Draft version January 13, 2016 PreprinttypesetusingLATEXstyleemulateapjv.12/16/11 DIRECT OBSERVATIONS OF PLASMA UPFLOWS AND CONDENSATION IN A CATASTROPHICALLY COOLING SOLAR TRANSITION REGION LOOP N. B. Orange, D. L. Chesny, H. M. Oluseyi, K. Hesterly, M. Patel, and P. R. Champey DepartmentofPhysics&SpaceSciences,FloridaInstituteofTechnology,Melbourne,FL32901 Draft version January 13, 2016 ABSTRACT Minimal observational evidence exists for fast transition region (TR) upflows in the presence of cool loops. Observations of such occurrences challenge notions of standard solar atmospheric heating 6 models, as well as their description of bright TR emission. Using the EUV Imaging Spectrometer 1 (EIS) onboard Hinode, we observe fast upflows (v ≤−10 km s−1) over multiple TR temperatures λ 0 (5.8≤logT ≤ 6.0) at the footpoint sites of a cool loop (logT ≤6.0). Prior to cool loop energizing, 2 asymmetric flows of +5 km s−1 and −60 km s−1 are observed at footpoint sites. These flows speeds n and patterns occur simultaneously with both magnetic flux cancellation (at site of upflows only) a derived from the Solar Dynamics Observatory’s (SDOs) Helioseismic Magnetic Imager’s (HMI) line- J of-sightmagnetogramimages,anda30%massin-fluxatcoronalheights. Theincurrednon-equilibrium 2 structure of the cool loop leads to a catastrophic cooling event, with subsequent plasma evaporation 1 indicating the TR as the heating site. From the magnetic flux evolution we conclude that magnetic reconnectionbetweenthefootpoint andbackgroundfieldareresponsibleforobservedfastTRplasma ] upflows. R Keywords: corona, transition region, upflows S . h 1. INTRODUCTION a response to cooling coronal plasma which was impul- p sively heated. The pervasively observed TR redshifts - Aparamountchallengeinsolarphysicsstillremains— o (Hansteen et al. 1996) and observations that cool loops thatisdefinitivelysolvingtheproblemofhowuppersolar r arecharacterizedbyplasmadownflows,atfootpointsand t atmospheric structures [i.e., transition region (TR) and s coronal]areheatedandmaintained(Tripathietal.2012). alongtheloopstructures(e.g.,DelZanna2008;Tripathi a et al. 2009), have provided significant support to such Studiesofplasmaloops,theprimarycomponentsofeach [ models. However, observational evidence is emerging levelofthesolaratmosphere(Hansonetal.1980;Walker (to our knowledge only those reported by Tripathi et al. 2 et al. 1993a,b; Golub et al. 1999; Oluseyi et al. 1999a,b), v have vastly improved and influenced our understanding 2012)thatrevealstheexistenceoffastTRupflowsinthe 9 of solar atmospheric heating, most notably that of the presence of cool loops. The significance of such results 9 corona (Aschwanden & Nightingale 2005; Warren et al. is the introduction of challenges to standard solar atmo- 9 2008). spheric heating models. Particularly these are their de- 4 scriptionsontheoriginofbrightTRemissionandheights These basic building blocks of the solar corona, i.e., 0 at which coronal heating occurs. Furthermore, emerg- plasmaloops,arecommonlyclassifiedviatheirpeaktem- . ing observational evidence for fast TR upflows, mainly 1 perature (Chitta et al. 2013). Hot loops (logT >6.0) associated with explosive events (Beckers 1968a,b) and 0 havebeenextensivelystudiedandmodeled(e.g.,Mackay spicules (De Pontieu et al. 2007; Langangen et al. 2008), 5 etal.2010;Aschwanden&Schrijver2002;Spadaroetal. 1 2006), while to a much lesser extent diffuse cool (TR) are providing increasing support that atmospheric heat- : loops (logT ≤6.0; e.g., Chitta et al. 2013; Tripathi et al. ing is not confined to the corona. v Though observational evidence exists indicating heat- 2012;Chesnyetal.2013;Mu¨lleretal.2003,2004;Oluseyi i X et al. 1999a,b). Recent observational and theoretical ad- ing occurs in cooler regions of the solar atmosphere; the heights, timescales, and mechanisms responsible remain r vancesontheheatingofplasmaconfinedwithinhotloop a structures have revealed that both steady-state (e.g., unclear (Tripathi et al. 2012). Moreover, a current topic of hot debate is whether fast TR upflows provide signif- Winebarger et al. 2011; Warren et al. 2010) and impul- icant mass-influx to coronal heights, as well as the role siveheating(e.g.,Viall&Klimchuk2012;Tripathietal. they play in the generation of observed coronal phenom- 2010) processes are consistent with their observed tem- ena(DePontieuetal.2009;Langangenetal.2008;Klim- perature and intensity structures. Impulsive heating has chuk 2012). It is also emphasized that minimal investi- been found to explain the properties