Synchrotron-Regulated Relativistic Magnetohydrodynamic Turbulence: Emission, Polarization, and Faraday Rotation

This paper presents the first relativistic magnetohydrodynamic (MHD) simulations of driven turbulence with self-consistent synchrotron cooling, using three-dimensional simulations with the Athena++ code. The study explores how the interplay between turbulent energy injection and radiative losses regulates plasma thermodynamics and produces observable signatures.

Numerical Setup: The authors implement a turbulence stirring force and synchrotron cooling in ideal special relativistic MHD. Energy is injected at large scales via an Ornstein-Uhlenbeck process, with the stirring force designed to conserve total momentum and avoid direct plasma heating. The synchrotron cooling power assumes an isotropic Maxwell-Jüttner electron distribution, with the cooling efficiency parameter ηsyn varied across three values: 6.4 × 10−2, 10−3, and 10−4, spanning strong- to weak-cooling regimes. Simulations use 10243 cells in a cubic periodic box with initial magnetization σ = 1 and adiabatic index Γ = 4/3.

Key Findings on Turbulence Properties:

1. Temperature regulation: The volume-averaged plasma temperature ⟨Θt⟩V drops as ηsyn increases, following ⟨Θt⟩V ∝ ηsyn(−1/2), consistent with theoretical expectations from Uzdensky (2018). The balance between energy injection and synchrotron cooling maintains a quasi-steady turbulent state.

2. Thermal instability and phase separation: Synchrotron cooling triggers the thermal instability (Simon & Axford 1967; Eilek & Caroff 1979), bifurcating the plasma into hot-dilute and cold-dense phases. In the most cooling-efficient case (ηsyn = 6.4 × 10−2), the lab-frame density distribution displays two peaks at ρ0/5 and 3ρ0. The phase separation is driven by a positive feedback loop: over-dense regions carry stronger magnetic fields due to flux freezing, increasing cooling losses, causing thermal pressure to drop, and drawing in surrounding plasma. Turbulent mixing suppresses the runaway instability, preventing sharp phase separation.

3. Power spectra: The magnetic power spectral density follows a Kolmogorov-like scaling (Ẽ(k) ∝ k(−5/3)) in the inertial range (4 ≲ kL/(2π) ≲ 96), while the kinetic PSD is slightly harder than k(−3/2). Magnetic fluctuations dominate over kinetic ones throughout the inertial range. The cascade is largely insensitive to the plasma thermodynamic state, with different cooling efficiencies producing broadly similar power spectra.

Observational Signatures:

1. Synchrotron spectra and polarization: The total synchrotron intensity exhibits a broad spectrum due to the wide range of plasma temperatures, especially in strongly cooled cases. The linear polarization degree remains at 5-10% at low and intermediate frequencies due to depolarization from randomly oriented magnetic fields, but can reach ≳50% at high frequencies where emission is dominated by rare, hot plasma columns with strong magnetic fields.

2. Spatial and temporal variability: The total intensity and linear polarization degree display stronger spatial and temporal variability at higher frequencies. At low frequencies, emission originates from the entire plasma with intensity close to the mean value; at high frequencies, emission becomes dominated by rare, hot, strongly-magnetized columns, producing broader non-Gaussian PDFs that vary significantly between time snapshots.

3. Faraday rotation: The rotation measure (RM) fluctuates in space and time, with distributions approaching Laplace distributions with exponentially decaying tails. The standard deviation of RM is systematically larger in cooling-efficient turbulence. The RM dependence is mainly controlled by the mean temperature, scaling as S.D.(RM) ∝ ⟨σ⟩V(1/2) × K0(⟨Θt⟩V(−1))/K2(⟨Θt⟩V(−1)). The mean-field contribution to Faraday rotation is comparable to the turbulent contribution, and RM viewed along the mean magnetic field direction is preferentially positive.

Astrophysical Implications: The results show qualitative similarities to observations of fast radio bursts (RM variability on timescales of days), pulsar wind nebulae (spatially inhomogeneous polarization increasing with photon energy), and blazars (stochastic flux and polarization variability with greater variability at higher frequencies). The RM scaling relation quantitatively constrains the Faraday screen temperature; for FRB 121102 with observed RM 1.46 × 105 rad m−2, electron density 100 cm−3, magnetization 0.1, and screen size 1017 cm, the average electron temperature would be ⟨Θt⟩V 3 × 103.

