ebook img

Observation of coronal loop torsional oscillation PDF

4 Pages·0.1 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Observation of coronal loop torsional oscillation

Astronomy & Astrophysics manuscript no. zaqarashvili February 2, 2008 (DOI: will be inserted by hand later) Observation of coronal loop torsional oscillation T.V. Zaqarashvili Abastumani Astrophysical Observatory,Al. Kazbegi ave. 2a, 380060 Tbilisi, Georgia e-mail: [email protected] 3 Received November22, 2002/ Accepted January 14, 2003 0 0 Abstract. Wesuggestthattheglobaltorsionaloscillation ofsolarcoronalloopmaybeobservedbytheperiodical 2 variation of aspectral linewidth.Theamplitudeof thevariation must bemaximal at thevelocity antinodesand n minimal at thenodes of thetorsional oscillation. Then thespectroscopic observation as a time series at different a heights above the active region at the solar limb may allow to determine the period and wavelength of global J torsional oscillation and consequently the Alfv´en speed in corona. From the analysis of early observation (Egan 6 & Schneeberger 1979) we suggest thevalueof coronal Alfv´en speed as ∼500km·s−1. 1 Key words.Sun:corona – Sun:oscillations 1 v 6 1 1. Introduction ofenergyfromthe kinkmode toAlfv´en(azimuthal)oscil- 3 lationswithinthe inhomogeneouslayer.Consequentlythe 1 Magnetohydrodynamic (MHD) waves and oscillations resonantabsorptionof globalkink modes may leadto the 0 play an important role in the dynamics of the solar global torsional oscillation of coronalloops. 3 corona (Roberts 2000). They can heat and accelerate the 0 Unfortunately, contrary to the sausage and kink coronal plasma (Goossens et al. 1995) and also offer an / modes, the global torsional oscillations do not cause nei- h unique tool for developing a coronal seismology (Edwin therintensityvariationnorthespatialdisplacementofthe p & Roberts 1983, Nakariakov & Ofman 2001). The so- - loopposition.Therefore they hardly subjectto the obser- o lar corona is highly structured into narrow coronal mag- vation in usual coronalspectral lines. r neticloopswhichareanchoredintothedensephotosphere. t The influence of torsional waves propagating along a s The loops permit the propagationof three kinds of MHD a waves: sausage, kink and torsional Alfv´en waves. The thin, vertical,photospheric flux tube on Zeeman-splitpo- : larized line profiles (Stokes profiles) was studied recently v closed boundary conditions at the loop footpoints lead to i the discrete spectrum of harmonics and consequently to by Ploner & Solanki (1999). In the presence of such a X wave spatially resolved Stokes profiles are found to oscil- the formation of global oscillations. r latestronglyinwavelength,amplitudeandblue-redasym- a The global MHD oscillation in the sausage mode metry. Also observation on the Solar and Heliospheric (Roberts et al. 1984) modulate loop cross section and Observatory (SOHO) showed an increase of spectral line causes the variation of the density and emission measure widthwithheight(Doyleetal.1998,Banerjeeetal.1998) of radiating plasma. Using coronal loop images obtained which can be interpreted as vertically propagating un- withtheYohkohsoftX-raytelescope(SXT),McKenzie& damped Alfv´en waves (Moran 2001). However no wave Mullan (1997) found the periodical modulation of the X- parameters (wavelength, frequency) were identified, be- ray brightness, which was interpreted as the signature of cause of the propagating pattern of waves. On the other globalmodeoscillation.Theglobalkinkoscillationscanbe hand,thestationarycharacterofstandingAlfv´enwavesin detected as transverse displacement of spatial loop posi- magnetic tubes (or torsional oscillations) allows to iden- tionsbetweenfixednodes.Thetypicalpropertyofrecently tify the velocity nodes and antinodes and thus may be observedkink oscillations by TRACE (Aschwanden et al. used for determining the wave parameters. 1999, Nakariakovet al.1999) is very shortdamping time. Wesuggestthattheglobaltorsionaloscillationofcoro- It was suggested that the oscillations may be damped ei- nalloopcanbedetectedbyperiodicalbroadeningofspec- ther due to the phase mixing with anomalously high vis- tral lines due to the periodic azimuthal velocity. The am- cosity(Ofman&Aschwanden2002)orduetotheresonant plitude of the azimuthal velocity in the torsional oscilla- absorption (Ruderman & Roberts 2002) acting in the in- tion is maximal at the antinodes and tends to zero at the homogeneousregionsofthetube,whichleadstoatransfer nodes. Therefore the torsional oscillation causes the peri- Send offprint requests to: T. Zaqarashvili odical variation of spectral line broadening with different 2 T.V. Zaqarashvili: Observation of coronal