Development of Solar Flares and Features of the Fine Structure of Solar Radio Emission
DDevelopment of Solar Flares and Features of the Fine Structureof Solar Radio Emission
G. P. Chernov a , b , *, V. V. Fomichev b , Y. Yan a , B. Tan a , Ch. Tan a , and Q. Fu a a Huairou Solar Observing Station, National Astronomical Observatory of China, Beijing, China b Institute ofTerrestrial Magnetism, Ionosphere and Radio Wave Propagation, Russian Academy of Sciences, Moscow, Russia * e-mail: [email protected] Received September 22, 2016; in final form, November 28, 2016
Abstract — The reason for the occurrence of different elements of the fine structure of solar radio bursts in thedecimeter and centimeter wavelength ranges has been determined based on all available data from terrestrial andsatellite observations. In some phenomena, fast pulsations, a zebra structure, fiber bursts, and spikes have beenobserved almost simultaneously. Two phenomena have been selected to show that the pulsations of radio emission arecaused by particles accelerated in the magnetic reconnection region and that the zebra structure is excited in a source,such as a magnetic trap for fast particles. The complex combination of unusual fiber bursts, zebra structure, and spikesin the phenomenon on December 1, 2004, is associated with a single source, a magnetic island formed after a coronalmass ejection.
1. INTRODUCTIONIn the continuum emission of type IV solar burstsfrom meter to centimeter wavelength ranges, the fol-lowing fine structure elements are usually observed: fastpulsations in a wide frequency range; fastest bursts ofmillisecond duration (spikes); narrow drifting bands inemission and absorption, among which there are fiberbursts with constant frequency drift and bands withvarious drift- a zebra structure (ZS) (Chernov, 2011). Dueto the variety of unusual band shapes, the ZS is the mostintriguing element of the fine structure. Therefore, itattracted special attention from researchers, and, after thefirst publication on ZS observations in the meter range(Elgaroy, 1959), more than a hundred works werepublished, not only on new ZS observations but alsotheoretical ones. Over the past 40 years, more than a dozen emissionmechanisms have been proposed to explain it. The mostpopular model is the emission mechanism under doubleplasma resonance (DPR) (Zheleznyakov and Zlotnik,1975), in which the upper hybrid frequency ω UH iscompared with an integer number of electron cyclotronharmonics ω UH = (ω Pe + ω Be ) = s ω Be , where ω Pe is theplasma frequency, ω Be is the cyclotron frequency ofelectrons, and s is the harmonic number provided thatω Be << ω Pe .The interaction of plasma waves with whistlers canbe considered an important alternative mechanism(Chernov, 1976): l + w → t . Kuijpers (1975) proposed itto explain fiber bursts. However, in a source such as amagnetic trap, whistlers excited on the anomalous Doppler effect can be generated in the form of periodicwave packets. They propagate at a group velocity anddetermine the frequency drift of the ZS bands, whichvaries synchronously with the spatial drift of the radiosource in the corona. Only the model with whistlersexplains a number of fine effects of the dynamics of theZS bands: the saw-tooth frequency drift, frequencysplitting of bands, and hyperfine structure of bands inthe form of millisecond spikes (Chernov, 2006).A number of other ZS models (Treuman et al., 2011;Ledenev et al., 2006; Laptukhov and Chernov, 2006,2012; Fomichev et al., 2009; LaBelle et al., 2003; etc.)have not yet been unequivocally recognized.The relevance of ZS research has increased after thediscovery of such bands in the radio emission of apulsar in the Crab nebula under extreme physicalparameters peculiar to pulsars (Zheleznyakov et al.,2012).In many solar phenomena, all of the marked elementsof the fine structure are present almost simultaneously onthe dynamic spectrum, and their emission mechanismscan be different. The use of all available optical and X-ray data often helps in the flare analysis. Thus, in thephenomenon on April 11, 2013, the use of simultaneousimages in several extreme ultraviolet lines from the SolarDynamic Observatory/Atmospheric Imaging Assembly(SDO/AIA) helped to understand the repeated change inthe sign of the circular polarization of the radio emission,when each new flare brightening occurred over regionswith dif Fig. 1.
