Observations of the failed eruption of a filament

We have observed the fine temporal and spatial structure of a filament eruption on 2002 May 27 following an M2-class flare. Our observations at Big Bear Solar Observatory were made at the wavelength of Hα 1.3 Å, with a cadence of 40 ms. The event was also observed by the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI) at X-ray energies from 3 to 50 keV and by the Transition Region and Coronal Explorer (TRACE) in poFe XII 195 Å. The event appears to be a "failed eruption," as the filament material, seen in absorption by TRACE, first accelerated then decelerated as it approached its peak height of ∼8 × 10⁴ km while the filament threads drained back to the Sun. The fact that the eruption did not lead to a coronal mass ejection indicates that the coronal magnetic field near ∼8 × 10⁴ km did not open during the flare. The height-time curve obtained from the TRACE 195 Å images during the deceleration phase shows that the deceleration of the filament exceeded the gravitational deceleration by more than a factor of 10, which suggests that the filament material was pulled back by magnetic tension. Also of importance are three sequential but cospatial features - brightenings in EUV, a loop-top hard X-ray emission, and "rupturing" of the Hα filament - that point to a release of energy (and probably magnetic reconnection) above the initial filament's location but well below its terminal height. Reconnection above a filament does not appear in most models, with the notable exception of quadrupolar and "breakout" models. These observations provide evidence that at least two conditions are required for a successful eruption: a reconnection very low in the corona (possibly above the filament) and open or opening fields above that point.