Three types of solar coronal rain during magnetic reconnection between open and closed magnetic structures

In two-dimensional models, magnetic reconnection occurs at an X-point, where oppositely directed magnetic field lines converge and undergo reconnection, rapidly releasing magnetic energy that is converted into kinetic and thermal energy through the acceleration or heating of charged particles (Yamada et al., 2010; Antolin et al., 2012; Ni et al., 2020). It reconfigures the topology of magnetic field, and plays an elemental role in the magnetized plasma systems throughout the universe including the planetary magnetospheres, magnetars, and accretion disks around black holes (Priest & Forbes, 2000). Numerous theoretical concepts of reconnection have been undertaken to explain various features in solar physics, such as the flares, filament eruptions, jets, and coronal mass ejections (Shibata, 1998; Lin & Forbes, 2000; Shen, 2021).

In-situ measurements are not feasible in the extremely hot solar atmosphere. Fortunately, the magnetic field pressure dominates the plasma pressure (low plasma beta) in the solar corona and the magnetic flux is frozen into the highly conductive coronal plasma (Aschwanden, 2006; Priest, 2014). The emitting plasma within coronal loops not only outlines the magnetic field’s geometry but also reflects changes in field connectivity. It enables observations through remote sensing of emission across the entire electromagnetic spectrum from radio to X-rays and gamma rays, and can serve as indirect indicators of magnetic field topology and evolution (Su et al., 2013; Li et al., 2016a).

So far, many signatures of reconnection have been reported using remote sensing observations. These include the reconnection inflows (Yokoyama et al., 2001; Lin et al., 2005; Li & Zhang, 2009a), current sheets (Ciaravella & Raymond, 2008; Liu et al., 2010; Song et al., 2012b; Li et al., 2016a, b; Cheng et al., 2018; Chen et al., 2019; Hong et al., 2019; Hou et al., 2021), reconnection outflows (Takasao et al., 2011; Tian et al., 2014; Chen et al., 2016; Li et al., 2016), plasmoid ejections (Song et al., 2012a; Kumar & Cho, 2013; Peter et al., 2019; Xue et al., 2020), loop-top hard X-ray sources (Masuda et al., 1994; Su et al., 2013), supra-arcade downflows (McKenzie, 2000; Innes et al., 2003; Li et al., 2016b; Samanta et al., 2019, 2021), cusp-shaped post-flare loops (Tsuneta et al., 1992; Yan et al., 2018), and coronal structural reconfigurations (Zhang et al., 2013; Li et al., 2014, 2018b, 2018a, 2019; Kong et al., 2018). However, reconnection events with clear structural evolution and evident complete process are still relatively rare in observations.

Over the past fifty years, dozens of extreme ultraviolet (EUV) imagers and spectrographs have been deployed into space for solar observation purposes. Significant advancements have been made in terms of spatial resolution and effective area, enhancing our ability to study the Sun. The Atmospheric Imaging Assembly (AIA; Lemen et al., 2012) onboard the Solar Dynamics Observatory (SDO; Pesnell et al., 2012) has been taking full-disk solar images in ten passbands including seven EUV passbands with a similar resolution. The Solar Upper Transition Region Imager (SUTRI), a narrow-band Ne vii 46.5 nm imager on board the Space Advanced Technology demonstration satellite (SATech-01), takes images of the plasma with a characteristic temperature of $\sim$0.5 MK (Tian, 2017; Bai et al., 2023).