Fast pulsations with an uneven emission cutoff boundary from the HF edge of the spectrum if there is no ZS.ferent magnetic polarities, and the ordinary wave moderemained (Chernov et al., 2016). Additional data in thecase of the hard X-ray emission (RHESSI) made itpossible to construct a probable radio source schemewithin the standard model of a flare with magneticreconnection, where the pulsations of radio emission inthe 2.6–3.4 GHz band were associated with the upwardacceleration of fast particles from the current sheet, andthe ZS emission at lower frequencies was associatedwith the capture of particles accelerated downward, intoa flare loop.However, in the following phenomenon on June 21,2013, a group of fast pulsations also developed around3 GHz, but there was no ZS. In the phenomenon onDecember 1, 2004, on the contrary, from the verybeginning of the phenomenon, there were differentfiber bursts, spikes, pulsations, and ZS almost simul-taneously on the spectrum in the decimeter range 1.1–1.34GHz. In this paper, it is attempted to understand thishierarchy of fine structure and to relate it to thedynamics of the flare process. The statistical analysis ofthe ZS is complicated by a wide variety of phenomena(Tan et al., 2014). In the absence of high-resolution positional observations of radio sources, it is firstimportant to understand the causes of the sequentialappearance of individual elements of the fine struture,and it is even more important to understand theirsimultaneous appearance, since no such analysis ofphenomena has been conducted so far.2. NEW OBSERVATIONSThe paper uses data from solar broadband radiospectrographs (SBRS) of the National AstronomicalObservatory of China (NAOC) installed at the Huairoustation near Beijing: for the phenomenon on June 21,2013, the spectrograph in the centimeter range 2.6–3.8GHz with a frequency resolution of 10 MHz and thetime resolution of 8 ms, and for the phenomenon onDecember 1, 2004, the spectrograph in the decimeterrange 1.1–1.34 GHz with a high resolution of 4 MHzand 1.25 ms (Fu et al., 2004). For an analysis ofphenomena in general, all available satellite data wereused: SOHO/LASCO C2, SOHO/MDI/EIT, RHESSI,and SDO/AIA. Fig. 2.
SDO 211 Å. Ejection along the open field lines goes high into the corona. There is no magnetic trap for particles.Soft X-ray emission profile from the GOES 15 (top panel). Start time is January 21, 2013, 01:55:00.
This phenomenon is remarkable in that pulsations and ZS appeared independently at different instants. A2.9 f lare occurred in the eastern region of NOAA 11777 (12 S, 75 E). The first fast pulsations in the 2.6–3.2 GHz range appeared right after the coronal mass ejection (CME) at 03:01:50 UT but without ZS bandsfrom the high-frequency (HF) edge, in contrast to the phenomenon on April 11, 2013, mentioned in theINTRODUCTION (Fig. 1).The CME starts at 02:45 UT. In SDO/AIA films, it starts with an explosion of a dark filament.
The emissiongoes high into the corona along the open field lines (Fig. 2), and, by the time of the pulsations at 03:01:50UT, the last small emissions in the tail of the CME (Fig. 3) are still visible. At the same time, the flarebrightening begins. It is probably related to the magnetic reconnection and, consequently, to the accelerationof particles, as a result of which fast pulsations occur (Fig. 1). As can be seen in Fig 3, there are no closedloops (magnetic trap) at this time, and therefore there are no conditions for the ZS excitation (within DPR orwhistler models).However, the ZS appeared about 1000 UT (Fig. 4). Exactly at this time, the formation of the flare loop canbe observed (Fig. 5), probably after the repeated brightening of the flare visible on the film. After about threeminutes, the flare loop fades.An even more intense ZS appeared only at 03:25:25 UT (Fig. 6). At this time, the new flare brighteningends with the appearance of a new flare loop in the same place. As can be seen in Figs. 5 and 7, the height ofthe flare loops corresponds to a centimeter range of the plasma frequency.In both cases, the ZS radio emission had a very weak polarization, which is characteristic of limb phe -nomena. In addition, it can be seen that the maximum energy release occurred at the tops of the loops. Thus,it can be clearly seen that the ZS appeared only at the time of the formation of flare loops, magnetic traps forfast particles. Fig. 3.
