Suzaku Detection of Solar Wind Charge Exchange Emission from a Variety of Highly-ionized Ions in an Interplanetary Coronal Mass Ejection
Kazunori Asakura, Hironori Matsumoto, Koki Okazaki, Tomokage Yoneyama, Hirofumi Noda, Kiyoshi Hayashida, Hiroshi Tsunemi, Hiroshi Nakajima, Satoru Katsuda, Daiki Ishi, Yuichiro Ezoe
aa r X i v : . [ a s t r o - ph . S R ] F e b Suzaku Detection of Solar Wind ChargeExchange Emission from a Variety ofHighly-ionized Ions in an Interplanetary CoronalMass Ejection
Kazunori
ASAKURA , Hironori MATSUMOTO , Koki
OKAZAKI , Tomokage YONEYAMA , Hirofumi NODA , Kiyoshi
HAYASHIDA , Hiroshi
TSUNEMI ,Hiroshi NAKAJIMA , Satoru KATSUDA , Daiki ISHI and Yuichiro EZOE Department of Earth and Space Science, Graduate School of Science, Osaka University,1-1 Machikaneyama, Toyonaka, Osaka 560-0043, Japan Project Research Center for Fundamental Sciences, Graduate School of Science, OsakaUniversity, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043, Japan Japan Aerospace Exploration Agency, Institute of Space and Astronautical Science, 3-1-1Yoshinodai, Chuo-ku,Sagamihara, Kanagawa 252-5210, Japan College of Science and Engineering, Kanto Gakuin University, 1-50-1 Mutsuura Higashi,Kanazawa-ku, Yokohama, Kanagawa 236-8501, Japan Graduate School of Science and Engineering, Saitama University, 255 Shimo-Ohkubo,Sakura, Saitama 338-8570, Japan Department of Physics, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji, Tokyo192-0397, Japan ∗ E-mail: [email protected]
Received ; Accepted
Abstract
X-ray emission generated through solar-wind charge exchange (SWCX) is known to contam-inate X-ray observation data, the amount of which is often significant or even dominant, par-ticularly in the soft X-ray band, when the main target is comparatively weak diffuse sources, epending on the space weather during the observation. In particular, SWCX events causedby interplanetary coronal mass ejections (ICMEs) tend to be spectrally rich and to providecritical information about the metal abundance in the ICME plasma. We analyzed the SN1006background data observed with Suzaku on 2005 September 11 shortly after an X6-class so-lar flare, signatures of which were separately detected together with an associated ICME.We found that the data include emission lines from a variety of highly ionized ions generatedthrough SWCX. The relative abundances of the detected ions were found to be consistent withthose in past ICME-driven SWCX events. Thus, we conclude that this event was ICME-driven.In addition, we detected a sulfur XVI line for the first time as one from the SWCX emission,the fact of which suggests that it is the most spectrally-rich SWCX event ever observed. Wesuggest that observations of ICME-driven SWCX events can provide a unique probe to studythe population of highly-ionized ions in the plasma, which is difficult to measure in currently-available in-situ observations.
Key words:
X-rays: diffuse background — solar wind — solar–terrestrial relations — Sun: coronal massejections (CMEs) — Sun: flares —
Coronal mass ejections (CMEs) are the most energetic eruptive phenomena in our solar sys-tem. The occurrence rate of CMEs depends on the solar activity. Notably, there is a linearcorrelation between the occurrence rate and sunspot number (Webb & Howard 1994). ACME transfers a large amount of plasma (typically 10 –10 g) from the low corona to the so-lar system (Colaninno & Vourlidas 2009). The CMEs which are directed toward the Earth areparticularly called “halo CMEs”, which were named after coronagraph images. The ejectedplasma of the halo CME tends to hit the Earth and is often observed as a disturbance of solarwind (interplanetary CME, called ICME). Whereas a variety of properties of the CMEs andICMEs have been identified (reviewed by, e.g., Chen 2011; Webb & Howard 2012), precisemechanisms of their initiation and energy-release procedure are not fully understood.Since the first detection of a CME in 1971 (Tousey 1973), thousands of CMEs havebeen detected, mainly in coronagraph observations. In addition to imaging observations in-cluding coronagraph ones, recently-developed ultraviolet spectrometers enable us to obtain2ew aspects of CMEs, providing essential diagnostic information about the plasma compo-nents. Alternatively, information of ICMEs can be obtained in-situ by spacecrafts. Arrival ofan ICME appears as various forms of observable signatures (see the review by Zurbuchen &Richardson 2006). In some cases, ICMEs have enhanced magnetic structures (termed “mag-netic cloud” in Burlaga et al. 1981). The magnetic clouds have low proton temperaturesand plasma beta (the ratio of the plasma pressure to the magnetic pressure), which are alsoindicators of the arrival of the ICMEs.ICMEs have also drawn a significant attention in X-ray astronomical observations be-cause propagating plasma causes geomagnetic storms, which affect observations with X-raysatellites. In addition to the geomagnetic storms, highly-ionized ions in the plasma are thesources of additional X-rays during observations of celestial objects. These additional X-raysare ascribed to so-called solar wind charge exchange (SWCX), i.e. , the phenomenon wherebyan electron in a neutral atom is transferred to a highly-ionized ion in solar wind during theircollisions. When an electron in an excited state falls back to the ground state, the energy isreleased as one or more photons in the extreme ultraviolet to soft X-ray range. This mecha-nism was first suggested by Cravens (1997) to explain the line emission spectrum discoveredin observations of C/Hyakutake 1996 B2 (Lisse et al. 1996). Freyberg (1998) and Cox (1998)suggested that solar wind could produce X-rays in the exosphere of the Earth (geocoro-nal SWCX) and in the heliosphere (heliospheric SWCX) and that the mysterious X-ray timevariation detected during the ROSAT All-Sky Survey, which was reported by Snowden et al.