Comparison of the Composition of ICMEs from Active Regions and Quiet-Sun Regions

Interplanetary coronal mass ejections (ICMEs; Manchester et al. 2017) are a counterpart of coronal mass ejections (CMEs; Forbes 2000; Forbes et al. 2006), which result from the eruption of magnetic flux ropes (MFRs; Chen 2011; Webb & Howard 2012) that can form prior to (Cheng et al. 2011; Patsourakos et al. 2013) and during (Song et al. 2014; Ouyang et al. 2015; Wang et al. 2017) eruptions. ICMEs can cause extreme space weather effects and affect the high-tech activities of humans (Gosling et al. 1991; Zhang et al. 2007; Xu et al. 2019). Investigating CMEs and ICMEs will help us predict space weather, and take timely precautions to reduce losses.

The composition of ICMEs, including the ionic charge states and relative abundances of heavy elements, can be obtained directly through in situ measurements. As the composition does not alter during ICME propagation in the interplanetary space (e.g., Owens 2018), ICMEs at 1 au and near the Sun have identical charge states and elemental abundances, which reflect the eruption characteristics and plasma origins of CMEs. Therefore, the in situ composition of ICMEs opens an important avenue for investigating CMEs (see the recent review by Song & Yao 2020).

According to the CME models (e.g., Lin & Forbes 2000), plasma heated by magnetic reconnections in the current sheet is transformed into the outer shell of MFRs along with the reconnected magnetic field lines (Song et al. 2015, 2016; Ye et al. 2021). As CMEs propagate outward, their electron density decreases rapidly. The ionization and recombination processes will be shut down when the density is low enough, making the ionic charge states freeze-in (Owocki et al. 1983). Therefore, the ionic charge states near 1 au can be used to infer information on electron temperature and density, as well as the ICME speed near the Sun (Landi et al. 2012); see also Shi et al. (2019) for a discussion in other contexts. For the metallic elements, such as Mg, Fe, and Si, their average values of ionic charge states, i.e., the 〈Q_Mg〉, 〈Q_Fe〉, and 〈Q_Si〉, are sensitive to the electron temperature (Lepri et al. 2001; Lepri & Zurbuchen 2004; Lepri et al. 2013). For the nonmetallic elements, such as C and O, the charge state ratios, e.g., C⁶⁺/C⁵⁺, C⁶⁺/C⁴⁺, and O⁷⁺/O⁶⁺, are used to indicate the temperature of coronal sources when freezing-in (Zhao et al. 2014, 2017a).

The elemental abundances of ICMEs are also employed to diagnose some issues related to CMEs, such as the plasma origin (Song et al. 2017), due to the first ionization potential (FIP) effect. The FIP effect is an elemental fractionation that occurs between the solar photosphere and the corona. The low-FIP elements (e.g., Mg, Fe, Si) are enhanced in abundance when they flow into the corona compared to the high-FIP elements (e.g., O). This leads to relative abundances, such as Fe/O, being higher in the corona (Laming 2004, 2009, 2015). Therefore, previous studies have explored the sources of ICME plasmas by comparing their elemental abundances with photospheric values. For example, Lepri & Rivera (2021) reported that the elemental abundances of prominences are similar to the photospheric values, which does not support the prominence material originating from the condensation of coronal plasma.

Many survey studies on the composition of ICMEs have been conducted in recent years. Zurbuchen et al. (2016) reported that the ICMEs with elevated Fe charge states have higher FIP fractionation than other ICMEs. Owens (2018) and Huang et al. (2020) found that fast ICMEs possess higher charge states and relative abundances compared to slow ones. Gu et al. (2020) and Song et al. (2021) demonstrated that the composition of ICMEs exhibits an obvious solar cycle dependence. Further, Song et al. (2022) compared the He abundance between ICMEs and solar wind, and concluded that both ICMEs and slow solar wind possess solar cycle dependence, while the fast solar wind does not show a dependence.

As both ICMEs and slow wind can originate from active regions and quiet-Sun regions, and more active regions appear in the solar atmosphere at the solar maximum, it is natural that more fractions of ICMEs and slow wind detected near 1 au originate from active regions around the solar maximum (e.g., Fu et al. 2017). If there exist obvious compositional differences between ICMEs from active regions and quiet-Sun regions, the solar cycle dependence of the composition of ICME can be explained reasonably by means of their source regions. Therefore, a comparative study on the composition of ICMEs between active regions and quiet-Sun regions is necessary. That is the major motivation for us to conduct this study. This paper is organized as follows. In Section 2, we introduce the instruments used in this study. In Section 3, the observations and results are displayed, which is followed by a summary and discussion in the final section.

