Using extreme-ultraviolet (EUV) spectroheliograms from the first intentional postflare observations with the Coronal Diagnostic Spectrometer (CDS) on board SOHO, we determine relative line-of-sight velocities and their temporal evolution during the gradual flare phase of an M6.8 two-ribbon flare that occurred on 1998 April 29. Dopplergrams in lines of O v, Fe xvi, and Fe xix, with formation temperatures of, respectively, 0.25, 2.0, T max and 8.0 MK show strong velocity gradients coincident with the Ha ribbons, visible in Big Bear Solar Observatory (BBSO) images. These gradients are perpendicular to and moving with the Ha ribbons. Bright downflowing plasma seems to be prevalent in the regions, between the ribbons and the magnetic neutral line, that coincide with the ends of postflare loops seen with the Extreme-Ultraviolet Imaging Telescope (EIT) on board SOHO. The plasma on the outer side of the ribbons is less bright in the EUV but shows strong relative blueshifts. This pattern of upflows and downflows demonstrates, for the first time in transition region and coronal lines, the existence of chromospheric evaporation during the late gradual phase of a flare and provides evidence for ongoing reconnection.

Abstract. During the solar flares on 9, 11, and 15 June 1991 the COMPTEL instrument measured extended γ-radiation in the 2.223 MeV neutron-capture line, in prompt nuclear deexcitation lines and in pion-decay radiation for several hours after the flares. The long-term time profiles can be described by a double exponential decay with decay constants on the order of 10 min for the fast and several 100 min for the slow components. We studied the 11 June 1991 flare in more detail and found that during the extended phase the accelerated proton and ion spectrum is harder, the e/p ratio is lower, and the emission profile is smoother, compared to those of the impulsive phase. Pion-decay radiation was not detected before the onset of the extended emission phase. When comparing the three flares to one another, we found a striking similarity in the time profiles of the nuclear line and the neutron capture line emission. However, the pion-decay radiation varied in intensity significantly from flare to flare. The impulsive-phase emissions of the flares show no such similarity. Our measurements indicate that the processes taking place during the extended phase differ from those during the impulsive phase, or in other γ-ray line flares. Based on these results long-term trapping of energetic particles from the impulsive phase seems unlikely, as opposed to continuous particle acceleration.

The COMPTEL Imaging Compton Telescope on-board the Compton Gamma Ray Observatory measured significant neutron and "/-ray fluxes from the solar flare of 9 June 1991. The "/-ray flux had an integrated intensity (> 1 MeV) of-30 cm-2, extending in time from 0136 UT to 0143 UT, while the time of energetic neutron emission extended approximately 10 minutes longer, indicating either extended proton acceleration to high energies or trapping and precipitation of energetic protons. The production of neutrons without accompanying "/-rays in the proper proportion indicates a significant hardening of the precipitating proton spectrum through either the trapping or extended acceleration process.

During the solar flare events on 11 and 15 June 1991, COMPTEL measured extended emission in the neutron capture line for about 5 hours after the impulsive phase. The time profiles can be described by a double exponential decay with decay constants on the order of 10 min for the fast and 200 min for the slow component. Within the statistical uncertainty both flares show the same long-term behaviour. The spectrum during the extended phase is significantly harder than during the impulsive phase and pions are not produced in significant numbers before the beginning of the extended emission. Our results with the measurements of others allow us to rule out long-term trapping of particles in non-turbulent loops to explain the extended emission of these two flares and our data favour models based on continued acceleration.

COMPTEL on the Compton Gamma Ray Observatory has measured the flux of y-rays and neutrons from several solar flares. These data have also been used to image the Sun in both forms of radiation. Unusually intense flares occurred during June 1991 yielding data sets that offer some new insight into of how energetic protons and electrons are accelerated and behave in the solar environment. We summarize here some of the essential features in the solar flare data as obtained by COMPTEL during June 1991.

The COMPTEL instrument onboard the Compton Gamma-Ray Observatory (CGRO) is sensitive to ),-rays in the energy range from 0.75 to 30 MeV and to neutrons in the energy range from 10 to 100 MeV.During the period of unexpectedly high solar activity in June 1991, several flares from active region 6659 were observed by COMPTEL. For the flare on June 11, we have analyzed the COMPTEL telescope data, finding strong 2.223 MeV line emission, that declines with a time constant of 11.8 minutes during the satellite orbit in which the flare occurs. It remains visible for at least 4 hours. We obtained preliminary values for the 2.2 MeV and 4-7 MeV fluences. Neutrons with energies above 20 MeV have been detected and their arrival time at the Earth is consistent with the v-ray emission during the impulsive phase. ~TRODUCTIONCOMPTEL is a useful instrument for studying solar flares. It is able to measure V -rays in the energy range from 0.75 MeV to 30 MeV and to derive the direction of the incoming radiation. When operating in its full-telescope mode, COMPTEL uses a double scattering of a v-quantum to get spectral and directional information. In this case, Compton scattering of the ),-ray is the fundamental process.

No abstract

Direct solar flare neutrons are a valuable diagnostic of high-energy ion acceleration in these events, and COMPTEL improves over all previous cosmic neutron detectors in its capacity for neutron energy measurement. Previous studies of COMPTEL neutron data have worked with an incomplete model of the instrumental response, applying energy-by-energy detection efficiencies. Here we employ statistical regularisation techniques with the full (Monte Carlo simulation derived) response matrix to produce improved estimates of neutron numbers and energy distribution. These techniques are applied to data from the well-observed 15 June 1991 flare. Our improved treatment of the instrumental response results in a reduction of 73% in total neutron numbers, compared with previously deduced values. Implications for the picture of primary ion acceleration in this flare are briefly discussed.