(1994) and dubbed as Long-Term Enhancement (LTE), might have originated in the SWCX.Since solar wind, which continually blows into the heliosphere, always interacts more orless with neutral atoms in the solar system, the SWCX is now recognized as a persistentbackground component in observations of X-ray diffuse sources.To date, plenty of SWCX detections with Chandra, XMM-Newton, and Suzaku havebeen reported (e.g., Wargelin et al. 2004; Snowden et al. 2004; Fujimoto et al. 2007; for moredetails see Kuntz 2019). Systematic studies of the time variations of the soft X-ray flux havebeen conducted with large archival datasets of XMM-Newton and Suzaku (Carter & Sembay2008; Carter et al. 2011; Ishi et al. 2017). Roughly a hundred observational datasets in each ofthe XMM-Newton and Suzaku archives have been found to show time variations caused bythe geocoronal SWCX. This implies that the data that contain time-variable SWCX compo-nents are not rare, hence the significance of the study of the SWCX to handle X-ray observa-tional data correctly. While an increasing number of the samples of the SWCX are availableby now, it still remains difficult to construct a comprehensive theoretical model that de-3cribes the observed properties of SWCX observations, such as times series and overall fluxlevels, for arbitrary spacecraft look directions and epochs (Kuntz 2019). One of the majordifficulties stems from the fact that the properties of ions in solar wind vary with both solarcycle and solar latitude; the intensity and fluctuation of SWCX emission are highly depen-dent on the viewing point. However, the global structure of the solar wind, especially ata high solar latitude, cannot be clarified with current in-situ measurements. Further X-rayobservation in a variety of directions is the only way to refine the entire model including thehigh solar latitude so far.In addition to the evaluation of the SWCX model, SWCX events associated withICMEs also provide us with ionic information of the plasma from various emission lines.The most spectrally-rich case of the SWCX events reported so far was that by Carter et al.(2010), where they made systematic study, using the XMM-Newton archive, and made un-equivocal detections of the lines from various ions including highly ionized neon, magne-sium, and silicon. Their work is a good precedent that the components of ICME plasmacan be identified indirectly with X-ray observations. One of the notable characteristics ofthe event is that the source ICME was accompanied with an X-class flare. Reinard (2005)suggested a positive correlation between the solar flare magnitude and ionization state ofthe CME plasma. Providing this correlation is correct, the ICMEs associated with largeflares should yield spectrally-rich SWCX events, which have plenty of information aboutthe ICME plasma. However, the opportunities to detect such spectrally-rich events rarelyarise because X-class flares are rare in the first place and because not all the X-class flaresare associated with CMEs. Thus, the SWCX events caused by ICMEs associated with X-classflares are valuable samples.In this paper, we report a newly discovered SWCX emission from the SN1006 back-ground data observed by Suzaku, which is likely to be associated with an ICME accom-panied with an X-class flare. This event could not be observed with XMM-Newton un-fortunately, due to the intense solar activity. Conventionally, SWCX emission can be simplyidentified as a temporal excess from the stable component in the observed X-ray light-curve.However, this simple method is not applicable to the dataset like the said data, where no X-ray time variation is observed within the observation period. Our strategy is to compare thedata with another data obtained in the same region at a different observation epoch to sub-tract the stable component and identify a potential excess. We summarize the observationsand our data analysis flow in §2 and §3, respectively. Discussion about the correlation witha CME and significance of this detection is given in §4.4 able 1. Summary of the observational parameters.
Target name SN 1006 SW BGObservation ID 100019040 100019060Observation start (UT) 2005-09-11 23:59 2006-01-26 17:16Effective exposure (ks) 25.2 17.0Target coordinates ( α , δ ) (J2000) (224 ◦ .6550, − ◦ .4005) (224 ◦ .6468, − ◦ .4025) We used the Suzaku data from the region “SN1006 background”, located at ( α , δ ) J2000.0 =( ◦ , − ◦ ) . The primary purpose of the original observations was to take datasetserved as the background for the nearby main target SN1006. This region was observedtwice by Suzaku, on 2005 September 11 and 2006 January 26. The average normalized point-ing vectors in the Geocentric Solar Ecliptic (GSE) coordinates were (0.7704, 0.8259, − − − Shortly before the first observation (i.e., in early September 2005), several successive X-classsolar flares occurred although it was close to solar-cycle minimum. In particular, a solar flareon 2005 September 7 was huge, classified as the X17 class. Subsequently, X6- and X2-classflares erupted on September 9 and 10. These X-class flares entailed CMEs, as reported inWang et al. (2006). We plot the time-series of some selected parameters of the solar wind5
Fig. 1.
