Solar Orbiter Catches External Reconnection Bringing a Filament Eruption to a Halt

(Solar Orbiter nugget #96 by Y. Fang¹, S. Tan^2,3, J. Guo⁴, L. Huang⁴, X. Cheng⁴, and J. Zhang¹)

Introduction

Solar filaments are the reservoirs of cool, dense plasma that, once they lose equilibrium, drive many of the corona's most energetic events: two-ribbon flares and coronal mass ejections (CMEs) [1]. Yet not every eruption escapes. A substantial fraction rise, decelerate, and fall back to the surface without ever producing a CME. Following the amount of mass and magnetic structure that escapes, filament eruptions are commonly classified as full, partial, or failed [2], and it is the failed (confined) eruptions, confined to the low corona with their material draining back [3], that remain the least understood.

Several mechanisms have been proposed to explain why an eruption fails: the flux rope reaching only into the torus-stable domain after lift-off [4]; a non-axisymmetric flux-rope geometry that produces a net downward Lorentz force [5]; or external magnetic reconnection between the erupting structure and the ambient field that gradually erodes the erupting flux until the driving force vanishes [6]. The last of these is attractive because it naturally links the failure to a clear energy-release signature, but it is also the hardest to demonstrate. Pinning it down requires simultaneously (i) resolving the fine EUV structure of the filament and its overlying field to see the reconnection geometry, and (ii) catching the timing and location of the associated hard X-ray (HXR) energy release. Direct observational evidence that external reconnection is the agent of eruption failure has therefore remained scarce.

This is precisely the science target that Solar Orbiter [7] was built for. By combining the ultra-high spatial resolution of the Extreme Ultraviolet Imager (EUI) [8] with the hard X-ray timing and imaging of the Spectrometer/Telescope for Imaging X-rays (STIX) [9], and adding the near-identical viewing angle of SDO on this particular day, we were able to watch external reconnection choke a filament eruption from start to finish, and then further confirm it in a data-constrained MHD simulation.

The event

On 2024 March 19, a filament in NOAA AR 13615 (southeastern quadrant of the disk) underwent a failed eruption associated with a GOES M2.1 flare. The filament first appeared as an elongated EUV filament channel (~40 Mm long, ~2 Mm wide) lying beneath a set of coronal loops. It rose, transitioned sharply from acceleration to deceleration, and its material ultimately drained back to the Sun, the textbook kinematic signature of a confined eruption.

Figure 1. Overview of AR 13615 and the M-class flare. (a1) Solar Orbiter/EUI-FSI 174 Å context image. (a2)–(a3) HRI_EUV 174 Å before and after the eruption. (b1) Time–distance plot along the eruption direction, with the leading edge outlined and the draining-back of material marked by arrows. (b2) Eruption velocity together with STIX HXR fluxes (10–15, 15–25, 25–50 keV) and the GOES 1–8 Å SXR flux, showing that the HXR peak coincides with the deceleration phase.

Results

What EUI's resolution revealed

Because HRI_EUV at 174 Å delivers a pixel size of 0.492″ (tens of km per pixel at perihelion) at a 16 s cadence, the fine architecture of this small, thin filament channel, invisible to lower-resolution imagers, could be tracked frame by frame:

• The channel was characterised by a strongly twisted flux rope, and as it rose its main axis developed a clear Y-shaped kink structure, the observational fingerprint of the ideal kink instability [10,11].

• Brightenings appeared at the footpoints of the overlying loops, spatially coincident (once cross-checked against the HMI magnetogram) with the opposite magnetic polarities in which those loops were rooted, a classic signature of external reconnection.

• After the eruption failed, multithermal bright material was seen streaming along the overlying loops, consistent with plasma transfer along newly reconfigured field lines.

Figure 2. HRI_EUV 174 Å sequence of the filament-channel evolution: brightening and two-ribbon flare (green arrows), the twisted threads and main axis (blue), the Y-shaped kink structure (blue contours), and the bright material moving along the overlying loops (green ellipses).

What STIX added — timing and the location of energy release

The HXR diagnostics from STIX were decisive in establishing the sequence of events:

• No significant HXR enhancement accompanied the early rise. This is a strong argument that the eruption was not initiated by reconnection; instead the initial rise was driven by the ideal kink instability. The HXR emission only began to rise well into the acceleration phase and peaked during the deceleration phase, exactly when the eruption was being braked.

• Reconstructing the 25–50 keV sources (MEM_GE [12], 60 s integrations) showed, in addition to a loop-top source, a distinct coronal HXR source at the leading edge of the erupting filament channel that intensified over time, direct evidence that energy was being released ahead of the rising structure, where the erupting flux met the overlying field. Two further sources near the northern ribbon mark where that energy was channelled down to the footpoints.

Figure 3. STIX 25–50 keV HXR contours (35%, 55%, 85% of peak) overlaid on HRI_EUV 174 Å, showing the emergence and intensification of the coronal source at the leading edge (white arrows) and the ribbon-associated sources.

SDO cross-check and MHD confirmation

On 2024 March 19 SDO and Solar Orbiter viewed the region from nearly identical angles, allowing a direct multi-temperature comparison. In SDO/AIA the overlying loops brightened strongly in 94 Å (~6.4 MK) after the eruption; they were heated, not merely dimmed, and remote brightening plus multithermal flows (304, 211 Å) confirmed the reconnection picture inferred from EUI.

