Hot X-Ray Onsets of Solar Flares

MNRAS 000, 000{000 (0000) Preprint 13 July 2020 Compiled using MNRAS LATEX style file v3.0 Hot X-ray Onsets of Solar Flares Hugh S. Hudson,1;2? Paulo J. A. Sim~oes,3;1 Lyndsay Fletcher,1;4 Laura A. Hayes,5 Iain G. Hannah1 1SUPA School of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ, UK 2Space Sciences Laboratory, U.C. Berkeley, CA USA 3Centro de Rádio Astronomia e Astrof´ısica Mackenzie, Escola de Engenharia, Universidade Presbiteriana Mackenzie, S~aoPaulo, Brazil 4Rosseland Centre for Solar Physics, University of Oslo, P.O.Box 1029 Blindern, NO-0315 Oslo, Norway 5Solar Physics Laboratory, Code 671, Heliophysics Science Division, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA June 23, 2020 ABSTRACT The study of the localized plasma conditions before the impulsive phase of a solar flare can help us understand the physical processes that occur leading up to the main flare energy release. Here, we present evidence of a hot X-ray ònsetâA˘Z´ interval of enhanced isothermal plasma temperatures in the range of 10-15 MK up to tens of seconds prior to the flareâA˘Zs´ impulsive phase. This `hot onsetâA˘Z´ interval occurs during the initial soft X-ray increase and prior to the detectable hard X-ray emission. The isothermal temperatures, estimated by the Geostationary Operational Environmental Satellite (GOES) X-ray sensor, and confirmed with data from the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI), show no signs of gradual increase, and the `hot onset' phenomenon occurs regardless of flare classification or configuration. In a small sample of four representative flare events we identify this early hot onset soft X-ray emission mainly within footpoint and low-lying loops, rather than with coronal structures, based on images from the Atmospheric Imaging Assembly (AIA). We confirm this via limb occultation of a flaring region. These hot X-ray onsets appear before there is evidence of collisional heating by non-thermal electrons, and hence they challenge the standard flare heating modeling techniques. 1 INTRODUCTION 12-25 keV for weaker ones) by the Reuven Ramaty High En- ergy Solar Spectroscopic Imager (RHESSI, Lin et al. 2002). Often flare initiation, as seen in soft X-ray (SXR) data from Our main finding is that the GOES isothermal temperatures the X-ray Sensor (XRS) on Geostationary Operational En- are significantly elevated from the very beginning of the on- vironmental Satellite (GOES), begins with a slow `precur- set phase, i.e. well before we have evidence for collisional sor' development phase. This can sometimes be identified heating by non-thermal electrons. We cannot preclude the with non-thermal activity (e.g. Fárn´ıket al. 2003) or with possibility of undetectable HXR emission, especially with a non-thermal velocity distributions (Harra et al. 2013). The softer spectrum, in the onset time interval. We note that preflare interval is often also taken as evidence for a `pre- Awasthi & Jain(2011) had already reported similar phe- heating' phase in which a gradual process heats a volume nomena via the independent dataset from the Solar X-Ray of flare plasma without a detectable hard X-ray signature Spectrometer (SOXS) spectrometer experiment (Jain et al. (e.g. Cheng et al. 1985), implying a very low flux of non- 2006), in a sample of 13 events. arXiv:2007.05310v1 [astro-ph.SR] 10 Jul 2020 thermal electrons, if any. Systematic studies of soft X-ray images suggested that in most cases any precursor source could not be directly identified with the main flare (Fárn´ık & Savy 1998; Hudson et al. 2008), appearing near but not exactly at the flare site. Previous conclusions about the relationship between `precursors' and flares mostly have dealt with image struc- We have used extreme ultraviolet (EUV) images to ture. Here we study the X-ray spectral evolution, focusing search out the spatial patterns of the onset sources (Sec- on a sample of four representative events. We examine flare tion 2.5). We have also studied the RHESSI data for the onset emission, where we define the term ònset' as the pre- four sample events (Section 3.2), finding satisfactory qual- flare interval during which elevated GOES soft X-ray flux itative agreement during the flare development, specifically is detected, but prior to the detection of any elevated hard in matching the isothermal-fit GOES parameters with the X-ray (HXR) emission (at >25 keV for stronger events, and more complete spectroscopy possible with RHESSI. c 0000 The Authors 2 H.S. Hudson et al. 2 DATA arbitrarily to represent fast, slow, strong, and weak flares respectively, crudely bracketing the parameter space of rise 2.1 GOES soft X-ray data time and flare energy. \Fast/slow" refers to the event rise The GOES