of cool loops (e.g., gations have been carried out on how plasma conden- Spadaroetal.2003),thatarenotinequilibriumandcon- sation relates to non-thermal equilibrium states of cool stantlyevolving(e.g.,Ugarte-Urraetal.2009). However, loops(Mu¨lleretal.2003,2004), whiletheroleoftheun- what remains unknown is the role plasma condensation derlying magnetic field plays in such scenarios remains plays in such processes (Chitta et al. 2013). unquantified (Chitta et al. 2013). Impulsive heating events, bundles of nanoflare heated In relation to the above discussion, we have on hand a loopstrands,occuratcoronalheightsandresultinchro- uniquedata-setderivedfromobservationstakenwiththe mosphericevaporation(e.g.,,Klimchuketal.2008;Klim- EUV Image Spectrometer (EIS; Culhane et al. 2007) on- chuk 2009). Bright TR emission is widely considered boardHinodethatprovidesdirectobservationalevidence 2 Orange et al. Table 1 ListofEISobservedemissionlineswithformatof: Ion, wavelength(˚A),andlogarithmicelectrontemperature. Ion Wavelength(˚A) logT Heii 256.32 4.9 Ov 192.90 5.4 Feviii 185.21 5.8 Feix 197.86 5.9 Fex 184.54 6.0 Fexii 186.88 6.2 Fexii 195.12 6.2 Fexiv 274.20 6.3 Fexv 284.16 6.4 of high-speed upflows at multiple TR temperatures oc- curring at the footpoint sites of a catastrophically cool- ing loop. We compliment these data with line-of-sight (LOS) magnetogram observations from the Heliosemic and Magnetic Imager (HMI; Schou et al. 2012) onboard SolarDynamicsObservatory(SDO)toinvestigatetheef- fects of magnetic flux evolution on both plasma upflows and the runaway cooling event. Inthatrespecttherestofthepaperisasfollows: obser- vationaldataitsprocessingandanalysisarepresentedin the next section, § 3 presents the measurement results, the discussion of these results and our conclusions are provided in §’s 4 and 5, respectively. 2. OBSERVATIONS,PROCESSING,ANDANALYSIS Observational data was obtained from EIS during the following time frame: 14:02:36 UT to 15:19:55 UT at ≈20 min time intervals on 18 October 2011. The data consistsofrasterscanswitha2(cid:48)(cid:48) slitwidthusing1(cid:48)(cid:48) steps resulting in a final field-of-view (FOV) of 100(cid:48)(cid:48)×130(cid:48)(cid:48) (Figure 1). Nine emission lines providing tempera- turecoveragefromthechromosphere(logT ≈4.9)tothe corona(logT ≈6.4)wereused. InTable1weprovidethe emittingions,theirrespectivewavelengths,andpeakfor- mation temperatures. Imagepre-processingofEISlevel-0datawasperformed with standard Solar SoftWare (SSW) to obtain flux cal- ibrated data. Additional corrections were made for the spatialoffsetoccurringbetweenitsshort(171–212˚A)and long (250–290˚A) wavelength bands (Young & Gallagher 2008), instrument and orbital jitter variations (Shimizu et al. 2007), CCD spectrum drift (Mariska et al. 2007), and tilt of the emission on the detector. The resultant EIS level-1 image data then possesses an absolute wave- Figure 1. Topandmiddle: EIS2(cid:48)(cid:48)intensityimages(ergcm−2s−1 length calibration of ±4.4 km s−1 (Kamio et al. 2010). sr−1),observed18October2011at14:59:21UT,ofOv(192.90˚A, logT≈5.4) and Feviii (185.20˚A, logT≈5.8), respectively, with The standard SSW routine eis auto fit.pro (Young green contours denoting the loop and boxed regions denoting it’s 2010) was used to derive integrated spectral line in- footpoints,respectively. Bottom: HMILOSmagnetogram(G)dis- tensities and their respective Doppler shifts, as well as played over the FOV of the EIS observation observed 18 October buildbothintensityandLOSvelocityimagesoftheloop 2011at14:59:21UTwithfootpointregionsagainidentified. (Figures 1 and 2). This routine fits a single Gaussian fits were consistent with the report of Del Zanna (2012) to each pixel forming the raster scan with the well- thatinon-diskquietSunregionsitcontributesover80% known mpfit.pro algorithm and propagates uncertain- of the observed intensity. Further support of this notion ties from 1σ fit uncertainties. However we note, during isfoundinadirectcomparisonofdistinctbrightnetwork this process multiple Gaussian fits were applied to both regionsoftheFexandHeiiintensityimagesinFigure3. the Heii256.32˚A and Ov192.90˚A spectrums due to Ov spectrum clearly indicated, and supported notions, their blending with coronal emission lines. Particulary, thatinmostregionstheFexi192.83˚Adominates(Young Heii256.32˚A is blended with that of the Six256.37˚A et al. 2007a), which is also observed in the similarity of and Fex256.41˚A lines, while Ov192.90˚A is blended Ov intensity images of Figure 3 to those of the corona. with that of the