Formation of Transient Coronal Holes during Eruption of a Quiescent Filament and its Overlying Sigmoid
aa r X i v : . [ a s t r o - ph . S R ] J a n Chin. J. Astron. Astrophys. Vol. 8 (2008), No. 3, 000–000( ) Chinese Journal ofAstronomy andAstrophysics
Formation of Transient Coronal Holes during Eruption of aQuiescent Filament and its Overlying Sigmoid ∗ Li-Heng Yang , , Yun-Chun Jiang and Dong-Bai Ren , National Astronomical Observatories/Yunnan Observatory, Chinese Academy of Sciences, Kunming650011; [email protected] Graduate school of Chinese Academy of Sciences, Beijing 100049
Abstract
By using H α , He I 10830, EUV and soft X-ray (SXR) data, we examined a filamenteruption that occurred on a quiet-sun region near the center of the solar disk on 2006 January12, which disturbed a sigmoid overlying the filament channel observed by the GOES-12
SXR Imager (SXI), and led to the eruption of the sigmoid. The event was associated witha partial halo coronal mass ejection (CME) observed by the Large Angle and SpectrometricCoronagraphs (LASCO) on board the Solar and Heliospheric Observatory (
SOHO ), and re-sulted in the formation of two flare-like ribbons, post-eruption coronal loops, and two tran-sient coronal holes (TCHs), but there were no significantly recorded
GOES or H α flares cor-responding to the eruption. The two TCHs were dominated by opposite magnetic polaritiesand were located on the two ends of the eruptive sigmoid. They showed similar locations andshapes in He I 10830, EUV and SXR observations. During the early eruption phase, bright-enings first appeared on the locations of the two subsequent TCHs, which could be clearlyidentified on He I 10830, EUV and SXR images. This eruption event could be explained bythe magnetic flux rope model, and the two TCHs were likely to be the feet of the flux rope. Key words:
Sun: filaments — Sun: chromosphere — Sun: coronal mass ejections (CMEs)
Coronal mass ejections (CMEs) are sudden eruptions of magnetized plasma from the solar corona into theheliosphere that can represent a large-scale rearrangement of the coronal magnetic field. CMEs have beenconsidered to be the cause of interplanetary shocks and the driver of geomagnetic storms. In particular, haloCMEs have a great influence on space weather, because they may be directed earthward and so can lead tosignificant geo-disturbances (jing et al. 2004). Although the halo CMEs have been extensively investigated,their origin and initiation are still not well understood. Therefore, identifying the on-disk CME signaturesis vital for detecting earth-directed CMEs and for an understanding of the ultimate driving mechanism ofCMEs, as well as for the forecast of space weather. Various warning signs of CMEs have been identifiedby in recent work, such as flares, eruptions of filaments and sigmoidal structures, transient coronal holes(TCHs), and so on (Hudson & Cliver 2001). However, the relationship between the on-disk surface activ-ities and CMEs needs to be further checked. In a few CME events, very weak surface activities led to anearthward CME. For example, the event on 1997 January 6 showed that a very weak and unimpressive solaractivity induced a halo CME and a geomagnetic storm (Webb 1998). Another similar event reported byjiang et al. (2006) showed that a filament eruption without associated H α and GOES flare was related to apartial halo CME. These weak surface activities are significant to predict the space weather, and so cannotbe ignored. ∗ Supported by the National Natural Science Foundation of China.