SDO 171 Å at the time of pulsations. Last CME traces along the open field lines. Flare brightening indicates particle accel-eration, but there is no magnetic trap, there are no conditions for the ZS generation, and, thus, only pulsations are observed.
The phenomenon on December 1, 2004, has been discussed for more than 10 years, but each work hasbeen devoted to a single selected effect. While Huang et al. (2007) considered the phenomenon as a whole,including an analysis of the fine structure of the radio emission (however, only on the pulse burst phase)(Ning et al., 2009; Gao et al., 2014), two unusual fiber burst groups at the onset of the phenomenon wereanalyzed (Fig. 8). In the first paper (Liu et al., 2006), in which only these unusual fiber burst groups werementioned, they were classified as fantastic patterns of the fine structure of bursts that resemble a hand withfingers outstretched (the term used in (Allaart et al., 1990)). The ZS and the so-called lace bursts wereobserved throughout the phenomenon for more than 30 min (Huang and Tan, 2012).The onset of the phenomenon is described in the literature (Huang et al., 2007). A small M1.1 flareoccurred in the active region of NOAA 10708 near the center of the disk (06 N, 20 E). It started at 07:06,reached the maximum at 07:15, and proceeded to 08:21 UT according to X-ray emission data (in softemission, GOES, and in hard emission, RHESSI). Even before the maximum, five separate peaks wereobserved. In the mentioned papers, the evolution of the fine structure and the mechanisms of radio emissionwere not analyzed.After two groups of unusual fiber bursts (Fig. 8), fiber bursts immersed in the developed ZS wereobserved (Fig. 9). The ZS lasted about a minute and ended with the limitation by HF fiber bursts (Fig. 10). Asimilar ZS limitation with fiber bursts from the HF edge was observed later in the microwave range 2.6–3.8 GHz in the phenomenon on August 1, 2010 (Chernov et al., 2014). The circular polarization sign changedseven times within 2.5 min (Fig. 11). Such polarization changes are consistent with the dynamics of sources n hard X-ray emission (Fig. 12): each new f lare brightening coincided with a new source ( A , B , C , D ) indifferent energy ranges. The first such brightening above the tail sunspot was observed at 07:05:55 UT and inthe extreme ultraviolet line 284 Å at the SOHO/EIT (Fig. 13).Ning et al. (2009) associate fine-structure emission with fast particles accelerating downward from themagnetic reconnection region. At the beginning, two groups of fiber bursts are associated with slow down -ward ejections, which explain the slow positive drift of the fiber bursts. One can only guess that the narrowfrequency band of the fiber bursts is associated with the size of these emissions. The ZS is considered in DPRand whistler models, in which the source is assumed to be in the same place of the flare loops where fiberburst sources were located several seconds earlier. Fig. 4.
First ZS appearance in the phenomenon on June 21, 2013.
SDO AIA_3 171 June 21, 2013 03:09:59.340 UT Fig. 5.
SDO 171 Å. Flare loop was formed at the time of the appearance of the first ZS. Fig. 6.
Second ZS appearance at the end of the phenomenon on June 21, 2013.
Fig. 7.
SDO 171 Å. Flare loop in the form of a magnetic trap appeared just at the time of the appearance of the ZS at the end of thephenomenon. Fig. 8.
Onset of the phenomenon on December 1, 2004, with a duration of 1 m 40 s. Moderate left-handed polarization of the fiber bursts and weak right-handed polarization of the continuum. Fig. 9.
Continuation of the phenomenon on December 1, 2004 (duration of 8 s) with the developed ZS with fiber bursts immersed init. The emission has a moderate (weak) right polarization (see the temporal profiles below). Fig. 10.
