Hot plasma in a quiescent solar active region as measured by RHESSI, XRT, and AIA

TL;DR

This study detects a hot plasma component in a quiescent solar active region using RHESSI, AIA, and XRT data, supporting nanoflare heating models with detailed temperature and emission measure analysis.

Contribution

It provides the first combined imaging and spectral analysis of hot plasma in a non-flaring active region across multiple wavelengths, revealing a multi-thermal structure consistent with nanoflare heating.

Findings

Hot plasma component detected with temperatures around 7 MK.

DEM analysis shows a broad temperature distribution with a peak between 2-3 MK.

Hot plasma has a small emission measure, indicating low filling factors.

Abstract

This paper investigates a quiescent (non-flaring) active region observed on July 13, 2010 in EUV, SXR, and HXRs to search for a hot component that is speculated to be a key signature of coronal heating. We use a combination of RHESSI imaging and long-duration time integration (up to 40 min) to detect the active regions in the 3-8 keV range during apparently non-flaring times. The RHESSI imaging reveals a hot component that originates from the entire active region, as speculated for a nanoflare scenario where the entire active region is filled with a large number of unresolved small energy releases. An isothermal fit to the RHESSI data gives temperatures around ~7 MK with emission measure of several times 10^46 cm^-3. Adding EUV and SXR observations taken by AIA and XRT, respectively, we derive a differential emission measure (DEM) that shows a peak between 2 and 3 MK with a steeply decreasing high-temperature tail, similar to what has been previously reported. The derived DEM reveals that a wide range of temperatures contributes to the RHESSI flux (e.g. 40 % of the 4 keV emission being produced by plasma below 5 MK, while emission at 7 keV is almost exclusively from plasmas above 5 MK) indicating that the RHESSI spectrum should not be fitted with an isothermal. The hot component has a rather small emission measure (~0.1 % of the total EM is above 5 MK), and the derived thermal energy content is of the order of 10 % for a filling factor of unity, or potentially below 1 % for smaller filling factors.

Tables

Table 1. Table 1: RHESSI HXR fluxes in each RHESSI orbit on July 13, 2010. Isothermal temperatures derived by those RHESSI fluxes are also shown.

Orbit	Time interval	Flux [photons/s/cm2/keV]				Temperature [MK]
No.	[UT]	3.5–4.5 keV	4.5–5.5 keV	5.5–6.5 keV	6.5–7.5 keV	with 3.5–6.5 keV	with 3.5–7.5 keV
1	05:31:48–05:48:22	16.1	1.95	0.222	0.0780	6.25	7.23
2	08:56:25–09:29:55	9.02	0.884	0.0912	–	5.31	–
3	13:30:30–14:18:10	15.6	2.44	0.328	0.0814	7.06	7.36
4	17:19:34–17:40:08	22.0	2.38	0.372	0.0828	6.53	7.02
5	23:04:40–23:55:00	3.0	0.1	0.4	–	–	–

Full text

Sun: corona; Sun: X-rays, gamma rays

††journal: ApJ

1 Introduction

To investigate energetics in the solar corona and explore how the corona maintains its high temperature, it is important to detect and quantitatively evaluate the hot ($>$5 MK) plasma component above the typical coronal temperature (2–3 MK) in active regions (e.g., Klimchuk, 2009). Although solar flares heat the coronal plasma well above 10 MK, it is known that the energy released by flares individually is not significant in the coronal energy balance (e.g., Shimizu, 1995). Therefore, the important next steps are studies of the hot component during quiescent times when no individually resolved flare is detected. Simulations suggest that a quasi quiescent hot plasma component could be produced by a superposition of the large number of discrete, but temporally overlapping, compact impulsive energy releases that are distributed over the entire Sun (Cargill & Klimchuk, 2004; Klimchuk et al., 2008; Bradshaw & Klimchuk, 2011), called nanoflares, originally introduced by Parker (1988).

From the observational side, temperature structures of active region plasmas have been investigated using observations at various wavelenghts, such as extreme ultraviolets (EUVs, i.e., Warren et al., 2012; Brosius et al., 2014; Parenti et al., 2017) and soft X-rays (SXR, i.e., Parkinson, 1975; Peres et al., 2000; Orlando et al., 2001, 2004; Reale et al., 2009a; Schmelz et al., 2009a; Del Zanna & Mason, 2014). In these studies the differential emission measures (DEMs), temperature derivatives of emission measures, are reconstructed from the observations. Deriving the DEM from this limited set of observations is not providing a unique solution, and calibration uncertainties can therefore significantly influence the result. The temperature range over which the DEM is reconstructed should be carefully selected. Hot plasmas above $>$5 MK emit several EUV lines (Young et al., 2007) and SXR emission (Golub et al., 2007) and therefore EUV and SXR observations in principle provide good diagnostics of hot plasmas. However, it has been pointed out that a combination of EUV and SXR observations is not enough to evaluate the hot plasma with current instruments such as EUV Imaging Spectrometer (EIS, Culhane et al., 2007) and X-ray Telescope (XRT, Golub et al., 2007) onboard the Hinode satellite, except at times when large flares occurred (Winebarger et al., 2012). During non-flaring times, EUV lines sensitive to the hot plasma are simply too faint to be detectable by EIS, and the temperature response of XRT is too wide and therefore dominated by the much larger emission measures of cooler plasma.

