Observation of Magnetic Switchback in the Solar Corona

Switchbacks are sudden, large radial deflections of the solar wind magnetic field, widely revealed in interplanetary space by the Parker Solar Probe. The switchbacks' formation mechanism and sources are still unresolved, although candidate mechanisms include Alfv\'enic turbulence, shear-driven Kelvin-Helmholtz instabilities, interchange reconnection, and geometrical effects related to the Parker spiral. This Letter presents observations from the Metis coronagraph onboard Solar Orbiter of a single large propagating S-shaped vortex, interpreted as first evidence of a switchback in the solar corona. It originated above an active region with the related loop system bounded by open-field regions to the East and West. Observations, modeling, and theory provide strong arguments in favor of the interchange reconnection origin of switchbacks. Metis measurements suggest that the initiation of the switchback may also be an indicator of the origin of slow solar wind.


INTRODUCTION
The solar wind is a continuous flow of charged particles streaming from the Sun's outermost atmosphere, the solar corona, into interplanetary space (Hundhausen 1972). It is characterized by the coexistence of large-scale structures of solar origin, turbulent fluctuations, magnetohydrodynamic (MHD) and kinetic plasma waves and instabilities, associated with physical processes such as magnetic reconnection, shocks, and a broad range of kinetic processes, which result in particle heating and energization (Bruno & Carbone 2013).
The solar wind acceleration mechanisms, its complex dynamics and interaction with the solar magnetic field, and the observed plasma heating are still outstanding questions in heliophysics. One interesting solar wind feature emerging from in-situ spacecraft measurements is the puzzling existence of abrupt, temporary magnetic field reversals, named magnetic switchbacks.
First observed in the outer heliosphere by the Ulysses spacecraft in high-latitude fast solar wind (Balogh et al. 1999;Yamauchi et al. 2004), Switchbacks are markedly Alfvénic, pressure-balanced structures, characterized by constant temperature and magnetic field magnitude, and associated with substantial acceleration of the plasma. Later analysis of Helios data in the ecliptic fast solar wind showed abundance of such structures in the inner heliosphere (Horbury et al. 2018). More recently, extensive measurements by Parker Solar Probe (PSP, Fox et al. 2016) confirmed that the presence of switchbacks increases dramatically near the Sun Kasper et al. 2019), and allowed a thorough study of their characteristics.
Switchbacks have been described as three-dimensional, elongated S-shaped structures with a high aspect ratio in the radial direction (Horbury et al. 2020). The typical deflection angle with respect to the heliospheric magnetic field is broadly distributed (Dudok de Wit et al. 2020), and the duration of the reversals ranges between seconds and hours (Dudok de Wit et al. 2020;Laker et al. 2021). At close distances from the Sun, switchbacks appear in dense, irregular clusters that alternate with quiet periods of stable field polarity and smaller magnetic fluctuations, with a modulation on a time scale of a few hours . The size and spatial distribution of the switchback patches, as well as the plasma characteristics within them, are compatible with the coronal magnetic structure determined by solar supergranulation . Switchbacks are associated with enhanced turbulence energy transfer (Bourouaine et al. 2020;Perrone et al. 2020;Hernández et al. 2021;Martinović et al. 2021) and magnetic reconnection was observed at their edges (Froment et al. 2021). On the other hand, because of their limited size, energetic particle propagation along the interplanetary magnetic field does not seem to be affected by their presence .
