high-resolution observations of siphon flows in a solar ... - IOPscience

The Astrophysical Journal Letters, 743:L9 (5pp), 2011 December 10 C 2011.

doi:10.1088/2041-8205/743/1/L9

The American Astronomical Society. All rights reserved. Printed in the U.S.A.

HIGH-RESOLUTION OBSERVATIONS OF SIPHON FLOWS IN A SOLAR MAGNETIC PORE 1

Salvo L. Guglielmino1,2 and Francesca Zuccarello3 Instituto de Astrof´ısica de Canarias, C/ V´ıa Láctea s/n, La Laguna, Tenerife E-38200, Spain; [email protected] 2 Departamento de Astrof´ısica, Univ. de La Laguna, La Laguna, Tenerife E-38205, Spain 3 Dipartimento di Fisica e Astronomia, Universit` a di Catania, I 95123 Catania, Italy Received 2011 May 2; accepted 2011 November 3; published 2011 November 16

ABSTRACT We investigate signatures of siphon flows in a region around a solar magnetic pore, observed in the photosphere at μ = 0.6, during its decay phase. We analyze high-resolution Stokes spectra acquired by Hinode/Solar Optical Telescope along the Fe i pair at 630.2 nm. We determine the vector magnetic field and the line-of-sight velocity by an inversion of the full Stokes vector using the SIR code. We also analyze photospheric G-band filtergrams. We find evidence of a transient siphon (counter)flow at the edge of the pore. An arch-shaped structure is found to have upflow motions of 4 km s−1 in the footpoint with a stronger magnetic field and positive polarity, and downflows of the same order of magnitude in the footpoint with opposite polarity and a weaker magnetic field. The event is different from those reported in previous observations of the Sun’s atmosphere and may represent a physical constraint for numerical models. Key words: Sun: photosphere – Sun: surface magnetism – techniques: high angular resolution

satisfies the conditions required for a siphon flow as predicted by Thomas & Montesinos (1991). Uitenbroek et al. (2006) analyzed observations over the Ca ii 854.2 nm infrared triplet line that showed evidence for a siphon-driven flow from a patch of weak magnetic field, with subsonic upflows, to the edge of a magnetic pore with a stronger field, with a supersonic downflow of well over 20 km s−1 . They also calculated synthetic profiles of the Ca ii 854.21 nm line which are consistent with the presence of a velocity discontinuity in the downflow, typically predicted for critical flows. Beck et al. (2010) reported on a dark-cored fibril observed in the chromosphere, analyzing spectropolarimetric observations taken simultaneously over the chromospheric Ca ii 854.2 nm line and the photospheric Fe i 630.25 nm line. The inversion of the spectra shows that the dark-cored fibril traces a siphon flow along the magnetic field lines. Recent investigations on the physical properties of the facular regions around magnetic pores in the solar atmosphere have been carried out by several groups using high-resolution data (Hirzberger 2003; Lagg et al. 2007; Giordano et al. 2008; Cho et al. 2010; Judge et al. 2010; Narayan & Scharmer 2010; Vargas Dom´ınguez et al. 2010). In this Letter, we report on the evidence for siphon counterflows at the edge of a pore, observed in the solar photosphere by the spectropolarimeter (SP) of the Solar Optical Telescope (SOT; Tsuneta et al. 2008) aboard the Hinode satellite (Kosugi et al. 2007). Thanks to the high spectral and spatial resolution achieved by SOT/SP, we are able to infer the full magnetic and dynamical configuration of the region around the pore, and to determine the motions within an arch-shaped structure found at an edge of the pore.