Zebra structure is limited from high frequencies by a series of fiber bursts in the phenomenon on December 1, 2004, in therange 1–1.34 GHz. Right polarization was changed to the left again.
Ning et al. (2009) associate two unusual groups of fiber bursts at the onset of the phenomenon with thechromospheric evaporation as a result of its bombardment by fast particles accelerated downward from themagnetic reconnection region. However, it can only be considered as an assumption, since there is no tem-poral coincidence of evaporation instants (according to the hard X-ray data) with the occurrence of the fiberbursts. In addition, there is no explanation for the narrow frequency band of the fiber bursts, their frequencydrift, and periodicity.Note that the most real explanation for unusual fiber bursts at the onset of the phenomenon was proposedin (Gao et al., 2014). The emission of fiber bursts was associated with the ejections of magnetic islands(clouds) from the region of magnetic reconnection and the rise of flare loops. However, the mechanism of thegeneration of fiber bursts was not analyzed. The possible role of termination shock was indicated, behind thefront of which particle acceleration must occur.
The beginning of the CME on December 1, 2004, is projected (approximated) at 07:00. This is the begin - ing of the flare in the soft X-ray at GOES. Therefore, the CME was caused by the initial explosion of thefilament . However, there was no type II burst. The HIRAISO spectrograph only gives a type IV burst. At thebeginning, there was only one HXR 5–6 keV source over the head sunspot, and it became double (at thebases of the loops) by 07:05 UT. This is the effect of the fast particles that reached dense plasma at the basesof the loops, which was used as evidence of chromospheric evaporations (Ning et al., 2009). After thedeparture of the CME, these particles could be accelerated during a prolonged magnetic reconnection withthe formation of the magnetic island. The magnetic island probably already existed by the time 07:07:30 UT,and its size determined the frequency range of two groups of unusual fiber bursts. Therefore, we propose anew flare scheme with a magnetic island (Fig. 14). The particles are accelerated in the current sheet at thebottom of the magnetic island, not only downward but also upward. Since the onset of particle acceleration,when there was no continuum, the upward accelerated particles Fig. 11.
Time profiles of channels in left and right polarization and degrees of polarization. Polarization changed the sign seventimes. should excite plasma waves and whistlers. Whistlers excited on the normal Doppler effect must propagatedownward. Thus, unusual groups of fiber bursts with positive frequency drift are fiber bursts excited by wavepackets of whistlers trapped in a magnetic cloud. They cannot spread below the reconnection point. The fre-quency drift is slowed down by this capture. articles accelerating downward excite whistlers that propagate upward. They cause fiber burstswith negative frequency drift. Therefore, on the spectrum they look like a high-frequency boundaryof groups of fiber bursts with positive drift. Simultaneously, the cloud could first descend and meet aclosed flare loop. Obvious plasma ejections down from the current sheet should cause a shock wavethat encounters a barrier in the form of a flare loop. Therefore, the shock wave should inevitablybecome a termination shock (TS), in which the particles are additionally actively accelerated andwhich was mentioned in the literature (Gao et al., 2014). In this paper, the authors are doubtful onlybecause of the fact that they have no other proof of the existence of the termination shock. However,a recent paper (Chen et al., 2015) provides such evidence for another phenomenon. The first effect of particles accelerating in the TS front is fast spikes like primary energy release (Chen etal., 2015). Some of the particles are captured in the magnetic cloud, where the velocity distribution with aloss cone (or a ring distribution) is formed; the ZS emission is associated with this. In the first seconds, theparticles (before capture) should trigger episodic fiber bursts immersed in the developed ZS.All of the above effects from particles accelerated in the TS front can be seen in Fig. 15. The fiber burst atthe beginning of the upper spectrum actually looks like a continuation of fiber bursts with negative frequencydrift limiting two groups of unusual fiber bursts from high-frequency bands at the beginning of thephenomenon. However, it is already immersing in a family of spikes and after six seconds it is divided intoseveral fiber bursts, which transform after 5 s to several ZS bands with a wave-like drift. The particles andwhistlers gradually fill the entire magnetic cloud, and on the third tab of the spectrum we observe the Fig. 12. Dynamics of hard X-ray sources (RHESSI). ig. 13. Flare brightening in line 284 Е (SOHO/EIT) over the tail sunspot at 0705:55 UT of the northern magneticpolarity. There is apparently just not one, which is the reason for the multiple polarization reversal.