Fexi192.83˚A line (Young et al. 2007b; However, in the cool loop regions (i.e., such as those Brown et al. 2008). Visual inspection of Heii256.32˚A denoted by contours and boxes on Figure 1) enhanced Catastrophic Cooling in Solar TR Loops 3 Figure 2. Topandbottom: LOSvelocityin(kms−1)andelectrondensity(incm−3)images,bothobservedat14:41:17UTon18October 2011,derivedfromtheintensityimageoftheFeix(197.86˚A,logT ≈5.9),andGaussianintensityratioofFexii186˚A/195˚A,respectively. Onbothpanelsgreencontoursareintensitylevelsof80%–90%ofthepeakintensityderivedfromFeixemissionlineobservedat14:59:21 UTdenotingthecompletefilledcoolloop. Ov192.90˚A emission was observed, and as such allowed ground/foreground emission is removed from the loop it’s spectrum to be resolved and fitted. Moreover, we structure by applying custom written software to im- point out Young et al. (2007a) observed similar physi- age thumbnails with the loop at their center and back- calcharacteristicsinthespectrumaround192.83˚Afora ground/foreground emission surrounding it. The tech- study of an active region TR brightening. niqueobtainsabackgroundestimateby: first,applyinga FulldiskLOSmagnetogramsfromHMIareutilizedto low-passfiltertoremovehigh-frequencynoisesuchasbad investigate the magnetic structure of the loops. HMI’s pixels;then,usesahistogramofimagefluxtoisolatethe magnetogramswereobtainedwithaspatialresolutionof lowest10%;andfinally,appliesaweightedaveragetothe ≈0(cid:48).(cid:48)5atacadenceof≈45sandcorrespondedtotheap- lowest 98% of the isolated flux. It is noted, care is taken proximatelyonehourofEISobservations. Magnetogram to reduce background overestimation (i.e., generation of data was pre-processed using standard SSW techniques negative pixel values) since previous studies have shown incorporating per-pixel noise subtraction (Brown et al. that coronal intensities of loop structures are ≈10% – 2011) and then averaged over ≈2.5 min intervals to in- 20% higher than background/foreground emission (e.g., crease the signal-to-noise ratio with additional pointing Viall & Klimchuk 2012; Del Zanna & Mason 2003). The correctionsusingthetechniquesofOrangeetal.(2013a). result is a background subtracted image where pixel val- Magnetograms were co-aligned to EIS scans, with obser- ues are reduced by (cid:46)10%. We then measure Fλ(s) by vational time differences of (cid:46)three minutes, using the averaging over cross-sections of loop width at each point SSW routine drot map.pro. A resultant alignment er- alongitsspine. Lightcurvesaregeneratedoftotalradia- ror of (cid:46)2(cid:48)(cid:48) was measured by cross-correlating visually tive flux for each respective region (i.e., loop core, NFP, brightcoronalstructurestostrongmagnetogramregions. and SFP) by integrating their 3σ brightest flux. We also note, the image’s vicinity to solar disk center Resultant spectral intensities of the Fexii emission ((cid:46)±150(cid:48)(cid:48) in both the solar x and y directions) result in lines(Table1)wereusedtogenerateelectrondensityim- negligible projection effects. ages of the loop (Figure 2) via the techniques discussed The loop plus minimal background emission was de- byYoung(2011). Itisnoted,thougheachFexiiemission fined by using a semi-supervised tracing algorithm ap- lineisblended,previousstudieshavesuggestedfordensi- plied at each observational time step to Ov192.9˚A ties<1010 cm−3,consistentwithresultsherein,blending (logT ≈5.4) intensity images (Figure 1). The resul- effects to the Fe xii195˚A line are not important (Dere tant region was then used to isolate the loop in all 2008). Moreover, Dere et al. (2007) have suggested a other EUV images. We assigned the loop a set coor- root-mean-squareerrorof≈1.6fordeterminationofden- dinates s, corresponding to pixels along its spine, and sities in the quiet Sun when this line ratio is utilized in segmentedthemintothreedistinctregionsofcore,north, theaforementionedconditions. Then,LOSvelocities(vλ; and south footpoints (NFP and SFP, respectively; Fig- km s−1) and electron density (N ; cm−3) are measured e ure1). Segmentationwasperformedusingloopfootpoint as a function of loop length by the techniques described regionsidentifiedinFeviii185.2˚A(logT ≈5.8)intensity above. Light curves of these parameters as a function of images at each time step. It is noted, the footpoint re- loopregionareobtainedbysmoothingoverthecoreand gionswerevisuallyverifiedincorrespondingHeii256.3˚A footpoint regions as function of time. Theloop’sfootpointregionsareusedtoaggregatemag- (logT ≈4.9). neticfielddataateachtimestep(Figure1). Thesemag- Prior