1. INTRODUCTION Siphon flows arise in flux tubes that connect magnetic elements of opposite polarity when there is a difference in gas pressure between the loop footpoints owing to different magnetic field strengths (Meyer & Schmidt 1968). In the solar atmosphere they have been studied in connection with the Evershed flow, but they might also occur in coronal loops (Noci 1981). Moreover, siphon flows have been proposed as a mechanism able to strengthen photospheric magnetic fields, alternative to convective collapse (Thomas 1988). Numerical investigations, which began during the late 1980s, led to the modeling of these flows over the next decade (e.g., Thomas 1988; Degenhardt 1989; Montesinos & Thomas 1989; Thomas & Montesinos 1990, 1991, 1993; Degenhardt & Kneer 1992; Degenhardt et al. 1993; Montesinos & Thomas 1997). The standard picture for siphon flows in an isolated magnetic flux tube, using the thin-flux tube approximation, depends on the critical speed in the tube, vC , given by 1/2 vC = vS vA / vS2 + vA2 , where vS is the internal sound speed and vA the Alfvén speed in the tube. The flow is subcritical if speeds are less than vC everywhere, critical if speeds are greater than vC in part of the tube, and supercritical if speeds are greater than vC everywhere (Thomas 1988; Degenhardt et al. 1993). In subcritical flows, the speed increases in the upstream footpoint and decreases in the downflowing footpoint, reaching a maximum value at the top of the arch. On the contrary, the magnetic field has a minimum at the top of the arch. Observational confirmations of this phenomenon have proved very elusive. Few examples of direct observations of siphon flows occurring in the low atmospheric layers of the Sun have been reported. Rüedi et al. (1992) found in infrared spectra, taken over the plage of an active region, a positive polarity field with an upflow of up to 2 km s−1 and an intensity of ≈1200 G at a side of the neutral line, while at the other side they found a negative field with a downflow of up to 1 km s−1 and an intensity of ≈1500 G. This configuration

2. OBSERVATIONS AND DATA ANALYSIS Two spectropolarimetric scans of a solar region located near the southeast solar limb at (−375 ,−675 ), corresponding to a heliocentric angle μ = 0.6, were taken on 2009 September 18 at 8:30 and 9:40 UT. The SOT/SP acquired Stokes I, Q, U, and V profiles along the pair of Fe i lines at 630.15 and 630.25 nm. The field of view (FoV) covered by these observations is of about 76 × 81 , with an effective pixel size of 0. 16, an integration 1

The Astrophysical Journal Letters, 743:L9 (5pp), 2011 December 10

Guglielmino & Zuccarello

time of 4.8 s per slit position, and a signal-to-noise ratio of 103 (Normal Map mode). Simultaneously with the beginning of each SP raster scan, two pairs of broadband filtergrams were acquired in the G band (λ = 430.56 nm ± 0.8 nm) and Ca ii H (λ = 396.85 nm ± 0.3 nm) by the SOT Broadband Filter Imager (BFI), with a delay of a minute. These images cover a FoV of 223 × 111. 5, with a spatial sampling of 0. 1 pixel−1 . We have corrected SOT/BFI images and raw SOT/SP data for dark current, flat field, and cosmic rays using the standard routines of the SolarSoft IDL package for Hinode/SOT observations. A quick look at some physical parameters deduced from the SP raster scans, i.e., the longitudinal and transverse apparent components of the vector magnetic field B and the line-ofsight (LOS) velocity, has been obtained from the analysis of the level 1D data maps generated by the ancillary SolarSoft routine stksimages_sbsp. The SP raster scan begun at 9:40 UT has been further analyzed carrying out an LTE inversion of the full Stokes vector with the SIR code (Ruiz Cobo & del Toro Iniesta 1992). First, we have found the 64 profiles with the lowest total polarization degree P = [(Q2 + U 2 + V 2 )/I 2 ]1/2 in a subset of the scan, ignoring the 16 pixels at the top and bottom of the slit. We have computed a reference profile by averaging those profiles with the lowest P signal. Then, all the spectra in the scan have been normalized to the continuum of the reference profile and corrected for limb darkening. The reference profile has also been used as a global stray-light profile. The SIR inversion of the spectra yields an estimate of the temperature stratification in the range 1.4 > log τc > −4.0, where τc is the optical depth of the continuum at 500 nm. We have used a one-component inversion, taking into account the magnetic filling factor due to stray-light contamination. The code provides an estimate of the magnetic field strength B, the inclination and azimuth angles in the LOS reference frame, γ and φ, the LOS velocity vLOS , as well as the micro-turbulent velocity. We have carried out the inversion with four iteration cycles in order to get the fitted values. The physical quantities are kept constant with respect to τc , except for the temperature which has up to four nodes in the last cycle (Ruiz Cobo & del Toro Iniesta 1992). The inclination and azimuth angles—γ and φ—have been transformed into the local solar frame zenith and azimuth angles—ψ and χ . To solve the ±180◦ ambiguity of φ, we have used the non-potential field calculation method described by Georgoulis (2005). The LOS velocity has been calibrated by assuming that the plasma belonging to the pore—defined in the scan as the pixels with intensity lesser than the 80% of the mean continuum value—was globally at rest. All the SOT/BFI filtergrams and SOT/SP continuum images have been aligned through cross-correlation algorithms and by using the iTools IDL package, which handles the image transparency and allows the user to check the correspondence of the common structures.