In the solar atmosphere, there are numerous fascinating phenomena, e.g., flares and filament eruptions. Among them, coronal rain (CR) is a particularly well-known and frequently observed phenomenon. First discovered in the 1970s (Kawaguchi, 1970; Leroy, 1972), CR plays an essential role in the mass cycle between the hot, tenuous corona and the cool, dense chromosphere (Li et al., 2018b, 2019; Kohutova et al., 2019; Li et al., 2020, 2021a). It is observationally characterized by the formation of cool, dense, and blob-like material within minutes in the hot corona, which subsequently descends along curved, loop-like trajectories to the solar surface (Antolin & Rouppe van der Voort, 2012). CR has been extensively observed both off-limb and on-disk, with its best observations typically achieved off-limb. Here, it appears in emission in the chromospheric and transition region lines, exhibiting high contrast against the dark background (Antolin, 2020). CR is a multithermal phenomenon, emitting in the chromospheric lines such as H$\alpha$, Ca ii h, and Mg ii k, as well as in the transition region lines, e.g., Si iv, with temperatures ranging from a few thousand to several hundred thousand Kelvin. Observations provide a comprehensive view of the multithermal coronal plasma, where emission of different wavelengths is cospatial (Antolin et al., 2015). As for the observed properties, the lengths of CR blobs range from a few hundred km to 1-20 Mm, with peak values of 0.7–1 Mm (Antolin & Rouppe van der Voort, 2012), and widths range from 150 to 300 km (Antolin et al., 2015, 2023; Li et al., 2022b; Şahin et al., 2023). The material with densities of 10¹⁰–10¹¹ cm^-3 falls to the solar surface with speeds in the range from a few 10 to a few 100 km s^-1, with the peak at 80–100 km s^-1 (Schrijver, 2001; Zhang & Li, 2009; Antolin & Rouppe van der Voort, 2012; Antolin et al., 2015).

While CR is usually seen falling towards the solar surface along the legs of a loop, there are instances of upward flow on one leg followed by downward flow on the other (partial siphon flow), as well as changes of direction (upward/downward) along the same leg (Antolin et al., 2010; Li et al., 2022b; Şahin et al., 2023). A particular aspect of CR dynamics is its downward acceleration being less than the free-fall acceleration (flow along magnetic field lines). Several studies, using various imaging and spectroscopic instruments capable of capturing the complete velocity vector, not just the apparent plane-of-sky component, have reported downward acceleration of about or less than 0.1 km s^-2 (Ahn et al., 2014; Antolin & Rouppe van der Voort, 2012; Schrijver, 2001; De Groof et al., 2004; Kohutova & Verwichte, 2016; Verwichte et al., 2017). Comparing with the effective solar gravitational acceleration, this acceleration is notably smaller. By restructuring the downstream of the CR, the gas pressure may result in the observed smaller acceleration (Oliver et al., 2014). Additionally, when transverse magnetohydrodynamic (MHD) waves are present, the transverse MHD waves’ suspension force does not play a major role (Kohutova & Verwichte, 2016; Antolin & Verwichte, 2011). While the heating mechanism of loops is expected to produce sustained gas pressure changes within the loops, the formation of a dense clump in a rarefied medium like the corona generates a leading gas pressure front that propagates along the magnetic field lines, enhancing the gas pressure at its rear and rapidly counteracting the downward acceleration (Antolin et al., 2010; Kohutova & Verwichte, 2017; Oliver et al., 2014, 2016). Recent studies have also found that the strong gas pressure gradient above and below the dense group acts as an antigravity levitation force that influences the dynamics of CR (Li et al., 2022b).

Thermal instability is widely recognized as a significant mechanism for the formation of condensed material in CR and prominences/filaments (Parker, 1953; Field, 1965; Claes & Keppens, 2019; Claes et al., 2020). The distinction between the two activities is largely influenced by their local magnetic field configurations. Condensation formed in an arched loop may descend along the loop legs and manifest as CR within a short period, while condensation in a prominence may linger for hours to days, supported by magnetic field depressions (Antiochos et al., 1994; Adrover-González et al., 2021; Jenkins & Keppens, 2021). Thermal instabilities occur when small primary perturbations disrupt the density and temperature equilibrium of the homogeneous coronal plasma, which is balanced between heating and cooling (Klimchuk, 2019). It is typically assumed that chromospheric plasma evaporates due to heating at the loop footpoints, filling the loop and making it denser and hotter (Antiochos et al., 1999; Aschwanden, 2001; Karpen et al., 2001; Müller et al., 2003, 2004, 2005; Xia et al., 2011). As a result, radiative losses become more efficient, and loops begin to cool when the heat transfer from heating is insufficient to balance the radiative losses. Since the radiative loss function is negatively correlated with temperature in the coronal temperature range, a decrease in temperature further increases radiation, leading to catastrophic cooling. The decrease in local temperature in the corona also reduces pressure, causing surrounding material to converge, forming localized material concentration regions. This dense matter is effectively cooled by radiation until the plasma becomes more optically thick and the cooling rate decreases. Modern views also consider the possibility that this thermal equilibrium may never be achieved, leading to a thermal non-equilibrium cycle (Antolin, 2020).