to measuring radiative flux (F ; arbitrary units) λ netogram data cubes are used to measure the magnetic as a function of loop length s, we apply a rigorous back- fluxdensity,i.e.,thenumberofpositiveandnegativepo- ground subtraction method. In this method, solar back- 4 Orange et al. Figure 3. EIS intensity images (erg cm−2 s−1 sr−1), covering electron temperatures in the range of logT≈4.9–6.2 (left to right, respectively),observedon18October2011from14:20:40UTto15:19:55UT(toptobottom,respectively)showingboththecool(logT≤6.0) and hot loop (logT>6.0) flux evolution. Note, an arrow identifies the filled cool loop corresponding to the time of peak emission in the coreregion. larity elements above and below a threshold value of 20 ShowninthetoprowofFigure3andcorrespondingto G, respectively. This threshold value is consistent with the14:20:40UT,temperatures≤1MKarecharacterized the average field-of-view strength of our magnetic field by visually bright footpoints and diffuse partially to not imagery. It is noted, analyzing magnetic field imagery filledloops. Thecoolloopcanthenbeseenbeginningto in this manner provides information on flux likely con- fill from the SFP to the NFP at logT ≈5.8 and 14:41:17 tributing to reconnection events given the notion that UT (i.e., 2nd row from top of Figure 3). The loop is stronger flux (i.e., > threshold) reconnects while weaker completelyfilled(logT ≤6.0),withatypicallength≈40 flux (i.e., < threshold) is “scattered” (Sakai et al. 1997; Mm,by14:59:21UT(i.e.,3rd rowfromtopFigure3and Chesny 2013, private communication). identified by arrow on Feviii185.2,˚A image). In the last Toexaminethetemperaturestructureasafunctionof observation time, 15:19:55 UT, the loop is returning to loop region, we use the aforementioned measurements of equilibriumandhascooledsufficientlygiventheobserved total radiative flux to execute an emission measure EM decreases in visually bright plasma. locianalysisviathetechniquesdiscussedbyOrangeetal. In Figure 4 we have provided the light curves of the (2013b). We employ the physical assumptions used by NFP, core, and SFP. In the NFP, the flux decreases un- Orange et al. (2013b), with exception of N which is de- e til the cool loop begins to fill (14:41 UT) and increases rivedfromourmeasurementsdiscussedabove. Note,loci thereafter, with exception of logT ≥6.2 which continu- curves represent the EM as it originates from isothermal ally decrease. The core region of the loop experiences plasmaatagiventemperaturetherebyrevealingisother- similar trends in flux evolution, as function of tempera- mal plasmas where all curves meet (Kamio et al. 2011). ture,asthatoftheNFP,withexceptionlogT ≥6.2which exhibit a triangular shape. SFP light curves for temper- 3. RESULTS Catastrophic Cooling in Solar TR Loops 5 Figure 4. FluxF (arbitraryunits)vstime(14:02UT–15:19UT)oftheloop’sNFP,core,andSFPregions(lefttoright,respectively) λ displayedfromtoptobottomasafunctionofincreasingtemperaturefortheemissionlinesofTable1,respectively. atures over 5.8≤logT ≤6.0 peak in irradiance at 14:41 logT ≈5.9 the plasma upflow speed peaked at ≈-60 km UT,whilebothcoolerandhotterregimespeak≈20min s−1,withupflowingplasmastilloccurringoveratemper- later (14:59 UT). aturerangeof5.8≤logT ≤6.2. However,attheSFPsite In relation to the cool loop, logT ≤6.0, irradiance and this same temperature range plasma was falling at peaks are consistent with the structural evolution, ob- a typical rate of (cid:46)5 km s−1 (Figure 5). Upon complete served in Figure 3 (i.e., consistent with previous results fillingofthecoolloop(i.e.,14:59UT)theplasmaflowdi- that loop fills from the SFP to the NFP). In terms of rectionsversustemperaturehadreturnedtotheirtypical the temporal evolution these flux peaks occur as follows: form,thatisupflowingplasmaintheupperTRandlower first SFP at 14:41 UT, then core at 14:59 UT, and fi- corona (5.8≤logT ≤6.2) with falling plasma at cooler nally NFP at 15:19 UT. Inspection of Figure 4 indicates andhottertemperatures. Wepointout,maximizedNFP that the cool loop is not a result of cooling coronal ma- plasma upflows, particularly over 5.8≤logT ≤6.0 tem- terial,basedoncomparisonsbetweenpeakTRandcoro- peratures,correspondedwiththetransitionfromupflows nal fluxes at the SFP site where said TR flux peaks pre- todownflows,atsimilartemperatures,intheSFPregion. cedethoseofhottertemperatures. Furthermore,theSFP peak in TR EUV flux transverses the loop back to the The coronal (logT ≈6.2) electron density evolution in NFP while hotter regions peak in the NFP and progress