It is however expected that switchbacks may play a fundamental role in driving processes that ultimately heat the solar wind. Despite the numerous and thorough experimental studies, the nature of the mechanisms generating the switchbacks is still being debated. For example, it is not clear if they are driven by processes in the lower solar atmosphere (Magyar et al. 2021) or self-consistently generated in the solar wind. While their overall occurrence characteristics observed by PSP so far better support a possible origin in the coronal transition region rather than in situ (Mozer et al. 2021;Bale et al. 2021;Fargette et al. 2021), several competing models have been proposed. For * These authors contributed equally to this work. example, switchbacks could be the signature of frequent interchange reconnection events in the solar corona (Veselovsky et al. 1999;Pariat et al. 2009;Fisk & Kasper 2020;Zank et al. 2020;Sterling & Moore 2020;Drake et al. 2021;Liang et al. 2021). According to a different proposed mechanism of coronal generation, they could be associated with the motion of coronal magnetic field footpoints from the slow to the fast wind sectors (Schwadron & McComas 2021). The observed coupling with solar supergranulation  seems to suggest that switchbacks are nonlinear Alfvénic structures somehow associated with the global circulation of open magnetic flux at the solar surface (Fisk & Kasper 2020;Zank et al. 2021). On the other hand, MHD numerical simulations suggest that switchbacks may be Alfvénic structures originating in the lower corona and propagating out into the heliosphere (Matteini et al. 2015;Tenerani et al. 2020;Jakab & Brandenburg 2021), or may be related with velocity-shear driven dynamics (Landi et al. 2006;Ruffolo et al. 2020). From the same perspective, switchbacks may be generated self-consistently during the solar wind expansion of turbulent fluctuations (Squire et al. 2020;Shoda et al. 2021).
The variety of proposed models highlights the complex nature of switchbacks, and the limited vantage point provided by single-spacecraft, in-situ experimental observations available so far. Remote observation of switchbacks in the corona could provide a more complete description of their space-time dynamics, and help validate the models. However, detecting switchbacks in the extended corona with remote-sensing instruments is challenging and involves high spatial resolution coronagraphs. Such an endeavor is further complicated by the absence of coronal magnetic field measurements, so switchbacks can only be revealed by seeking their associated plasma counterparts.
This Letter reports observations from Metis (Antonucci et al. 2020), the coronagraph onboard Solar Orbiter (SO, Müller et al. 2020), showing the first evidence of a magnetic switchback in the solar corona. A thorough data analysis, complemented by a stringent theoretical description and detailed modeling of the event and of the coronal magnetic field, strongly supports the interpretation of the switchback being generated by interchange reconnection between the coronal loops formed above an active region and the nearby open-field regions. This is also particularly relevant to the slow solar wind, whose origin, still widely debated, and seems likely to be related to interchange reconnection events continuously occurring at the boundaries between coronal holes and loop systems.  Figure 1(b) shows the same view of the corona but with the Metis image reprocessed by the Simple Radial Gradient Filter (SiRGraF) algorithm proposed by Patel et al. (2022) to bring out the short-scale dynamic coronal structures. A peculiar S-shaped kink in the stream of plasma flowing off the Sun is clearly visible against the darker coronal background, above the southeastern solar limb. It is worth noting that this coronal feature also emerges by applying independent and different methods, such as a wavelet-based image enhancement algorithm (Telloni et al. 2013) and the Proper Orthogonal Decomposition (POD, Lumley 1981). A zoom of the S-shaped structure is provided in Figure 1(c) and delimited by the rectangular region of interest (ROI). The structure formed above the complex loop system related to the Active Region (AR) 12972 (marked with a white square in the figure). Finally, Figure 1(d) shows the velocity of coronal flows in the same FOV as Figure 1(c). The large-scale coronal flow pattern is derived by tracking outwardly moving density enhancements in the Metis white-light images, via the Fourier Local Correlation Tracking (FLCT, Fisher & Welsch 2008) technique. Noteworthy is the slower plasma stream at latitudes (indicated by a white dotted line) where the S-structure occurs.
The sequence of frames in the ROI, displayed in Figure 2 every 400 seconds from 20:11 UT on March 25, 2022, shows the spatiotemporal evolution of the coronal feature under study. The direction along the structure is oriented horizontally and indicated with altitude above the Sun. The bright structure is observed initially well above the Metis occulted region, forming at a height of about 2.6 R , and then propagates outward and folds back somewhat. At its full development, the structure acquires the peculiar "S" shape. The time-height plot allows an estimate of the propagation speed of the main ripple, which expands at a constant speed of about 80 km s −1 (red dashed line in Figure 2). This is, however, only a lower limit. Indeed, Metis can only estimate the structure's velocity component in the observed plane of the sky (POS), so that possible projection effects cannot be ruled out. In addition, the whole structure does not seem to propagate at the same rate, but rather to stretch slightly in the radial direction. This could be again due to projection on the POS of a very warped 3D structure, or to actual velocity dispersion along the structure. Finally, the Metis images seem to suggest not only an outgoing but also an incoming flow (see, in particular, the frames at 800 and 1200 seconds after event initiation, where plasma filaments appear to protrude from the main ripple downward), although no firm conclusion can be drawn.