The in situ compositional data of ICMEs used in this study are provided by the Solar Wind Ion Composition Spectrometer (SWICS; Gloeckler et al. 1998) aboard the Advanced Composition Explorer (ACE) located at the Lagrangian L1 point between the Sun and Earth. SWICS performs electrostatic selection, time-of-flight analysis, and total energy measurements for each solar wind ion. Through this triple combination measurement, SWICS can determine the mass, charge, and energy of solar wind ions of He, C, N, O, Ne, Mg, Si, S, and Fe, in the energy-per-charge range of 0.49–60 keV e⁻¹. The SWICS data used in this paper have been recalibrated with an improved algorithm (Shearer et al. 2014), which resolves ion species with greater accuracy and includes charge state distributions with uncertainties less than 25%. The temporal resolutions are 1 and 2 hr for ionic charge states and elemental abundances, respectively.

The corresponding CMEs and their source regions are examined through the Large Angle and Spectrometric Coronagraph (LASCO; Brueckner et al. 1995) and the Extreme Ultraviolet Imaging Telescope (EIT; Delaboudinière et al. 1995) on board the Solar and Heliospheric Observatory. The soft X-ray (1–8 Å) fluxes of associated flares are from the Geostationary Operational Environment Satellite (GOES), which provides the integrated X-ray emission of the whole solar disk.

3.1. ICMEs and Classifications

3.2. Comparison between ICMEs from Different Regions

Figure 1 shows histograms of ionic charge states for ICMEs from active regions (red) and quiet-Sun regions (blue). In panels (a)–(f), the 〈Q_Mg〉, 〈Q_Fe〉, 〈Q_Si〉, C⁶⁺/C⁴⁺, C⁶⁺/C⁵⁺, and O⁷⁺/O⁶⁺ are displayed sequentially. Based on all of the corresponding data points we calculate the average values for each parameter, which are 9.424 (9.277), 12.510 (11.531), 10.279 (10.006), 9.630 (5.401), 2.055 (1.302), and 0.956 (0.606) correspondingly for ICMEs from active regions (quiet-Sun regions). These values are also marked with arrows in each panel. Overall, all of the ionic charge states are higher in ICMEs from active regions compared to those from quiet-Sun regions. This agrees with the expectation as the temperatures of reconnection regions are higher for active-region CMEs.

Zoom In Zoom Out Reset image size

The histograms of the charge state of metallic elements (〈Q_Mg〉, 〈Q_Fe〉, and 〈Q_Si〉) show that they are obviously different for the two groups of ICMEs. The active-region ICMEs contain higher charge states. For example, the active-region ICMEs include a fraction of Fe ions with charge states higher than 16+, while the quiet-region ICMEs do not at all. For the nonmetallic elements, the average values of C⁶⁺/C⁴⁺, C⁶⁺/C⁵⁺, and O⁷⁺/O⁶⁺ within ICMEs from active regions are obviously higher than those from quiet-Sun regions by 78.3%, 57.8%, and 57.8%, respectively.

Figure 2 shows histograms of relative elemental abundances for ICMEs from active regions (red) and quiet-Sun regions (blue). The six panels display Mg/O, Fe/O, Si/O, C/O, Ne/O, and He/O sequentially, with the corresponding average values being displayed in each panel. It is obvious that the elemental abundances of active-region ICMEs contain a substantial fraction of higher abundances, and the average values of each parameter for active-region ICMEs are higher than those of the quiet-region ICMEs by 59.8% (from 0.184–0.294), 61.1% (from 0.185–0.298), 44.1% (from 0.179–0.258), 10.5% (from 0.486–0.537), 56.9% (from 0.167–0.262), and 71.7% (from 78.418–134.677), respectively. Note that large overlaps exist between the composition histograms of active-region and quiet-region ICMEs, although there are significant differences.

Zoom In Zoom Out Reset image size

3.3. Comparison between ICMEs with Different Flare Classes

The comparison between ICMEs from different regions clearly demonstrates that the active-region ICMEs have elevated ionic charge states and enriched elemental abundances, which should correlate with the high temperatures of the reconnection regions and magnetic structures of active regions. To further confirm this speculation, we examine the correlations between the composition of ICMEs and flare classes. As mentioned, we find the associated flares for 72 active-region ICMEs, including 21 (29, 20) events associated with C-class (M-class, X-class) flares, except for two ICMEs associated with B-class flares, which are not considered here due to their limited number.