A data-constrained, zero-β MHD simulation [14] (an RBSL flux rope [13] inserted into an NLFFF extrapolation) closed the loop: as the flux rope rose to ~25 Mm it began to reconnect with a fan-spine overlying field; its negative footpoint progressively drifted into the polarity region where the outer spine was rooted, until the erupting flux was completely eroded. Crucially, the computed decay index gave a torus-instability critical height of ~70 Mm, above the interaction site, so torus instability can be excluded as the cause of failure.

Figure 4. Magnetic configuration from the MHD simulation: the twisted flux rope and its overlying fan-spine field, the side-view drift and eventual erosion of the flux-rope footpoint, and the decay-index profile showing the torus-stable domain below ~70 Mm.

Discussion

Putting the EUI, STIX, SDO, and MHD pieces together yields a coherent, well-constrained sequence for this failed eruption:

1. The ideal kink instability initiates the rise (strong twist, Y-shaped kink, and no early HXR emission).

2. The rising flux rope then undergoes external reconnection with its overlying loops (footpoint brightenings, the leading-edge coronal HXR source, loop heating, and drifting flux in the simulation).

3. This reconnection erodes the poloidal flux that drives the eruption, producing the deceleration seen exactly when the HXR flux peaks, and the eruption ultimately fails.

This ordering, kink first and external reconnection second, differs from scenarios in which breakout reconnection precedes and facilitates the eruption [15]. It provides one of the clearest observational cases to date in which external reconnection is caught in the act of stopping an eruption, complementing simulation and statistical work that had argued the outcome (success vs. failure) hinges on the flux ratio between the erupting rope and the overlying field [6].

For the broad solar-physics community, the take-away is methodological as much as physical: distinguishing “the eruption never had enough drive” from “the drive was actively removed” requires catching the reconnection while it happens, in both EUV morphology and HXR energetics. The EUI + STIX combination on Solar Orbiter makes that possible, resolving structures a few Mm across while simultaneously timing and locating the non-thermal energy release, and this event shows what that pairing can deliver for the long-standing problem of what determines whether a filament eruption escapes or fails.

Publication

This nugget is based on the following work: Fang, Y., Tan, S., Guo, J., Huang, L., Cheng, X., & Zhang, J., “Observations and Magnetohydrodynamic Simulation of External Reconnection Resulting in a Failed Filament Eruption”, The Astrophysical Journal, 1006, 81 (2026). DOI: 10.3847/1538-4357/ae8020

Affiliations

(1) School of Physics, Anhui University, Hefei 230601, People's Republic of China

(2) Leibniz-Institut für Astrophysik Potsdam (AIP), An der Sternwarte 16, 14482 Potsdam, Germany

(3) Institut für Physik und Astronomie, Universität Potsdam, Karl-Liebknecht-Straße 24/25, 14476 Potsdam, Germany

(4) School of Astronomy and Space Science, Nanjing University, Nanjing 210046, People's Republic of China

References

[1] Lin, J., & Forbes, T. G. 2000, JGR, 105, 2375. https://doi.org/10.1029/1999JA900477

[2] Gilbert, H. R., Alexander, D., & Liu, R. 2007, SoPh, 245, 287. https://doi.org/10.1007/s11207-007-9045-z

[3] Ji, H., Wang, H., Schmahl, E. J., Moon, Y. J., & Jiang, Y. 2003, ApJL, 595, L135. https://doi.org/10.1086/378178

[4] Guo, Y., Ding, M. D., Schmieder, B., et al. 2010, ApJL, 725, L38. https://doi.org/10.1088/2041-8205/725/1/L38

[5] Zhong, Z., Guo, Y., & Ding, M. D. 2021, NatCo, 12, 2734. https://doi.org/10.1038/s41467-021-23037-8

[6] Chen, J., Cheng, X., Kliem, B., & Ding, M. 2023a, ApJL, 951, L35. https://doi.org/10.3847/2041-8213/acdef5

[7] Müller, D., St. Cyr, O. C., Zouganelis, I., et al. 2020, A&A, 642, A1. https://doi.org/10.1051/0004-6361/202038467

[8] Rochus, P., Auchère, F., Berghmans, D., et al. 2020, A&A, 642, A8. https://doi.org/10.1051/0004-6361/201936663

[9] Krucker, S., Hurford, G. J., Grimm, O., et al. 2020, A&A, 642, A15. https://doi.org/10.1051/0004-6361/201937362

[10] Kliem, B., & Török, T. 2006, PhRvL, 96, 255002. https://doi.org/10.1103/PhysRevLett.96.255002

[11] Török, T., & Kliem, B. 2005, ApJL, 630, L97. https://doi.org/10.1086/462412

[12] Massa, P., Schwartz, R., Tolbert, A. K., et al. 2020, ApJ, 894, 46. https://doi.org/10.3847/1538-4357/ab8637

[13] Titov, V. S., Downs, C., Mikić, Z., et al. 2018, ApJL, 852, L21. https://doi.org/10.3847/2041-8213/aaa3da

[14] Guo, Y., Zhong, Z., Ding, M. D., et al. 2021, ApJ, 919, 39. https://doi.org/10.3847/1538-4357/ac10c8

[15] Chen, Y., Cheng, X., Chen, J., Dai, Y., & Ding, M. 2023b, ApJ, 959, 67. https://doi.org/10.3847/1538-4357/ad09d8