Line of sight directions in the GSE X-Y and X-Z planes during each Suzaku observation. Dashed and solid curves show the positions of the magne-topause and the bow shock at the end of each observational period, respectively. during the Suzaku observations in figure 2: the solar X-ray flux in the 0.1–0.8 nm band ob-served by Geostationary Operational Environmental Satellite (GOES), solar proton flux, in-terplanetary magnetic field (IMF), proton temperature, plasma beta measured by WIND ,and Dst Index provided by the World Data Center for Geomagnetism, Kyoto, Japan. Graypoints in the fourth panel of figure 2 indicate the temperature T exp expected from the solarwind velocity, calculated with the empirical formula in Lopez (1987).At the beginning of day of year (DOY) 254 in 2005, the plasma beta decreased andmeasured proton temperature dropped below T exp for a while, which followed discontin-uous rises of the proton flux and proton temperature at DOY 254.0 (indicated by the redshaded region in figure 2). This phenomenon of a sudden drop of the plasma beta is a well- We obtained the data from OMNIWeb; https://omniweb.gsfc.nasa.gov
005 September 07-12 2006 January 22-26
X17 Flare X6 Flare X2 Flare
Fig. 2.
From the top panel, solar X-ray light curve ( – ), solar proton flux, the IMF, proton temperature, plasma beta, and Dst Index. Gray points inthe fourth panel indicate the temperature T exp expected from the solar wind velocity, calculated with the empirical formula in Lopez (1987). The orange areasindicate the periods of the Suzaku observations. The plotted points in the first and other panels are the average values over 5-min and 1-hour periods,respectively. Red shaded region shows the period of a lower plasma beta, starting on DOY 254 in 2005. known indicator of a front of an ICME passing the observer’s location; the enhancements ofthe proton flux and temperature are due to a CME-driven front shock, whereas the low pro-ton temperature and low plasma beta are caused by a magnetic cloud. This CME event wassuspected to be caused by the X6 flare on September 9, which propagated from the sun tothe L1 point (where the GOES and WIND satellites stayed) with a velocity of ∼ − (Wang et al. 2006). The mass of the CME is ∼ × g, which is ranked the most massivehalo CME in the SOHO LASCO CME catalog (Gopalswamy et al. 2009). From these facts,we conclude that the Suzaku observation on 2005 September 11 was affected by the CME,whereas that on 2006 January 26 was not. Hereafter, we refer to the first and second Suzakuobservation periods as the active and stable periods, respectively. https://cdaw.gsfc.nasa.gov/CME_list/ ig. 3. XIS1 – images in the (a) active and stable periods. The color scale represents counts per pixel on a linear scale. The green circles show theregion where we extracted the X-ray events (we excluded the hatched regions to get rid of the few point sources). We analyzed the data obtained with the XISs in both the active and stable periods. In our datareduction and analysis, we utilized the software included in the HEAsoft package version6.25. We used the 3 × × ; processing script version is 3.0.22.43). Hereafter,quoted errors refer to 90% uncertainties unless otherwise noted. We examined the data of the BI CCD (XIS1), which is more sensitive than the FI CCDs (XIS0,2, and 3) in the soft ( < active period was found to be clearly brighter than that of the stable period. Therefore, some additional X-ray emission must exist in the active period. We ex-tracted the X-ray events from a circular region with a radius of 8.5 arcmin, excluding regionsencompassing a few point sources (hatched regions in figure 3). We confirmed the same X-ray enhancements in the FI CCDs (XIS0, XIS2, and XIS3). Hence, the X-ray enhancementwas not instrumental. https://heasarc.gsfc.nasa.gov/docs/suzaku/processing/criteria_xis.html ig. 4. XIS1 480 s-bin light curves in the (a) active and (b) stable periods in the (black) – and (red) – bands. Orange and gray shadedregions show the “flare” and a few rapid-rise periods, respectively, the data from which are excluded in the analyses. We extracted 480 s-bin light curves in the 0.4–2.0 keV and 2.0–10.0 keV bands for each XIS in-dividually, and plot those of the XIS1 in figure 4. The 2.0–10.0 keV light curves of any of theXISs showed a distinctive X-ray enhancement between DOYs 255.35 and 255.40 (indicatedwith the orange-shaded area in figure 4). The timing coincided with a known M6-class solarflare (we can confirm the flare in the top panel of figure 2). Thus, we excluded the data inthis time region in the following analyses. The 2.0–10.0 keV light-curves of the XIS1 and 2showed also a couple of sharp spikes with a shorter time scale than that of the typical solarflare. However, the spikes did not appear in the XIS0 or XIS3 light curves. The inconsistencyamong the detectors implied that the spikes originated in something instrumental. To iden-tify quantitatively the time regions of the instrumental spikes, we calculated the averageand the standard deviation of the count rate in the 2.0–10.0 keV band, excluding the above-mentioned solar-flare period. Then, we selected the time bins in the light curves the countrate of which were out of the 90% confidence interval as the anomaly bins and excludedthem in the following analyses. The excluded time bins are indicated with gray-shaded ar-eas in figure 4.We found that the average count rate of the 0.4–2.0 keV light curves in the active pe-riod was apparently nearly twice as high as that in the stable period (figure 4). We can alsoconfirm the slight increase of the count rate of the 2.0–10.0 keV light curves, even excludingthe solar-flare period (a possible origin of the increase are described in §3.3).9 .3 Spectral Fitting