Limitations: The study acknowledges the choice of a fixed driving scheme, lack of explicit resistivity or thermal conduction, and the need for more realistic radiation models including nonthermal electrons, anisotropic electron distributions, two-temperature plasmas, and additional radiative processes such as synchrotron self-absorption and inverse Compton scattering.

Improvements to AI systems based on this paper:

1. Physics-constrained turbulence emulator: Train a neural network to predict the quasi-steady plasma temperature, density phase separation, and magnetic/kinetic power spectra from input parameters (ηsyn, σ, Γ, injection scale). The AI can interpolate across cooling regimes (strong to weak) and magnetization values, enabling rapid parameter sweeps without running expensive 10243 MHD simulations.

2. Subgrid-scale cooling model for coarse simulations: Develop a machine-learned closure that maps resolved turbulent statistics (e.g., local density variance, magnetic field strength) to effective synchrotron cooling rates and phase-separation tendencies. This allows lower-resolution astrophysical simulations (e.g., galaxy clusters, pulsar wind nebulae) to capture the thermal instability and temperature bifurcation without resolving the full inertial range.

3. Synthetic observation generator: Build a generative model that takes simulation snapshots (density, temperature, magnetic field) and outputs synthetic synchrotron intensity maps, polarization degree, and Faraday rotation measure (RM) maps across frequencies. The AI can learn the mapping from plasma state to observable signatures, including the non-Gaussian PDFs and high-frequency hot-column emission, enabling direct comparison with FRB, blazar, and pulsar wind nebula observations.

4. RM variability forecaster: Train a time-series model (e.g., transformer or LSTM) on the simulated RM fluctuations to predict the temporal evolution of RM standard deviation and sign asymmetry given initial conditions. This can be used to forecast FRB RM variability on timescales of days and constrain Faraday screen properties (temperature, density, magnetization) from sparse observational data.

5. Inverse solver for plasma parameters: Develop an AI-based inverse model that, given observed synchrotron spectra, polarization frequency dependence, and RM statistics, infers the underlying turbulent plasma properties (mean temperature, cooling efficiency, magnetization, phase-separation degree). This leverages the paper’s scaling relations (e.g., S.D.(RM) ∝ ⟨σ⟩V(1/2) × K0/K2) and the polarization–frequency trend to provide probabilistic constraints on astrophysical sources.

6. Anomaly detector for non-thermal effects: Use the simulated thermal-cooling baseline to train an anomaly detection system that flags observational deviations (e.g., unexpected spectral hardening, polarization excess) indicative of nonthermal electron populations, anisotropic distributions, or two-temperature effects—phenomena not captured in the current simulations but expected in real sources.

7. Adaptive mesh refinement (AMR) controller: Implement a reinforcement learning agent that dynamically adjusts grid resolution based on local cooling-to-injection ratio and turbulent mixing activity, focusing computational resources on regions prone to thermal instability and phase separation, thereby reducing cost for future simulations with more realistic physics (e.g., resistivity, conduction).

Relativistic magnetized plasmas in many high-energy astrophysical systems are both turbulent and strongly radiative, yet their nonlinear dynamics and radiative outcomes remain poorly understood. Here we present results from three-dimensional driven turbulence simulations in relativistic magnetohydrodynamics with synchrotron cooling. We compute Faraday rotation measures, synthetic synchrotron spectra and linear polarization maps from the simulated turbulence. The balance between energy injection from turbulent driving and synchrotron cooling keeps the plasma, on average, relativistically hot, thereby influencing the rotation measure. Synchrotron cooling triggers the thermal instability and drives the plasma into hot dilute and cold dense phases, which enhances the spatial and temporal variability of synchrotron emission, especially at high frequencies. These diagnostics show qualitative similarities to observations of fast radio bursts, pulsar wind nebulae, and blazars, suggesting that turbulence may play an important role in shaping emission and propagation effects around high-energy sources.