Solar Tachocline

A gear-clutch buried inside the Sun

Picture the Sun as two rotating bodies welded together at an awkward seam. The outer third by radius — the convection zone — does not spin like a solid object. Its equator laps its poles: a parcel at the equator circles the Sun in about 25 days, while one near 60° latitude takes closer to 31. This is solar differential rotation, and it persists with depth all the way down through the boiling convective envelope.

The radiative interior beneath behaves completely differently. Down there, with no convection to enforce a latitude-dependent profile, the gas rotates almost as a rigid body — one spin rate, period about 27 days, independent of latitude. So somewhere the differentially-rotating envelope has to hand off to the solid-body core. It does so not gradually over a wide buffer, but across a startlingly thin transition. That transition layer is the tachocline — from the Greek tachos (speed) and klinein (to slope): the layer where the rotation speed slopes from one regime to the other. Edward Spiegel and Jean-Paul Zahn coined the name in 1992.

Two things make this seam matter enormously. First, packing a ~30 percent change in angular velocity into a layer only a few percent of the solar radius thick produces an intense velocity shear. Second, the layer straddles the boundary between turbulent, convecting gas above and quiet, stably-stratified gas below — exactly the kind of interface where magnetic fields can be both generated and stored. The tachocline is, in the standard picture, where the Sun builds its magnetism.

The internal rotation profile

Helioseismology reconstructs the Sun's internal angular velocity Ω as a function of radius r and colatitude θ. The headline result, established with GONG, SOHO/MDI, and later SDO/HMI data, is remarkably clean:

Convection zone (0.71 ≲ r/R☉ ≲ 1):   Ω(r, θ) ≈ Ω_eq − a·cos²θ − b·cos⁴θ   (differential)
Radiative interior (r/R☉ ≲ 0.69):     Ω(r) ≈ const ≈ 2.7 × 10⁻⁶ rad/s     (solid body)
Tachocline (r/R☉ ≈ 0.69):             Ω transitions across Δr ≈ 0.04 R☉

In rotation-frequency units that the helioseismology community uses, the equatorial surface rotates near 470 nHz, the high-latitude surface near 330 nHz, and the radiative interior settles to about 432 nHz — close to the rotation rate of the convection zone at mid-latitudes (~35°). That intermediate value is a clue: the radiative interior appears to have been spun down to roughly the latitudinally-averaged rate of the layer above it.

The shear has two components. The radial shear ∂Ω/∂r is the change in rotation rate with depth across the layer; it is positive (rotation increasing outward) at the equator and negative at high latitudes, with a sign reversal near 35°. The latitudinal shear is the residual equator-to-pole contrast that survives into the layer. The radial shear is the term that the dynamo's Ω-effect exploits.

The math: how shear winds up a field

The reason astronomers fixate on the tachocline is the induction equation of magnetohydrodynamics, which governs how a flow v reshapes a magnetic field B:

∂B/∂t = ∇ × (v × B) + η ∇²B          (η = magnetic diffusivity)

The first term on the right is induction by flow; the second is ohmic decay. In a strongly sheared, high-conductivity plasma the flow term dominates, and field lines are effectively frozen into the gas (Alfvén's theorem). Now take a weak poloidal (north–south) field threading the tachocline and let the differential rotation act on it. A field line anchored at one depth gets dragged forward faster than the same line at an adjacent depth, so each turn of the Sun stretches it out in the azimuthal (east–west) direction. This is the Ω-effect: it converts poloidal field B_p into toroidal field B_φ at a rate set by the shear,

∂B_φ/∂t ≈ r sinθ (B_p · ∇)Ω        (Ω-effect — toroidal generation)

Because B_φ grows roughly linearly with time under steady shear, even a modest seed poloidal field of a few gauss can be wound up into kilogauss-to-tens-of-kilogauss toroidal bands over a few years — a fair fraction of the ~11-year cycle. The companion step, regenerating poloidal field from the toroidal field (the α-effect, supplied either by helical convection or, in flux-transport models, by the decay of tilted active regions at the surface), closes the dynamo loop. Shear builds toroidal field; helicity rebuilds poloidal field; repeat.

Why store the field down there at all?

A buried, stably-stratified layer solves a problem that plagues dynamos seated higher up. A horizontal magnetic flux tube of field strength B in a gas of pressure P is buoyant: the magnetic pressure B²/8π lowers the internal gas pressure and density, so the tube tends to rise. In the violently convecting envelope, a kilogauss tube would be torn apart and flung to the surface in weeks — far too fast to be amplified to the needed strength or to keep cycle memory.

The tachocline sits partly in the overshoot region and the stably-stratified top of the radiative zone, where the entropy gradient is subadiabatic and actively opposes vertical motion. A flux tube there is held down by the stratification and can be pumped to high field strength before magnetic buoyancy finally wins, lifts a loop through the convection zone, and erupts at the surface as a bipolar active region — a pair of sunspots. The latitude and tilt at which spots emerge (Spörer's law and Joy's law) are clues about the field strength at the storage layer, and reproducing them is what pushes estimates toward the strong-field end of 10⁴–10⁵ gauss.