Figure 1. Region around the pore during the SOT/SP scan begun at 09:40 UT: G-band filtergram (top panel) and the Stokes V signal integrated in the 630.15 nm Fe i line (bottom panel) [V /Ic = ±0.8%]. The square indicates the part of the FoV analyzed throughout the Letter. Here and in Figure 2, north is up and west is to the right.

We display a closer view of the region around the pore in Figure 2 (first row). A comparison of the pore size at 8:30 UT and 9:40 UT indicates that it has been observed during its decay phase. The intensity increase found at the southeastern edge of the pore in the red continuum of the Fe i 630.25 nm line and in the G band in Figure 1 may be attributed to the brightening of the limbward granules adjacent to the pore, known from both observations and numerical simulations (e.g., Cameron et al. 2007). We note an emission enhancement at 9:40 UT in the G band at the southern edge of the pore. In Figure 2 (second row), we display the maps of the LOS velocity vLOS and of the magnetic flux Φ, given by Φ = ε f B cos ψ, where f is the filling factor and ε = 3.72 × 1014 cm2 is a conversion factor, which corresponds to the flux for a pixel with vertical field of intensity B = 1 G and f = 1 for these SP measurements at μ = 0.6. An arch-shaped structure is found at [−367 , 667 ] in the Φ map. This feature corresponds to a magnetic region with two adjacent, positive–negative polarities, characterized by upflows and downflows in the vLOS map, respectively. This arch-shaped structure is not clearly recognizable in the level 1D maps for

3. RESULTS Figure 1 shows the solar region observed by Hinode/SOT in the photosphere. A pore is clearly visible in the G-band filtergram in the left part of the FoV, surrounded by a bright facular region. The map of Stokes V, which can be regarded as a LOS magnetogram, indicates that the pore is embedded in the negative polarity field of the region. 2



Figure 2. Region around the pore during the two SOT/SP scans: images refer to the red continuum of the Fe i 630.25 nm line (first row). In the bottom row, maps of the physical parameters inferred by the SIR inversion for the scan begun at 9:40 are displayed: LOS velocity (left panel) and magnetic flux in the local solar frame (right panel). The cutoff of the logarithmic scale is at 3.72 × 1014 Mx. Shown are only the pixels with total polarization degree P > 1%. The segments over the magnetic flux map pinpoint the pixels analyzed more in detail in Figure 4.

the scan begun at 8:30 UT, and at that time there are no peculiar motions in this region. Its total length is about 4 Mm. There is an unbalance of about a factor of two in the total flux content for the two polarities: Φ+ ≈ 3.×1018 Mx and Φ− ≈ 1.5×1018 Mx. The feature is also visible in Figure 3, where we display the map of the red continuum of the Fe i 630.25 nm line, the magnetic field strength B, the zenith angle ψ, i.e., the inclination angle, and the azimuth angle χ , for the second scan. All these maps have been transformed into the heliographic plane to get rid of projection effects due to the relatively limbward position of the observations. Note that there is a slight misalignment between the pore’s center in the continuum image and the pore’s center in ψ and χ due to second-order effects in the deprojection into the heliographic plane (Gary & Hagyard 1990). The pore has a strong vertical field up to ≈2200 G, with a slightly more horizontal region of field strength of ≈1800 around the inner part. The feature consists of two main magnetic elements joined by plasma with a weaker field. Such a structure is very reminiscent of a magnetic flux tube. The upflowing footpoint has a field strength of ≈1000 G, while the downstreaming footpoint of weaker field strength (≈500 G) is cospatial with

