💡 Suggest / Edit
← Wiki Index / Astronomy & Cosmology / Zeeman–Doppler Imaging
🔭 Astronomy & Cosmology

Zeeman–Doppler Imaging

Zeeman–Doppler imaging (ZDI) is a tomographic spectropolarimetric technique that reconstructs the large-scale vector magnetic field topology on the surfaces of rotating stars by inverting time series of circularly and linearly polarized spectral line profiles.

10 min read • Updated October 2026 • Knowledge Wiki • AI-Assisted & Human-Supervised

Zeeman–Doppler imaging (ZDI) is a high-resolution spectropolarimetric tomographic technique in stellar astrophysics that reconstructs two-dimensional maps of the vector magnetic field (B = (Br, Bθ, Bφ) in radial, meridional, and azimuthal components) and surface brightness or chemical abundance distributions across the photospheres of rotating stars. Pioneered theoretically by Meir Semel in 1989 and developed into a practical maximum-entropy vector tomography code by Jean-François Donati and Stephen F. Brown in 1991–1997, ZDI overcomes a fundamental obstacle in stellar magnetism: on a spatially unresolved stellar disk, nearby magnetic regions of opposite polarity produce opposite circular Zeeman polarization signatures that cancel out to nearly zero when integrated over the disk at a single instant. By observing a rotating star across a full rotation cycle with high-resolution echelle spectropolarimeters, ZDI exploits the rotational Doppler shift across the stellar disk to separate the polarized Zeeman signatures of distinct longitudes in velocity space, while tracking how their velocity trajectories evolve with rotational phase to determine their stellar latitude and magnetic vector orientation (Semel, 1989; Brown et al., 1991; Donati & Brown, 1997; Donati & Landstreet, 2009).

Physical Principles: Resolving Polarity Cancellation via Rotational Doppler Shifts

When an atomic spectral line forms in a stellar photosphere permeated by a magnetic field B, the Zeeman effect splits the degenerate magnetic sublevels (mJ) into three groups of components shifted from the unperturbed central wavelength λ0 by:

ΔλB = (e / (4π me c2)) geff λ02 B ≈ 4.67 × 10−13 geff (λ0 [Å])2 (B [G]) Å

where geff is the effective Landé factor (typically geff ≈ 1.0–2.5). The split components carry distinct polarization states described by the four Stokes parameters (I, Q, U, V):

  • Stokes I (Unpolarized Total Intensity): Measures the total line depth and equivalent width, sensitive to local temperature (starspots) and chemical abundance patches.
  • Stokes V (Circular Polarization, IRHC − ILHC): Generated by the longitudinal magnetic field component (Bℓ = B · n̂) projected along the observer's line of sight n̂. In the weak-field regime (ΔλB ≪ ΔλD, typical of cool stars where B ∼ 10–1,000 G), the Stokes V profile is antisymmetric and proportional to the first wavelength derivative of the intensity profile: V(λ) ∝ −geff λ02 Bℓ (dI / dλ).
  • Stokes Q and U (Linear Polarization): Generated by the transverse magnetic field component (B⊥) perpendicular to the line of sight. Because linear Zeeman polarization scales quadratically with field strength (Q, U ∝ ΔλB2 (d2I / dλ2)), Stokes Q and U signatures in cool stars are roughly 10 to 50 times weaker than Stokes V, though readily detected in strongly magnetic Ap/Bp stars (B ∼ 1–30 kG).

Even without Stokes Q and U, Donati & Brown (1997) demonstrated that Stokes V time series alone can reconstruct all three vector components (Br, Bθ, Bφ) of the stellar magnetic field through rotational projection geometry:

  1. Radial Field (Br, perpendicular to the stellar surface): Projects most strongly toward the observer when the magnetic spot crosses the central meridian near the sub-observer latitude; its Stokes V signature maintains the same sign throughout the spot's passage across the visible hemisphere, peaking at meridian transit (φrot = 0).
  2. Azimuthal (Toroidal) Field (Bφ, parallel to lines of constant latitude): Projects toward the observer on the approaching limb (cos ψ > 0), vanishes at central meridian passage (φrot = 0, where Bφ ⊥ n̂), and projects away from the observer on the receding limb; thus an azimuthal field ring or spot produces a Stokes V signature that reverses sign at central meridian transit.
  3. Velocity-Space Latitude Sorting: Because a spot at stellar latitude θ has a rotational velocity amplitude vspot = v sin i cos θ, equatorial magnetic features (θ = 0°) sweep rapidly across the entire width of the rotationally broadened line profile (−v sin i to +v sin i), whereas high-latitude / circumpolar features (θ → 90°) remain confined near the line center (v ≈ 0) throughout the entire rotation period (Semel, 1989; Donati & Brown, 1997).