We extracted X-ray events from each XIS data and made BI (XIS1) and FI (XIS0 + XIS2 + XIS3)spectra in each observation period. The FI and BI spectra are binned with minima of 50 and30 counts, respectively. Then we subtracted from each of the FI and BI spectra a simulatednon X-ray background (NXB) spectrum, which was created with xisnxbgen (Tawa et al.2008). We made the redistribution matrix files (RMFs), using xisrmfgen , and the auxiliaryresponse files (ARFs), using xissimarfgen (Ishisaki et al. 2007) under the assumption thatthe emission region is a circular region with a radius of 20 arcmin, given that the X-rayemission from SN1006 background area appeared uniform. For spectral fitting in this work,we used
XSPEC package (version 12.10.1). stable period
We analyzed the BI and FI spectra in the stable period with model fitting. In general, theNXB-subtracted X-ray spectrum of blank sky consists mainly of three components: theGalactic Halo (GH), the Local Hot Bubble (LHB), and the cosmic X-ray background (CXB)(e.g., Kushino et al. 2002). The GH and LHB are thin thermal plasma (with typical tem-peratures of kT ∼ ∼ − cm − keV − str − at 1 keV, which is consistent with the past result obtained with ASCA (Kushino et al. 2002).Then, we extended the energy ranges to 0.3–5.0 keV and 0.4–5.0 keV for the BI and FI spec-tra, respectively, for the further analysis. We adopted the two thin-thermal models and apower-law ( phabs*(vapec+powerlaw)+vapec in XSPEC ). The parameters for the power-lawcomponent were fixed to the best-fit values of the result described above in the model-fitting. We note that the CXB and GH are affected by the Galactic hydrogen absorptionwhile the LHB is not. In our observed regions, the Galactic absorption column density wasn H = × cm − according to nh in the HEASoft package, which we adopted. Weadopted the solar abundance table given by Anders & Grevesse (1989); the abundances ofcarbon, nitrogen, neon and iron were set to be free while those of the other elements werefixed to unity in our model fitting. All the abundance parameters were linked between the10 " ! "$"!"$!! %& ’ ( ) *+ ,- . / &1% / ! ! / - ! ! !"$6 7"$6!!$677$6 ’ ) & GHLHBCXB
Fig. 5.
Spectra obtained with the (black) BI-CCD and (red) FI CCDs in the stable period. Solid lines show the best-fit models of the (dashed-dotted lines) GH,(dotted lines) LHB, and (dashed lines) CXB components. two thermal plasma models. Figure 5 shows the spectra and resultant best-fit models andtable 2 summarizes the fitting results. The fitting was satisfactory with a reduced chi-squareof 1.06. active period
We then analyzed the spectra in the active period to evaluate quantitatively the differencefrom the spectra in the stable period (hereafter, the spectra and best-fit model of the latterare referred to as the 2006 spectra and model, respectively). Since these event data wereextracted from an identical celestial region, we can get rid of the ambiguity due to the dif-ference of the observational directions in the GSE coordinates. Following the method forthe model fitting of the 2006 spectra, we first fitted the spectra of the active period in the2.0–5.0 keV only with a power-law of which the parameters were fixed to the same as thoseof the 2006 model. We found that whereas the best-fit parameters were consistent with thoseof the 2006 model with the FI-CCD spectrum, those of the BI-CCD spectrum showed an ex-cess over the 2006 model. A similar excess in the BI-CCD spectrum was also reported inthe past SWCX studies with Suzaku (e.g., Fujimoto et al. 2007; Ezoe et al. 2011). Since theexcess is only seen in the BI spectrum, it is probably due to particle backgrounds such as11oft protons, as discussed in Ezoe et al. (2011). Regarding spectral fitting, Fujimoto et al.(2007) dealt with this feature by adding an unabsorbed power-law to the blank-sky model.Accordingly, we added an unabsorbed power-law to the 2006 model in order to address thediscrepancy. Figure 6 shows the BI- and FI-CCD spectra and the 2006 model plus the addi-tional power-law. Many Gaussian-like residuals, mainly below 2keV, are still clearly visible,after including the additional power-law component.Then, we added Gaussian models one by one to the spectral model to eliminate thesignificant residuals and repeated model-fitting. In the fittings, the center energy and nor-malization of each Gaussian were allowed to vary but were set to be common ( i.e. , linked) be-tween the BI- and FI-CCD spectra. Consequently, we obtained an acceptable fit by adding 17Gaussian models. Figure 7 shows the best-fit model with individual components. Residualsthat are still seen in figure 7 may be due to systematic uncertainties of the line profiles. Theparameters of the additional power-law and Gaussians are summarized in tables 3 and 4. Itshould be noted that the additional power-law component is response-folded in our analy-sis. On the other hand, the soft proton contamination in the XMM-Newton observations canbe approximated by a broken power-law model which is not folded through the detectorresponse (Kuntz & Snowden 2008). We also confirm that our results shown in table 4 remainalmost unchanged (even for the C lines the decrease of the normalization is less than 30%),even if we substitute the response-unfolded power-law for the response-folded one. Thecenter energy of many of the Gaussians was found to be consistent with the energies of thecharacteristic X-rays from H-like or He-like ions of one of the relatively abundant elements.Notably, we identified the 459 eV line (C VI
4p to 1s), which is one of the main features of theSWCX (Fujimoto et al. 2007). Therefore we concluded that the residuals in the fitting withthe 2006 model were attributed to the SWCX emission.
