed and applied to estimate $τ_{LR}$ based on 0D engineering parameters for different tokamaks and for different operating points.

Simply Connected Topology in Perturbed Vortices and Field-Reversed Configurations

2026-03-13T10:13:17Z

Zero-helicity vortices, such as Hill's vortex and field-reversed configurations (FRCs), have long been assumed to be toroidal in topology. This paper proves this assumption false: under arbitrarily small odd-parity (with respect to the symmetry axis) transverse field perturbations, interior flux surfaces become simply connected. The previous topological categorization--open and closed field lines separated by an ellipsoid separatrix--is updated to three distinct categories: open field lines in the outermost region, closed field lines on torus flux surfaces in an intermediate region, and closed field lines on simply connected flux surfaces in the innermost region. In addition to a shifted ellipsoid outer separatrix separating closed and open field lines, a new crescent-shaped inner separatrix separates the torus and simply connected surfaces. The simply connected region is significant even for small perturbations; e.g., in a spherical vortex with a perturbation 10% of the background field strength, it occupies 40% of the outer separatrix. The analysis also proves the conjecture regarding field line closure under odd-parity perturbation in the full three-dimensional context. Preliminary numerical simulations of charged particle trajectories in FRC magnetic confinement under odd-parity perturbation were also conducted; crescent-like simply connected volumes were observed even when gyro-radii were small compared to the system size. Since FRCs are sustained by a rotating magnetic field with odd parity, these results motivate a revision of FRC-related fusion confinement physics. Given the mathematical equivalence to Hill's vortex, this also updates our topological understanding of fluid flow in a wide array of phenomena.