loop torsional oscillation amplitudes atdifferent heights fromthe solar surface:the where v is the rotation velocity at x=R (R is the radius place of stronger variation corresponds to the antinode of the tube). So the Doppler shifts at different x from the and the place of constant broadening corresponds to the tube axis cause the spectral line broadening by 2 ∆λ. It | | node. Then the time series of spectroscopic observations must be mentioned that the broadening of spectral lines at different heights above the active region at the solar can be caused by thermal motion of ions and turbulent limb may allow to determine the period and wavelength motionofgas,howeverhereweconsideronlythenonther- ofglobaltorsionaloscillationandconsequentlytheAlfv´en mal broadening of spectral lines caused by the azimuthal speed in the coronalloops. velocity of the tube. Recently, strongly damped Doppler shift oscillations Now suppose that the tube undergoes the torsional of hot (T>6 MK) coronal loops were observed with oscillationduetotheclosedboundaryconditionsimposed SUMER on SOHO (Kliem et al. 2002, Wang et al. 2002). at the tube ends. For simplicity the density ρ and the These oscillations were interpreted as signatures of slow- magnetic field B=(0,0,B0), directed along the axis z of modemagnetosonicwavesexcitedimpulsivelyintheloops acylindricalcoordinatesystemr,φ,z,areconsideredtobe (Ofman&Wang2002).Itispossiblethattheslowmagne- homogeneousthroughoutthe tube.Thenthe velocityand tosonic waves (as well as the kink waves) may also cause the magnetic field components of the torsional oscillation the oscillation of Doppler broadening in given direction can be expressed as of propagation. However the velocity field of waves is in- homogeneous(which is necessary for the line broadening) Bφ =−B0αcos(ωnt)cos(knz), (2) only at wave nodes and thus may lead only to negligible V =V αsin(ω t)sin(k z), (3) effect.Ontheotherhand,thevelocityfieldatwaveantin- φ A n n odes will cause the Doppler shift oscillation as was inter- where VA =B0/√4πρ is the Alfv´en speed, kn =nπ/L (L pretedbyOfman&Wang(2002).Howeveritisalsopossi- is the length of the tube and n=1,2,3...) is eigenvalue of blethattheDopplerbroadeningofthelineisduetosuper- wavenumber,ω isthecorrespondingeigenfrequencyand n imposeofseveralloopsintheline ofsightcontainingslow αistherelativeamplitudeofoscillations.Eigenvaluesand wavesatdifferent phases.But in this case the broadening eigenfrequencies satisfy the dispersion relation of Alfv´en must be nearlyconstantor it will showrather chaotic be- waves haviourintime thanthe periodicalone.Recently Sakurai ω n etal.(2002)presentedatimesequenceover80minofcoro- =V . (4) A k nalgreen-linespectraobtainedwitha ground-basedcoro- n nagraph at the Norikura Solar Observatory. They found Thevelocityfieldofeachnthmodehasdifferentamplitude the oscillations of the spectral line Doppler shifts and in- of oscillation, but the same frequency ω along the tube: n terpretedthemaspropagatingslowMHDwaves.Noclear the amplitude is maximal at the antinodes (sin(k z) = n evidence of the Doppler width oscillations was found. It 1) and tends to zero at the nodes (sin(k z)=0). n ± mayindicatethateither they couldnotcatchthe velocity Thenthe velocityfield(3)leadsto the periodicalvari- antinodes of the global torsionaloscillation or the oscilla- ationofDopplerwidthateachpointz alongthetubeaxis. tions are not common in corona and they are related to It means that the torsional oscillation causes the period- the flareormassejection. Future spaceandground-based ical variation of spectral line broadening (expressed by a spectroscopicobservationsarenecessaryto makethecon- halfoflinewidth,∆λ ,hereafterHW)withthesamefre- B clusion. quency,butwithdifferentamplitudesalongthetube.The In the next section we briefly describe the method of amplitude of HW variation will be maximal at the veloc- observation and interpret the early observation of Eagen ity antinodes and minimal atthe nodes.According to the & Schneeberger (1979) as the global torsional oscillation equations (1) and (3) HW can be expressed by of coronal loop. αV λ A ∆λ = sin(ω t)sin(k z). (5) B n n c | | 2. Observation of the torsional oscillation The distance between the points of maximal variation Consider a coronal loop as straight magnetic tube, the i.e. between the velocity antinodes will be the half wave- axis of which is perpendicular to the line of sight. If the lengthoftorsionaloscillation.Thentheobservedperiodof tube rotates about its axis with an angular velocity Ω, HW variation allows to determine the Alfv´en speed from then the Doppler shift of the observed spectral line with thedispersionrelation(4).AlsotheamplitudeofHWvari- wavelength