L. H. Yang, Y. C. Jiang & D. B. Ren
TCHs have been identified as key on-disk indicators of CMEs in some halo events with observationsfrom
Yohkoh
Soft X-ray (SXR) Telescope and the Extreme Ultraviolet Imaging Telescope (EIT) on boardthe
Solar and Heliospheric Observatory ( SOHO ) (Thompson et al. 1998; Wang et al. 2000). It has beenfound that they are often associated with eruptions of filaments or sigmoidal structures (Sterling et al. 2000;Jiang et al. 2007b), and appear as darkenings in SXR and EUV and brightenings in He I 10830 ˚A (Toma etal. 2005) in regions of unipolar opening magnetic field, with lifetimes of about a few hours to days generally,and can give rise to transient high-speed wind (Rust 1983).Since their typical timescale of formation is less than an hour, much shorter than the typical radiativecooling timescale of about 36 hours in the corona (Hudson et al. 1996), the TCHs are often interpretedas due to density depletions, rather than temperature variation. A pair of compact and symmetric TCHsdominated by opposite polarities can form immediately close to the two ends of an eruptive filament or,sometimes, a sigmoid. It is believed that they mark the positions of the footpoints of the flux rope andthe mass loss from them is expelled into the CMEs (Sterling & Hudson 1997; Webb et al. 2000; jiang etal. 2006).Moreover, such dual TCHs can have H α counterparts, which means that they could extend from thecorona deep into the chromosphere (Jiang et al. 2003). More recently, Toma et al.(2005) and Jiang etal.(2007a) found that some twin TCHs showed up above a clear chromospheric network pattern consist-ing of patchy bright/dark H α /He I 10830 ˚A plages with strongly concentrated magnetic network elements,as darkenings/brightenings of these plages during eruptions of the associated filament. It was further foundthat these TCHs were often preceded by chromospheric and coronal brightenings in the rising phase of theassociated flares. Up to now, however, only a few such cases were observed, and the cause and role of theplages and preceding brightenings in the formation of TCHs are still not clear. To understand clearly therelationship of the brightenings and the following TCHs, and the eruption process and the associated CMEs,further observations are strongly needed.In this paper, we present multi-wavelength observations of the eruption of an S-shaped filament thatoccurred on 2006 January 12, in which twin TCHs were formed in He I 10830 ˚A, EUV, and SXR. Thiseruption was not associated with any recorded GOES or H α flare but was directly related to a partial haloCME observed by the Large Angle and Spectrometic Coronagraphs (LASCO) on board SOHO . We willshow that the brightenings clearly appeared at the locations of the following TCHs and the TCHs werepossibly part of the on-disk proxy of the CME source region.
For the present work, the following data are used:1 Full-disk He I 10830 ˚A intensity and velocity images from the Chromospheric Helium ImagingPhotometer (CHIP, MacQueen et al. 1998) and H α disk images from the Polarimeter for Inner CoronalStudies (PICS) at the Mauna Loa Solar Observatory (MLSO). The CHIP data were acquired using a tunableLyot filter ( ≈ s − . The cadenceof these images is 3 minutes and the pixel size is . ′′ . The PICS disk images were acquired by using anarrowband filter of ± s − . These imageshave the same cadence with the CHIP data and a resolution of . ′′ pixel − (Toma et al. 2005).2. Full-disk EUV images from the EIT (Delaboudini ´ e re 1995) on SOHO . EIT images are taken in fourspectral bands centered on Fe IX / X (171 ˚A), Fe XII (195 ˚A), Fe XV (284 ˚A), He II (304 ˚A), which allowimaging of the solar plasma at temperatures ranging from 6 × K to 3 × K. For the present work, theEIT provided 195 ˚A images with cadence of 12 minutes and pixel size of . ′′ , while the 304 ˚A and 284 ˚Aimages were taken only once every 6 hours.3. Full-disk SXR images from the GOES-12
Solar X-ray Imager (SXI, Hill et al. 2005). For the cur-rent work, SXR images were taken every two minutes using thin polyimide filter. These images cover awavelength range of 0.6-60 nm (sensitive to the temperature of 0.9-20 MK), at a pixel size of ′′ .4. Full-disk longitudinal magnetograms from the Michelson Doppler Imager (MDI, Scherrer etal. 1995) on SOHO , with cadence of 96 minutes and pixel size about ′′ . ilament Eruption, Transient Coronal Holes and CME 3 Fig. 1