In §3, we have evaluated the SWCX lines from several ions with 17 Gaussian models.Notably, we have found a line at around 2.62 keV, the energy of which coincides with theS
XVI Ly α line. No previous studies have reported a positive detection of a S XVI line from theSWCX emission. Our detection is a significant discovery if it is genuinely a S
XVI line. Herewe discuss how plausible it is the case. 12 " ! "$"!"$!! %& ’ ( ) *+ ,- . / &1% / ! ! / - ! ! !"$6 7!7 ’ ) & Fig. 6.
Spectra of the (black) BI CCD and (red) FI CCDs in the active period. The solid lines show the model consisting of the 2006 model and, for the BI-CCDspectrum only, an unabsorbed power-law (dashed line).
The SWCX emission is known to contain not only He α and Ly α lines but also vari-ous lines caused by other transitions. In case of the 2.62 keV line, there is a possibility thatthis line is attributed to Si XIV Ly ζ and Ly η , of which the energies are at around 2.62 keV.In order to evaluate the contribution of the characteristic cascades, we incorporate a set ofthe AtomDB Charge Exchange (ACX) models (Smith et al. 2014) instead of a set of simpleGaussians, fit the spectra with the model, and see how the residuals in figure 6 are explained,as follows. The residuals in figure 6 can be mostly attributed to the characteristic linesfrom seven elements of carbon, oxygen, neon, magnesium, silicon, sulfur, and iron (table4). Accordingly, our new model consists of, in addition to the continuum of the 2006 modeland an additional power-law, 7 vacx models (the variable abundance version of the ACXmodel), each of which represents the charge exchange emission from a separate element( n.b. , we set the other abundances to a fixed value of 0). Since we have no way to measurethe absolute hydrogen abundance in the plasma (the vacx model has no continuum), onlythe relative abundances can be obtained in principle. Therefore, we link the normalizationsof all the vacx components and fix the oxygen abundance to unity in the fitting. We perform " ! "$"!"$!! %& ’ ( ) *+ ,- . / &1% / ! ! / - ! ! !"$6 7!7 ’ ) & Fig. 7.
Same as figure 6, but the model contains additional 17 Gaussians (dotted lines). The slight offset between the Gaussians for the BI- and FI-CCDspectra are due to the difference of energy gain. model-fitting of the spectra in the active period and obtain the relative abundances of the sixelements to oxygen and ion population T z , which indicates the ion population as though itwere in collisional ionization equilibrium created by electrons at temperature T z . These ionpopulation derived by SWCX do not necessarily coincide with the real ion population; ourobservation in the X-ray band is only sensitive to highly charged ions.Table 5 lists the fitting results and Figure 8 shows the best-fit model. The relativeabundances of the elements to oxygen are found to be close to the solar abundances, ex-cept for that of carbon, which is significantly higher (carbon-rich) than the solar abundance.We conjecture that the apparent high-carbon abundance is probably due to an effect of thecontamination from the heliospheric SWCX emission. We also note that the tendency ofthe low abundances of silicon and sulfur are consistent with an other observational resultof the same solar flare with Suzaku, using Earth X-ray albedo (Katsuda et al. 2020). Thisresult, where the X-rays from the Si XIV line are fully taken into account, implies that theACX model of sulfur is needed to explain the observed spectra at a confidence level of morethan 3 σ . Hence we conclude that the detected 2.62 keV line is S XVI Ly α , which is the firstdetection of the line from the SWCX emission.14 " ! "$"!"$!! %& ’ ( ) *+ ,- . / &1% / ! ! / - ! ! !"$6 7"$6!!$677$6 ’ ) & Fig. 8.
Same as figure 7, but the model has 7 ACX models instead of Gaussians (only those for the BI spectrum are depicted). The line color distinguishesthe element of each ACX model: (green) C, (orange) O, (gray) Fe, (blue) Ne, (lime-green) Mg, (red) Si, and (magenta) S.
