727 papers
Plasma physics explores the behavior of the fourth state of matter, a superheated soup of charged particles that makes up most of the visible universe. From the fusion power we hope to harness on Earth to the glowing auroras and distant stars above, this field investigates how these energetic gases interact with magnetic fields and light. It is a dynamic area where extreme conditions reveal fundamental laws of nature in ways solid matter never can.
At Gist.Science, we bridge the gap between these complex discoveries and curious minds by processing every new preprint from arXiv in this category. We transform dense, technical research into clear, plain-language explanations alongside detailed summaries, ensuring that breakthroughs in plasma dynamics and fusion energy are accessible to everyone. Below are the latest papers in plasma physics, curated and simplified for your reading.
By combining an analytical framework with numerical simulations using a supercomoving MHD formulation, this study demonstrates that turbulent dynamos in collapsing clouds drive super-exponential magnetic field growth, enabling fields to become dynamically relevant much earlier in star and galaxy formation than previously predicted.
This paper demonstrates that rapidly changing all-dielectric metasurfaces via optically injected free carriers can act as temporal interfaces to generate persistent, nanoscale quasistatic magnetic fields supported by residual circulating currents, alongside frequency-shifted and time-refracted metasurface-guided waves.
This paper introduces SOLPS-NN, a deep-learning surrogate model trained on extensive SOLPS-ITER simulations that utilizes simple fully connected neural networks to efficiently and accurately predict plasma exhaust conditions and detachment access, demonstrating that independent models for specific observables yield higher accuracy and that transfer learning offers no significant advantage over training from scratch.
This paper presents a robust, generalized potential model for ion wake-mediated dust interactions under PK-4 conditions, which uses a minimal set of coefficients derived from molecular dynamics simulations to accurately capture potential distributions across various dust arrangements beyond traditional string-like configurations.
This paper establishes a deep connection between quasisymmetric magnetic fields in stellarators and soliton theory by demonstrating that periodic Korteweg-de Vries soliton potentials generate these fields, a relationship validated through non-perturbative mathematical analysis and machine learning that reveals hidden lower-dimensional symmetries and potential divertor configurations.
This paper introduces TGLF-WINN, a data-efficient deep learning surrogate for turbulent transport modeling in fusion that combines physics-guided feature engineering, wavenumber-resolved regularization, and Bayesian Active Learning to achieve high accuracy with significantly reduced training data requirements while enabling a 45x speedup over the traditional TGLF model.
This paper presents a methodology that leverages differentiable kinetic simulators and plasma phase space data to learn time-dependent and integro-differential collision operators, demonstrating their ability to accurately reproduce complex non-equilibrium plasma dynamics more effectively than traditional particle track statistics.
This Vlasov-Poisson study investigates the thermal effects on Buneman instability, revealing that while the maximum growth rate retains its dependence, it remains largely independent of the temperature ratio, and that ion density inhomogeneity self-consistently governs the efficiency of electron beam energy transfer to bulk plasma heating.
This review argues that a complete understanding of cosmic particle acceleration requires shifting focus from traditional cascade-based descriptions of magnetohydrodynamic turbulence to the central role of self-consistently emerging coherent structures, such as current sheets and flux ropes, as the primary sites of localized energization and dissipation.
This study demonstrates that a convolutional neural network (CNN) trained on MAVEN SWIA ion energy spectra outperforms a multilayer perceptron in accurately and automatically classifying solar wind, magnetosheath, and induced magnetosphere regions around Mars, offering an efficient framework for large-scale plasma analysis.