the photospheric brightening seen in the G band and in the Fe i line. Keep in mind that the G-band images are not simultaneous with the SP measurements at those positions. We plot in Figure 4 the quantities B and vLOS along the three segments shown in Figure 3, which follow the apparent path of the plasma along the arch-shaped structure. We indicate with arrows the inclination angle ψ of each pixel. The LOS velocity decreases in the upflowing stream and increases again in the part of the structure with downward motions. The absolute values of vLOS are similar for the two footpoints, reaching up to 4 km s−1 at their ends. The magnetic field intensity is stronger in the first segment, being on average ≈1200 G, then drops to ≈250 G in the more horizontal second segment, and rises again to ≈400 G in the third segment. 4. DISCUSSION AND CONCLUSIONS We have analyzed both filtergrams and polarimetric measurements performed by Hinode/SOT. These observations of the solar photosphere show peculiar flows within an arch-shaped structure at the edge of a magnetic pore. The topology inferred 3



Figure 3. Magnetic configuration of the region around the pore during the SOT/SP scan begun at 09:40 UT. Maps of continuum, magnetic field strength, ψ, and χ angles are shown in the heliographic plane (only the pixels with a total polarization degree P > 1%). The rhomboid indicates the deprojection of the FoV displayed in Figure 2. Contours pinpoint the location of the pore.

Figure 4. Magnetic field strength and LOS velocity of the pixels lying along the three segments indicated in Figure 2. The arrows indicate the inclination angle, with their colors also referring to the colors of the segments drawn in Figure 2.

by the physical parameters suggests that we have observed a magnetic flux tube characterized by plasma motions from a footpoint toward the opposite one, following the path of the field lines along the tube.

An important difference from the usual siphon flow models is that the field is stronger in the upflowing stream and weaker in the downstreaming footpoint. In this sense, we refer to it as a countersiphon flow. The arch-shaped structure was not 4



combine an excellent spectral and spatial resolution, like that achieved in these measurements, with higher temporal cadence.

present in the scan acquired at 08:30 UT, therefore it might have emerged during this short timescale. In this case, we could expect downflows in both legs, as generally observed, e.g., in the Arch Filament Systems (Spadaro et al. 2004). In these structures, some possible asymmetries may be due to the action of the Coriolis force or to a remnant longitudinal motion in the original mechanical equilibrium which leads to counterflows, but this does not seem the case. Another interpretation of such countermotions involves mass flow along field lines which bend down, as proposed in a different context by Zirker et al. (1998). The presence of the countersiphon flow could also be related to a dynamical instability related with a shock or a possible reconnection. We speculate therefore that some kind of energy-supplying mechanism may act at the footpoints, as photospheric brightenings could suggest (see, e.g., Lagg et al. 2007). We are not able to classify the flow as subcritical or critical. Recall that, owing to the position of the region on the solar disk (μ = 0.6), the vLOS values retrieved by SIR represent only a lower limit for the real velocities, as we lack information about the transverse component of v. The sound speed, for typical photospheric values of density and temperature, is about 8 km s−1 , while the Alfvén speed is 6.5 km s−1 for a magnetic field of 1000 G. We obtain vC ≈ 5 km s−1 . Therefore it is likely that, somewhere in part of the tube, v can be greater than vC , so the flow might be critical. Although the topology of the magnetic field described here is self-consistent, we actually lack simultaneous observations of the overlying upper atmospheric layers which can definitely prove the existence of the flux tube containing the siphon flows. Because of the relatively large values of the force-free α parameter, which are ≈10 times larger in the region of interest than in the pore itself, a potential or even non-linear force-free extrapolations would be of little help in confirming the connectivity between the up- and downflowing patches, particularly at such a small spatial scale. Thus, we cannot rule out other possibilities for the onset of the observed peculiar flows. Our findings may provide a new issue for the modeling of siphon-driven flows, to be clarified with the help of more realistic, three-dimensional MHD simulations. Unfortunately, the timing of these observations is very poor, and no scans were acquired by SOT/SP after that in which siphon flows have been detected. Thus, we cannot make inferences about the duration and the evolution of the flow. Further investigations should