CR is basically classified into two flavours based on its relation with the flare: the flare-driven CR, labeled as the type I CR, and the quiescent CR (Antolin, 2020; Li et al., 2021b). Additionally, the quiescent CR is sorted into two kinds according to the magnetic field structures where it forms (Li et al., 2020, 2021a). One takes place along the non-flaring active region (AR) closed loops (Schrijver, 2001; Antolin et al., 2010), labeled as the type II CR, and the other occurs along the open magnetic structures in both quiet-Sun regions (Li et al., 2018b) and ARs (Mason et al., 2019; Li et al., 2021b), labeled as the type III CR. For these three kinds of CR, different formation mechanisms have been widely investigated (Müller et al., 2003, 2004; Scullion et al., 2014, 2016; Li et al., 2018b, 2019, 2020, 2021a; Mason et al., 2019; Antolin, 2020).

The type I CR often occurs in post-flare loops (Brosius, 2003; Scullion et al., 2014, 2016; Lacatus et al., 2017). It has been suggested that the condensed material of this type of CR forms because the directed non-thermal particles and/or thermal conduction fronts produced during the flare rapidly heat the chromospheric material, causing the heated chromospheric material to evaporate to fill the post-flare loops, increasing their mass (Li & Zhang, 2009b; Tian et al., 2015; Zhang et al., 2016). At the top of the post-flare loops, thermal instability occurs when the radiative loss exceeds the heating input (Parker, 1953; Field, 1965). As a result, the thermally evaporating plasma rapidly cools and condenses. The condensation then falls toward the solar surface, under the effect of gravity, along both legs of post-flare loops as the type I CR. Here, the electron beam heating alone cannot directly produce the CR in post-flare loops (Reep et al., 2020).

For the formation of condensation in the type II CR (Müller et al., 2003, 2004), heating events believed to be concentrated at/near the loop footpoints lead to chromospheric material evaporation and direct mass ejections into the AR closed loops (Patsourakos & Klimchuk, 2006; Li et al., 2015). Consequently, the loops quickly become hotter and denser, and then rapidly cool and condense due to thermal instability within a loop in a state of thermal non-equilibrium (Antolin et al., 2010; Antolin, 2020). The condensation falls toward the solar surface along both legs of the loops as the type II CR under the effect of gravity. Moreover, along loops, the formation of the type II CR, possibly triggered by impulsive heating associated with reconnection, has also been observed (Kohutova et al., 2019). However, similar event showing the type II CR associated with reconnection is rarely reported. Unlike the type I CR, the type II CR is a (quasi-) periodic phenomenon with intervals of a few hours (Antolin et al., 2012; Auchère et al., 2018; Kohutova & Verwichte, 2016; Froment et al., 2020). This periodicity of stationary CR may be attributed to thermal non-equilibrium (Auchère et al., 2018; Froment et al., 2020).

A recent model was proposed to elucidate the generation of condensation in the quiescent CR along open field lines, i.e., the type III CR, facilitated by interchange reconnection (Li et al., 2018b). In this model, the curved higher-lying open structures descend toward the solar surface and undergo interchange reconnection with the lower-lying closed loops, leading to the creation of a magnetic dip in the former. Subsequently, the newly-formed open structures and closed loops appear and recede from the reconnection region. In the higher-lying open structures, the coronal plasma around the dip converges into the dip, leading to an increase in plasma density in the dip. This density enhancement triggers thermal instability, leading to the cooling and condensation of hot coronal plasma within the dip. A prominence is thus formed (Antiochos & Klimchuk, 1991; Yan et al., 2015; Chen et al., 2020), facilitating a speedup of the reconnection. Due to the successive reconnection, under the effect of gravity, the condensed material falls to the solar surface along the legs of the newly-reconnected closed loops and the higher-lying open structures as the type III CR (Li et al., 2018a).