Least-Squares Deconvolution (LSD) and Spherical Harmonic Inversion

In solar-type, T Tauri, and M-dwarf stars, the Stokes V circular polarization signal inside an individual spectral line has an amplitude of only V / Ic ∼ 10−4 to 10−5 (requiring an unattainable single-line signal-to-noise ratio of SNR > 20,000). To overcome photon noise, Donati et al. (1997) introduced Least-Squares Deconvolution (LSD), a multi-line cross-correlation technique that simultaneously combines 3,000 to 10,000 atomic or molecular absorption lines across a broadband echelle spectrum (370–1,050 nm), weighting each line by its central depth di, wavelength λi, and Landé factor gi. LSD boosts the effective polarimetric sensitivity by a factor of ~30 to 70, routinely detecting mean longitudinal magnetic fields smaller than 0.5 Gauss with polarimetric SNR > 20,000.

In modern ZDI codes (Donati et al., 2006; Hussain et al., 2000; Kochukhov et al., 2014), the surface vector magnetic field B(θ, φ) is parameterized as the sum of a poloidal field and a toroidal field expanded in complex spherical harmonics Yℓm(θ, φ) up to degree ℓmax ∼ 10–30:

  • Radial Component: Br(θ, φ) = −∑ αℓm Yℓm(θ, φ)
  • Meridional & Azimuthal Poloidal/Toroidal Components: Expressed via coefficients βℓm (horizontal poloidal field) and γℓm (toroidal field, ensuring ∇ · B = 0 identically everywhere).

Because tomographic inversion is an ill-posed inverse problem, the spherical harmonic coefficients (αℓm, βℓm, γℓm) are solved by maximizing the Shannon–Jaynes configurational entropy S (or minimizing a Tikhonov regularization penalty ∑ ℓ2(|αℓm|2 + |βℓm|2 + |γℓm|2)) subject to fitting the observed Stokes I, V (and Q, U) LSD profiles to reduced chi-squared χr2 = 1.0. Furthermore, by incorporating a latitudinal surface differential rotation law—Ω(θ) = Ωeq − dΩ sin2 θ—directly into the ZDI phase mapping (Petit et al., 2002), ZDI measures stellar differential rotation shear dΩ to better than ±1 mrad day−1.

Major Spectropolarimetric Instruments and Discoveries

Carried out primarily with high-resolution twin-fiber echelle spectropolarimeters—including ESPaDOnS (R = 65,000, 370–1,050 nm on the 3.6 m Canada–France–Hawaii Telescope), NARVAL and Neo-NARVAL (2 m Telescope Bernard Lyot, Pic du Midi), HARPSpol (3.6 m ESO La Silla), PEPSI (R = 130,000–250,000, 2×8.4 m Large Binocular Telescope), and the near-infrared cryogenic spectropolarimeter SPIRou (R = 70,000, 0.98–2.35 μm, CFHT)—Zeeman–Doppler imaging has transformed stellar dynamo physics across the Hertzsprung–Russell diagram (Donati & Landstreet, 2009; Reiners, 2012):