Fig. 14.
Scheme of the flare for the phenomenon on December 1, 200 4. This is a long-lasting phenomenon. Magnetic islands form above the flare loop. The scheme refers to the instant after the interaction of the island with the flare loop with the formation of the X point of magnetic reconnection and acceleration of the particles in both directions from the current sheet. Fig. 15.
Gradual transformation of fiber bursts into zebra and its evolution within 46 s. expansion of the ZS frequency band almost to the entire range, which is only partially surrounded by spikes.On the lower tab, the ZS is mixed with spikes (or spikes are organized into ZS bands, which in part look likefiber bursts with negative drift). All of these processes occur in a rather narrow frequency band of 1.1–1.34GHz, in a magnetic island source. he polarization of the continuum changed sign seven times during the first 10 min of the burst (Fig. 11).Even a cursory glance at the images of hard X-ray sources (Fig. 16) is enough to understand the effect of thealternating predominance of the northern (with the left sign of polarization) and southern (with the right signof polarization) sources. The 25–30 keV HXR source appeared on the side, in the southern part of AR 10708.After ~07:11 UT, there was only one southern source for several minutes (C in Fig. 12), and the polarizationat that time preserved the right sign. According to MDI’s AR magnetograms (Fig. 6, (Ning et al., 2009)),radio emission always corresponded to the ordinary mode.Later, up to 07:17 UT, episodic ruptured fiber bursts appeared (Fig. 17) against the background of acontinuum consisting of spikes with a duration at the instrument resolution limit of 1.25 ms. That is, theacceleration in the TS front continued. It is difficult to say Fig. 16.
Hard X-ray contours (RHESSI) against the background of a magnetic map (MDI) for six instants of the phenomenon on December 1, 2004. The levels of X -ray contours correspond to 40, 60, and 80% for 25–50 keV (dotted line), 50, 60, 70, and 90% for 10–15 keV (thin lines), and 70, 80, and 90% for 6–10 keV (thick lines). (Fig. 6 in (Ning et al., 2010)). hether the turbulence of the plasma inside the magnetic island increased or, conversely, decreased. How -ever, fiber bursts and clouds of spikes indicate the presence of whistlers and ion-acoustic waves. It is possiblethat there was no rapid rupture of the TS (as in the phenomenon of March 3, 2012 (Chen et al., 2015)). Mostlikely, the TS front along with the top of the flare loop experienced slow shifts (oscillations) up and down,and the particles and waves continued to be captured in the magnetic island. The gradual evolution of the finestructure (the initial unusual fiber bursts in the ZS, spikes, new fiber bursts, and lace bursts) is associated withthe evolution of the distribution function of fast particles inside the magnetic island.The fine structure at the damped phase of the burst was analyzed (Huang and Tan, 2012), and specialattention was paid to the so-called lace bursts, which appeared at 07:25 UT (Fig. 18) and continued to07:33:30 UT.Lace bursts were interpreted (Karlický et al., 2001) as the radiation from turbulent plasma under condi -tions of double plasma resonance, when the paramters of the magnetic field and density change abruptly andoften even violate the DPR condition. It is difficult to expect sharp fluctuations of the magnetic field anddensity at the damped phase of the f lare. Huang and Tan (2012) associate lace bursts with the emission ofBernshtein modes, although there is little evidence for this, both because of the chaotic form of the bands andbecause of the strong polarization of the bursts. However, estimates of the strength of the magnetic field byBernstein modes coincided with the values obtained for fiber bursts and the ZS, ~ 70 Gs.In this phenomenon, it was more obvious to associate the lace bursts with the wave packets of whistlerstrapped in the magnetic island, since throughout the phenomenon the fiber bursts transforming into ZS bandsand back are clearly associated with the whistlers.3. CONCLUSIONSIn the absence of high-resolution positional observations of radio sources, we tried to understand thecauses of the sequential appearance of individual elements of the fine structure and to explain their simul -taneous appearance in the decimeter and microwave wavelength ranges, since no such analysis of the phe-nomena has been conducted so far. All available terrestrial and satellite observations were used. Twophenomena were selected to show that the pulsations of radio emission are caused by particles accelerated inthe magnetic reconnection region, and the zebra structure is excited in the source such as a magnetic trap forfast particles. The complex combination of unusual fiber bursts, ZS, and spikes in the phenomenon onDecember 1, 2004, is associated with the development of instabilities in one source, a magnetic island formedafter a coronal mass ejection. Fig. 17.