MLSO/PICS H α line center (a), MLSO/CHIP He I 10830 ˚A intensity (b), EIT 195 ˚A (c),and SXI SXR (d), showing the appearance of the filament, “F”, before its eruption. The field ofview (FOV) is ′′ × ′′ .5. C2 and C3 white-light coronagraph data from LASCO, which cover the range of 2 to 6 and 4 to 32solar radii, respectively (Brueckner et al. 1995). The eruptive filament, “F”, was located on a quiet-sun region in the southern hemisphere centered atS03W16, on January 12. Figure 1 shows the general appearance of the eruptive region before the F eruptionin the H α , He I 10830 ˚A, EIT 195 ˚A, and SXR images, respectively. In the H α image (Fig. 1a), F, indicatedby the two black arrows, clearly shows an inverted S shape. As expected, it lies along a polarity inversionzone of the photospheric magnetic fields, and the corresponding positive/negative polarity on either sideis marked by plus/minus sign. When F can be identified as a dark feature in He I 10830 ˚A (Fig. 1b), itsEUV counterpart was clearly seen and showed a similar inverted S shape in the EIT 195 ˚A image (Fig. 1c).However, F in EUV appeared more complicated and bifurcated, and the same bifurcation was seen in theSXR image. In the higher temperature wavebands, F appeared increasingly broader. From SXI/SXR image(Fig. 1d), we also note an inverted S sigmoid located just above the F, indicating that they could belong tothe same topological structure. This is consistent with the observation of Pevtsov (2002) that the chromo-spheric filament and the coronal sigmoid had a close spatial association. Although we can not determine F’schirality from its barb orientations due to the poor spatial resolution of the Hα observations, we concludethat the axial field of F before its eruption was directed eastward since the sign of the photospheric magneticfield in the F ends was negative in the southeast and positive in the northwest (see Fig. 3a). According tothe definition given by Martin et al. (1994), F was dextral. However, it is noted that the chirality patternconsistent with both the dextral F and the inverted S sigmoid disobeyed the hemispheric chirality rule forFs and sigmoids in the southern hemisphere (Zirker et al. 1997).Figure 2 shows the evolution of the eruption in H α , He I 10830 ˚A, EIT 195 ˚A, and SXI/SXR. The erup-tion showed up as a total disappearance of F in H α . We see that its inverted S shape was clearly discernibleat 23:16 UT, but parts of it became invisible at 23:25 UT, and the whole F disappeared at 23:37 UT. In theHe I 10830 ˚A velocity observations, however, F mainly showed blueshift signature, clearly indicating thatthe disappearance was due to its eruption. We also note that the blueshift signature could be detected at23:44 UT, lagging the complete disappearance of F in H α by 7 minutes. This possibly suggests that theerupting F was fast enough to be Doppler shifted out of the filter of the MLSO/PICS instrument but wasstill in the detectable range of the MLSO/CHIP instrument. In EIT 195 ˚A images, we see that the darkEUV counterpart of the H α F also showed a consistent eruptive process. As a distinct feature of the event,however, the SXR sigmoid was strongly disturbed during the eruption, and it appeared that the sigmoidalso underwent an eruption, which was made manifest by the clear disappearance of some loops that con-tained it after the F eruption. Thus, this event involved the eruptions of both the F and the sigmoid. This isgreatly different from the cases studied by Pevtsov (2002) and jiang et al. (2007b), in which coronal sig-moids underwent activations and eruptions while the underlying H α filaments were left largely untouched.The F eruption was accompanied by the occurrence of two flare-like ribbons, “R1” and “R2”, in H α , He I L. H. Yang, Y. C. Jiang & D. B. Ren
Fig. 2
MLSO/PICS H α line center (a), MLSO/CHIP He I 10830 ˚A velocity (b) and intensity(c), EIT 195 ˚A (d) and SXI SXR (e), showing the evolution of the event. To clearly exhibit thebrightenings in the early formation phase of the TCHs, a difference image is presented in thesecond column of (d)/(e). The FOV is the same as in Fig. 1.10830 ˚A and EIT 195 ˚A lines. They were located on opposite sides of the eruptive F, and were dark in He I10830 ˚A due to the increased absorption. Notably, a bright flare-like kernel, “Fk”, was visible from 23:35 to23:44 UT in He I 10830 ˚A intensity images, this means that the energy release was large enough to changethe He I 10830 ˚A absorption line into emission (Penn & Kuhn 1995).After the F eruption, post-eruptive loops, “PELs”, gradually appeared in EUV and SXR observations.They connected the two flare-like ribbons, and their footprints expanded with the increasing separation ofthe ribbons. All of these observations indicated typical features of flares associated with F eruptions, but nooptical flare was reported by the Solar Geophysical Data online around the time of the event. Furthermore,despite the GOES
SXR flux showing an increase relative to the background level after the F eruption (seeFig. 5), no
GOES flare above X-ray class B1 was recorded. Therefore, it seems that the flare-like ribbonswere too weak to be regarded as an H α or GOES flare. As the SXR and H α wavebands each contain only ilament Eruption, Transient Coronal Holes and CME 5 Fig. 3
MDI magnetogram (a), MLSO/CHIP He I 10830 ˚A (b), EIT 304 (c), 195 (d), and 284 ˚A(e), and SXI/SXR (f) difference images. The two TCHs, “TCH1” and “TCH2” are predominatedby opposite polarities and are located on the two F ends. The TCHs and the pre-eruptive H α Fare depicted by the black and white contours of (a). The FOV is the same as in Fig. 1.