Stellar Class & Prototype TargetsZDI Magnetic Field Topology & StrengthAstrophysical Mechanism & Discovery Impact
Rapidly Rotating Solar-Type & RS CVn Stars (HR 1099, AB Doradus, LQ Hydrae)Large-scale fields of 100–500 G featuring prominent unidirectional azimuthal (toroidal) magnetic rings encircling the star at mid-to-high latitudes, plus dark polar starspots.Directly on the stellar surface, strong toroidal magnetic flux belts appear that on the Sun remain buried at the base of the convective zone (the tachocline), proving that rapid rotation promotes distributed convective-zone dynamos (Donati et al., 1992, 2003).
Solar-Twin Magnetic Cycle Reversals (τ Boötis, 61 Cygni A, ε Eridani)Weak large-scale poloidal/toroidal fields (2–15 G) exhibiting complete global magnetic polarity reversals every 120 days (τ Boo, a 240 d Hale cycle synchronized 1:1 with its hot Jupiter τ Boo b) or every 3.6 yr (61 Cyg A, a 7.2 yr solar-like Hale cycle).First direct imaging of complete global magnetic polarity reversals on stars other than the Sun (Donati et al., 2008; Boro Saikia et al., 2016).
Fully Convective M Dwarfs (M < 0.35 M☉: V374 Peg, AD Leo, WX UMa, Proxima Cen)Bistable Dynamo Topologies: Fully convective M dwarfs lacking a radiative core or tachocline generate large, stable axisymmetric dipolar fields of 1,000–4,000 G alongside a second branch with weak, multipolar non-axisymmetric fields (100–300 G) at identical mass and rotation rate.Proved that a solar-type tachocline is not required to generate kilogauss axisymmetric dipolar fields; M-dwarf convective dynamos exhibit bistability across the Rossby number boundary Ro ≤ 0.1 (Donati et al., 2006; Morin et al., 2008, 2010).
Classical T Tauri Pre-Main-Sequence Stars (BP Tau, V2129 Oph, TW Hya)Strong 1–3 kG dipolar and octupolar fields anchoring magnetospheric accretion funnels from the inner protoplanetary disk onto high-latitude chromospheric accretion shocks.Confirmed magnetospheric accretion theory and showed that disk-locking torques mediated by ZDI-measured dipolar components spin down young stars (Donati et al., 2007, 2011).
Chemically Peculiar Ap/Bp & Massive OB Stars (α2 CVn, 53 Cam, τ Sco)Ultra-stable, non-axisymmetric fossil fields (300 G to 30 kG) harvested in full four-Stokes (I, Q, U, V) ZDI, revealing small-scale multipolar distortions superimposed on tilted oblique dipoles alongside stratified horizontal elemental abundance rings (Cr, Fe, Eu, Si).Full four-Stokes vector ZDI showed that fossil fields deviate significantly from pure dipoles and channel radiatively driven chemical diffusion and wind confinement (Kochukhov et al., 2004, 2010).

References

  1. Semel, M. (1989). Zeeman–Doppler imaging of active stars. I. Basic principles. Astronomy and Astrophysics, 225(2), 456–466. https://ui.adsabs.harvard.edu/abs/1989A%26A...225..456S
  2. Brown, S. F., Donati, J.-F., Rees, D. E., & Semel, M. (1991). Zeeman–Doppler imaging of solar-type and Ap stars. IV. Maximum entropy reconstruction of 2D magnetic topologies. Astronomy and Astrophysics, 250(2), 463–474. https://ui.adsabs.harvard.edu/abs/1991A%26A...250..463B
  3. Donati, J.-F., & Brown, S. F. (1997). Zeeman–Doppler imaging of active stars. V. Sensitivity of maximum-entropy magnetic maps to field orientation. Astronomy and Astrophysics, 326, 1135–1142. https://ui.adsabs.harvard.edu/abs/1997A%26A...326.1135D
  4. Donati, J.-F., Semel, M., Carter, B. D., Rees, D. E., & Cameron, A. C. (1997). Spectropolarimetric observations of active stars. Monthly Notices of the Royal Astronomical Society, 291(4), 658–682. https://doi.org/10.1093/mnras/291.4.658
  5. Donati, J.-F., Forveille, T., Cameron, A. C., Barnes, J. R., Delfosse, X., Jardine, M. M., & Valenti, J. A. (2006). The large-scale axisymmetric magnetic topology of a very-low-mass fully convective star. Science, 311(5761), 633–635. https://doi.org/10.1126/science.1121102
  6. Donati, J.-F., & Landstreet, J. D. (2009). Magnetic fields of nondegenerate stars. Annual Review of Astronomy and Astrophysics, 47, 333–370. https://doi.org/10.1146/annurev-astro-082708-101833
  7. Kochukhov, O., Lüftinger, T., Neiner, C., Alecian, E., & MiMeS Collaboration. (2014). Magnetic field topology of the unique chemically peculiar star CU Virginis. Astronomy and Astrophysics, 565, A83. https://doi.org/10.1051/0004-6361/201423472
🧬 Support Open Science
Explore interactive science games on LowKey Games or browse scientific apparel & accessories in the CRV Science Shop.
💡 Spot an inaccuracy or have primary literature to expand this article?
Submit a citation or revision note using your shared CRV Science / LowKey Games profile ID.
Suggest Revision