Shafranov-Shift Diagnostic

The invisible fingerprint of plasma pressure: the hotter and denser the plasma, the further its magnetic core drifts outward — and a ring of coils outside the vessel can read that drift to within millimetres.

Measuring the Shafranov shift — the radial outward displacement of the magnetic axis relative to the geometric axis of the torus — gives operators a non-invasive handle on the plasma's total pressure energy (proportional to the poloidal beta, β_p). Magnetic probes (pick-up coils) and flux loops installed on the outer surface of the vacuum vessel record the poloidal field and poloidal flux. The measured asymmetry between the high-field side (inboard) and low-field side (outboard) of the plasma cross-section encodes the shift. Feeding these signals into equilibrium reconstruction codes (notably EFIT, first described by Lao et al. in 1985) allows the two-dimensional current density and flux surfaces to be reconstructed in real time, including the precise location of the magnetic axis.

In a tokamak, the magnetic field is stronger on the inside of the torus (inboard) than the outside (outboard). A plasma column under pressure pushes outward against this field gradient — the so-called hoop force. The toroidal geometry requires the plasma equilibrium to satisfy the Grad–Shafranov equation, derived independently and equivalently by Vitaly Shafranov (Soviet Physics JETP 6, 1958, p. 545) and Harold Grad (Geneva Atoms for Peace Conference, 1958). For a circular plasma cross-section in the large-aspect-ratio limit, Shafranov's analytic solution predicts that the displacement Δ of the magnetic axis from the geometric centre scales approximately as Δ ∝ a²(β_p + l_i/2) / R, where a is the minor radius, R the major radius, β_p the poloidal beta, and l_i the internal inductance. A suite of magnetic diagnostics measures the asymmetric poloidal field distribution. EFIT (and its real-time variants) solve the nonlinear Grad–Shafranov boundary-value problem iteratively to find the flux function ψ(R,Z) consistent with all the coil-current and probe measurements. The shift of the computed magnetic axis directly yields an integrated measure of the plasma pressure, complementary to Thomson scattering and charge-exchange spectroscopy measurements.

In plain terms

Think of a car tyre. When it is flat, its centre is where you expect. When it is over-inflated, the tube bulges outward on one side. A tokamak plasma does the same thing: the harder you squeeze and heat it, the more the whole magnetic core shifts outward in the torus. By measuring how far off-centre it drifts, scientists can read the plasma's internal pressure without sticking anything into the hot gas.

Also called: Shafranov shift measurement, magnetic-axis displacement diagnostic, equilibrium reconstruction diagnostic.

Why it matters

The Shafranov shift is the cleanest external observable of the plasma's total stored thermal and fast-ion energy, accessible with sensors outside the plasma that survive the harsh radiation environment indefinitely.
EFIT-based equilibrium reconstruction is the standard tool on virtually every tokamak in operation — it provides the safety factor profile q(r), the plasma boundary shape, and the current density profile needed for real-time position and shape control.
The shift gives a cross-check on energy confinement time τ_E: the first modern estimates of τ_E on the T-3 tokamak in the late 1960s used the Shafranov equilibrium equations — measurements whose surprisingly good results led directly to the worldwide tokamak programme.
Deviations between the predicted and measured shift signal departures from ideal MHD equilibrium, including fast-ion driven instabilities (energetic-particle modes) that redistribute plasma energy.

Where it lives

Every modern tokamak with an equilibrium reconstruction system: JET, ASDEX Upgrade, DIII-D, KSTAR, JT-60SA, C-Mod (MIT, decommissioned 2016)
ITER — the 55-channel magnetic diagnostic suite (pick-up coils + flux loops) is designed to support real-time EFIT reconstruction for plasma position and shape control
KSTAR (Daejeon, South Korea) — validated GS-DeepNet neural equilibrium reconstruction using standard Shafranov-shift-sensitive magnetic diagnostics (Nature Scientific Reports, 2023)
T-3 tokamak (Kurchatov Institute, Moscow, 1967) — first device on which Shafranov shift measurements were used to infer energy confinement time, establishing the tokamak confinement advantage

Key numbers

Shift magnitude (typical high-β discharge): 1–10 cm (Scales with a²β_p/R; detectable by outer magnetic probes)
Poloidal beta range in present tokamaks: 0.1–2 (β_p ~ 1 is typical H-mode; higher in some advanced scenarios)
EFIT first publication: 1985 (Lao et al., Nuclear Fusion 25 (1985) 1611; still the reference algorithm)
Magnetic probe accuracy for equilibrium: ~5–10 mm (position) (GS-DeepNet study on KSTAR (2023) confirms EFIT position uncertainty ~5–10 mm)
Shafranov original paper: Soviet Physics JETP 6 (1958) 545 (Submitted March 1957; equivalent result by Grad at Geneva conference, 1958)
Number of magnetic diagnostics (ITER): ~55 channels (Pick-up coils + flux loops supporting real-time equilibrium reconstruction)