The types I and III CR are generally associated with reconnection. Both of them are separately observed during the interchange reconnection between open and closed structures (e.g., Sun et al., 2016; Li et al., 2018b). However, it is still not clear under what conditions the types I and III CR form during the interchange reconnection. In this study, we utilize the SDO and SUTRI data to analyze an interchange reconnection event with a distinctive structure and clear evolution. Notably, for the first time, three types of CR are observed in the reconnection event. The conditions under which the types I and III CR form are then investigated. The observations, results, and summary and discussions are described in Sections 2, 3, and 4, respectively.

The SDO/AIA comprises a set of normal-incidence imaging telescopes that capture solar atmospheric images across ten wavelength bands, including seven EUV channels and three ultraviolet (UV) channels (Lemen et al., 2012). The AIA EUV and UV images are taken with a time cadence of 12 and 24 s, respectively, and have a spatial sampling of 0.6″ pixel^-1. Each AIA channel corresponds to plasma at different temperatures, e.g., the 94 Å channel captures plasma at $\sim$7.2 MK (Fe xviii), the 335 Å channel captures plasma at $\sim$2.5 MK (Fe xvi), the 211 Å channel captures plasma at $\sim$1.9 MK (Fe xiv), the 193 Å channel captures plasma at $\sim$1.5 MK (Fe xii), the 171 Å channel captures plasma at $\sim$0.9 MK (Fe ix), the 131 Å channel captures plasma at $\sim$0.6 MK (Fe viii) and $\sim$10 MK (Fe xxi), and the 304 Å channel captures plasma at $\sim$0.05 MK (He ii).

The scientific objective of the science and technological demonstration payload SUTRI is to establish a connection between the lower solar atmosphere and the structure of the corona (Bai et al., 2023) . It operates in a Sun-synchronous orbit at an altitude of approximately 500 km and aims to enhance our understanding of various solar activities such as flares, jets, and coronal mass ejections. SUTRI primarily focuses on the Ne vii 465 Å spectral line, which corresponds to a characteristic temperature of $\sim$0.5 MK in the solar atmosphere (Tian, 2017). With a filter width of about 30 Å, it achieves a moderate spatial resolution of $\sim$8″, allowing for full-surface solar atmospheric images.

In this study, we employ all available SDO/AIA EUV, i.e., 94, 335, 211, 193, 171, 131, and 304 Å, images, and SUTRI 465 Å images to conduct the comprehensive investigation of the interchange reconnection event on 2022 October 11. The alignment of AIA EUV and SUTRI 465 Å images is achieved using a rapid subpixel image registration method (Feng et al., 2012; Yang et al., 2015). Data from SUTRI 465 Å has been aligned by optimizing cross-correlation with AIA 304 Å images, as they show the most similar characteristic features.

Between 05:00 and 16:00 UT on October 11, 2022, a sustained interchange reconnection event occurs above the southeastern solar limb. By investigating the evolution of the region, where the event happens, in the following 10 days using AIA EUV images and line-of-sight (LOS) magnetograms of the Helioseismic and Magnetic Imager (HMI; Schou et al., 2012) onboard the SDO, we notice that the reconnection takes place in a large bipolar magnetic region with positive field, PF1, in the north and negative field, NF1, in the south. Moreover, there is another positive field, PF2, located to the south of the bipolar region. At 00:00 UT on 2022 October 13, about two days after the event, the magnetic fields PF1, NF1, and PF2 are located at the heliographic positions S15 W60, S18 W57, and S27 W60, respectively. We use SDO/AIA EUV and SUTRI 465 Å images to study the magnetic structures and their evolution in the solar corona. We find that a set of open structures, denoted as L2, move downward, while a set of closed loops, labeled as L1, move upward (Figure 1). Here, the closed loops L1 connect the negative and positive fields NF1 and PF2, while the open structures L2 root in the positive field PF1.