λ is ∆λ = (λΩ/c)x, where x is the distance ation at the velocity antinodes (maximal value of ∆λ B from the tube axis in the plane perpendicular to the line along the tube axis) gives the value of α i.e. the ampli- ofsightandcisthespeedoflight(theexpressionmustbe tude of torsional oscillations. multipliedbysiniifthetube isinclinedwiththeanglei). However we have to be sure that the periodical varia- The largest shift occurs at the edges of the tube and has tionofspectralline broadeningcanbe causedonlyby the a value torsionaloscillations.The velocityfieldofotherwavemo- tionsmaycausetheDopplershiftvariations.Forexample, λ ∆λ= v, (1) theslowmagnetosonicwavespropagatingalonganapplied c T.V. Zaqarashvili: Observation of coronal loop torsional oscillation 3 magnetic field may lead to this effect recently observed by SUMER in given arbitrary loop orientation (Kliem B et al. 2002, Wang et al. 2002, Ofman & Wang 2002). * A Howeverneitherslowmagnetosonicwaves,norkinkwaves * 66 000 km can cause the significant periodical variation of Doppler 88 000 km broadening at given height from the solar surface. The velocity field of waves must be highly inhomogeneous to produce the line broadening. The velocity field of slow magnetosonicwaves(aswellasthekinkwaves)isinhomo- geneous at velocity nodes, where they can produce only negligible Doppler broadening. On the other hand, they can produce the Doppler shift oscillations at antinodes where the velocity is maximal and almost homogeneous. HoweveritisalsopossiblethattheDopplerbroadeningof the line is due to superimpose of several loops in the line of sight containing slow waves at different phases. But in thiscasethebroadeningmustbenearlyconstantoritwill showratherchaoticbehaviourintime thanthe periodical Fig.1. Schematic picture of a coronal loop at the solar one. Therefore from our point of view the torsionalwaves limb is presented. The points A and B correspond to the are best candidates for producing the periodicalvariation velocity antinodes of second mode of global torsional os- of spectral line broadening. cillation. Both points are located at 88 000 km from ∼ Thus the observation as a time series at different the surface.Howeveronly 66 000km partis visible in the heightsabovetheactiveregionatthesolarlimbmayallow caseofpointB,becausethefootpointisanchoredintothe to determine the wavelength and the period of torsional reverse side of the solar surface. oscillation.Asanexampleofproposedmethodletanalyse the observation made almost two decades ago by Egan & The period of torsional oscillations will be 12.2 min i.e. Schneeberger(1979).Theypresentedthe time seriesofFe ∼ double of observed period. Then we easily calculate the XIV coronal emission line spectra above both active and Alfv´enspeed incorona(here the constantAlfv´enspeedis quiet regions at the solar limb. They found the Doppler assumedalongtheloop,howeverinrealconditionsitmay widthtemporalvariationswiththeperiodof6.1 0.6min ± be inhomogeneous which complicates the problem) above the active region. The width fluctuations were am- plified in two 5500 km long segments at heights of 66 000 V 500kms−1. (7) A ∼ · kmand88000kmabovethe limb.The amplitudeofvari- ation was 0.125 ˚A at 88 000 km, while at lowest heights If the regions of amplified Doppler width oscillations itreduces to 0.04˚A. They alsofound amarginalevidence belong to two different loops, then the similar calculation for 6 min intensity oscillations at these two positions.For for the loop with velocity antinodes at 66 000 km gives quite region they found neither Doppler width nor inten- −1 V 360kms , (8) sity variations. A ∼ · We suggestthatthisobservationcanbeinterpretedas which is below the expected value of the Alfv´en speed in the global torsional oscillation of coronal loop. The two corona. Also it seems unreal that the two different loops regions of amplified Doppler width variation may corre- havethesameperiodsofglobalmode.Ontheotherhand, spond to the velocity antinodes of torsional oscillation. the global oscillation of coronal loop in other wave mode They likely belong to the second mode of oscillation, be- may not cause the periodic oscillations of Doppler width. cause it has two velocity antinodes at the middle of both Only the temperature changing due to the nonadiabatic footpoints, while first mode has only one at the top (the slow MHD wavesmay leadto this variation.But it seems velocityantinodeatthelooptophardlyundergototheob- again unreal that the nonadiabaticity leads to the varia- servationinthisspectrallinebecauseofrapidlydecreasing tion of line broadeningwith 25%.Therefore we suggest ∼ intensity in the corona). thatthetorsionaloscillationofoneloopcanbethereason