Surprising facts

Vitaly Shafranov's equilibrium formula, submitted to a Soviet journal in March 1957, was used in experiments on the T-3 tokamak a decade later to produce the first credible estimates of electron temperature (~1 keV) — results so good that British scientists were sent to Moscow to verify them independently, triggering the global tokamak era.
In a high-performance tokamak discharge (poloidal beta ~ 0.5–1), the Shafranov shift of the magnetic axis can reach several centimetres — a significant fraction of the plasma minor radius — making it easily detectable by the outer magnetic probes without any internal measurement.
EFIT's core algorithm was published by Lao et al. in Nuclear Fusion in 1985. It has been adapted and run on every major tokamak since; no alternative has supplanted it as the standard equilibrium reconstruction tool, making it arguably the most-used fusion diagnostic code ever written.
The Grad–Shafranov equation has a curious history: Harold Grad and Vitaly Shafranov independently derived mathematically equivalent results in the same year (1957–1958) using entirely different approaches — variational energy minimisation versus direct toroidal force balance — without knowledge of each other's work.

Open questions

Ill-posedness of current profile inference from boundary measurements alone: The inverse problem of deducing the internal current density j(R,Z) — and hence the safety factor profile q(r) — from magnetic measurements taken only at the tokamak wall boundary is fundamentally ill-posed: small measurement errors in the outer probes inflate dramatically when propagated inward toward the plasma axis. EFIT and its variants require regularisation assumptions (parameterised current profile models, additional internal diagnostics such as Motional Stark Effect or Faraday rotation) to produce stable solutions. The exact amount of information the boundary uniquely specifies — and how much is genuinely indeterminate without internal constraints — remains an active topic in the mathematical plasma physics literature, with uniqueness proofs available only for special geometries.

A short history

Vitaly Dmitrievich Shafranov of the Kurchatov Institute submitted his derivation of the toroidal equilibrium force balance to the Soviet journal JETP in March 1957, published in Russian in September 1957 and translated into English in Soviet Physics JETP 6 (1958), p. 545. Harold Grad independently obtained equivalent results using a variational approach, presenting them at the Second United Nations Atoms for Peace Conference in Geneva in 1958. The two results are now combined under the name Grad–Shafranov equation. Shafranov and Vladimir Mukhovatov subsequently applied the analytic solution to circular tokamak plasmas, deriving formulas for the magnetic-axis displacement and the vertical field needed for equilibrium — the foundations of tokamak position control. Experimental verification came on the T-3 and T-5 tokamaks at the Kurchatov Institute in the mid-1960s, and the measured good confinement results were confirmed by a British team (from Culham) who visited Moscow in 1969. The standard numerical implementation — EFIT, by Lao, St. John, Stambaugh and colleagues at GA Technologies — was published in Nuclear Fusion 25 (1985) 1611 and became universal across the tokamak community.

Sources

Shafranov, V.D. — On Magnetohydrodynamical Equilibrium Configurations, Soviet Physics JETP 6 (1958) 545–554 (http://jetp.ras.ru/cgi-bin/dn/e_006_03_0545.pdf)
ITER Organization — The Grad–Shafranov Equation (history and role in plasma control, 2026) (https://www.iter.org/print/whatsnew/487)
Lao et al. — Reconstruction of current profile parameters and plasma shapes in tokamaks, Nuclear Fusion 25 (1985) 1611 (EFIT reference; via MIT C-Mod wiki) (https://cmodwiki.psfc.mit.edu/index.php/EFIT)
Kim et al. — GS-DeepNet: mastering tokamak plasma equilibria with deep neural networks, Scientific Reports 13 (2023) (https://www.nature.com/articles/s41598-023-42991-5)
Mukhovatov & Shafranov — V.D. Shafranov and Tokamaks, Journal of Plasma Physics 82 (2016), Cambridge Core (https://www.cambridge.org/core/journals/journal-of-plasma-physics/article/v-d-shafranov-and-tokamaks/809508C83F56D2E49F05F9A278B8FACD)
De Blank — Plasma Equilibrium in Tokamaks (lecture notes), Forschungszentrum Jülich (JuSER) (https://juser.fz-juelich.de/record/283634/files/DeBlank_EI-1.pdf)