Ruptured fiber bursts in the decreasing phase of the flare. ig. 18. Lace bursts in the decreasing phase of the flare. ACKNOWLEDGMENTSWe are grateful to the teams of SOHO, RHESSI, and SDO, as well as the STEREO experiment, for openaccess to databases. This work was supported by the Russian Foundation for Basic Research, project no. 17-02-00308. We thank Robert Sych for their assistance in the preparation of the SDO films and participation intheir discussion. EFERENCES
Allaart, M.A.F. and Van Nieukoop, J., Fine structure in solar microwave bursts,
Sol. Phys. , 1990, vol. 130, pp. 183–199.Chen, B., Bastian, T.S., Shen, Ch., Gary, D.E., and Krucker, S., Particle acceleration by a solar flare termination shock,
Science AAAS , 2015, vol. 350, no. 6265, pp. 1248–1260.Chernov, G.P., Modulation of continuous emission by wave packets of whistlers,
Astron. Zh. , 1976, vol. 53, no. 5, pp.1027–1040.Chernov, G.P., Solar radio burst with drifting stripes in emission and absorption,
Space Sci. Rev. , 2006, vol. 127, pp.195–326. doi 10.10007/s11214-006-9141-7Chernov, G.P.,
Fine Structure of Solar Radio Bursts , Berlin: Springer, 2011. doi 10.1007/978-3-642-20015-1Chernov, G.P., Yan, Y.H., and Fu, Q.J., The importance of source positions during radio fine structures observations,
Res. Astron. Astrophys ., 2014, vol. 14, no. 7, pp. 831–845. doi 10.1088/1674-4527/14/7/005Chernov, G.P., Sych, R., Tan, B., Yan, Y., Tan, Ch., Fu, Q., Karlicky, M., and Fomichev, V.V., Flare evolution andpolarization changes in fine structures of solar radio emission in the 2013 April 11 event,
Res. Astron. Astrophys .,2016, vol. 16, no. 2, pp. 28–40. doi 10.1088/1674-4527/16/2/028Elgaröy, O., Observations of the fine structure of enhanced solar radio radiation with a narrow-band spectrum analyser,
Nature , 1959, vol. 184, pp. 887–888.Fomichev, V.V., Fainshtein, S.M., and Chernov, G.P., A possible interpretation of the zebra pattern in solar radiation,
Plasma Phys. Rep. , 2009, vol. 35, no. 12, pp. 1032–1035. doi 10.1134/S1063780X09120058Fu, Q.J., Ji, H.R., and Qin, Z.H., A new solar broadband radio spectrometer (SBRS) in China,
Sol. Phys. , 2004, vol. 222,pp. 167–175.Gao, G.-N., Wang, M., Lin, J., Wu, N., Tan, C.M., Kliem, B., and Su, Y., Radio observations of the fine structure inside apost-CME current sheet,
Res. Astron. Astrophys ., 2014, vol. 14, no. 7, pp. 843–854. doi 10.1088/1674-4527/14/7/006Huang, J. and Tan, B.L., Microwave bursts with the fine structures in the decay phase of a solar flare,
Astrophys. J. ,2012, vol. 745, pp. 186–196. doi 10.1088/0004-637X/745/2/186Huang, J., Yan, Y., and Liu, Y., An analysis of solar radio burst events on December 1, 2004,
Adv. Space Res. , 2007, vol.39, pp. 1441–1446. doi 10.1016/j.asr.2007.03.061Karlický, M., Bárta, M., Jiřička, K., et al., Radio bursts with rapid frequency variations—Lace bursts,