To investigate the vortical properties of the main ripple, which visual inspection seems to suggest, a spectral analysis was performed. Positioning in the rest reference frame of the evolving structure (i.e., at its barycenter) as it propagates outward, its transverse displacements with respect to the radial direction were evaluated and related to temporal fluctuations in velocity perpendicular to the structure. The corresponding Fourier power spectral density ( Figure  3) shows clearly that there is a single dominant frequency (at 12 mHz) with some sideband noise or less important fluctuations. Since 12 mHz is not a common frequency, it is likely to be an intrinsic kink mode of the structure. It follows that the characteristic timescale of the vortex roll is about 80 seconds. Hence, the ripple is essentially non-turbulent in nature, because otherwise a broadband spectrum should have been found.

Coronal Magnetic Field Modeling
The large-scale configuration of the coronal magnetic field during the observation period, and, most importantly, the local magnetic topology around the active region above which the S-shaped structure appears to emerge, is of crucial importance for context and for proper interpretation of the results obtained from the Metis data. A magnetic field  Figure 4(a). The squashing factor Q (Titov et al. 2011), which highlights the separatrix arcs of the 3D coronal magnetic field, is imaged in Figure 4(b). The separatrix web, made up of the separatrix and quasi-separatrix layers surrounding the heliospheric current sheet (HCS), can be better visualized by building maps of log Q at 3 R (Figure 4(c)). In these layers, the photospheric dynamics stresses the separatrices inducing current sheets and reconnection processes with subsequent release of coronal plasma. The released plasma is proposed to contribute to the slow wind observed in the heliosphere in the zones around the current sheet (Antiochos et al. 2012). West of AR 12972 (located exactly at the heliographic coordinates of the Metis-observed S-shaped feature), the separatrix between a tiny positive coronal hole, which is connected to the active region and touches the HCS, and a larger positive coronal hole has the signature of a null where they collide with the HCS. A pseudostreamer (PS) separatrix curtain lies on the East of the active region. Figures 4(d) and (e) finally show the synoptic maps of the photospheric radial magnetic field and of the footprints of the field lines associated with outward (red) and inward (blue) open flux at 3 R . A small patch of open field lines can be identified at the active region.
What clearly emerges from the MHD extrapolation of the global and local magnetic field is the complex loop system above the active region and the large magnetic arcs extending up to 2 − 3 R in the FOV of Metis, just where the S-shaped, vortical structure forms. More interestingly, the active region appears to be surrounded, both East and West, by open-field regions injecting plasma along the coronal plasma sheet. Essentially, the large helmet streamer formed above AR 12972 is broken up slightly by the fluxes/thermodynamics of the strong ARs. This is vital to the interpretation of the nature and source mechanism of the kink-like structure observed with Metis, as discussed below.

Switchback Driven by Interchange Reconnection
Based on Metis measurements, the previous section has provided observational evidence of the shape of a plasma structure in the solar corona, the height at which it was formed, the way it was generated, its spatiotemporal evolution, its vortical properties, the local magnetic field topology associated with an underlying active region, and the coupling of  the structure with a slow coronal plasma flow. As will be argued in the following, all this evidence concurs in supporting the interpretation of the observed structure as a magnetic switchback generated in the solar corona through interchange reconnection.
The process of magnetic reconnection of a loop and an open field line, known as interchange reconnection, may occur at the boundaries between closed-and open-field regions at the streamer-coronal hole interface, and at the edges of active regions surrounded by unipolar regions or small coronal holes. The mechanism is depicted in the cartoon of Figure 3 in Zank et al. (2020). Interchange reconnection events can trigger the formation of magnetic field deflections, which propagate both outward and inward, whenever they are launched in the sub-Alfvénic flow (Zank et al. 2020). The most extreme switchbacks take the characteristic S-shape and are detected in situ as magnetic field reversals. The launch of switchbacks could also be accompanied by coronal jets (Sterling & Moore 2020). As a result of interchange reconnection, coronal plasma can be additionally released as blobs or plasmoids in the solar wind (Sheeley et al. 1997;Wang et al. 1998). The switchback generated by such a mechanism (likely higher up in the corona, between large-scale loops and the adjacent open-field lines surrounding an active region) will hence be a single structure characterized by a perturbed magnetic field having a wavy shape. It will be associated with upward and potentially also downward flow of plasma that is different from the coronal background. In addition, such a switchback will consist of a single characteristic mode.