These structures eventually converge and transform into new sets of closed loops, denoted as L3, and open structures, labeled as L4 (Figure 1). These newly-formed structures L3 and L4 then separate from each other. Details are presented in Section 3.1. Moreover, abundant CR activities are observed in AIA 304 Å and SUTRI 465 Å images. We analyze the morphology and evolution of the CR activities and identify a total of three types of CR. This is the first time where all three types of CR have been observed in a single reconnection event; see details in Section 3.2.

In this paper, we utilize EUV images from the SDO and the SUTRI to clearly describe a process of interchange reconnection and the evolution of three types of CR. On 2022 October 11, the closed loops L1 and their nearby open structures L2 are observed beyond the southeastern solar limb. They converge and interact, with the formation of current sheet at the interface. Reconnection then occurs, and the newly-formed closed loops L3 and open structures L4 recede from the reconnection region. Before the reconnection, condensation appears in the reconnecting closed loops L1, and falls toward the solar surface as the type II CR. During the reconnection, condensation occurs in the newly-formed closed loops L3, and rains down along both loop legs as the type I CR. Moreover, condensation forms in the dips of reconnecting open structures L2, and propagates along the leg of open structures toward the solar surface as the type III CR. For all three types of CR, the imaging observations and light curves of the AIA and SUTRI EUV channels clearly show the cooling and condensation process of coronal plasma.

Schematic diagrams, which are rotated $\sim$90^∘ clockwise compared to the observations, are presented in Figure 8 to explain the interchange reconnection and its associated three types of CR. PF1 and NF1 separately represent the positive and negative fields of the bipolar region, while PF2 represents the positive field, located to the south of the bipolar region. Before the reconnection, the heating events at the footpoints of loops L1; see the green lines; lead to the chromospheric evaporation and direct mass ejections into the loops. Consequently, the loops L1 are rapidly filled with hotter and denser plasma. Due to thermal nonequilibrium, they then cool and condense quickly; see the red ellipse in Figure 8(a). The condensation falls to the solar surface along both legs of loops L1; see Figure 8(b), under the effect of gravity as the type II CR. Subsequently, the closed loops L1 and open structures L2 (pink lines in Figure 8), converge and reconnect (blue star in Figure 8(b)). The newly-formed closed loops L3 and open structures L4 (green-pink lines in Figures 8(c) and (d)), then appear and move away from the reconnection region. The thermal conduction front and/or beamed nonthermal particles, produced during the reconnection, heat the chromospheric material quickly, leading to the chromospheric evaporation and then the greater mass loading of the newly-reconnected loops L3. At the top of loops L3, thermal instability takes place. Cooling and condensation of hot plasma (red ellipse on loops L3 in Figure 8(c)), then occur rapidly. Under the effect of gravity, the condensation moves down to the solar surface along both legs of loops L3 (Figure 8(d)) as the type I CR. Here, it should be noted that, according to the simulation results of Reep et al. (2020), electron beam heating alone cannot directly result in the formation of CR in loops L3. During the reconnection, magnetic dips of open structures L2 form upon the reconnection region. Material around the dips flows into the dips. This causes an increase in plasma density in the dips. This increase in density triggers thermal instability, which leads to the cooling and condensation of hot plasma within the dips; see the red ellipse on open structures L2 in Figure 8(c). The condensation falls to the solar surface under the effect of gravity along the leg of open structures L2 as the type III CR (Figure 8(d)).

Interchange reconnection between closed and open structures is reported. The converging motions of structures L1 and L2, the appearance and disappearance of current sheet at the interface, and the subsequent formation of newly-reconnected structures L3 and L4 clearly support that interchange reconnection takes place between the closed and open structures L1 and L2. The moving speeds of structures L1 and L2 (2-3 km s^-1) toward the reconnection region and those of structures L3 and L4 (4-13 km s^-1) away from the reconnection region are consistent with those in Li et al. (2018b, 2019), but smaller than those in flares in Li et al. (2016b) and Sun et al. (2016). The width of current sheet (5.3 Mm) is consistent with those in Liu et al. (2010) and Li et al. (2016b), but larger than those in Li et al. (2016a, 2021c, 2022a) and Xue et al. (2020). The length of current sheet (13 Mm) is smaller than those in Li et al. (2016a), but larger than those in Li et al. (2021c, 2022a). The reconnection rate (0.4) is similar to those in Xue et al. (2018) and Li et al. (2021c), but large than those in Xue et al. (2020). The temperature (1.3 MK) and density (1.2$\times$10⁹ cm^-3) of current sheet are smaller than those in Li et al. (2016a, 2021c, 2022a). This indicates that less thermal energy may be converted during the reconnection.

