Magnetic return flux in a sunspot penumbra

The left panel shows the scanned map of a sunspot (AR12045). The black arrow points at solar disk center. A rectangular field-of-view of the limb side penumbra is shown in the right panel: The colours indicate the line-of-sight velocity (red indicates red-shifts), the small black arrows outline the projected horizontal magnetic field lines. The black contours coincide with B(vertical) = 0 G and mark the patches of magnetic return flux, i.e. of negative vertical field components. The white contours are drawn at B(vertical) = -250 G.

The spectro-polarimetric data cube was acquired by scanning a spot with a slit. For each slit position, the net exposure time for all polarimetric states was about 1.2 sec at a wavelength of 1564.8 nm. If one aims at reaching the highest possible spatial resolution, a shorter wavelength is chosen and a burst of images is recorded, each image having an exposure time as small as 1 ms. Then it is possible with GREGOR to reconstruct images with a spatial resolution of 0.08 arcsec. Here we show an example of a spot image taken at 589 nm with the Broad Band Imager at GREGOR on May 31, 2013.

A highlight of first science results with GREGOR has been achieved by detecting the small-scale geometry of the magnetic field in a sunspot penumbra. Spectro-polarimetric data with GRIS@GREGOR reveals that 35 % of the penumbra area is covered with return flux, i.e., locations where the magnetic field has opposite polarity. Since the field strength of this downward-pointing component is small, this return flux only makes up 10% of the total magnetic flux in the penumbra. A careful analysis of the depth dependence of the magnetic field reveals that the return flux is only present in the deepest layers of the photosphere, which are probed using the four iron lines in the spectral vicinity of 1564.8 nm. This is the reason why the spectropolarimeter onboard HINODE using higher-forming iron lines around 630.2 nm cannot detect the return flux, while previous observations at 1564.8 nm with the VTT did not have the necessary spatial resolution.

This finding has important implications on the velocity and magnetic field structure of penumbral magneto-convection: These patches tend to coincide with the locations where the Evershed outflow is largest. Since the Evershed flow is magnetized, one can assume that the flow is aligned with the magnetic field. Hence these return flux patches should be associated with downflows.