All these expected characteristics are actually depicted in the Metis images investigated here. First, the observed structure is S-shaped (Figure 1(c)). If the direction of the flow is thought to be controlled by the strong coronal magnetic field, the white-light enhancement can be used as a way to infer the magnetic field direction. Thus, an S-shaped enhancement can be associated with an S-shaped magnetic field region. Second, the time-distance image plot of Figure 2 indicates the observation of both switchback-related forward and inward flows, as predicted by Zank et al. (2020). Third, the observed coronal feature originates exactly above AR 12972 (Figure 4(d)), whose loop system, extending well above the Metis occulter, is surrounded by open-field regions, as shown by magnetic field extrapolations (Figure 4(e)). Thus, it appears to be the ideal site for an interchange reconnection event to launch a switchback. Furthermore, the switchback-associated plasma is denser than the surrounding ambient corona (Figure  1(c)), probably due to the compressions associated with the magnetic reconnection process and/or the entrapment of denser loop material. And, finally, the observation that the presumed switchback forms at high altitude in the solar corona 1 is suggested by the possible location of the separatrix field lines at the null point (as well as by the squashing factor synoptic map displayed in Figure 4(c)). Fourth, observation of the single well-defined structure in Metis images rules out a shear-related Kelvin-Helmholtz (KH) type of driving (as proposed by Ruffolo et al. 2020) and favors the interchange reconnection scenario. Indeed, a KH instability would not generate one large single propagating vortex, but rather multiple spatially distributed structures along the shear flow, many of which could be expected to be closely spaced or even interacting (as at the CME flank observed by Foullon et al. 2011, and in striking contrast to Metis observations). Fifth, the presence of a single dominant timescale (Figure 3) characterizing the observed ripple is consistent with a non-turbulent description of the switchback (as in the model by Zank et al. 2020;Liang et al. 2021). This is not compatible with the KH origin of switchbacks, or with the idea that switchbacks are a natural outcome of evolving turbulence in an expanding flow (as advanced by Squire et al. 2020). Indeed, in both cases, switchbacks would include a superposition of modes mutually interacting nonlinearly, thus generating a turbulent flow, without a characteristic dominant frequency. According to this analysis, the structure observed by Metis looks like a very clear switchback initiated by an interchange reconnection event occurring above an active region higher up in corona.
To conclude, adopting the perspective that interchange reconnection generates a switchback and then asking what the observational and theoretical support might be once that hypothesis is adopted, support for the existence of switchbacks in the solar corona generated by interchange reconnection can be argued from both observational and theoretical perspectives.

Switchback Modeling
To provide further clues to support the proposed interpretation and possibly to empirically validate the theoretical predictions by Zank et al. (2020) against the Metis observations, the spatiotemporal evolution of the observed switchback structure was modeled using the Markov Chain Monte Carlo (MCMC) method. The shape of the structure was initially modeled with a sinusoidal function n − n 0 = A sin [b (r − r 0 )], where n and r are the positions in heliographic coordinates, n 0 and r 0 are the shifts in the n and r directions, respectively, A is the amplitude of the perturbation in n, b is the wave number of the sinusoidal function, and d is a factor adjusting the width of the structure along the r direction, i.e., r − r 0 − 1 b arcsin n−n0 A < d 2 . In addition, to simulate the flow shear in the surrounding solar wind, a velocity gradient varying as n (U r (n) = kn + U 0 , where U r (n) is the background solar wind speed, U 0 is a reference speed, and k denotes the slope of the flow velocity) was also included. For simplicity, only the propagation in r was considered. The initial structure was then allowed to evolve, based on the linear theory propagation function by Zank et al. (2020). The shape of the propagated structure was finally compared with Metis observations at 2400 seconds after the switchback initiation, using the MCMC technique. The MCMC method is based on Bayesian inference (e.g., Bonamente 2013) and was used to provide the most probable values of the free parameters n 0 , r 0 , A, b, d, and k that lead to the optimal fit of the model to the observations (this approach has been successfully applied in heliospheric studies by Zhao et al. 2019;Liang et al. 2021, where the interested reader can find more details on the application of the MCMC technique). The results are displayed in Figure 5, where the modeled switchback shape before and after propagation (white shape) is overlaid on the Metis image 800 and 2400 seconds after the switchback formation. 800 s 2400 s Figure 5. MCMC-modeled switchback (white shape) before and after linear theory-based propagation, compared with Metis observations at 800 and 2400 seconds after initiation.
