Free Access
Issue
A&A
Volume 562, February 2014
Article Number L6
Number of page(s) 4
Section Letters
DOI https://doi.org/10.1051/0004-6361/201323224
Published online 10 February 2014

© ESO, 2014

1. Introduction

In some cases, a sunspot penumbra contains umbrae of opposite magnetic polarities, as first described by Künzel (1960). A few years later, Künzel (1965) introduced the new category “δ-spot” extending Hale’s spot classification in terms of the magnetic field. Although δ-spots are frequent and often associated with flares (see Sammis et al. 2000), accurate measurements of the magnetic field vector with high resolution are still rare, especially in the near infrared (NIR). In the context of flares and δ-spots, observations beyond 1 μm have been limited to white-light flare emission at 1.56 μm after major X-class flares (Xu et al. 2004, 2006) and sporadic long-wavelength infrared spectropolarimetry in the Mg i λ12.32 μm line (Jennings et al. 2002).

Velocity investigations in the visible are more frequent. Persistent downflows of up to 14 km s-1 within the penumbra close to the polarity inversion line (PIL) have been reported by Martínez Pillet et al. (1994). At lower spectral and spatial resolution, similar downflows of 1.5–1.7 km s-1 have been detected by Takizawa et al. (2012) in regions of penumbral decay. Evershed flows of opposite polarity umbrae converge near the PIL and have to pass each other, which is indicative of an interleaved system of magnetic field lines (Lites et al. 2002; Hirzberger et al. 2009). Moreover, significant up- and downflows in excess of 10 km s-1 are encountered during flares (Fischer et al. 2012).

The observed δ-spot shares a remarkable visual resemblance with its counterpart in the flare-prolific active region NOAA 10930, which by contrast exhibits significant rotation (up to 8° h-1) of the δ-umbra (Min & Chae 2009). Shear flows along the PIL potentially contribute to the build-up of non-potential magnetic field configurations (cf., Denker et al. 2007). Conversely, Tan et al. (2009) noted that the shear flow between opposite-polarity umbrae decreases from 0.6 km s-1 before to 0.3 km s-1 after an X3.4 flare.

In this Letter, we introduce NIR spectropolarimetry as a diagnostic tool for studying the intriguing properties of δ-spots.

2. Observations

On 2012 June 17, not far from the maximum of solar cycle No. 24, we observed a sunspot with a δ-configuration in active region NOAA 11504 at 18° S and 29° W (μ = cosϑ = 0.82). A map of the full Stokes vector was obtained with the Tenerife Infrared Polarimeter (TIP; Collados et al. 2007) and adaptive optics’ correction (Berkefeld et al. 2010) at the Vacuum Tower Telescope (VTT; von der Lühe 1998). The dispersion of the Fe iλ1078.3 nm and Si iλ1078.6 nm spectra is 2.19 pm pixel-1. The map (10:00–10:38 UT) consists of 180 scan positions with a step size of 036. The image scale of 0175 pixel-1 along the slit direction was resampled to 035 pixel-1 to reduce the noise in preparation for the spectral inversions. For each Stokes parameter, we accumulated ten exposures of 250 ms (see Balthasar & Gömöry 2008).

The Fe i and Si i lines sample different atmospheric layers of the quiet Sun, only in dark umbrae they originate at almost the same height (see Balthasar & Gömöry 2008). We derived the magnetic field vector using “Stokes inversions based on response functions” (SIR; Ruiz Cobo & del Toro Iniesta 1992). The two lines were inverted separately. The magnetic field strength, the magnetic inclination and azimuth as well as the Doppler velocity were kept constant with height, whereas we use three nodes for the temperature. Initially, the magnetic azimuth ambiguity was resolved by selecting directions that matched a radial configuration with a common center (see Balthasar 2006). The subsequently applied minimum-energy method of Leka et al. (2009) yields a more reliable configuration. Finally, we rotated the magnetic vector to the local reference frame and corrected the images for projection effects (see Verma et al. 2012). The resulting field of view (FOV) of 60 Mm × 40 Mm served as a reference for all auxiliary data.