We have shown that Suzaku detected various kinds of the SWCX lines during the active period. Given that indicators of an ICME were observed by in-situ measurements at thesame time, this SWCX emission is likely to be associated with the ICME. Some past similarstudies examined the correlation between the observed X-ray light curves and arrival of anICME to establish their association (e.g. Ezoe et al. 2010; Ishi et al. 2019). However, sincethe observed X-rays during the observation in the active period did not show much timefluctuation, the method would yield nothing significant in our case. Instead, we evaluate theplasma metal-abundance in the same way as in a previous study of spectrally-rich ICME-driven SWCX emission (Carter et al. 2010). They took account of 33 emission lines fromcarbon, nitrogen, and oxygen, of which the emission cross sections are listed in table 2 inBodewits et al. (2007). The normalization of the line with the largest cross section in eachion species was set as a free parameter in the fitting, while those of the other lines weredetermined according to the relative emission cross-sections. In our case, we adopt the crosssection at the solar wind speed of 1000 km s − which is close to the average speed duringour observation period (915 km s − ). As for the other line parameters, we add Gaussians for15he ions at the fixed transition energies which were detected in Carter et al. (2010) and forthe S XVI at 2.62 keV. Then we fit the spectra in the active period with the model. Table 6summarizes the fitting result and figure 9 shows the best-fit spectra. We find that this modelalso can explain the observed spectra.From the fitting result, we calculate the energy flux of each ion species and plot theenergy flux ratios of the lines to the O
VIII line (654 eV) in figure 10. In the figure, we also plotfor reference the results by Carter et al. (2010) and the calculated energy flux ratio for the he-liospheric SWCX, where the same cross sections and slow equatorial solar-wind abundanceas given by Schwadron & Cravens (2000) are assumed. Our result is found to roughly agreewith that of the ICME-driven case (Carter et al. 2010) except for carbon, nitrogen, and oxy-gen. We conjecture that the discrepancies with regard to carbon, nitrogen, and oxygen aredue to a potential increase of the heliospheric SWCX component, given that the intensities ofthe heliospheric SWCX can be different between the stable and active periods. The detectionof the lines from Fe
XVII , Fe
XVIII , and Fe XX indicates the existence of highly-ionized iron; itis consistent with the characteristic of the iron state in ICMEs (summarized in Zurbuchen &Richardson 2006), whereas the stable slow solar wind barely has such ions. These featuresalso support that this event is ICME-driven.The first detection of the emission line from S XVI suggests this to be the mostspectrally-rich SWCX event to date ever reported ( n.b. , there is a possibility that the en-ergy range above 2keV was not fully investigated in some of the SWCX studies and that linefeatures were overlooked). The flare which triggered the ICME associated with the event isof the X6 class, which is one of the most intense flares in the ICME-driven SWCX samples.This fact may suggest that the solar flare magnitude is related to the ionization states in theICME, as pointed out by Reinard (2005).
From the eruption at the low corona to propagation to interplanetary space, a CME evolvesand its properties change. In particular, the CME plasma in the early phase of the evolution(up to a few R ⊙ ) undergoes a continuous heating process; signs of heating have been ob-served by Ultraviolet Coronagraph Spectrometer (UVCS) onboard SOHO, which providesvaluable information to constrain the heating model (e.g. Akmal et al. 2001, Landi et al.2010). Even though the evolution of CMEs have been studied extensively with observationsand simulations for a few decades, the heating mechanism still remains an open question.16 " ! "$"!"$!! %& ’ ( ) *+ ,- . / &1% / ! ! / - ! ! !"$6 7"$6!!$677$6 ’ ) & Fig. 9.
Same as figure 6, but the model additionally contains the same SWCX model as in Carter et al. (2010) and the S
XVI line (only those for the BI spectrumare depicted). The center energy of each Gaussian is fixed to the transition energy of each ion. The line color shows the type of the element species: (green)C, (cyan) N, (orange) O, (gray) Fe, (blue) Ne, (lime-green) Mg, (black) Al, (red) Si, and (magenta) S. S XV I S i X I V : Carter et al. 2010: This work S i X III A l X III M g X I M g X II M g X I N e X N e I X N e X N e I X F e XX F e XV III F e XV II F e XV II : Heliospheric SWCX O V III O V II N V II N V I C V I C V △●♢ Fig. 10.
Energy flux ratio of the line to the O
VIII line ( ). Circles and triangles indicate our results and those by Carter et al. (2010), respectively. Diamondsindicate the heliospheric SWCX case calculated with the slow solar wind abundance given by Schwadron & Cravens (2000). X X X Fig. 11.
Detectable charge states of C, N, O, Ne, Mg, Si, S, and Fe with the currently available observations. Gray shaded areas show the available chargestate distributions in the SWICS standard dataset. Red and black hatched areas show the ranges obtained in our SWCX observation and Gilbert et al.(2012), respectively. ter. The X-Ray Imaging and Spectroscopy Mission (XRISM), scheduled to be launched inJapanese fiscal year 2022, has a non-dispersive soft X-ray spectrometer with a high spectralresolution (the full-width at half-maximum < We analyzed the SN1006 background data observed with Suzaku on 2005 September 11,using another observation of the same field at a different epoch as the reference. We foundthat the data contained an additional soft-X-ray component compared with the referencedata, which were taken when the solar activity was stable. With spectral model fitting, we19evealed the additional component to be multiple emission lines, which can be explained asSWCX emission. The observation period coincided with the passage of an ICME associatedwith a X6 flare that had happened shortly before the Suzaku observation. The coincidence,as well as the fact the characteristics of the emission lines were consistent with those of theICME-driven SWCX events reported in the past, strongly implies that this SWCX event wasICME-driven. Notably, we detected an emission line from S
XVI with a confidence level ofmore than 3 σ for the first time as one from a SWCX, the fact of which suggests that thisis the most spectrally-rich SWCX event ever reported to date. Note that some of the ions,including highly-ionized sulfur, that we detected cannot be detected by currently-availablein-situ solar-wind monitoring due to poor statistics. We propose that the SWCX is a newtool for diagnosing the ICME plasma, the understanding of which is important not only instudying X-ray diffuse background but also understanding the coronal heating mechanism. Acknowledgments
The authors gratefully acknowledge the use of data obtained from the Suzaku and ACE satellites, NASA/GSFC’s Space Physics DataFacility’s OMNIWeb service, and SOHO LASCO CME catalog. The CME catalog is generated and maintained at the CDAW Data Centerby NASA and The Catholic University of America in cooperation with the Naval Research Laboratory. SOHO is a project of internationalcooperation between ESA and NASA. This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI GrantNos. 20J20685 (KA), 20H00175 (HM), 18J20523 (TY), 19K21884, 20H01941, 20H01947, 20KK0071 (HN), 18K18767, 19H00696, 19H01908,20H00176 (KH), 20K20935 (SK), 19J20910 (DI) and 20H00177 (YE). This work was also partly supported by the Mitsubishi FoundationResearch Grants in the Natural Sciences 201910033 (KH), Leading Initiative for Excellent Young Researchers, MEXT, Japan (SK), and TorayScience and Technology Grant (YE).