To properly detect the high-temperature tail, hard X-rays (HXR, X-ray emissions with a few keV and above hereafter) are an essential diagnostic tool. HXRs are produced by bremsstrahlung, but only the tail of the electron distribution has enough energy to make photons in the HXR range. Hence, HXR bremsstrahlung observations are biased towards the hottest plasma making them the ideal diagnostic of the hottest temperatures. Schmelz et al. (2009b) showed that the hot component is highly constrained by an HXR upper limit derived from an instrument background of the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI, Lin et al., 2002) satellite. Therefore, a combination of HXR with other wavelength bands is essential to accurately evaluate a DEM with a wide temperature range that includes the hot component. McTiernan (2009) analyzed RHESSI observations of day-night transitions during non-flaring times, and found HXR emissions from 5–10 MK plasma demonstrating that a hot component is typically present, at least within the sensitivity range of RHESSI. Their work was a spectroscopic analysis using the day-night transition for an accurate non-solar background subtraction, but no imaging was performed. The detected emissions, however, are thought to come from active regions. Reale et al. (2009b) compared a RHESSI observations with SXR observations by Hinode/XRT, and found that RHESSI mainly detects emission from a hot plasma around 6–8 MK and XRT could mainly observe cooler plasmas around 2–2.5 MK, indicating the existence of a hot component as envisioned by nanoflare coronal heating models.

More recently, advances have been made with high sensitivity HXR focusing optics telescopes. HXR focusing optics has much improved sensitivity compared to RHESSI’s indirect imaging method by achieving large effective areas and a low non-solar background. The Focusing Optics Solar X-ray Imager (FOXSI) sounding rocket experiment applied this technique for the first time to the solar observation (Krucker et al., 2014). The first flight had a moderate sensitivity that was not high enough to detect HXR emissions from a non-flaring active region, but the observations nevertheless strongly constrained the hot component in active regions (Ishikawa et al., 2014). During the second launch of the FOXSI sounding rocket which provided a much improved sensitivity, HXR emissions from a quiescent active region have been detected (Ishikawa et al., 2017), corroborating the existence of the hot component even above 10 MK. The NuSTAR satellite (Harrison et al., 2013) has provided further hard X-ray focusing observations giving new insights into the existence of a hot component. As NuSTAR has not been designed for solar observations, the observations so far are limited by short effective exposure times, and only upper limits of the quiescent component above 5 MK have been derived so far (Hannah et al., 2016; Grefenstette et al., 2016). The availability of all these new hard X-ray focusing observations have triggered several simulation studies, and initial results of nanoflare models are constructed to explain those observations (Barnes et al., 2016a, b; Marsh et al., 2018).

In this paper we revisit the RHESSI data taking advantage of the possibility to image and integrate in time by almost an hour to compensate for RHESSI’s moderate effective area. We report on hard X-ray emissions in the range from 3 to 8 keV from an active region on July 13, 2010 during time intervals without any individual X-ray flares. RHESSI was able to successfully obtain HXR images of this quiescent active region, and by combining with EUV and SXR observations, it was possible to measure the differential emission measure distribution and compare the core component with the hot tail of the distribution.