Figure 3 presents histograms of 〈Q_Mg〉, 〈Q_Fe〉, 〈Q_Si〉, C⁶⁺/C⁴⁺, C⁶⁺/C⁵⁺, and O⁷⁺/O⁶⁺ for ICMEs associated with C-class (blue), M-class (orange), and X-class (red) flares. The arrows and numbers in each panel represent the corresponding mean values. To demonstrate the variations between ICMEs with weak and strong flares, we focus on the values of C-class and X-class flares. The average values for C-class flares are 9.323, 11.951, 10.062, 8.217, 1.832, and 0.679, respectively, while for X-class flares they are 9.570, 13.296, 10.557, 11.988, 2.322, and 1.319, respectively. It is clear that the ionic charge states of ICMEs associated with X-class flares are higher than those associated with C-class flares, in agreement with our expectation.

Zoom In Zoom Out Reset image size

Figure 4 displays histograms of Mg/O, Fe/O, Si/O, C/O, Ne/O, and He/O for ICMEs associated with C-class (blue), M-class (orange), and X-class (red) flares. Also, the arrows and numbers represent the corresponding means in each panel. The average values of each parameter associated with C-class (X-class) flares are 0.255 (0.336), 0.245 (0.320), 0.236 (0.276), 0.534 (0.536), 0.199 (0.388), and 122.649 (147.362), respectively. Again it is clear that the ICMEs associated with X-class flares possess higher elemental abundances compared to those associated with C-class flares. The statistical results shown in Figures 3 and 4 support that ICMEs associated with stronger flares possess more elevated ionic charge states and enriched elemental abundances. As both the M-class and X-class flares belong to strong events, the histograms of their composition are relatively similar.

Zoom In Zoom Out Reset image size

In addition, we perform a t-test using the Python routine stats.ttest_ind to examine whether the average values of the two groups are significantly different. The function runs the independent sample t-test and outputs a p-value and the test statistic. The p-value ranges from 0.0–1.0, and a small value (typically less than 0.05 or 0.01) indicates that the averages of the two groups are truly different. The t-test is conducted on all the compositional parameters as shown in Figures 1–4, i.e., between ICMEs from active regions and quiet-Sun regions, as well as ICMEs associated with C-class and X-class flares. The results show that almost all of the p-values are less than 0.01 except for that of the C/O of ICMEs associated with C-class and X-class flares, whose p-value is 0.928. This illustrates that most averages of the compositional parameters are significantly different at the 99% confidence level.

Based on the ICME catalog available at the ACE science center (Richardson & Cane 2010) and the CME catalog at the CDAW (Gopalswamy et al. 2009), we manually found the source regions for 96 ICMEs from 1998–2011, including 82 from active regions, 11 from quiet-Sun regions, and three from the intermediate regions. The comparison of the composition of active-region and quiet-region ICMEs showed that the events from active regions possess elevated ionic charge states and enriched elemental abundances compared to those from quiet-Sun regions. For the 82 active-region ICMEs, we found the corresponding flares for 72 of them, including 21 (29, 20) events associated with C-class (M-class, X-class) flares, except for two events associated with B-class flares. The comparison between ICMEs with different flare classes demonstrated that ICMEs associated with X-class flares exhibited higher charge states and elemental abundances (except the C/O) compared to C-class flares.

Our survey studies support that the elevated charge states within ICMEs mainly result from the higher temperatures of reconnection regions during solar eruptions (Lepri & Zurbuchen 2004; Song et al. 2016), and agree with the expectations of popular CME models (Mikic & Linker 1994; Antiochos et al. 1999; Lin & Forbes 2000; Moore et al. 2001; Török et al. 2004; Kliem & Török 2006; Chen 2008). Our studies also support that the elemental abundances within ICMEs depend on both the magnetic structures of source regions and the X-ray features of associated flares, which are mainly correlated with the FIP effect (Laming 2004, 2015) and photoionization (Shemi 1991; Schmelz 1993), respectively. Further discussion can be found in Song et al. (2021) and Song et al. (2022).