Multiframe blind deconvolution (MFBD; van Noort et al. 2005) was applied to high-resolution (004 pixel-1) G-band context images (see Fig. 1), which were recorded in the VTT’s optical laboratory. Additional Ca ii λ854.2 nm spectra with a dispersion of 0.82 pm pixel-1 and an image scale of 035 pixel-1 were acquired at the Echelle spectrograph with a camera mounted next to TIP. Integration times of 9 s matched the TIP accumulation cycle. The spectral range also covered the Si iλ853.6 nm line, which is weak, has a high excitation potential of 6.15 eV, and which represents deep photospheric layers.

thumbnail Fig. 1

G-band image of the δ-spot in active region NOAA 11504.

Open with DEXTER

Magnetograms of the Helioseismic and Magnetic Imager (HMI; Schou et al. 2012) onboard of the Solar Dynamics Observatory (SDO; Pesnell et al. 2012) accompany the VTT data and provide information of the magnetic field evolution and are used to determine horizontal velocities.

3. Results

Penumbral filaments related to the δ-umbra display a counterclockwise twist in the central part of Fig. 1, while those of the main umbra show an opposite twist in the vicinity of the δ-umbra. The braided strands of penumbral filaments are parallel to the PIL and roughly perpendicular to the line connecting the centers of the main and δ-umbrae.

thumbnail Fig. 2

Maps of continuum intensities at λ1078 and λ854 nm, Ca iiλ854.2 nm line-core intensity, and maximum of the central Ca ii emission (top-left to bottom right). The ordinate is along the solar north-south direction. The black arrow points toward the disk center, and the white arrow indicates the bright feature BF near the PIL (thick red line). Contour lines denote the umbral and penumbral boundaries. Three white crosses mark the positions of the spectral profiles in Fig. 3.

Open with DEXTER

thumbnail Fig. 3

Observed (solid) and inverted (dashed) Stokes profiles in the umbra of the main spot (left), the δ-spot umbra (middle), and in the penumbra near BF (right). The wavelength is given with respect to the center of the Fe i λ1078.3 nm (left) and Si i λ1078.6 nm (right) lines, respectively. I0 is the quiet-Sun continuum intensity at disk center.

Open with DEXTER

Figure 2 shows continuum-intensity images at λ1078 nm and λ854 nm. In addition to the main umbra, several dark features reside in the common penumbra. But only one dark umbral core close to the right border of the penumbra at the location (x,y) ~ (28,15) Mm has an opposite magnetic polarity. Central emission of the Ca ii line is mainly present in the δ-umbra and along a diagonal starting just to the north of the δ-umbra and extending for about 30 Mm to the southeast. We refer to this alignment of brightenings as the “central emission line” (CEL), which is in some locations, but not everywhere, co-spatial with the PIL. The CEL is inconspicuous in continuum images.

A comparison of observed and fitted Stokes-profiles for three selected pixels (marked in Fig. 2) is shown in Fig. 3. Observed and fitted profiles agree very well for the main and the δ-umbra, but for a point on the PIL near the bright feature (BF) indicated in Fig. 2, a single-component inversion is not able to reproduce the observed multi-lobe profiles.

thumbnail Fig. 4

Total magnetic field Btot and Cartesian components Bx, By, and Bz (top to bottom) derived from inversions of the Fe i λ1078.3 nm (left) and Si i λ1078.6 nm (right) lines, respectively.

Open with DEXTER

The total magnetic field strength Btot = |B| and Cartesian components of the magnetic field vector B = (Bx,By,Bz) were determined with SIR for the Fe iλ1078.3 nm and Si iλ1078.6 nm lines. The vertical component Bz in Fig. 4 clearly reveals opposite polarities of the main and δ-umbrae. The parasitic polarity is positive like that of the leading sunspot in the active region. Inversions of the Fe i and Si i lines yield field strengths in the main umbra of up to 2600 G and 2570 G, respectively. Corresponding values in the δ-umbra only reach 2250 G and 2030 G. The height difference between the formation layers of the two NIR spectral lines is larger above the δ-umbra, because it is hotter than the main umbra. Nevertheless, dividing the difference in Btot by the difference in height covered by the two lines (see Balthasar & Gömöry 2008) indicates that the magnetic field decreases much faster with height in the δ-umbra than in the main umbra. In the δ-umbra, the mean formation height of the Fe i line is 150 km and that of the Si i line is 210 km. The corresponding values at the center of the main umbra are 140 km and 180 km, respectively. With these values, we find magnetic gradients of ΔBtotz ≈ 4.5 ± 1.4 G km-1 and 2 ± 0.5 G km-1, respectively. Btot is very low below the southeastern part of the CEL.

thumbnail Fig. 5

Doppler velocities obtained with SIR (top) based on the the Fe i λ1078.3 nm (left) and Si i λ1078.6 nm (right) lines, respectively. Line-fitting (bottom) yields LOS velocities of the chromospheric Ca iiλ854.2 nm (left) and photospheric Si i λ853.6 nm (right) lines. Note the different scale bar for the velocities from the Ca ii line.