2 Observations

Our study presented here does not intend to be a statistical search, but we wanted to find an apparently quiescent, non-flaring active region that is detected close to the RHESSI sensitivity limit. To be able to compare the RHESSI results with extreme UV and soft X-ray observations, we restricted our search for periods after February 2010 when observations of the Atmospheric Imaging Assembly onboard the Solar Dynamic Observatory satellite (SDO/AIA, Lemen et al., 2012) started. In addition, only times when Hinode/XRT was running in a multiple filter configurations especially designed to observe active region were considered. To have good calibration of the RHESSI data and minimal effects of radiation damage, we further restricted the search to times within a few months after the second RHESSI anneal that was completed by the end of April 2010. To facilitate RHESSI imaging, we looked further at times when a single active region was dominating the total X-ray flux. This simplified RHESSI imaging as a single source is much easier to image for RHESSI than two widely separated sources from different active regions. A good candidate for our study was found to be NOAA AR 11087, a single active region seen on the disk on July 13, 2010 that has good coverage by RHESSI, AIA and XRT. The upper panel of Fig. 1 shows the GOES X-ray lightcurve in linear scale for part of the selected day. The solar activity was low with the X-ray background flux being around B1 level. Several A- and B-class microflares occurred throughout the day, but there are also times in between microflares without obvious activity in the GOES light curve. The RHESSI spacecraft has a day-night cycle providing uninterrupted solar observations of up to 60 minutes per 96 minute orbit. Avoiding times when RHESSI crossed the South Atlantic Anomaly, we selected 5 RHESSI orbits when XRT ran a standard active region observation program with two filter configurations, the Al-mesh and Ti-poly filters, and an image with the Al-thick filter was taken every hour. GOES and RHESSI data for the 5 selected orbits are shown in the bottom panel of Fig. 1. The RHESSI data is shown as a lightcurve (4-8 keV), but also as spectrogram plots covering the energy range from 3 to 200 keV. In the spectrogram plot where the observed HXR count rate is shown in color in a 2D representation as a function of time and energy, the RHESSI non-solar background variations can be clearly identified by following the high-energy time variations. The observed variations reflect the spacecraft latitude with the background increasing towards higher latitudes. We selected intervals for RHESSI imaging spectroscopy analysis (blue lines) that exclude times of obvious X-ray microflares seen by GOES and RHESSI. The emissions in the selected time intervals are dominated by non-solar background without an obvious solar signal. In the following we describe our efforts to use long-time integration RHESSI imaging to search for a hidden solar signal within these intervals.

The selected time intervals for RHESSI imaging have a duration between 17 and 55 minutes, much longer than typically used when imaging flare emissions where duration below 1 minute are generally used. However, similar duration integrations have been used for RHESSI imaging in the gamma-ray range (e.g., Hurford et al., 2006). The long time integration allows us to increase statistics for these times of very weak emissions. We reconstructed RHESSI images for those intervals with the standard CLEAN algorithm, as it appears to be the most robust algorithm for low-statistics imaging. Despite the long integration, counting statistics are still low with typically a few thousand counts per detector per interval between 3.5–6.5 keV, of which at least half are non-solar background counts (i.e. unmodulated counts). These are low numbers compared to RHESSI images with excellent statistics ($\sim$105 counts/detector), but similar to statistics available for the largest gamma-ray flare (Hurford et al., 2006), although at much higher background.

For all but interval 5, RHESSI 3.5–6.5 keV images show an extended source covering the entire active region (interval 3 and 4), or at least a large part of the active region (interval 1 and 2). For interval 5, no significant modulation was seen in the RHESSI data, indicating that for this interval the solar signal relative to the background was too weak for imaging, or even absent. RHESSI images for the other intervals are shown as blue contours in Fig. 2 overlayed on the XRT images. Each column corresponds to one of the selected time intervals, and rows correspond to the different XRT filters. As the source is extended, only the coarse subcollimators 6 to 9 with natural weighting (equal weighting resulting in an angular resolution of 61*′′* FWHM) are used in the reconstruction of the images shown in Fig. 2.

To further corroborate the extended nature of the quiescence source, we compared the imaging result of interval 4 to the microflare that occurred a few minutes earlier (see Fig. 3). The microflare comes from a compact source located in the southern part of the quiescent source and it is co-spatial with an XRT brightening. In addition to the high-resolution image of the microflare, the dashed contours show the microflare image reconstructed with the same parameters as we use for imaging the quiescent active region. This clearly confirms that the quiescent source is not related to the microflare occurring early, but it is an independent component from the entire active region.

In a second step, we used RHESSI imaging to determine the X-ray flux of the quiescent active region at 1 keV bin size. As the non-solar background signal is not modulated by the RHESSI imaging system, using imaging to determine the flux has the advantage that the non-solar background is not contributing to the reconstructed image. Hence, imaging provides a simple background subtraction without selecting a pre-flare time interval to determine the background. We obtain RHESSI images at 1 keV bin size using subcollimator 9 and then calculate the fluxes by summing all pixels within the 50% contours and multiplying by a factor of 2 to account for the wing of the clean beam (Table 2). Before we present a differential emission measure analysis in the next section, we briefly mention here the standard approach of fitting the RHESSI fluxes with a single-temperature model. Using only the lowest three energy bins (3.5–6.5 keV) gives values between 5.3–7.1 MK with EM between $\sim$3 to 13$\times 10^{45}$ cm*-3*, while including the 6.5 to 7.5 keV bin for the three intervals with good statistics gives values slightly higher 7.1–7.4 MK at lower EM ($\sim$2 to 4$\times 10^{45}$ cm*-3*). This indicates that there is a distribution of temperature that contributes to the RHESSI observations, but a single temperature approximation is apparently representing the data reasonably well. As we show in the following section, this will no longer be the case after we include EUV and SXR to the analysis.