theNFPwasroughlyconstantoverthetimeframestud- towards the SFP site. ied here (Figure 6). However, in the SFP region sig- At the NFP prior to the appearance of the cool loop nificant density fluctuations occurred and are described (14:20UT;≈40min),temperaturesover5.8≤logT ≤6.2 as follows (Figure 6). During the first ≈20 mins (14:20 were characterized by plasma upflow speeds of ≈8.0 UT–14:41 UT) a mass in-flux of ≈30% is found. Next, – 40 km s−1 (Figure 5) with plasma falling at hotter mass-loss of ≈40% is witnessed over the next ≈20 mins and cooler temperatures. In the NFP at 14:41 UT and (14:41UT–14:59UT).Finally,duringthetimeinwhich 6 Orange et al. Figure 6. Normalizedelectrondensity(logNe,atlogT≈6.2)as a function of loop position (s ). The NFP and SFP regions are λ denoted on the plot, while the observation times of 14:20:40 UT, 14:41:17UT,14:59:21UT,and15:19:55UTarerepresentedbyx’s, squares,triangles,anddiamonds,respectively. as well as the approximately constant TR EM distribu- tion,aresuggestiveofunresolvedstructure(Brooksetal. 2012) and are expected given the varying visual nature observed in emission lines formed at these temperatures (Figure 3). The EM loci analysis also indicates the SFP is the site of condensation, based on observed differences be- tween EM’s of the two footpoint regions (logT’s≤6.0; Figure 7). This notion is expected when a single foot- point acts as the dominate energization site (Craig & McClymont 1986; McClymont & Craig 1987). These re- sults are further consistent with previous observations that the cool loop is filling from the SFP to the NFP. Moreover, they are consistent with both the evolution of flux and velocity observed in each of the loop regions as a function of temperature. The evolution of the normalized magnetic flux density at both footpoint sites is shown in Figure 8. We note, theNFPandSFPsitescorrespondedtothepositiveand negative polarity magnetic flux, respectively. The posi- tive magnetic flux density evolution, for intensities >20 Figure 5. LOS velocity (km s−1) versus electron temperature G, is described in detail as follows (i.e., NFP). The flux (logT) for the NFP (asterisks) and SFP (pluses) regions derived densitywasdecreasingatarateof≈1%min−1 untilthe from observational data on 18 October 2011 at the observational coolloopstartedtofill(14:41UT).Magneticfluxdensity times of14:20:40 UT–14:59:21 UT(top to bottom, respectively). was increasing at a similar rate during 14:41 UT–14:59 Note, v <0 and > indicate upflows and downflows, respectively, λ whilev =0isdenotedbydashedline. UT,thuscorrelatingwiththecompletefillingofthecool λ loop. While the cool loop drained and returned to equi- theloopcooledcompletely(14:59UT–15:19UT)itcon- librium the magnetic flux density was again decreasing tinued to lose mass with a total loss of ≈30%. at≈1%min−1. FortheSFPsite,themagneticfluxden- In Figure 7 we have provided EM loci curves for both sity, intensities <−20 G, were approximately constant footpoints as a function of observation time. It is ob- with minimal fluctuations, ≈±5%, during the observa- served,adistinctiveisothermalcomponentatlogT ≈6.2 tiontimeframestudied(Figure8). InFigure8amagne- (i.e., NFP and SFP; Figure 7). We note, though not togramsequencecenteredontheNFPhasbeenprovided shownhere,theseresultsareindicativeoftheloopcore’s toshowthemajorfluxelementscontributedtomagnetic EM loci analysis as well. The isothermal component is fluxdensitymeasurementsarecontainedanddonotdrift expected given the hot loop’s consistent visually bright out of the FOV. nature both throughout our observational sequence and emission lines with such formation temperatures (Fig- 4. DISCUSSION ure 3). Furthermore, the EM loci analysis indicates that We have presented observations of a cool loop thecoolloopanditsfootpoints,peakformationtempera- (logT ≤6.0)directlybelowathermallyisolatedhotcoro- tures≤1.0MK,werenon-isothermalduringouranalysis nal loop (logT ≈6.2) recorded on 18 October 2011 by givenitsbroaddistributionoflocicurves. Theseresults, EIS near solar center (≤100(cid:48)(cid:48) in both the solar x and y Catastrophic Cooling in Solar TR Loops 7 Figure 7. EMlocicurves(blueandgreenrepresentemissionlineswithpeakformationtemperaturesintheTRandcorona,respectively) for observation times of 14:20:40 UT–14:59:21 UT (top to bottom, respectively) of the NFP and SFP regions, left and right columns, respectively,derivedfromEISemissionlineintensities(Table1). directions). Our study, to best of our knowledge, pro- ure 7). This result is expected in the presence of impul- vides the first