SDO/AIA OBSERVATIONS OF LARGE-AMPLITUDE LONGITUDINAL OSCILLATIONS IN A SOLAR FILAMENT

Oscillations of filaments have been observed using Hα spectrograms since 1930s (Dyson 1930). They are broadly classified into two groups according to their velocity amplitudes, namely, the "large-amplitude oscillations" and "small-amplitude oscillations" (Thompson & Schmieder 1991; Oliver & Ballester 2002). The large-amplitude oscillations occur when the entire filament oscillates with a velocity amplitude of the order of 20 km s⁻¹ or larger. The small-amplitude oscillations are localized to part of the filament and have a velocity amplitude of the order of 2–3 km s⁻¹ or even less. There have been many studies about small-amplitude oscillations in the last few decades (Oliver 2009; Ning et al. 2009; Soler et al. 2011). However, the large-amplitude oscillations are less widely reported (Okamoto et al. 2004; Tripathi et al. 2009; Arregui et al. 2012).

Moreover, most reported large-amplitude oscillations are perpendicular to the filament axis. This kind of transverse oscillations is often triggered by fast-mode magnetohydrodynamic (MHD) waves coming from distant flares (Ramsey & Smith 1966; Isobe & Tripathi 2006; Gilbert et al. 2008; Chen et al. 2008; Hershaw et al. 2011; Liu et al. 2012). The restoring force in this type of oscillations is generally thought to be the magnetic tension.

Unlike transverse oscillations, Jing et al. (2003) first reported periodic motions along a solar filament axis by using high-cadence Hα observations. The longitudinal oscillations were triggered by a subflare occurring near the filament footpoint. They considered the restoring force due to magnetic tension. Afterward, Jing et al. (2006) and Vršnak et al. (2007) observed similar large-amplitude longitudinal (LAL) oscillations. Vršnak et al. (2007) proposed that the LAL oscillations were triggered by an injection of poloidal magnetic field into the flux rope, and the restoring force was caused by the magnetic pressure gradient along the filament axis. Recently, (Luna & Karpen 2012; Luna et al. 2012) first simulated the LAL oscillations and suggested that the restoring force was the projected gravity in the flux tube dips where the threads oscillate. The oscillations were strongly damped by the mass accretion of the threads. Zhang et al. (2012) analyzed the LAL oscillations above the solar limb by using multi-wavelength data. Their study suggested that the oscillations were excited due to plasma injection and the restoring force may be the gravity.

Since there are different viewpoints on the physical nature of LAL oscillations, further investigations are needed. In this Letter, we present the first observations of the LAL oscillations by the new Atmospheric Imaging Assembly (AIA; Lemen et al. 2012) on board the Solar Dynamics Observatory (SDO; Pesnell et al. 2012).

On 2012 April 7, SDO/AIA observed LAL oscillations of an active region (AR) filament in NOAA AR 11451 (N17W45). The LAL oscillations occurred in the south and north parts (SP and NP in Figure 1) of the filament (see Animations 1 and 3 in the online journal). Ultimately, the SP of the filament erupted and the eruption was associated with a B6.8 flare and a coronal mass ejection (CME). The oscillations of the NP damped obviously with time and died out at last.

Zoom In Zoom Out Reset image size

The SDO/AIA takes full-disk images in 10 (E)UV channels at 15 resolution and high cadence of 12 s. Among the 10 wavelengths of AIA, the 171 and 304 Å channels best show the LAL oscillations and we focus on these channels in this study. The 171 Å bandpass channel is dominated by emission of Fe ix formed at ∼0.6 MK and the 304 Å channel is dominated by He ii emission at 0.05 MK (O'Dwyer et al. 2010; Boerner et al. 2012; Parenti et al. 2012). The LAL oscillations were also observed by the Extreme Ultraviolet Imager (EUVI; see Wuelser et al. 2004) on board Solar-Terrestrial Relations Observatory (STEREO; Kaiser et al. 2008) that was 111° ahead of the Earth. EUVI observes the chromosphere and corona in four spectral channels (304 Å, 171 Å, 195 Å, and 284 Å) out to 1.7 R_sun with a pixel size of 16. All the data used here are corrected for solar rotation.

3.1. LAL Oscillations of the SP

The SP of the analyzed filament mainly consists of two parts (Parts 1 and 2 in Figure 2(a)), which have different axis directions (slices A–B and C–D in Figure 2(b)). By examining the STEREO A data, we find that Part 2 lies above Part 1 and extends longer than Part 1.

