Magnetic Reconnection Power

Plasmoid-mediated Penrose extraction as a parallel energy channel — compare reconnection power to BZ jet power at the same spin and field (Meringolo et al. 2025)

🔬 Established MHD/GRMHD ✦ Engineering application
The Meringolo, Camilloni & Rezzolla (2025) GRPIC simulations showed that plasmoid instabilities in Kerr magnetospheres generate a reconnection-driven energy extraction channel parallel to the Blandford-Znajek mechanism. At spin a★ > 0.9, this channel contributes ~10–30% of total BZ power. Important: this is distinct from the Comisso & Asenjo (2021) ergospheric reconnection mechanism; the two should not be conflated.
Parameters
8,200 M☉
0.900
10⁶ T
η = 0.15
Power outputs
P_BZ — BZ jet power
P_rec — reconnection power
P_rec / P_BZ
Reconnection fraction of BZ power
0%
P_total = P_BZ + P_rec
r_plus — outer horizon
Ω_H — horizon angular velocity
η_Penrose max (theoretical)

The BZ formula

The Blandford-Znajek (1977) mechanism extracts rotational energy from a Kerr black hole electromagnetically via a large-scale field threading the horizon. In SI units:

r_g = GM/c² (gravitational radius; the Schwarzschild radius is 2r_g)
r_+ = r_g(1 + √(1 − a²)) (outer horizon)
Ω_H = ac/(2r_+) (horizon angular velocity)
P_BZ = (κ/4πμ₀) B² r_+⁴ Ω_H²/c, where κ = 0.044 (numerical factor)

The reconnection formula (Meringolo et al. 2025)

In plasmoid-mediated reconnection the current sheet near the ergosphere boundary breaks into a chain of magnetic islands (plasmoids), dramatically accelerating the reconnection rate. The energy tap is approximately:

P_rec ≈ 1.646 × η_rec × a*² × P_BZ

This is a phenomenological calibration to the Meringolo et al. (2025) GRPIC result that reconnection contributes ~10–30% of P_BZ at spin a* > 0.9. The factor 1.646 is fixed so that η_rec = 0.15 at a* = 0.9 gives exactly 20% of P_BZ. The reconnection efficiency η_rec is the reconnection inflow speed as a fraction of the Alfvén speed; the Meringolo 2025 GRPIC runs find a median of ~0.15 at high spin. (An earlier analytic formula P_rec = η(B²/2μ₀)πrg²c overestimated by ~160× because the effective reconnecting layer area is far smaller than the full ergosphere cross-section.)

Why two channels add

BZ extracts energy via the Poynting flux of the global magnetosphere threaded through the horizon. Reconnection extracts energy locally in the current sheet where field lines of opposite polarity annihilate; the liberated magnetic energy goes partly into particle kinetic energy (some of which escapes) and partly into electromagnetic radiation. In the Kerr geometry both channels draw on the same reservoir — the spin energy of the hole — but via distinct pathways that operate simultaneously. For engineering purposes they sum as parallel power channels with the same magnetic field as fuel.

η_Penrose bound

The theoretical maximum fraction of rest-mass energy extractable via Penrose-class processes (Penrose 1969) is η_max = 1 − √((1 + √(1−a²))/2). At a=0.998 this reaches ~20.7%; at a=0.9 it is ~15.6%. The reconnection efficiency η_rec is a different quantity — it is the rate parameter of the reconnection layer, not a thermodynamic bound — but both quantities are shown here for context.

Meringolo vs. Comisso & Asenjo

Comisso & Asenjo (2021, Phys Rev D 103:023014) studied reconnection driven by field-line ergospheric crossing in the equatorial plane — a topological mechanism. Meringolo et al. (2025) studied GRPIC simulations of plasmoid-instability reconnection in the magnetosphere above the ergosphere. The two papers describe different physical processes and should not be combined or confused. This calculator implements the Meringolo 2025 prescription.

Field strengths in context

10⁴ T: extreme laboratory (pulsed). 10⁶–10⁸ T: neutron-star interior estimates. 10⁸–10¹¹ T: magnetar surface. The equatorial field threading an accreting IMBH horizon is model-dependent but GRMHD magnetically-arrested disk (MAD) simulations suggest B ~ (10⁴–10⁶) T for an 8,200 M☉ hole accreting near the Eddington rate. The default 10⁶ T is optimistic but not physically excluded.

v1.0 — 2026-06-01 · Code MIT · Prose CC BY 4.0 · Meringolo, Camilloni & Rezzolla (2025) ApJL 992, L8 arXiv:2507.08942 · Blandford & Znajek (1977) MNRAS 179:433 · Comisso & Asenjo (2021) Phys Rev D 103:023014