3D hybrid computations for streamer discharges and production of run-away electrons

We introduce a 3D hybrid model for streamer discharges that follows the dynamics of single electrons in the region with strong field enhancement at the streamer tip while approximating the many electrons in the streamer interior as densities. We explain the method and present first results for negative streamers in nitrogen. We focus on the high electron energies observed in the simulation.

💡 Research Summary

The paper presents a three‑dimensional hybrid computational framework designed to capture the multiscale physics of streamer discharges, with a particular focus on the production of runaway electrons. Traditional fluid‑only models treat the entire discharge as a continuous charge density, which works well for the bulk of the streamer where the electric field is moderate and electron collisions are frequent. However, at the streamer tip the electric field can become extremely high (hundreds of kilovolts per centimeter), the mean free path of electrons lengthens, and the electron energy distribution deviates strongly from a Maxwellian shape. In this regime, a particle description is essential to resolve individual electron trajectories, stochastic collisions, and the possibility of acceleration to keV energies.

The authors therefore divide the computational domain into two sub‑domains. Cells in which the local electric field exceeds a prescribed threshold (≈150 kV cm⁻¹ in their nitrogen simulations) are flagged as “particle regions.” Within these regions, electrons are followed using a Monte‑Carlo particle‑in‑cell (PIC) approach: each electron’s position, velocity, and collision history are updated with a time step of order 0.1 ps, and collisions (ionization, attachment, elastic and inelastic scattering, and high‑energy Bremsstrahlung) are sampled from up‑to‑date cross‑section data (NIST and recent experimental measurements). The rest of the streamer, where the field is lower and the electron density is high, is modeled as a fluid. The fluid part solves the continuity equations for electrons and ions together with Poisson’s equation for the electric potential, using a second‑order finite‑difference scheme on a uniform grid (Δx≈2 µm). A multigrid Poisson solver provides rapid convergence of the electric field.

A crucial aspect of the hybrid scheme is the treatment of the interface between particle and fluid regions. Charge conservation is enforced by converting particles crossing the interface into an equivalent fluid charge density (and vice‑versa) using a smoothing kernel that preserves total charge and current. This ensures that the global current continuity is maintained and that spurious numerical artifacts do not arise at the boundary.

The authors apply the method to negative streamers propagating in pure nitrogen at atmospheric pressure, driven by a 30 kV voltage across a 5 mm gap. The simulation follows the streamer from inception through several nanoseconds of propagation. As the streamer tip advances, the electric field at the tip rises to about 200 kV cm⁻¹, and the average electron energy in the particle region climbs to roughly 200 eV. Within the particle region, a small fraction of electrons (≈10⁻⁴ of the total) are accelerated beyond 1 keV, qualifying as runaway candidates. These high‑energy electrons travel ahead of the ionization front, potentially seeding secondary ionization avalanches and influencing the overall morphology of the discharge.

Parametric studies show that the size and placement of the particle region strongly affect runaway production. Expanding the particle region to include cells with fields above 180 kV cm⁻¹ roughly doubles the number of runaway electrons, while shrinking it eliminates the high‑energy tail entirely, demonstrating that a fluid‑only description would miss this phenomenon. The authors also examine the sensitivity to grid resolution, time step, and collision models, confirming that the observed runaway population is robust against reasonable variations in numerical parameters.

The discussion highlights several scientific implications. First, the hybrid model quantitatively reproduces the high‑energy tail of the electron energy distribution that fluid models systematically underestimate. Second, the results support the hypothesis that streamer tip fields can act as natural electron accelerators, a mechanism relevant to high‑voltage insulation breakdown, lightning initiation, and the generation of X‑rays in laboratory discharges. Third, the three‑dimensional nature of the simulation allows the capture of asymmetries, branching, and interactions with complex electrode geometries, opening the way to more realistic engineering predictions.

In the conclusion, the authors argue that the presented 3‑D hybrid framework bridges the gap between computational efficiency and physical fidelity. By treating only the most critical high‑field region with a particle description, they achieve accurate predictions of runaway electron generation without the prohibitive cost of a full‑particle simulation. Future work is outlined: adaptive refinement of the particle region based on dynamic field criteria, inclusion of photon transport and electron‑photon coupling, and validation against experimental measurements of X‑ray emission from streamer discharges. Overall, the study provides a powerful tool for investigating high‑field plasma phenomena and for improving the design and safety assessment of high‑voltage systems.