Petschek Fast Reconnection: The Slow-Shock X-Point That Beats Sweet-Parker

What Petschek reconnection is and the physical problem it solves

Magnetic reconnection is the process by which oppositely directed magnetic field lines break and re-join, releasing stored magnetic energy as heat, bulk kinetic flow, and accelerated particles. The first quantitative model, Sweet-Parker (1957–58), assumed a single long, thin current sheet of macroscopic length L where plasma is squeezed out the sides. That geometry gives a reconnection rate M = v_in/v_A ∝ S^(-1/2), where S is the Lundquist number S = v_A L / η (Alfvén speed × length over magnetic diffusivity).

The trouble: in the solar corona S ≈ 10^12–10^14, so the Sweet-Parker rate is around 10^(-6)–10^(-7) of the Alfvén speed. A flare that is observed to release its energy in ~100–1000 seconds would instead require weeks to months. In 1964 Harry E. Petschek proposed a radically different geometry: confine the slow, resistive diffusion to a tiny central X-point, and let two pairs of standing slow-mode MHD shocks fan out to do most of the energy conversion. This decouples the rate from L and makes reconnection genuinely fast.

The mechanism: X-point diffusion plus four standing slow shocks

In Petschek's picture the reconnection region has a nested structure:

• The diffusion region is a small central patch of length δ (much smaller than the system size L) where the frozen-in condition breaks and field lines actually reconnect. Its size is set by resistivity, roughly δ ~ η/v_in.
• Four slow-mode shocks emanate from that X-point in a wide-open configuration, forming a shallow wedge. Most of the inflowing magnetic energy is dissipated at these shocks, not in the diffusion region.

The slow shocks are of the switch-off variety: the tangential magnetic field component is almost entirely annihilated across them, its energy going into heating and into accelerating plasma to near the upstream Alfvén speed, v_out ≈ v_A. Because the shocks open outward, the outflow can be evacuated over a wide channel rather than through a long, throttling sheet. Crucially, the reconnection rate is now set by how fast the shocks can consume inflow, giving M ≈ π/(8 ln S) — only a logarithmic dependence on S, so the rate stays around 0.01–0.1 even for astronomically large S.

Key quantities and a worked example

The governing scaling law is the maximum dimensionless reconnection rate:

• M_max ≈ π / (8 ln S), where M = v_in/v_A is the inflow speed in units of the Alfvén speed.
• Outflow speed: v_out ≈ v_A (near the upstream Alfvén speed).
• Lundquist number: S = v_A L / η.

Worked example (solar corona). Take B ≈ 100 G = 0.01 T, number density n ≈ 10^9 cm^(-3), giving an Alfvén speed v_A ≈ 1000 km/s. For a flare with L ≈ 10^4 km and coronal resistivity, S ≈ 10^12, so ln S ≈ 27.6 and M ≈ π/(8 × 27.6) ≈ 0.014. The inflow speed is then v_in ≈ 0.014 × 1000 ≈ 14 km/s. Sweet-Parker at the same S would give M ≈ S^(-1/2) ≈ 10^(-6), or v_in ≈ 1 m/s. At S = 10^6, Petschek is roughly 28× faster than Sweet-Parker; at coronal S the advantage is a factor of ~10^4–10^5.

How it is observed and where it appears

Petschek-type reconnection has three main observational arenas:

• Solar flares and coronal mass ejections. Missions such as Yohkoh, RHESSI, Hinode, and the Solar Dynamics Observatory (SDO) show cusp-shaped flare loops, above-the-looptop hard X-ray sources, and reconnection outflow jets at ~10^2–10^3 km/s — signatures consistent with a localized X-point feeding standing shocks. Studies (e.g., the Global Energetics of Solar Flares series) infer Petschek Mach numbers of order 0.01–0.1.
• Earth's magnetotail and magnetopause. In-situ spacecraft — Cluster, THEMIS, and especially NASA's Magnetospheric Multiscale (MMS) mission — have directly sampled diffusion regions and detected slow-shock-like structures and Alfvénic reconnection jets.
• Laboratory plasmas. The Magnetic Reconnection Experiment (MRX) at Princeton and similar devices measure reconnection rates and confirm that localized (anomalous) resistivity can produce open X-point, Petschek-like configurations rather than a long Sweet-Parker sheet.

How it differs from Sweet-Parker, plasmoids, and collisionless reconnection

Petschek sits between two other regimes and is often confused with them:

• vs Sweet-Parker: Same physics of resistive diffusion, but opposite geometry. Sweet-Parker uses one long sheet (rate ∝ S^(-1/2)); Petschek uses a point-like diffusion region plus shocks (rate ∝ 1/ln S). Same S can differ by 10^4–10^5 in rate.
• vs plasmoid-mediated reconnection: For S above a critical ~10^4, a long Sweet-Parker sheet is tearing-unstable and fragments into a chain of magnetic islands (plasmoids). This gives a fast, nearly S-independent rate (~0.01) without needing Petschek's special resistivity — the modern resolution for many collisional systems.
• vs collisionless (Hall) reconnection: In thin enough sheets, ion and electron scales decouple; Hall physics and kinetic effects produce fast reconnection at ~0.1 v_A. This is the relevant regime in the magnetosphere and is what MMS actually resolves.

Petschek remains the archetype of fast, steady, single-X-point reconnection in resistive MHD.

Significance, the collapse controversy, and open questions

Petschek's model was a landmark because it showed that fast, steady reconnection is at least possible in MHD. But its status is subtle. In a famous result, Biskamp (1986) found that in resistive MHD simulations with spatially uniform resistivity, the Petschek configuration does not survive: the current sheet elongates and the system collapses back to a long Sweet-Parker layer with rate ∝ S^(-1/2). The slow shocks fail to anchor.

The resolution, developed by Biskamp & Schwarz (2001), Kulsrud (2001), and Uzdensky & Kulsrud (2000, 2003), is that Petschek-like reconnection can be sustained if the resistivity is locally enhanced at the X-point — so-called anomalous or current-dependent resistivity. Then a stable open X-point with standing shocks reforms and the fast rate returns.

Open questions remain: What physically produces the required localized resistivity in nearly collisionless space plasmas? How do Petschek shocks, plasmoids, and kinetic effects coexist in real 3D systems? And why does nature so robustly converge on a rate near 0.1 — the still-unexplained "0.1 reconnection rate problem"? Petschek's slow-shock X-point is thus both a textbook solution and a live research frontier.

Sweet-Parker vs Petschek vs plasmoid-mediated reconnection: rate, geometry, and where each applies.

Property	Sweet-Parker (1957)	Petschek (1964)	Plasmoid-mediated (2000s)
Reconnection rate M = v_in/v_A	∝ S^(-1/2)	π/(8 ln S), ~0.01–0.1	~0.01, nearly S-independent
Diffusion-region length	L (macroscopic, ~10^4 km on Sun)	δ << L (microscopic X-point)	Fragmented into many small sheets
Energy-conversion structure	Long thin resistive current sheet	Two pairs of standing slow shocks	Chain of plasmoids + mini-sheets
Rate at S = 10^6	~10^(-3) v_A (far too slow)	~0.03 v_A (~28× faster)	~0.01 v_A
Stability / realizability	Robust but too slow	Needs localized/anomalous resistivity	Emerges naturally at S > ~10^4