Astheobservationswereperformedoffthesolarlimbit for the observed phenomenon. isreasonabletosupposethattheantinodeswerelocatedat We also easily calculate the amplitude of torsional os- 88000kmheightfromthephotosphere,butthesecond cillation for Fe XIV (λ5303) spectral line from the equa- ∼ footpoint was anchoredat the reverseinvisible side ofthe tion (5) Sun, so that only 66 000 km part of footpoint was visible −1 7kms . (9) (Fig.1).Thenthewavelengthofglobaltorsionaloscillation ∼ · can be This result is calculated when the loop axis is perpendic- ular to the line of sight. The amplitude of oscillation will λ=4 88000km 352000km. (6) be even greater in the case of loop inclination. · ≈ 4 T.V. Zaqarashvili: Observation of coronal loop torsional oscillation Evaluated Alfv´en speed, as well as the amplitude of References torsional oscillations, have the suggestive values for the Aschwanden, M.J., Fletcher, L., Schrijver, C.J. & Alexander, low corona. This gives the idea that the observation of D.1999, ApJ,520, 880 Egan & Schneeberger (1979) may be interpreted as the Banerjee,D.,Teriaca,L.,Doyle,J.,&Wilhelm,K.1998,A&A, global torsional oscillation of coronal loop. Observed in- 339, 208 tensity oscillations can be interpreted as the density per- Doyle, J., Banerjee, D., & Perez, M. 1998, Sol. Phys., 181, 91 turbationscausedbythenonlinearmagneticpressureact- Edwin, P., & Roberts, B. 1983, Sol. Phys., 88, 179 ing at velocity antinodes of torsional oscillations. Egan, T.F., & Schneeberger, T.J. 1979, Sol. Phys., 64, 223 Goossens, M., Ruderman, M.S., & Hollweg, J.V. 1995, Sol. Another problemis thatthe amplitude of the Doppler Phys., 157, 75 width oscillations in the paper of Egan & Schneeberger Kliem, B., Dammasch, I.E., Curdt, W., & Wilhelm, K. 2002, (1979) is comparable or below the instrumental spectral ApJ,568, L61 resolution.Thereforethe validityofthe observationisun- McKenzie, D.E., & Mullan, D.J. 1997, Sol. Phys.,176, 127 clear. On the other hand, Sakurai et al. (2002) could not Moran, T.G. 2001, A&A,374, L9 findclearperiodicalvariationofthe linewidth.Itmayin- Nakariakov, V.M., Ofman, L., Deluca, E.E., Roberts, B. & dicate that either they couldnotcatchthe velocityantin- Davila, J.M. 1999, Science, 285, 862 odes of the global torsional oscillation or the oscillations Nakariakov, V.M., & Ofman, L. 2001, A&A,372, L53 are not common in corona and they are related to the Ofman, L. & Aschwanden,M. 2002, ApJ,576, L1530 flare or mass ejection like the transverse loop oscillations Ofman, L. & Wang,T. 2002, ApJ, 580, L85 (Schrijver et al. 2002). Therefore the future observations Ploner, S.R.O., & Solanki, S.K. 1999, A&A,345, 986 Roberts, B., Edwin, P., & Benz, A.O.1984, ApJ, 279, 857 are necessary for making the final conclusion. Also anal- Roberts, B. 2000, Sol. Phys., 193, 139 ysis of recent space based spectroscopic observations on Ruderman,M. & Roberts, B. 2002, ApJ,577, 475 SOHO and TRACE may reveal more precise data which Sakurai, T., Ichimoto, K., Raju, K.P. & Singh, J. 2002, Sol. can test the validity of this method. Phys., 209, 265 Schrijver, C.J., Aschwanden, M.J. & Title, A.M. 2002, Sol. Phys., 206, 69 Wang,T.,Solanki,S.K.,Curdt,W.,Innes,D.E.&Dammasch, 3. Conclusion I.E. 2002, ApJ, 574, L101 Wesuggestthattheglobaltorsionaloscillationofthecoro- nal loop, axis of which is not parallel to the line of sight, canbeobservedbytheperiodicalvariationofspectralline width. The amplitude of the variation must be maximal at certain regions which correspondto the velocity antin- odesoftorsionaloscillations.Thenthe periodofvariation will be the half of oscillation period and the distance be- tweentheregionswithmaximalvariationswillbethehalf ofwavelength.Thespectroscopicobservationasatimese- riesatdifferentheightabovetheactiveregionatthesolar limb may allow to determine the wavelength and the pe- riodofthetorsionaloscillation.ThentheAlfv´enspeedcan becalculatedwhichisveryimportantfordeterminationof magnetic field strength in corona. The observation of the spectrallineswithlargerwavelengthswillbemoresuccess- ful,becauseoftheirsensitivitytotheDopplershift.Both, ground based coronagraphs and space based instruments can be used. The analysis of early observation (Egan & Schneeberger 1979) give the values of Alfv´en speed and the amplitude of torsional oscillations as 500kms−1 and >7kms−1 respectively. ∼ · · We hope that the observation of torsional oscillations after the damping of coronal loop kink oscillations will provide the clue for the damping mechanism and con- sequently for the coronal heating (Ruderman & Roberts 2002). Acknowledgements. Theauthorthankstheanonymousreferee for helpful suggestions.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.