Although the model obviously cannot capture all features of the imaged switchback, the agreement with observations is striking. In particular, the modeled switchback elongates during propagation, which is consistent with the observations. Even more remarkable is that although the free parameters were set to best reproduce the shape of the 2400-second switchback, the 800-second structure so modeled also fits satisfactorily the observations. This is a strong indication of the validity of the linear theory proposed by Zank et al. (2020) to describe the propagation of switchbacks in the corona, and greatly supports the scenario of having identified a magnetic switchback low in the solar corona for the first time.

Switchbacks and the Slow Solar Wind
The observation of switchbacks in the solar corona is closely related to the still unsolved problem of the origin of the slow solar wind. Indeed, one of the possible mechanisms for the origin of the slow coronal flows is that initially proposed by Fisk et al. (1999);Fisk (2003) and later extended by Zank et al. (2021) by incorporating a mechanism for plasma heating. In this model (well illustrated in Figure 1 of Zank et al. 2021), the large-scale loop plasma is heated to high temperatures via the dissipation of quasi-2D turbulence generated by the magnetic carpet. An interchange reconnection event with an open-field line at altitudes of 2 − 4 R liberates the already hot loop plasma onto an open-field coronal region, which then expands to reach supersonic and super-Alfvénic speeds. It appears thus evident that switchbacks and slow solar wind streams can be viewed as two manifestations of the same physical process, i.e., interchange reconnection. A number of clues supporting this possible scenario are present in the literature. For instance, the speed of the coronal blobs released by reconnection events (Sheeley et al. 1997;Wang et al. 1998) was found to be consistent with that of the slow wind (Antonucci et al. 2005). Actually, quantifying the contribution of reconnection-related coronal blobs to the slow wind plasma has been the subject of several studies over the past decade (e.g., Antiochos et al. 2012;Wang 2012Wang , 2020. Furthermore, Harra et al. (2008); Brooks & Warren (2011);Del Zanna et al. (2011);Fazakerley et al. (2016); Harra et al. (2017) observed persistent coronal plasma upflows in association with quiescent active regions (first identified by Uchida et al. 1992) and interpreted this evidence as an indication of magnetic reconnection being the main driver of the slow wind. Furthermore, slow coronal flows have been observed to originate at the edges of coronal holes adjacent to closed-configuration equatorial streamers, and then stream along their flanks (Antonucci et al. 2005). Finally, switchbacks are not observed in the middle of low-speed streams because, again in the interchange reconnection picture, these can occur only at the boundaries of slow and fast wind. Since both the loop-system and the open-field region consist of multiple magnetic field lines, interchange reconnection can occur in rapid succession, resulting in clustering of switchbacks (consistent with PSP measurements, e.g., Dudok de Wit et al. 2020) and in steady slow coronal flows. This process will continue while the loop system keeps emerging from the active region and until the magnetic footpoint motions ensure that the closed-and open-field regions are close enough to magnetically reconnect. Hence, the timing of the coronal switchback observed with Metis, as well as the height of its initiation, may be indicators of the origin of slow solar wind. As discussed in the previous section (Figures 1(d) and 2), the switchback is observed to form at about 2.6 R and then expand at the low speed of ∼ 80 km s −1 in a plasma stream slower than the adjacent ones. A remarkable relation of the switchback formation with the coronal region of slow plasma outflows has therefore been found, interestingly converging towards the interchange reconnection interpretation. Therefore, the study of switchbacks in the corona, especially where, how, and how often they are formed, seems to be very promising in disclosing the origin of the slow solar wind.