Zoom In Zoom Out Reset image size

At 13:26 UT, a weak flare-like brightening appeared in the vicinity of the SP at STEREO A 304 Å (see the white circle in Figure 1(a)). The observations indicate that the subflare caused the activation of Part 1. The north end of Part 1 started to rise slowly with a velocity of ∼20 km s⁻¹ since 13:27 UT and the rise lasted about 5 minutes (Figure 2(a)). After 13:32 UT, the filament shrank southward along the axis of Part 1 (Figure 2(b)). To illustrate the motion of filament, Figure 3 shows the intensity evolution along slices A–B and C–D in Figure 2(b) (stack plots). Seen from the stack plot along slice A–B, the mass motion of Part 1 has a velocity of 70 ± 10 km s⁻¹ (blue dotted lines in Figure 3(a)). Furthermore, the main body of Part 1 shows a helical motion in the image plane during the interval of the rise and shrinkage (refer to images between 13:27 and 13:34 UT and Animations 2 and 4 in the online journal).

Zoom In Zoom Out Reset image size

Following the mass motion of the filament, the oscillations along Part 1 were clearly observed since 13:45 UT (Figure 3(a)). At about 14:06 UT, Part 2 started to oscillate along its axis (Figure 3(b)). The arrows in Figure 3 indicate the times at which the direction of the oscillating motions seems to reverse. It can be seen that the oscillatory period varies with filamentary threads. Two oscillatory profiles along slice A–B have oscillatory periods of ∼44 and 54 minutes, respectively. Their oscillatory amplitudes of the displacement obtained are ∼2.5 × 10⁴ and 2.1 × 10⁴ km, and the corresponding velocity amplitudes are 60 and 40 km s⁻¹, respectively. The LAL oscillation along slice C–D shows a longer period of ∼67 minutes than that of Part 1. Its maximum displacement of the oscillation is ∼2.0 × 10⁴ km, and the displacement shows no significant damping. The oscillation in Figure 3(b) has a velocity amplitude of about 30 km s⁻¹ and completes two cycles before it could be traced clearly.

By comparing the three oscillatory profiles in Figure 3, we find that the oscillations of different filamentary threads start approximately in phase. However, due to different initiation times and oscillatory periods, their phases are not synchronous. At 16:23 UT, partial filament material of Part 1 separated from the main body and moved toward the south. The velocity of the mass motion is the same as that at 13:32 UT (Figure 3(a)). Meanwhile, similar helical motion phenomenon was observed until 16:43 UT (see Animations 2 and 4 in the online journal).

At the late phase of LAL oscillations, bi-directional flows along the filament axis are observed. In order to display the flows clearly, we select two parallel slices with a separation of 6.1 Mm (slices E–F and G–H in Figure 2(c)) and obtain their corresponding stack plots in Figure 4. Between 16:34 and 16:50 UT, the filament material located at slice E–F moved toward the north with a velocity of ∼30 km s⁻¹, while some material at slice G–H moved toward the south with velocities of 20–40 km s⁻¹ (the region between two vertical dashed lines in Figure 4(a)). Along slice G–H, the flow toward the north was also observed, and the velocity was about 50 km s⁻¹. Between 17:12 and 17:36 UT, the oscillating material moved southward along slice E–F and its velocity was about 20 km s⁻¹, while the material at slice G–H moved northwardly with a velocity of ∼10 km s⁻¹ (the region between two blue dashed lines in Figure 4(a)). The oscillations of the SP last about 4 hr, almost four times the corresponding periods, before the onset of the filament eruption.

Zoom In Zoom Out Reset image size

3.2. Eruption of the SP

3.3. LAL Oscillations of the NP

About two minutes after the C2.4 flare occurring at the middle part of the entire filament (17:02 UT), the brightenings of filamentary threads nearby the flare ribbons were clearly observed (see Animations 1 and 3 in the online journal). Subsequently, the north filament became activated and the northward mass motion along the filament axis was formed since 17:14 UT. From ∼17:20 UT, the great bulk of filament material coming from the south was injected to the NP of the filament (Figure 5(a)). Initially, the injected material moved toward the east and the average velocity was about 60 km s⁻¹. Then it changed its direction and turned back to the west upon reaching its maximum displacement at ∼18:24 UT (Figure 5(c)).

Zoom In Zoom Out Reset image size

The movement of the oscillating material is nearly along the axis of the NP, which is also the LAL oscillation. The oscillations have a period of ∼57 minutes and last about 3 hr, almost three cycles before it finally damps out.

In this Letter, we report the first simultaneous observations of the LAL oscillations of the filament and its later eruption. About 30 minutes before the filament eruption, a C2.4 flare occurred at the north of the SP. The flare resulted in the southward mass motion along the axis of the SP. This may suggest that the flare causes the transition from oscillations to the eruption. The flare may impose a strong impulse on the oscillating filament and leads to its loss of equilibrium that results in the filament eruption. The filament eruption was also associated with a CME and a B6.8 flare. Chen et al. (2008) and Bocchialini et al. (2011) also presented observations of vertical and transverse oscillations prior to eruption. They suggested that filament oscillations could be considered to be another precursor of CMEs.