Open with DEXTER

Bx varies smoothly with negative values in the region between the two umbrae. The perpendicular component By has a positive sign south of the main umbra and north of the δ-umbra, while it is negative on the other side. In the northern part of the spot, these opposite signs are clearly separated, but in the southern part, at the edge of BF, high values of By with opposite signs occur in close proximity. Inside BF, Bz is positive. The PIL bends at BF and separates from the CEL. At these locations, anomalous multilobed Stokes profiles arise (see right column in Fig. 3), which cannot be reproduced by a single-component inversion. (A two-component inversion is beyond the scope of this Letter.) Changing the magnetic azimuth by 180° even compounds this problem. Lites et al. (2002) explained such profiles in terms of interleaved systems of magnetic field lines. Such interleaved systems might facilitate reconnections, but in our case, the region was small enough not to cause a major flare.

The Doppler velocities derived from different lines are summarized in Fig. 5. Photospheric lines show the Evershed effect as blueshifts in the upper left part of the spots and as redshifts on the right side. The Evershed effect is most pronounced in the Si i λ853.6 nm line that forms at deep layers. The Evershed flow from the main umbra is interrupted near to the area where it reaches the CEL, and where the filaments end that come from the main umbra. Beyond this line, it drops and finally reaches another maximum at the outer edge of the penumbra. A strong blueshift of about 1 km s-1 is apparent just to the north of the δ-umbra along the PIL. We cannot finally conclude whether this blueshift corresponds to an upflow or an Evershed-like flow away from the δ-umbra because the spot is at some distance from disk center. This blueshift encompasses the PIL, where the magnetic field is horizontal, implying that the flow is more horizontal as well. The Evershed effect is also present in the Ca iiλ854.2 nm line. Patches of high velocities of up to 8 km s-1 exist along the CEL occurring both as blue- and as redshifts. High redshifts appear also in the periphery of the spot (cf., Balthasar et al. 2013). Because these patches are small, we interpret them as vertical flows, in agreement with downflows found by Martínez Pillet et al. (1994), although we do not find supersonic flows in photospheric layers (see Takizawa et al. 2012).

thumbnail Fig. 6

One-hour averaged HMI magnetogram (clipped to ±500 G in gray scale) with superimposed vectors of vDAVE. Length (see arrow below the figure) and color of the vectors represent the magnitude of vDAVE. Emission along the CEL is indicated by the red background.

Open with DEXTER

The differential affine velocity estimator (DAVE, Schuck 2005, 2006) estimates the magnetic flux transport velocity vDAVE by using an affine velocity profile and minimizing the deviation in the magnetic induction equation. The input HMI magnetograms were aligned, corrected for geometrical foreshortening, and adjusted for the μ-dependence of the magnetic field strength assuming that the field lines are perpendicular to the solar surface. DAVE depends on the temporal derivative of the magnetograms implemented as a five-point stencil, spatial derivatives carried out with the Scharr (2007) operator, and a sampling window of prescribed size (11 pixels ~ 3520 km). Arrows overlaid on the magnetogram in Fig. 6 indicate the one-hour averaged (09:30–10:30 UT) vDAVE in the δ-umbra’s neighborhood. The horizontal velocity structure differs from a regular sunspot. The most distinct velocity feature is a counterclockwise spiral motion centered between the δ-umbra and a neighboring dark feature. In addition, shear flows parallel to the CEL prevail locally with an imbalance of 0.2 km s-1 toward stronger flows on the δ-umbra side. This value is similar to the 0.3 km s-1 found by Tan et al. (2009) after a flare. Denker et al. (2007) discussed that these velocities do not necessarily build up magnetic shear, and the observed drift does not destabilize the magnetic configuration. Between main and δ-umbra, vDAVE is 0.2 km s-1 and crosses the PIL in the opposite direction as vLOS.