Fig. 4 (a) Composite image of the inner EIT 195 ˚A and the outer LASCO C2 difference images.(b) Height of the CME front as the function of time.5–10% of the total radiated energy, it may not be proper to identify a flare with only these wavebands (Zhouet al. 2003). The EIT 195 ˚A light curve measured in the flare-like region is plotted in the Figure 5(a), and wenote that the brightness increased about 125% from the starting time to the time of maximum enhancement.Due to the brightness at the start time was greater than that of the background, the enhancement could beconsidered as an EUV flare according to the criterion given by Zhou et al. (2003).
L. H. Yang, Y. C. Jiang & D. B. Ren
Fig. 5
Time profiles of
GOES-12 /SXR in the energy channel of 1-8 ˚A (dashed thin lines in Panelsa and c), in arbitrary unites and in unites normalized to one, respectively. The light curves of EIT195 ˚A (a), SXI/SXR (b), He I 10830 ˚A (c) intensities in areas centered on the TCH1, TCH2,and R(indicated by the boxes in Fig. 3(d)), and H α intensities (c) in an area centered on the Rare computed from the intensity averaged over these regions. EIT 195 ˚A is normalized to themaximum value, and the SXI/SXR, He I 10830 and H α are normalized to one.As another remarkable feature of the event, two TCHs, labelled “TCH1” and “TCH2”, were formedduring the eruptions of the F and the sigmoid. After the eruptions, they were quite obvious in He I 10830 ˚Aintensity, EIT 195 ˚A, and SXI/SXR images, see Figure 2. As in the cases investigated by jiang et al. (2003),Toma et al. (2005) and jiang et al. (2007a), it is noted that the dual TCHs were also preceded by faintbrightenings in their early formation phase, which can be clearly seen in the He I 10830 ˚A intensity image,and the EVU and SXR difference images; see the second row of Figure 2 (indicated by the white arrows). ilament Eruption, Transient Coronal Holes and CME 7 They were located at the sites of the following TCHs, and were kept away from the two flare-like ribbons, sowere not the result of a spreading and expansion of these ribbons. Thus it appears that formation of TCHs isoften associated with certain brightenings. Figure 3 shows the difference images from the pre-event imagesat He I 10830 ˚A, 304 ˚A, 195 ˚A, 284 ˚A and SXI/SXR lines. The TCHs, appearing as brightening regionsin He I 10830 ˚A and dimming regions in EUV and SXR, had similar locations and shapes in these linesformed over a temperature range from several K to 2 × K, so indicating that they are caused bydensity depletions rather than by temperature variations (Thompson et al. 1998). Moreover, they occurredin opposite polarity regions near the two ends of the erupted sigmoid and flanked the central PEL.A partial halo CME was observed by SOHO/LASCO around the time of the event. The CME was firstdetected by LASCO C2 at 00:54 UT on January 13, and later became visible in C3 at 02:42 UT locatedabove the W limb. According to the measurements of Seiji Yashiro, its center position angle (PA) was ◦ with a width of ◦ . Figure 4a presents a composite image of an inner EIT 195 ˚A with an outer LASCOC2 difference image. We see that the eruptive region was nearly located along the direction of the centralPA of the CME. The height-time (H-T) plot of the CME front at ◦ is shown in Figure 4b. The averagespeed of the CME front given by a linear fitting was 433 ± s − , and the average acceleration givenby a second-order polynomial fitting was -1.7 ± s − (shown in the upper left corner of Fig. 4b),indicating that the CME was a slow one, which is consistent with the results given by sheeley et al. (1999)that slow CMEs were often associated with filament eruptions. From the back extrapolation of the second-order polynomial fitting of the CME H-T plot to the solar disk center, we estimate that the onset time ofthe CME was about 23:45 UT on January 12 (marked by a vertical bar in Fig. 4 (b)), which was very closeto the start time of the GOES
GOES α intensities of the flare-likeregion, respectively. Beginning at 23:45 UT on January 12 (marked by the vertical line in Fig. 5), the GOES
SXR flux showed a weak increase, taking about 1 hour to reach the peak, which was below the X-ray classB1 level. So the enhancement was too weak to be recorded as a
GOES flare. Furthermore, we see that EIT195 ˚A, SXI/SXR light curves measured in the flare-like region are similar to the
GOES α light curves was earlier the timeof disappearance of F. Especially, with the appearance of the flare-like kernel, the light curve showed anincrease between 23:35 to 23:44 UT, January 12. We also note that during the formation of the two TCHs,the EIT 195 ˚A and SXR intensities obviously decreased and the He I 10830 ˚A intensity distinctly increased.In addition, we found that during the early eruption phase there was an increase in the SXR flux curvesand a decrease in the He I 10830 ˚A flux curves due to the appearance of the brightenings, and the increase(decrease) lasted only during the rise phase of GOES