References
Akmal, A., Raymond, J. C., Vourlidas, A., Thompson, B., Ciaravella, A., Ko, Y. -K., Uzzo, M. and Wu,R. 2001, ApJ, 553, 922Anders, E. and Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197Bodewits, D., et al. 2007, A&A, 469, 1183Burlaga, L., Sittler, E., Mariani, F. and Schwenn, R. 1981, J. Geophys. Res., 86, 6673Carter, J. A. and Sembay, S. 2008, A&A, 489, 837Carter, J. A., Sembay, S. and Read, A. M. 2010, MNRAS, 402, 867Carter, J. A., Sembay, S. and Read, A. M. 2011, A&A, 527, A115Chen, P. F. 2011, Living Rev. Sol. Phys., 8, 1 olaninno, Robin C. and Vourlidas, Angelos 2009, ApJ, 698, 852Cox, D. P. 1998, in IAU Collq. 166, the local bubble and beyond, ed. Breitschwerdt, D., Freyberg, M.J., & Truemper, J., Lecture Notes in Physics, 506 (Berlin: Springer Verlag), 121Cravens, T. E. 1997, Geophysical Research Letters, 24, 105De Luca, A. and Molendi, S. 2004, A&A, 419, 837Ezoe, Y., Ebisawa, K., Yamasaki, N. Y., Mitsuda, K., Yoshitake, H., Terada, N., Miyoshi, Y. andFujimoto, R. 2010, PASJ, 62, 981Ezoe, Y., Miyoshi, Y., Yoshitake, H., Mitsuda, K., Terada, N., Oishi, S. and Ohashi, T. 2011, PASJ, 63,S691Freyberg, M. J. 1998, in IAU Collq. 166, the local bubble and beyond, ed. Breitschwerdt, D., Freyberg,M. J., & Truemper, J., Lecture Notes in Physics, 506 (Berlin: Springer Verlag), 113Fujimoto, R., et al. 2007, PASJ, 59, S133Gilbert, J. A., Lepri, S. T., Landi, E. and Zurbuchen, T. H. 2012, ApJ, 751, 20Gloeckler, G., et al. 1998, Space Sci. Rev., 86, 497Gopalswamy, N., Yashiro, S., Michalek, G., Stenborg, G., Vourlidas, A., Freeland, S. and Howard, R.2009, Earth Moon and Planets, 104, 295Gruesbeck, J. R., Lepri, S. T., Zurbuchen, T. H. and Antiochos, S. K. 2011, ApJ, 730, 103Ishi, D., Ishikawa, K., Ezoe, Y., Ohashi, T., Miyoshi, Y. and Terada, N. 2017, in The X-ray Universe2017, ed. Ness, J-U, Migliari, S. 103Ishi, D., Ishikawa, K., Numazawa, M., Miyoshi, Y., Terada, N., Mitsuda, K., Ohashi, T. and Ezoe, Y.2019, PASJ, 71, 23Ishisaki, Y., et al. 2007, PASJ, 59, S113Katsuda, S., et al. 2020, ApJ, 891, 126Koyama, K., et al. 2007, PASJ, 59, S23Kuntz, K. D. and Snowden, S. L. 2008, A&A, 478, 575Kuntz, K. D. 2019, A&AR, 27, 1Kushino, A., Ishisaki, Y., Morita, U., Yamasaki, N. Y., Ishida, M., Ohashi, T. and Ueda, Y. 2002, PASJ,54, 327Landi, E., Raymond, J. C., Miralles, M. P. and Hara, H. 2010, ApJ, 711, 75Lisse, C. M., et al. 1996, Science, 274, 205Lopez, R. E. 1987, J. Geophys. Res., 92, 11189Merka, J., Szabo, A., Slavin, J. A. and Peredo, M. 2005, J. Geophys. Res., 110, A04202Mitsuda, K., et al. 2007, PASJ, 59, S1Reinard, A. 2005, ApJ, 620, 501 ivera, Y. J., Landi, E., Lepri, S. T. and Gilbert, J. A. 2019, ApJ, 874, 164Schwadron, N.A. and Cravens, T. E. 2000, ApJ, 544, 558Serlemitsos, P. J., et al. 2007, PASJ, 59, S9Shue, J. H., et al. 1998, J. Geophys. Res., 103, 17691Smith, R. K., Foster, A. R., Edgar, R. J. and Brickhouse, N. S. 2014, ApJ, 787, 77Snowden, S. L., McCammon, D., Burrows, D. N., and Mendenhall, J. A. 1994, ApJ, 424, 714Snowden, S. L., Collier, M. R. and Kuntz, K. D. 2004, ApJ, 610, 1182Takahashi, T., et al. 2007, PASJ, 59, S35Tashiro, M., et al. 2020, Proc. SPIE, 11444, 293Tawa, N., et al. 2008, PASJ, 60, S11Tousey, R. 1973, Space Res. XIII, 2, 713Wang, Y., Xue, X., Shen, C., Ye, P., Wang, S. and Zhang, J. 2006, ApJ, 646, 625Wargelin, B. J., Markevitch, M., Juda, M., Kharchenko, V., Edgar, R. and Dalgarno, A. 2004, ApJ, 607,596Webb, D. F. and Howard, R. A. 1994, J. Geophys. Res., 99, 4201Webb, D. F. and Howard, T. A. 2012, Living Rev. Sol. Phys., 9, 3XRISM Science Team. 2020, arXiv e-prints, arXiv:2003.04962Zurbuchen, T. H. and Richardson, I. G. 2006, Space Sci. Rev., 123, 31 able 2. Results of the model fitting to the stable -periodspectra.