Three types of CR are observed in the interchange reconnection. The speed and acceleration of these CR are consistent with those previously reported (see reviews in Antolin, 2020; Li et al., 2021b, and references therein). The newly-reconnected loops L3, where the type I CR forms, appear in AIA 211, 193, 171, and 131 Å, rather than AIA 94 and 335 Å, images; see Section 3.2.1. The plasma in the loops L3 is hence heated up to $\sim$1.9 MK, the characteristic temperature of AIA 211 Å channel, rather than higher temperatures. It then cools down, and appears sequentially in the lower temperature, e.g., AIA 193, 171, 131 (SUTRI 465), and 304 Å, channels. The reconnecting loops L1, where the type II CR forms, however, first appears in AIA 94 Å images; see Section 3.2.2. Compared with that in the newly-reconnected loops L3, the plasma in the reconnecting loops L1 is thus heated up to a much higher temperature of $\sim$7.2 MK, the characteristic temperature of AIA 94 Å channel. It then cools down, and appears sequentially in the lower temperature, e.g., AIA 335, 211, 193, 171, 131 (SUTRI 465), and 304 Å, channels. Comparing with those in the loops L3 and L1 where the types I and II CR form, the coronal plasma with a lower temperature of $\sim$0.9 MK, the characteristic temperature of AIA 171 Å channel, in the dips of open structures L2 above the reconnection region quickly cools and condenses, forming the type III CR; see Section 3.2.3. It is hardly heated by the reconnection, compared with that of the newly-reconnected loops L3, as the thermal conduction, produced during the reconnection, is inefficient across the field lines. Even though the formation mechanisms of three types of CR are different, the cooling times of them are comparable. For example, coronal plasma cools down from $\sim$0.9 MK (AIA 171 Å) to $\sim$0.6 MK (AIA 131 Å) in 10, 6, and 5 minutes, and then to $\sim$0.05 MK (AIA 304 Å) in 9, 21, and 13 minutes for the types I, II, and III CR, respectively.

The types I and III CR are commonly associated with reconnection (Li et al., 2018b, 2019, 2020, 2021b; Antolin, 2020). In this study, more type I CR and a few type III CR are observed during the interchange reconnection. This is different from Li et al. (2018b, 2019, 2020) in which the type III CR, rather than the type I CR, is detected. Comparing the interchange reconnection event here with those in Li et al. (2018b, 2019, 2020), we may answer the question that under what conditions the types I and III CR form during the interchange reconnection. Compared with the reconnection events in Li et al. (2018b, 2019, 2020), in this study, similar converging speeds of reconnecting structures toward the reconnection region and moving speeds of newly-reconnected structures away from the reconnection region are observed. However, the reconnecting closed loops L1, where the type II CR that always occurs in the non-flaring AR closed loops forms, and the newly-reconnected closed loops L3, where the type I CR takes place, are clearly identified. Moreover, the current sheet evidently appears at the interface between the reconnecting open and closed structures. These observations indicate that more magnetic flux participates in the interchange reconnection region with a similar inflowing speed in this study. More magnetic energy is thus converted into more other, e.g., thermal and kinetic, energy in a similar or even short time, showing a larger reconnection rate. Furthermore, we calculate the reconnection rate, the ratio of the width and the length of the current sheet, in the interchange reconnection events in Li et al. (2018b) and Li et al. (2020), and get similar values of $\sim$0.2, half of the reconnection rate (0.4) in this study. We hence propose that the reconnection rate plays an essential role in the formation of types I and III CR during the interchange reconnection. The type I (III) CR tends to form in the interchange reconnection between open and closed structures with a larger (smaller) reconnection rate.