The LAL oscillations occur in the SP and NP of the filament, and they have some common characteristics. They are both triggered by flare activities close to the filament (Jing et al. 2003; Vršnak et al. 2007). These flares may excite the mass motion along the filament axis and then the longitudinal oscillations are formed. The oscillations of different threads are out of phase and their velocity amplitudes vary from 30 to 60 km s⁻¹, with a maximum displacement of about 25 Mm. The oscillations of the SP repeat for about four cycles without any significant damping before the onset of the filament eruption. For the NP, the oscillations damp with time and finally die out.

Generally, there are several physical mechanisms that have been proposed to explain various oscillations. Now we compare these wave modes (such as magnetic kink-mode wave, Alfvénic oscillations, and longitudinal slow magentosonic waves) with our observational results. The transverse oscillations of filaments and coronal loops are interpreted as standing fundamental kink modes (Aschwanden et al. 1999; Ofman & Wang 2008; Hershaw et al. 2011), and the oscillating amplitude of the location close to the filament/loop footpoints is smaller than that of other locations. However, the variation of the amplitude with locations is not observed in our event (Figure 3(a)) and the kink-mode oscillations can be ruled out. The direction of Alfvénic oscillations is perpendicular to the direction of the magnetic field. However, the LAL oscillations of the filaments are approximately along the direction of magnetic field. Thus, the LAL oscillations should not be Alfvénic oscillations. The propagating intensity disturbances in polar plumes and coronal loops are generally interpreted as slow magentosonic waves (Ofman et al. 1999, 2000, 2012; De Moortel et al. 2000; Wang 2011). In our event, we find no significant phase propagation along the filament and thus the longitudinal slow magentosonic waves are also ruled out.

We suggest that the restoring force is the coupling of the magnetic tension and gravity. Our observations show that the activated part of the SP repeatedly shows a helical motion. This indicates that the magnetic structure of the filament is possibly modified during this process. The twisted field of the filament has the component perpendicular to the direction of the filament axis, although it is mostly parallel to the filament axis. The motions of filamentary threads lead to distortion of the twisting magnetic field, and the field component perpendicular to the direction of the filament axis generates magnetic tension along the filament axis as the restoring force, similar to the analysis of Jing et al. (2003). The gravity could also be considered as the restoring force. Recently, Luna & Karpen (2012) and Zhang et al. (2012) simulated LAL oscillations and suggested that the restoring force was the projected gravity in the flux tube dips where the threads oscillate. A comprehensive understanding of the physical nature on the LAL oscillations of filaments requires more observations and further analysis in the future. Assuming the restoring force is only the gravity, we estimate the minimum magnetic field strength in the filament as 28–55 G according to the analysis of Luna & Karpen (2012). If the restoring force is only the magnetic tension, the estimated magnetic field strength is about 32 G based on the study of Kleczek & Kuperus (1969). These estimated magnetic field strengths of the filament are both similar to the observations of active region filament magnetic fields (Mackay et al. 2010), so it is difficult to determine which kind of force is dominant as the restoring force. We consider great uncertainty to exist if we employ LAL oscillations of filaments for corona seismology.

We find that the oscillatory period varies with filamentary threads, ranging from 44 to 67 minutes. This discrepancy may suggest that the magnetic field and plasma condition of filamentary threads are different (Hershaw et al. 2011). Due to different initiation times and oscillatory periods, the oscillatory phases of different filamentary threads are not synchronous, which indicates that the excited threads lose their coherence from the beginning.

Bi-directional flows along the filament axis are observed at the late phase of LAL oscillations with a speed range of 10–40 km s⁻¹ (Figures 4(a) and (b)). These speeds are a factor of two greater than those of the counterstreaming flows reported before (5–20 km s⁻¹; Zirker et al. 1998; Schmieder et al. 2010; Lin 2011). Since the oscillations of different filamentary threads are not synchronous, the interaction between filamentary threads is ultimately destroyed at the late phase. Thus, some threads have opposite movement directions, which shows the characteristic of bi-directional flows. Our observations show that the bi-directional flows occur before the filament eruption, similar to the observations of Lin et al. (2003) and Schmieder et al. (2008).

We are grateful to Dr. P. F. Chen for useful discussions. We acknowledge SECCHI and AIA for providing data. This work is supported by the National Basic Research Program of China under grant 2011CB811403, the National Natural Science Foundations of China (11025315, 40890161, 10921303, and 11003026), the CAS Project KJCX2-EW-T07, and the Young Researcher Grant of National Astronomical Observatories, Chinese Academy of Sciences.