We are grateful to M. Georgoulis for kind discussions during data analysis. Financial support by the European Commission through the SOLAIRE Network (MTRN-CT-2006-035484) is gratefully acknowledged. Hinode is a Japanese mission developed and launched by ISAS/JAXA, with NAOJ as a domestic partner and NASA and STFC (UK) as international partners. It is operated by these agencies in co-operation with ESA and NSC (Norway). Use of NASA’s Astrophysical Data System is gratefully acknowledged. Facility: Hinode REFERENCES Beck, C., Tritschler, A., & Wöger, F. 2010, Astron. Nachr., 331, 574 Cameron, R., Schüssler, M., Vögler, A., & Zakharov, V. 2007, A&A, 474, 261 Cho, K.-S., Bong, S.-C., Chae, J., Kim, Y.-H., & Park, Y.-D. 2010, ApJ, 723, 440 Degenhardt, D. 1989, A&A, 222, 297 Degenhardt, D., & Kneer, F. 1992, A&A, 260, 411 Degenhardt, D., Solanki, S. K., Montesinos, B., & Thomas, J. H. 1993, A&A, 279, L29 Gary, G. A., & Hagyard, M. J. 1990, Sol. Phys., 126, 21 Georgoulis, M. K. 2005, ApJ, 629, L69 Giordano, S., Berrilli, F., Del Moro, D., & Penza, V. 2008, A&A, 489, 747 Hirzberger, J. 2003, A&A, 405, 331 Judge, P. G., Tritschler, A., Uitenbroek, H., et al. 2010, ApJ, 710, 1486 Kosugi, T., Matsuzaki, K., Sakao, T., et al. 2007, Sol. Phys., 243, 3 Lagg, A., Woch, J., Solanki, S. K., & Krupp, N. 2007, A&A, 462, 1147 Meyer, F., & Schmidt, H. U. 1968, Z. Angew. Math. Mech., 48, 218 Montesinos, B., & Thomas, J. H. 1989, ApJ, 337, 977 Montesinos, B., & Thomas, J. H. 1997, Nature, 390, 485 Narayan, G., & Scharmer, G. B. 2010, A&A, 524, A3 Noci, G. 1981, Sol. Phys., 69, 63 Rüedi, I., Solanki, S. K., & Rabin, D. 1992, A&A, 261, L21 Ruiz Cobo, B., & del Toro Iniesta, J. C. 1992, ApJ, 398, 375 Spadaro, D., Billotta, S., Contarino, L., Romano, P., & Zuccarello, F. 2004, A&A, 425, 309 Thomas, J. H. 1988, ApJ, 333, 407 Thomas, J. H., & Montesinos, B. 1990, ApJ, 359, 550 Thomas, J. H., & Montesinos, B. 1991, ApJ, 375, 404 Thomas, J. H., & Montesinos, B. 1993, ApJ, 407, 398 Tsuneta, S., Ichimoto, K., Katsukawa, Y., et al. 2008, Sol. Phys., 249, 167 Uitenbroek, H., Balasubramaniam, K. S., & Tritschler, A. 2006, ApJ, 645, 776 Vargas Dom´ınguez, S., de Vicente, A., Bonet, J. A., & Mart´ınez Pillet, V. 2010, A&A, 516, A91 Zirker, J. B., Engvold, O., & Martin, S. F. 1998, Nature, 396, 440

5