series of missions has provided soft X-ray mea- time and \strong/weak" refers to the GOES 1-8 A˚ peak flux surements via its X-ray Sensor (XRS) instrument in two values. These typical events are from the 2011-2014 time nominal wavelength bands, (1{8 A˚ and 0.5-4 A),˚ for many frame and do not represent different flare classes as such, decades now. Such observations began as early as 1960 with since the parameters generally have broad, continuous dis- ionization chambers on board SOLRAD and other satel- tributions (e.g. Lee et al. 1993). lites (Dere et al. 1974; Thomas et al. 1985; White et al. The hot onset sources appear substantially before the 2005). The two passbands of the GOES/XRS, 1{8 and 0.5{ beginning of the impulsive phase (indicated by the dotted 4 A,˚ allow for the determination of an isothermal temper- vertical lines in the left column panels in Fig.1), as deter- ature and emission measure, interpreted here in terms of mined by RHESSI hard X-rays above 25 keV where possible, the CHIANTI atomic-physics package (Dere et al. 1997) and above 12 keV where not. The correlation between tem- as implemented in the SolarSoft (Freeland & Handy 1998) perature and emission measure (right column in Figure1) code GOES TEM.pro. These parameters usually describe shows a roughly clockwise circulation during the main phase the coronal part of the flare, and specifically the plasma of the flare, ending with the cooling of the coronal loops. The trapped in the system of magnetic loops made visible in soft hot onset emission precedes these features, appearing at the X-rays by the injection of new plasma expanding upwards lowest emission measure but an elevated temperature. The from the lower atmosphere due to the sudden energy re- cooling phase passes through the onset temperature range lease. The GOES data also have sufficient sensitivity and smoothly, establishing that the 10-15 MK level is not an ar- signal contrast to study the onset phase of a flare, often tifact. The data points are colour-coded and mapped to the many minutes prior to the impulsive phase (Kane & Ander- temperature curves in the middle column of Figure1 to indi- son 1970). In this paper we use these simple GOES/XRS cate the time-evolution of this correlation. These panels in- observations to characterize the onset temperatures, at the dicate that the hot onset (with temperatures around 10 ∼ 15 earliest possible times permitted by the observations, and MK) is associated with a low amount of plasma, with emis- 47 −3 then follow up with EUV images, taken by the Atmospheric sion measure values below 10 cm . In the flare sample Imaging Assembly (AIA, Lemen et al. 2012) on board the discussed here, the \fast strong" event SOL2014-01-07 (bot- Solar Dynamics Observatory (SDO, Pesnell et al. 2012). tom row of Fig.1), for example, has a hot onset detectable We also examine a flare series in which limb occultation dis- more than a minute prior to the detectable HXR emission. tinguishes the coronal and chromospheric components (e.g. The GOES isothermal onset temperatures, i.e. the first ob- Hudson 1978). servable measurements lie well above the low-temperature Because the GOES data integrate over the whole disk range of the these data (e.g. Sterling et al. 1997; White et al. (Sun-as-a-star), all of the concurrent soft X-ray sources will 2005). So far as the data permit us to tell, the first detected contribute to the background level for a given flare. In prin- emission at these hot onset times already has a measurable ciple there is no exact way to estimate this background level temperature significantly above any observational limit. for such Sun-as-a-star observations, since an independent source(s) could occur at any time, in any active region that 2.3 Uncertainties on GOES temperature might be coincidentally present. In the present work we es- measurements timate the flare background level by simply taking the local minimum of the 0.5-4 A˚ channel immediately prior to the The error bars in Figure1 reflect both random errors, as flare onset; the nearer the better. The actual epoch of the estimated from the scatter of data at an intermediate flux hot onset will depend upon flare brightness and detection level, and the digital uncertainty resulting from undersam- threshold; for an X-class flare occurring in low-background pling the true background noise, as discussed in Sim~oes et al. conditions, the GOES photometers can already detect the (2015a). The digital step size varies from epoch to epoch, source at a level 0.1-1% of flare maximum flux. since the different GOES satellites have different properties, We can also check the background source locations via For GOES-15 in the 0.5-4 A˚ channel it was 6 × 10−10 W/m2 the EUV images from SDO/AIA.

Load more