observational evidence of plasma upflows siveheatingtypeevents(Brooksetal.2012). Weexplain in the presence of cool loops in quiet Sun regions. HMI these results by first noting that at the cool loop onset LOS magnetic field observations were utilized to investi- (i.e., initial filling) non-steady symmetrical flows indi- gate the relationship between EUV flux evolutions and cate an asymmetric loop structure (Mariska et al. 1982; plasma motion compared to that of the underlying mag- Craig & McClymont 1986; McClymont & Craig 1987), netic field. The intensity and velocity structure of the while simultaneously plasma condensation is occurring cool loop showed characteristics similar to those pub- at the SFP site (Figure 7). The runaway cooling ob- lished by Tripathi et al. (2012) for active region loops, served ≈20 min later, 14:59 UT, in lower coronal and andcontrastthosemosttypicallyfoundinthepresenceof upper TR EUV images is then indicative of the move- suchstructures,e.g.,redshiftsalongandatthefootpoints ment of the condensation region to the less heated foot- of these structures (e.g., Del Zanna 2008; Tripathi et al. point (i.e., NFP). Craig & McClymont (1986) noted the 2009;Chesnyetal.2013). Bothfootpointswerepredom- temperature gradient of a condensing loop leg (i.e., our inately blueshifted throughout the upper TR and lower SFP region) is shallower than that of the evaporating corona (5.8≤logT ≤6.2) during the cool loop’s lifetime, leg (i.e., our NFP region), and as such is characterized except when filling began (14:41 UT). At this time a by larger EMs. Therefore, using the minima of our foot- symmetrical flow was observed that corresponded with pointEMdistributions,overlogT=5.8-6.0(Figure7), maximal upflow speeds in the NFP region (v ≈60 km a heating rate asymmetry of ≈2% existed between the λ s−1)peakingatlogT ≈5.9anddecreasedwithincreasing NFPandSFPregions. Wepointout,Mu¨lleretal.(2003) temperature. reported a 1% energy asymmetry between loop legs dic- As discussed in § 1, cool loops have often been consid- tates the draining direction which supports our observa- eredtobearesultofcoolingandcondensingcoronalma- tions of the condensation being driven from the SFP to terialwhichwasheatedimpulsivelyovermanystrandsat the NFP. These results provide significant evidence that coronalheights. However,likethesuggestionsofTripathi the catastrophic cooling event occurred from the loop’s et al. (2012), our observations of plasma motions do not non-equilibrium state. Moreover, the non-equilibrium support models which predict upper TR and lower coro- state formed when plasma condensation began in a sin- nalemissionlinesthataredominatedbyredshiftedemis- gle footpoint. Below we hypothesize on the mechanism sion. Our EM loci analysis indicates the presence of un- responsible for initiating the condensation event. resolved structure and non-isothermal plasma through- Hegglandetal.(2009)suggestobservationsofsolarat- outthe coolerlayersof theatmosphere (logT ≤6.0; Fig- mospheric bi-directional jets are useful tools for probing 8 Orange et al. Figure 8. Right: evolution of the normalized magnetic surface flux density (ρΦ±; arbitrary units) for the NFP (asterisks, Φ+ > 20 G) andSFP(triangles,Φ− <20G)regionsovertheobservationaltimeframestudiedherein. Left: magnetogramtemporalevolutionofNFP regionovertheobservationaltimeframeof14:20UT–15:19UTfromtoprighttobottomrightinclockwisefashion,respectively. the heights in which magnetic energy is being converted was cooler regions of the solar atmosphere, particularly to thermal energy. As such, bi-directional jets provide theupperTR.Moreover,thesenotionssupportfootpoint uniquediagnostictoolsforconstrainingtheheightsofat- heating scenarios, possibly at higher temperatures than mospheric heating. Inspecting our velocity versus tem- typically considered, (Tripathi et al. 2012; Aschwanden perature profiles prior to and at the onset of loop fill- et al. 2007). Finally, our results also point to the fact ing (Figure 5), a bi-directional jet is occurring between that this whole process can be considered a domino ef- logT ≈5.4–5.8 at the NFP site. Combining these ob- fectasadirectresultofTRmagneticreconnectioninthe servationswithsignificantdropsinmagneticsurfaceflux opposing loop leg. Therefore, we have presented strong density (Figure 8) occurring simultaneously both spa- evidence supporting notions that coronal heating is not tially with the NFP and temporally with peak plasma confined to only coronal temperatures, while tracing the upflows, we find support that impulsive magnetic