We have made an investigation on a filament eruption near the center of the solar disk on 2006 January12, and on the associated brightenings, TCHs, and a halo CME, as well as the eruptive sigmoid above thefilament. The main observational results are as follows: (1) The disappearance of the inverted S-shapedfilament was followed by two flare-like ribbons, two TCHs, and post-eruption coronal loops, but no evidentX-ray and H α flares were recorded. (2) During the early eruption phase, brightenings first appeared on thesites of the two TCHs, and were short lived during the rise phase of the flare, then the two TCHs wereformed near the two ends of the sigmoid, dominated by opposite polarities, which were clearly visible inHe I 10830 ˚A, EUV and SXR, with similar locations and shapes. (3) The partial halo CME showed a closespatial and temporal relation to the filament eruption and the two TCHs.This event was not related to any recorded H α and GOES flares, while the usual feature of flares oftwo flaring ribbons, was identified in multi-wavelength observations (H α , He I 10830 ˚A, EIT 195 ˚A), but itis hard to associate such a weak surface activity to the partial halo CME. Fortunately, the filament and thesigmiod eruption, and the formation of the two TCHs, which are considered to be predictors of the CME, L. H. Yang, Y. C. Jiang & D. B. Ren directed our attention to this event. It is incredible that such a very minor change in the chromosphere couldbring about such a major coronal perturbation. Similar results have been obtained by Webb et al. (1998),and they found that some weak on-disk activities on 1997 January 6 led to a CME and the “problem”geomagnetic storms. Later, Shakhovskaya et al. (2002) showed that a prominence eruption on 2000 August11 with no associated flares resulted in a faint CME. More recently, jiang et al. (2006) reported a filamenteruption on 1999 March 21 without flares recorded, which led to a partial halo CME. These events indicatethat CMEs may or may not be relate to flares in the chromosphere ( ˇ S vestka 2001) and that weak eventsare valuable in identifying the source regions of the earth-directed CMEs, important for predicting spaceweather.We have described a rare event where the eruptions of a quiescent filament and an overlying sigmoidled to the formation of the two TCHs, which is not related to any recorded flare events, which is very similarto the event of 1997 October 23 presented by Hudson and Cliver (2001). The H α filament and sigmoid areconsidered to be the signature of a flux rope system (Pevtsov 2002), and the eruption event can be explainedby the flux model of CME, the two TCHs representing the two footprints of the flux rope (Sterling andHudson 1997; Jiang et al. 2003; Jiang et al. 2006). However, during this event both the filament and thesigmoid erupted, unlike the observations by Pevtsov (2002) where the sigmoid activated and disappearedwithout a corresponding filament eruption.According to the analytic 3-D MHD model of flux ropes proposed by Gibson and Low (2000), the twoTCHs would appear anti-symmetrically on either side of the photospheric neutral line, and interpreted tobe the counterpart of the CME cavity seen at the limb. The formation of the two TCHs could be explainedby the opening of previous closed field lines or by the expansion of those related to the flux rope. The twoTCHs clearly appeared on the 304 ˚A difference image, indicating that the feet of the flux rope were deeplyrooted in the chromosphere, and that the mass loss may originate from the cooler and denser chromosphere.Meanwhile, the two TCHs had a temperature range from several K to 2 × K, suggesting that theappearance of the two TCHs was due to the density depletion rather than the temperature decrease. The massloss probably provided a supplement to the CME. In addition, the brightenings appeared on the locations ofthe two TCHs at the initial phase of the flare, and their appearance may be the result of a flux rope eruption(jiang et al. 2007a). This has not been described by any theoretical models, and further observational andtheoretical investigations are needed for further understanding.
Acknowledgements
The authors thank an anonymous referee for valuable comments. The H α and HeI 10830 ˚A data are provided by the High Altitude Observatory, which is part of the National Center forAtmospheric Research under sponsorship of the National Science Foundation. The authors are indebted tothe SOHO /EIT, MDI and LASCO teams,
GOES /SXI team for free access to the wonderful data. This workis supported by the NSFC under grants 10573033 and 40636031, and by the 973 program (2006CB806303).
References