Astron. Astrophys. ,2001, vol. 375, pp. 638–643. doi 10.1051/0004-6361:20010888Kuijpers, J.,
Collective wave particle interaction in solar type IV radio sources, Ph.D. Dissertation , Utrecht: UtrechtObservatory, 1975. http://adsabs.harvard.edu/abs/ 1975PhDT………1K.LaBelle, J., Treumann, R.A., Yoon, P.H., and Karlický, M., A model of zebra emission in solar type IV radio bursts,
Astrophys. J. , 2003, vol. 593, pp. 1195–1207. doi 10.1086/376732Laptukhov, A.I. and Chernov, G.P., New mechanism for the formation of discrete stripes in the solar radio spectrum,
Plasma Phys. Rep. , 2006, vol. 32, no. 10, pp. 866– 871. doi 10.1134/S1063780X06100060Laptukhov, A.I. and Chernov, G.P., Ref lection and transmission coefficients for radio waves incident obliquely on N identical plane plasma layers and zebra patterns in the spectrum of solar radio emission, Plasma Phys. Rep. , 2012,vol. 38, no. 7, pp. 560–565. doi 10.1134/S1063780X12060049Ledenev, V.G., Yan, Y., and Fu, Q., Interference mechanism of “zebra-pattern” formation in solar radio emission,
Sol.Phys. , 2006, vol. 233, pp. 129–138. doi 10.1007/s11207-006-2099-5Liu, Y.Y., Fu, Q.J., Yan, Y.H., and Tan, Ch.M., New result obtained from the solar radio spectrometer in decimeterwavelength with super-high temporal resolution,
Publ. Nat. Astron. Obs. China , 2006, vol. 3, pp. 119–140.Ning, Z., Cao, W., Huang, J., Huang, G., Yan, Y., and Feng, H., Evidence of chromospheric evaporation in the 2004December 1 solar f lare,
Astrophys. J. , 2009, vol. 699, pp. 15–22. doi 10.1088/0004-637X/699/1/15Tan, B.L., Tan, Ch.M., Zhang, Y., Mészárosová, H., and Karlický, M., Statistics and classification of the microwavezebra patterns associated with solar flares,
Astrophys. J. , 2014, vol. 780, pp. 129–138. doi 10.1088/0004-637X/780/2/129Treumann, R.A., Nakamura, R., and Baumjohann, W., A model of so-called “Zebra” emissions in solar f lare radio burstcontinua,
Ann. Geophys. , 2011, vol. 29, pp. 1673–1682. doi 10.5194/angeo-29-1673-2011Zheleznyakov, V.V. and Zlotnik, E.Ya., Cyclotron wave instability in the corona and origin of solar radio emission withfine structure. III. Origin of zebra-pattern,
Sol. Phys. , 1975, vol. 44, pp. 461–470. doi 10.1007/BF00153225Zheleznyakov, V.V., Zaitsev, V.V., and Zlotnik, E.Ya., On the analogy between the zebra patterns in radio emission fromthe sun and the crab pulsar, 2012, vol. 38, no. 9, pp. 589–604. doi 10.1134/S106377371209006X, 1975, vol. 44, pp. 461–470. doi 10.1007/BF00153225Zheleznyakov, V.V., Zaitsev, V.V., and Zlotnik, E.Ya., On the analogy between the zebra patterns in radio emission fromthe sun and the crab pulsar, 2012, vol. 38, no. 9, pp. 589–604. doi 10.1134/S106377371209006X