The study, published in the journal Solar Physics, was led by Akie Moritsuka of the University of Tokyo and the National Astronomical Observatory of Japan, together with Yukio Katsukawa and Ryohtaroh T. Ishikawa. The team mined the archives of the Hinode satellite, a Japanese mission launched in 2006 and operated in collaboration with NASA and the UK Space Agency. Hinode’s Solar Optical Telescope, a 50-centimeter instrument with a diffraction-limited resolution of about 0.31 arcseconds at 630 nanometers, carries a Spectro-Polarimeter that records the full polarization state of light across two iron absorption lines at 630.15 and 630.25 nanometers. Crucially, the team selected 49 polar-observation datasets in which the north or south limb fell inside the field of view, allowing them to trace spectral line shifts from heliocentric angles of about 63 degrees all the way to the limb at 90 degrees, and even slightly beyond it.

Observing this close to the limb is notoriously difficult. The line of sight grazes the solar surface at an extremely shallow angle, so the light from any single pixel samples a long, slanted path through the atmosphere, blending structures that would be cleanly separated at disk center. The researchers first had to pin down the exact position of the limb in every scan by finding the pixel with the steepest gradient in continuum intensity, then fitting a circle to the limb to recover the coordinates of the disk center. From there they could assign a precise heliocentric angle to every pixel on the disk. The spatial resolution of the instrument, roughly 0.15 to 0.16 arcseconds along each axis, corresponds to about 600 kilometers on the Sun near disk center, degrading toward the limb, which sets a hard limit on how finely the convective pattern can be resolved at grazing angles.

To disentangle the contributions of different parts of the convective pattern, the team classified every observed spectrum into three categories. Pixels showing a total polarization degree of 0.6 percent or more, three times the internetwork average, were labeled magnetic regions. Among the non-magnetic pixels, those brighter than the local average continuum intensity, corrected for limb darkening, were classified as granules, while darker pixels were classified as intergranules. This classification matters because the physics of the two regions differs fundamentally: granules are the sites of hot, rising, blueshifted plasma, while intergranular lanes host cool, sinking, redshifted material. Because the polar fields observed here are quiet, magnetic regions occupied less than one percent of the area at 66 degrees, so the analysis of convective velocities focused on the granule and intergranule populations.

The headline result concerns how the averaged Doppler velocity of the iron line at 630.15 nanometers changes across the disk. The spatially averaged profiles showed a steady increase in redshift from about 65 degrees up to roughly 80 degrees, then the trend flattened, and finally the velocities fell back to nearly zero at the very limb. The increase of about 70 meters per second between 66 and 72 degrees agrees well with high-accuracy ground-based measurements made with the Laser Absolute Reference Spectrograph, which anchored the team’s velocity calibration using laboratory-grade laser frequency comb references. But the real insight came from splitting the signal by region. Granules showed a dramatic evolution: their Doppler velocities shifted by about 240 meters per second toward the red between 65 and 80 degrees, ending up redshifted by roughly 120 meters per second. Intergranules, by contrast, stayed almost perfectly constant at around 120 meters per second of redshift, varying by no more than about 20 meters per second across the same range.

This asymmetry is the key to the limb effect. The traditional explanation invokes the corrugated, three-dimensional surface of the photosphere: when we look at the limb, the line of sight skims across the tops of the granular bumps, and the redshifted horizontal flow on the near side of each granule appears brighter than the blueshifted flow on the far side, because the back granule’s wall provides the background illumination. When a front granule partially hides a back one, the blueshifted material is concealed while the redshifted material remains visible. The new observations confirm this picture quantitatively: since the overall average shift is dominated by the bright granules, and it is the granules whose redshift grows toward the limb, the limb effect is essentially a granule phenomenon. The intergranular lanes, whose sinking flows are already redshifted, contribute a nearly constant baseline.

The team went a step further by analyzing the bisectors of the spectral lines, the locus of wavelength midpoints measured at different depths within the absorption profile. Because the line core forms higher in the photosphere while the wings form deeper down, the bisector encodes the vertical gradient of the line-of-sight velocity. At 66 degrees, the overall bisector showed the classic backslash shape, with a velocity difference of roughly 500 meters per second between core and wing, consistent with non-magnetic numerical simulations. Toward the limb, this shape gradually straightened into an I-shape by about 84 degrees, indicating that the line-forming region samples a much smaller range of velocities. Most tellingly, the bisectors of granules changed far more dramatically than those of intergranules, which remained nearly symmetric throughout. The authors interpret this as evidence that the atmosphere within granules, where the line forms, is geometrically thicker than in intergranules, so that the velocity gradient along the line of sight is larger and its apparent compression toward the limb is more pronounced.

The observations also captured something rarer still: emission lines just beyond the limb. Previous work had shown that within about one arcsecond above the limb, the iron line at 630.15 nanometers can flip from absorption to emission, a phenomenon linked to localized temperature enhancements and torsional flows in magnetic flux concentrations. The team averaged the emission profiles from the second and third pixels above the limb and found small blueshifts, averaging around 130 meters per second relative to the disk absorption lines. The authors are careful to note that this difference is only about one standard deviation and could partly reflect limitations of their velocity correction, which relied on a linear fit below 80 degrees. If the emission lines were taken as the true zero-velocity reference, the absorption lines near the limb would instead be redshifted by a comparable amount, a distinction that future observations with better absolute calibration will need to settle.

Comparisons with three-dimensional radiative magnetohydrodynamic simulations reveal both agreement and tension. The granular trends match simulations that model magnetoconvection, while the flat intergranular velocities do not, likely because intergranular lanes are only about 300 kilometers wide and fall below the effective spatial resolution at large heliocentric angles, where granules and lanes blur together. The bisector velocity difference of about 500 meters per second aligns better with non-magnetic simulations than with magnetic ones, consistent with the quiet, weakly magnetized polar regions studied here. The authors argue that the next leap requires instruments such as the Visible Spectro-Polarimeter on the Daniel K. Inouye Solar Telescope, which can resolve granules and lanes up to the limb, combined with telluric-line or laser-comb velocity calibration and multi-wavelength observations probing from the deep photosphere to the low chromosphere. For now, the Hinode archive has done what it does best: turning a century-old puzzle about crooked spectral lines into a concrete, measurable statement about the three-dimensional architecture of the Sun’s boiling surface.

Subject of Research: Center-to-limb variation of Doppler velocities and spectral line bisectors in the solar photosphere observed with Hinode

Article Title: Doppler Velocity Variation Near the Solar Limb in the Solar Photosphere Observed with Hinode

Article References: Moritsuka, A., Katsukawa, Y., & Ishikawa, R. T. (2026). Doppler Velocity Variation Near the Solar Limb in the Solar Photosphere Observed with Hinode. Solar Physics, 301(10), Article 153. https://doi.org/10.1007/s11207-026-02747-6

Image Credits: AI Generated

DOI: 10.1007/s11207-026-02747-6

Keywords: solar physics, photosphere, granulation, Doppler velocity, limb effect, spectral line bisector, Hinode, Solar Optical Telescope, spectro-polarimetry, convective blueshift, iron spectral lines, solar limb