Tachocline by the numbers

Quantity	Value	Note
Centre radius	0.693 ± 0.003 R☉	≈ 482,000 km from Sun's centre
Base of convection zone	0.713 R☉	Just above the tachocline centre
Radial thickness	~0.04 R☉ (0.01–0.05)	≈ 7,000–35,000 km
Local temperature	≈ 2.2 × 10⁶ K	From standard solar model
Local density	≈ 0.2 g/cm³	Denser than water; still a gas
Radiative-interior period	≈ 27 days (432 nHz)	Latitude-independent
Equator-to-pole CZ contrast	~30 % (25 d vs 35 d)	Source of the shear
Estimated toroidal field	10⁴–10⁵ G	Stored / amplified here
Prolateness	Deeper & thicker at poles	By ~0.02–0.05 R☉ vs equator

To make the thinness vivid: if you shrank the Sun to a 2-metre-radius ball, the whole tachocline would be a shell about 8 centimetres thick, sitting roughly 60 centimetres below the surface — and the entire machinery of the 11-year cycle would be threaded through that thin shell.

How we see a layer we can't reach

We can't sample the tachocline directly; the deepest probe ever sent sunward, Parker Solar Probe, only grazes the corona at a few solar radii. Everything we know about the layer comes from helioseismology — reading the Sun's internal structure from the acoustic waves that make it ring.

The Sun supports roughly ten million standing acoustic modes (p-modes), trapped between the surface and a lower turning point that depends on the mode's frequency and angular degree. Each mode samples a specific range of depths. Rotation breaks the degeneracy between modes of the same degree but different azimuthal order m, splitting their frequencies by an amount proportional to a depth-and-latitude-weighted average of Ω. Measure thousands of these rotational splittings and you can invert them — a classic ill-posed inverse problem solved with regularised least squares or optimally-localised averaging — to recover Ω(r, θ) layer by layer.

When those inversions were first done with high-quality data in the mid-1990s, the result was a surprise: the differential rotation did not continue smoothly inward, nor did it blend over a broad region. It collapsed onto solid-body rotation across a layer the inversions could barely resolve — only a few percent of the radius thick. That thinness is itself one of the layer's defining and most puzzling properties.

Where it shows up in dynamo theory

• Interface dynamo (Parker 1993). Places the Ω-effect (toroidal generation) just below the convection-zone base, in the tachocline, and the α-effect (poloidal regeneration) just above it, in the convection zone. The two layers communicate across the interface, and the configuration naturally produces dynamo waves that migrate equatorward — matching the butterfly diagram.
• Flux-transport / Babcock–Leighton dynamo. The dominant modern paradigm. The Ω-effect still winds up toroidal field in the tachocline, but the poloidal field is regenerated at the surface from the decay of tilted bipolar active regions, then carried back down by meridional circulation. The tachocline remains the toroidal-field factory and storage vault; the conveyor belt sets the cycle period.
• Tachocline α-effect and instabilities. The layer can host its own instabilities — magnetic buoyancy instability, the joint instability of latitudinal shear and toroidal field — that may supply part of the α-effect and seed the structure of emerging active regions.
• Distributed dynamos. Because fully convective M dwarfs lack a tachocline yet still cycle, some models distribute field generation throughout the convection zone. The Sun's tachocline is then an organising layer, not an absolute requirement — a debate kept alive by global MHD simulations that struggle to reproduce solar-like cycles.

The confinement problem and open questions

The single deepest puzzle is why the tachocline is thin. Spiegel and Zahn's own 1992 analysis showed the trouble: the latitudinal differential rotation imposed from the convection zone above should, through viscous and thermal diffusion, burrow downward into the radiative interior over time, thickening the shear layer. Their estimate had it spreading to roughly 0.3 R☉ over the Sun's 4.6-billion-year life — vastly thicker than observed. Something stops it.

The leading candidates each have advocates and difficulties:

• A primordial internal magnetic field. A weak fossil field locked into the radiative zone would couple it into solid-body rotation and resist the imposed shear, confining the tachocline. Gough and McIntyre's 1998 model is the classic version. The catch: the field must thread the radiative zone yet not leak up into the convection zone, which requires a delicate confinement of its own.
• Anisotropic turbulence. Strongly stratified turbulence transports angular momentum mainly horizontally, smoothing latitudinal shear without letting it diffuse deep. Whether such turbulence exists and acts the right way is debated.
• Meridional flows. A thin circulation confined to the layer can balance the inward burrowing and pin the tachocline at its observed depth.

Open questions extend beyond confinement: the exact field strength stored there, whether the layer's mild prolateness reflects its formation, how much it contributes to the α-effect, and whether subtle long-term changes in the tachocline drive grand minima like the Maunder Minimum. These are active targets for the next generation of helioseismic monitoring and global simulation.

Common misconceptions and edge cases

• The tachocline is not the core, and not the surface. It sits at the bottom of the convection zone, around 0.7 R☉ — well above the fusing core (inner ~0.25 R☉) and far below the visible photosphere. Sunspots are its surface symptoms, not the layer itself.
• "Solid-body" does not mean solid. The radiative interior is gas, denser than water but still a plasma. "Solid-body rotation" means only that its angular velocity is independent of latitude and radius — it turns as one piece, like a record on a turntable.
• The tachocline is not strictly required for magnetic activity. Fully convective M dwarfs have no radiative–convective boundary yet still show strong, sometimes cyclic, fields. The tachocline is the Sun's preferred dynamo seat, not a universal one.
• Shear is not the same as turbulence. The intense gradient in the tachocline is an ordered, large-scale velocity difference. Whether genuine turbulence lives inside the layer — and how anisotropic it is — is one of the unsettled questions, not an established fact.
• The field strength is inferred, not measured. No instrument has measured the magnetic field inside the tachocline. The 10⁴–10⁵ gauss figure is a modelling inference from where and how active regions emerge; quoting it as observed overstates the case.