As mentioned, previous studies have shown that the composition of ICMEs (Gu et al. 2020; Song et al. 2021, 2022) and slow solar wind (Aellig et al. 2001; Kasper et al. 2007; McIntosh et al. 2011; Kasper et al. 2012; Lepri et al. 2013; Alterman & Kasper 2019; Alterman et al. 2021) have an obvious solar cycle dependence. The solar cycle is well represented by the sunspot number, and the number of active regions increases from the solar minimum to the maximum. Therefore, it is natural that more CMEs occur in active regions around the solar maximum, which leads to more fractions of ICMEs detected near 1 au originating from active regions and are associated with stronger flares. This might answer why the composition of ICMEs measured near 1 au exhibits a solar cycle dependence (Song et al. 2021, 2022).

The situation is similar for slow solar wind, which can emanate from both active regions and quiet-Sun regions (e.g., Zhao et al. 2017a). More fractions of slow wind emanate from active (quiet-Sun) regions around the solar maximum (minimum; e.g., Fu et al. 2017); thus, the composition of slow wind measured near 1 au also exhibits the solar cycle dependence (Lepri et al. 2013; Song et al. 2022). On the contrary, the fast solar wind always originates from coronal holes, which leads to their composition, such as He abundance, detected near 1 au not exhibiting variation as obvious as that of the slow wind (Alterman & Kasper 2019; Alterman et al. 2021; Song et al. 2022) and ICMEs (Song et al. 2022) during the solar cycle.

The FIP effect is the main feature of the elemental abundance of the slow wind and ICMEs, while the gravitational settling (Raymond et al. 1997; Lenz et al. 1998; Raymond et al.1998; Raymond 1999) can modify signatures of the FIP effect in both the slow wind (Weberg et al. 2012) and ICMEs (Rivera et al. 2022). Rivera et al. (2022) reported that the ICMEs between 1998 and 2011 exhibited some gravitational settling effects in ∼33% of the ICME periods. They also found that the effect is most prominent during the solar minimum. The gravitational settling induces mass-proportional depletion of elements in the corona, and leads to the decrease of relative abundances such as Fe/O. This could also be a factor responsible for the solar cycle dependence of the relative abundances of partial heavy elements (Song et al. 2021).

The histograms of Mg/O, Fe/O, and Si/O shown in Figure 2 support the idea of strong FIP enhancement in active regions, while the Ne/O and He/O distributions do not as the FIPs of Ne (21.56 eV) and He (24.58 eV) are higher than that of O (13.61 eV). Song et al. (2021) suggested that the enhanced Ne/O in ICMEs from active regions could result from the photoionization effect of flares (Shemi 1991; Schmelz 1993). The pre-flare soft X-ray can create a slab-like region of nonthermal ionization ratios at the chromosphere base (Shemi 1991). As the photoionization cross-section ratio of Ne and O is 9:4 (Yeh & Lindau 1985) and the photoionized O can recombine through the charge transfer reaction with the neutral H, the Ne ions are mixed with the neutral O in the slab region and selected with the thermally ionized low-FIP elements for preferential transfer to the solar corona and ICMEs (Song et al. 2021).

Besides, there exist bare ion anomalies in both slow wind (Zhao et al. 2017b; Raymond et al. 2022) and ICMEs (Kocher et al. 2017), in which C⁶⁺ and other fully stripped ions are unusually low and they are called outlier wind or ICMEs. Rivera et al. (2021) reported that Ne/O is enhanced in the outlier ICMEs, which can also contribute to the high Ne/O of ICMEs from active regions. The number of outlier events is positively correlated with the solar cycle, so it could be another factor responsible for the solar cycle dependence of the relative abundances of some elements (Song et al. 2021).

However, both the photoionization effect and outlier can not explain the high He/O in active-region ICMEs because O has a larger photoionization cross section than He (Yeh & Lindau 1985) and the unusually low He²⁺ density of outlier ICMEs reduces the He/O (not shown). Since the mass of O is heavier than that of He, the gravitational settling might be adopted to explain the high He/O in active regions. More studies are necessary to clarify this speculation.

We thank the referee (Dr. J. Raymond) for the helpful comments and suggestions. Hongqiang Song thanks Drs. Xin Cheng and Bo Li for their discussions. We acknowledge the use of the ICME catalog at the ACE science center (Richardson & Cane 2010) and the CME catalog at CDAW (Gopalswamy et al. 2009). All the SWICS data are downloaded from the ACE science center. This work is supported by the NSFC grants U2031109, 11973031, and 12073042, and the National Key R&D Program of China 2022YFF0503003 (2022YFF0503000). H.S. is also supported by the CAS grant XDA-17040507.