recon- origin of its magnetic energy conversion source. nection events between the photospheric footpoint and It is worth noting, for logT ≥6.2 light curves indicate surrounding background field are the source of the jet. footpoint heating of the hot loop in the NFP region, Consistencies of our observational reports to Heggland which peaked at ≈14:20 UT (Figure 4). Coronal EM etal.’s(2009)TRsimulatedreconnection,leadustosug- loci provide evidence of condensation occurring simulta- gestthereconnectioneventpropelledacooldenseblobof neously in time and spatial region (Figure 7). Using the plasma upwards along the field lines to the region of the techniques discussed previously, a heating asymmetry of SFP. Thereby, the SFP’s sudden density enhancement ≈5% was measured between the footpoints for tempera- (Figure 6), and most likely the conversion of magnetic turesoflogT ≈6.2–6.4(Figure7). Theseresultsfurther waveenergytoheat(Hollweg&Yang1988;Poedts&de support our suggestion of footpoint heating for the hot Groof 2004), initiated plasma condensation. However, loop(i.e.,NFPregion),aswellasthereportsofTripathi it cannot be ruled out that a dip in the magnetic field etal.(2012)andAschwandenetal.(2007). Finally,sym- topology,atorveryneartheSFP,wasresponsibleforthe metricalflowsinthehotloopatlogT ≈6.2and14:20UT initiation of plasma condensation (Muller et al. 2003). (Figure 5) provide evidence that the heating event was Thepreviousdiscussionpointstothefactthatthecool occurring at coronal heights with high enough frequency loop was heated in a single footpoint (i.e., SFP) which to maintain its visually and isothermally stable nature lead to a runaway cooling based on the non-equilibrium (Figures 3 and 7, respectively). structure of the loop. The SFP heating event is consid- eredtobelowfrequencyinnature(e.g.,nanoflare;Chitta 5. CONCLUSION etal.2013)asthisexplainsourEMlociresults(Figure7) This work has presented a play-by-play of the life cy- while remaining consistent with previous notions of cool cle of a cool loop that emphasized the importance of the loops (Ugarte-Urra et al. 2009; Spadaro et al. 2003). TR as the site of coronal heating. Our results have pro- Our pervasive blueshifts immediately after catastrophic videdsignificantinsightontherelationshipbetweennon- cooling, are indicative of plasma evaporation (Tripathi equilibriumstructuringandcondensationincoolplasma et al. 2012), suggest the origin of the nanoflare storm loops. It has also provided the first observational evi- Catastrophic Cooling in Solar TR Loops 9 dence,tobestofourknowledge,ofplasmaupflowsinthe —.2012,ArXive-prints presenceofcoolloopsinquietSunregionswhilesupport- DelZanna,G.,&Mason,H.E.2003,A&A,406,1089 ing the findings of Tripathi et al. (2012) for cool active Dere,K.P.2008,A&A,491,561 Dere,K.P.,Doschek,G.A.,Mariska,J.T.,etal.2007,PASJ,59, region loops. We have built upon the work of Tripathi 721 et al. (2012) by including complimentary LOS magne- Golub,L.,Bookbinder,J.,Deluca,E.,etal.1999,Physicsof togramdataandcoronaldensityevolution. Weconclude Plasmas,6,2205 from the evolution of the underlying magnetic flux that Hanson,J.M.,Roelof,E.C.,&Gold,R.E.1980,NASA observed TR upflows are an indication of magnetic re- STI/ReconTechnicalReportN,81,28032 Hansteen,V.,Maltby,P.,&Malagoli,A.1996,inAstronomical connection at similar atmospheric heights. SocietyofthePacificConferenceSeries,Vol.111,Astronomical We recognize, more observations are required to SocietyofthePacificConferenceSeries,ed.R.D.Bentley& provide conclusive statements about the specifics of cool J.T.Mariska,116–121 loop heating and the shared relationship between their Heggland,L.,DePontieu,B.,&Hansteen,V.H.2009,ApJ,702, continuous evolution and plasma condensation, which 1 Hollweg,J.V.,&Yang,G.1988,J.Geophys.Res.,93,5423 is planned for in a forthcoming paper. Moreover, these Kamio,S.,Curdt,W.,Teriaca,L.,&Innes,D.E.2011,A&A, are required to better understand the cool loop’s non- 529,A21 equilibrium nature as it relates to both asymmetrical Kamio,S.,Hara,H.,Watanabe,T.,Fredvik,T.,&Hansteen, and symmetrical flow patterns. Finally, plasma upflows V.H.2010,Sol.Phys.,266,209 in cool loops require further observational evidence Klimchuk,J.A.2009,inAstronomicalSocietyofthePacific ConferenceSeries,Vol.415,TheSecondHinodeScience to be used as constraints on the models proposed by Meeting:BeyondDiscovery-TowardUnderstanding,ed. Mu¨ller et al. (2003, 2004) to determine their validity. B.Lites,M.Cheung,T.Magara,J.Mariska,&K.Reeves,221 Lastly, we have not made suggestions on whether