The CEL seems to be of much higher importance for this spot than the PIL. An explanation might be that new bipolar flux emerged southwest of the main umbra, and the CEL separates the new flux from the pre-existing one. Between main and δ-umbra, PIL and CEL are identical, but then the PIL bends at BF and crosses the new bipolar flux. This scenario also explains that the Evershed flow related to the main umbra ends at the CEL. In the range of repelling field lines from the two flux systems, chromospheric brightenings occur as well as between attracting field lines.

4. Conclusions

The magnetic field strength Btot at deep photospheric layers is 2250 G in the δ-umbra, somewhat weaker than the 2600 G of the main umbra. It decreases twice as fast with height in the δ-umbra than in the main umbra. We found a smooth transition of the magnetic vector field between the two umbrae of opposite polarity. At some distance from the δ-umbra, near BF,

we found a discontinuity of the horizontal magnetic field due to an interleaved system of magnetic field lines.

The δ-umbra causes its own Evershed flow, which is separated from that of the main umbra. Along the CEL, we found flows of ±8 km s-1 in chromospheric layers. We observed a horizontal flow of magnetic features (“magnetic flux transport velocities”) on one side parallel to the CEL with an imbalance of 0.2 km s-1. The magnetic configuration of the δ-spot is seemingly stable for at least 10 h. The group produced an M1.9-flare on June 14 and several C-flares during the previous day, but no flare of class C or larger was recorded by the GOES satellite during 19 h before and 6 h after our observation.

The central emission patches in the Ca ii line at 854.2 nm indicate that important physical processes happen in the chromospheric layers above δ-spots. For future observations, for instance, with the GREGOR Infrared Spectrograph (GRIS, Collados et al. 2012) at the GREGOR solar telescope (Schmidt et al. 2012), the chromospheric magnetic field should also be investigated, for example, to search for probable current sheets.

Acknowledgments

We are grateful to L. Bellot Rubio for carefully reading the manuscript and his valuable comments. The VTT is operated by the Kiepenheuer-Institut für Sonnenphysik (Germany) at the Spanish Observatorio del Teide of the Instituto de Astrofísica de Canarias. SDO/HMI data are provided by the Joint Science Operations Center – Science Data Processing. M.V. expresses her gratitude for the generous financial support by the German Academic Exchange Service (DAAD) in the form of a PhD scholarship. C.D. and R.E.L. were supported by grant DE 787/3-1 of the German Science Foundation (DFG).

References

All Figures

thumbnail Fig. 1

G-band image of the δ-spot in active region NOAA 11504.

Open with DEXTER
In the text
thumbnail Fig. 2

Maps of continuum intensities at λ1078 and λ854 nm, Ca iiλ854.2 nm line-core intensity, and maximum of the central Ca ii emission (top-left to bottom right). The ordinate is along the solar north-south direction. The black arrow points toward the disk center, and the white arrow indicates the bright feature BF near the PIL (thick red line). Contour lines denote the umbral and penumbral boundaries. Three white crosses mark the positions of the spectral profiles in Fig. 3.

Open with DEXTER
In the text
thumbnail Fig. 3

Observed (solid) and inverted (dashed) Stokes profiles in the umbra of the main spot (left), the δ-spot umbra (middle), and in the penumbra near BF (right). The wavelength is given with respect to the center of the Fe i λ1078.3 nm (left) and Si i λ1078.6 nm (right) lines, respectively. I0 is the quiet-Sun continuum intensity at disk center.

Open with DEXTER
In the text
thumbnail Fig. 4

Total magnetic field Btot and Cartesian components Bx, By, and Bz (top to bottom) derived from inversions of the Fe i λ1078.3 nm (left) and Si i λ1078.6 nm (right) lines, respectively.

Open with DEXTER
In the text
thumbnail Fig. 5

Doppler velocities obtained with SIR (top) based on the the Fe i λ1078.3 nm (left) and Si i λ1078.6 nm (right) lines, respectively. Line-fitting (bottom) yields LOS velocities of the chromospheric Ca iiλ854.2 nm (left) and photospheric Si i λ853.6 nm (right) lines. Note the different scale bar for the velocities from the Ca ii line.

Open with DEXTER
In the text
thumbnail Fig. 6

One-hour averaged HMI magnetogram (clipped to ±500 G in gray scale) with superimposed vectors of vDAVE. Length (see arrow below the figure) and color of the vectors represent the magnitude of vDAVE. Emission along the CEL is indicated by the red background.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.