Model ParametersCXB (power-law) N H [ cm − ] Γ ∗ N H [ cm − ] kT [keV] 0.297 + − abundance † C 1.26 + − N 2.70 + − O 1.00 (frozen)Ne 1.26 + − Fe 0.45 + − Normalization ‡ + − LHB (vapec) § kT [keV] 0.130 + − Normalization ‡ + − χ /d.o.f 325.02 / 308 ∗ In units of photonss − cm − keV − str − at 1keV. † These values are abundances with respect to solar. ‡ In units of ( π ) − D − ( + z ) − − R n e n H dV per steradian, where D A is the angular size distance to the source (cm), and n e and n H are theelectron and hydrogen densities (cm − ), respectively. § Each abundance parameter is linked to that of the GH model.
Table 3.
Results of the model fitting to the active -period spectra.
Model Parameters2006 model Described in table 2.Additional power-law Photon Index Γ + − Normalization ∗ + − Additional 17 Gaussians Described in table 4. χ /d.o.f 524.60 / 369 ∗ In units of photonscm − s − keV − str − at 1keV. able 4. Best-fit parameters of the Gaussians and identification of the lines.
Gaussian ∗ Center Energy [eV] Normalization † f X‡ Line Identification1 374 + − + − + C VI
2p to 1s (368 eV)2 468 + − + − + − C VI
4p to 1s (459 eV)3 571 + − + − + − O VII + − + − + − O VIII
2p to 1s (653 eV)5 814 + − + − + − O VIII or Fe
XVII + − + − + − Ne IX or Fe XVIII + − + − + − Ne X
2p to 1s (1022 eV)8 1060 + − + − + − Ne IX ?9 1172 + − + − + − -10 1270 + − + − + − -11 1339 + − + − + − Mg XI
12 1468 + − + − + − Mg XII
2p to 1s (1473 eV)13 1564 + − + − + − Mg XI
14 1722 + − + − + − Mg XII
3p to 1s (1745 eV)15 1836 + − + − + − Si XIII
16 2006 + − + − + − Si XIV
2p to 1s (2006 eV)17 2616 + − + − + − S XVI
2p to 1s (2623 eV) ∗ The width of the lines are fixed to 0. † In units of photonscm − s − str − . ‡ f X is energy flux in units of 10 − ergs − cm − . able 5. Best-fit parameters with the 2006 model plus additional power-lawand seven ACX models.
Model Parameters2006 model Described in table 2.Photon Index Γ Normalization ∗ Additional power-law 1.21 + − + − abundance † Ion population T z [keV]vacx C > + − O 1 (fixed) 0.210 + − Ne 1.22 + − + − Mg 1.67 + − + − Si 0.729 + − + − S 0.619 + − > + − + − χ /d.o.f 595.56 / 389 ∗ In units of photonscm − s − keV − str − at 1keV. † These values are abundances with respect to solar. able 6. Best-fit parameters in model fitting with the fixed-center-energyGaussians.
Gaussian ∗ Center Energy [eV] Normalization † Line Identification1 299 33.6 + − C V + − C VI VI + − N VII + − O VII + − O VIII + − Fe XVII + − Fe XVII + − Fe XVIII
10 920 1.04 + − Ne IX
11 960 0.39 + − Fe XX
12 1022 2.53 + − Ne X
13 1100 0.60 + − Ne IX
14 1220 0.56 + − Ne X
15 1330 1.32 + − Mg XI
16 1470 1.31 + − Mg XII
17 1600 0.20 + − Mg XI
18 1730 0.26 + − Al XIII
19 1850 0.69 + − Si XIII
20 2000 0.30 + − Si XIV
21 2623 0.32 + − S XVI ∗ The width of the lines are fixed to 0. † In units of photonscm − s − str − ..