any Klimchuk,J.A.2012,JournalofGeophysicalResearch(Space heating connection exists between the cool and hot Physics),117,12102 Klimchuk,J.A.,Patsourakos,S.,&Cargill,P.J.2008,ApJ,682, loop. However, significant coronal mass-influx peaking 1351 simultaneously with TR plasma upflows suggest such a Langangen,Ø.,DePontieu,B.,Carlsson,M.,etal.2008,ApJ, connection exists. These notions give rise to questions 679,L167 of the mass and energy coupling relationship shared Mackay,D.H.,Karpen,J.T.,Ballester,J.L.,Schmieder,B.,& betweenthesetwoloopstructures,theTRandcoronain Aulanier,G.2010,SpaceSci.Rev.,151,333 Mariska,J.T.,Doschek,G.A.,Boris,J.P.,Oran,E.S.,& general, and whether the high frequency heating events Young,Jr.,T.R.1982,ApJ,255,783 of the hot loop are a direct result of the conversion of Mariska,J.T.,Warren,H.P.,Ugarte-Urra,I.,etal.2007,PASJ, magnetic to thermal energy at TR heights. 59,713 McClymont,A.N.,&Craig,I.J.D.1987,ApJ,312,402 Mu¨ller,D.A.N.,Hansteen,V.H.,&Peter,H.2003,A&A,411, 605 The authors greatly appreciate the reviewers con- Mu¨ller,D.A.N.,Peter,H.,&Hansteen,V.H.2004,A&A,424, structive comments on the manuscript. This research 289 was supported by National Aeronautics and Space Ad- Oluseyi,H.M.,Walker,II,A.B.C.,Porter,J.,Hoover,R.B.,& ministration (NASA) grant NNX-07AT01G and Na- Barbee,Jr.,T.W.1999a,ApJ,524,1105 tional Science Foundation (NSF) grant AST-0736479. Oluseyi,H.M.,Walker,II,A.B.C.,Santiago,D.I.,Hoover, R.B.,&Barbee,Jr.,T.W.1999b,ApJ,527,992 Any opinions, findings and conclusions or recommen- Orange,N.,Oluseyi,H.,Chesny,D.,etal.2013a,Sol.Phys. dations expressed in this material are those of the au- —.2013b,Sol.Phys. thor(s) and do not necessarily reflect the views of the Poedts,S.,&deGroof,A.2004,inESASpecialPublication,Vol. NSF or NASA. The final publication is available at 575,SOHO15CoronalHeating,ed.R.W.Walsh,J.Ireland, http://iopscience.iop.org/0004-637X/778/2/90/ D.Danesy,&B.Fleck,62 Sakai,J.I.,Ryutova,M.,Schrijver,K.,etal.1997,inBulletinof theAmericanAstronomicalSociety,Vol.29,AAS/Solar REFERENCES PhysicsDivisionMeeting#28,904 Schou,J.,Scherrer,P.H.,Bush,R.I.,etal.2012,Sol.Phys.,275, Aschwanden,M.J.,&Nightingale,R.W.2005,ApJ,633,499 229 Aschwanden,M.J.,&Schrijver,C.J.2002,ApJS,142,269 Shimizu,T.,Katsukawa,Y.,Matsuzaki,K.,etal.2007,PASJ,59, Aschwanden,M.J.,Winebarger,A.,Tsiklauri,D.,&Peter,H. 845 2007,ApJ,659,1673 Spadaro,D.,Lanza,A.F.,Karpen,J.T.,&Antiochos,S.K. Beckers,J.M.1968a,Sol.Phys.,3,367 2006,ApJ,642,579 —.1968b,Sol.Phys.,3,367 Spadaro,D.,Lanza,A.F.,Lanzafame,A.C.,etal.2003,ApJ, Brooks,D.H.,Warren,H.P.,&Ugarte-Urra,I.2012,ApJ,755, 582,486 L33 Tripathi,D.,Mason,H.E.,DelZanna,G.,&Bradshaw,S.2012, Brown,C.M.,Feldman,U.,Seely,J.F.,Korendyke,C.M.,& ApJ,754,L4 Hara,H.2008,ApJS,176,511 Tripathi,D.,Mason,H.E.,Dwivedi,B.N.,delZanna,G.,& Brown,D.,Regnier,S.,Marsh,M.,&Bewsher,D.2011,version Young,P.R.2009,ApJ,694,1256 1.2 Tripathi,D.,Mason,H.E.,&Klimchuk,J.A.2010,ApJ,723,713 Chesny,D.L.,Orange,N.B.,&Oluseyi,H.M.2013,Submitted Ugarte-Urra,I.,Warren,H.P.,&Brooks,D.H.2009,ApJ,695, Chitta,L.,Kariyappa,R.,vanBallegooijen,A.A.,etal.2013, 642 ArXive-prints Viall,N.M.,&Klimchuk,J.A.2012,ApJ,753,35 Craig,I.J.D.,&McClymont,A.N.1986,ApJ,307,367 Walker,Jr.,A.B.C.,Hoover,R.B.,&Barbee,Jr.,T.W.1993a, Culhane,J.L.,Harra,L.K.,James,A.M.,etal.2007, inSocietyofPhoto-OpticalInstrumentationEngineers(SPIE) Sol.Phys.,243,19 ConferenceSeries,Vol.1742,SocietyofPhoto-Optical DePontieu,B.,McIntosh,S.W.,Hansteen,V.H.,&Schrijver, InstrumentationEngineers(SPIE)ConferenceSeries,ed.R.B. C.J.2009,ApJ,701,L1 Hoover&A.B.C.Walker,Jr.,515–526 DePontieu,B.,McIntosh,S.,Hansteen,V.H.,etal.2007,PASJ, 59,655 DelZanna,G.2008,A&A,481,L49 10 Orange et al. Walker,Jr.,A.B.C.,Hoover,R.B.,&Barbee,Jr.,T.W.1993b, Young,P.2010,SSWsoftwareforautomaticallyfitting inSocietyofPhoto-OpticalInstrumentationEngineers(SPIE) Hinode/EISspectra, ConferenceSeries,Vol.1742,SocietyofPhoto-Optical http://msslxr.mssl.ucl.ac.uk:8080/Wiki.jsp?page=eis auto fit.pro InstrumentationEngineers(SPIE)ConferenceSeries,ed.R.B. —.2011,Derivingdensities,columndepths,andfillingfactors Hoover&A.B.C.Walker,Jr.,500–514 fromHinode/EISdata, Warren,H.P.,Ugarte-Urra,I.,Doschek,G.A.,Brooks,D.H.,& http://msslxr.mssl.ucl.ac.uk:8080/Wiki.jsp?,eISSOFTWARE Williams,D.R.2008,ApJ,686,L131 NOTENo.15,Version2.3 Warren,H.P.,Winebarger,A.R.,&Brooks,D.H.2010,ApJ, Young,P.R.,DelZanna,G.,Mason,H.E.,etal.2007a,PASJ, 711,228 59,727 Winebarger,A.R.,Schmelz,J.T.,Warren,H.P.,Saar,S.H.,& —.2007b,PASJ,59,857 Kashyap,V.L.2011,ApJ,740,2 Young,C.A.,&Gallagher,P.T.2008,Sol.Phys.,248,457