Particles from Impulsive and Gradual Events

For a number of reasons, nearly all of the early SEP observations were made in events we now call gradual events or simply ``proton'' events. Element abundances of the MeV particles in these events were found to be simply related to those of the solar corona [e.g., Meyer 1985; Reames 1992]. In 1970, however, Hsieh and Simpson [1970] first reported He-rich events. Such events would be found to have He/He ratios in the range of 0.1 to 10, while the corresponding ratio in the corona and solar wind is 5x10. Order of magnitude enhancements in heavy element abundances (e.g. Fe/O) were also seen (see reviews by Reames [1990a, 1993, 1994]). As more and more He-rich events were seen, it became clear that they had a different behavior from the proton events.

Figure 1a shows the time history of a typical large gradual proton event. Protons dominate electrons in these events. Low energy (1 MeV) protons reach a plateau in intensity at the same value in many large events, although their intensity may rise again several days later as the shock passes. Profiles of electrons and higher-energy protons decline slowly with time. Figure 1b shows a series of He-rich events at the same scale. These events are often dominated by electrons and the intensities of all species decay rapidly with time. If the time profiles were controlled by pre-existing interplanetary scattering, why would particles of the same species and energy always have different profiles for He-rich and proton events?

When He-rich events have high intensities of nearly relativistic electrons, which arrive within minutes of the photons, it is relatively easy to associate the events with impulsive X-ray and H flares and the with type III radio bursts [see Reames et al. 1988, 1990]. The longitude dependence of the associated flares fall within a longitude band of <30 width about the footpoint of the nominal spiral field line, as seen in Figure 2. Gradual proton events [from Cane et al. 1988] show a more uniform longitude pattern. Why would particles from the He-rich events be unable to share the ``coronal diffusion'' of their brethren in the gradual events? Shocks cross field lines much more easily than particles and thus can accelerate particles over a wide longitude interval.

A completely different sort of information comes from the ionization state of the energetic Fe ions. In the gradual events, Fe has a mean ionization state of 14.1 0.2 [ Luhn et al. 1987], similar to that of Fe ions in the solar wind, and indicative of ambient (unheated) coronal material at a temperature of 2 MK. Even at energies above 200 MeV/amu Fe ions were found to have charge Q12.5 [ Adams et al. 1993] in three large gradual events. If this Fe had been processed at a temperature of 20-30 MK, typical of a large flare, it would have Q24 and lighter elements would be fully ionized. Meanwhile, Fe ions in He-rich events have an average ionization state of 20.5 0.2 [ Luhn et al. 1987], indicating either heating to 10 MK, or, more probably, stripping of the ions in the intense electron beams in impulsive flares [ Miller and Viñas 1993]. Fully ionized Fe has Q=26.

A final indication that the particle acceleration in the large proton events is related to CMEs rather than flares comes from event associations. According to Kahler et al. [1984], 96% of the large proton events have CMEs associated with them. Some of the proton events are associated with ``disappearing filament'' events on the Sun. In these events a filament and surrounding magnetic structure rises from the Sun to form a CME, but there is no associated impulsive flare event [see Kahler et al. 1986]. In fact, the ``typical'' large proton event I have shown in Figure 1a comes from a disappearing filament event.

Impulsive-flare (He-rich) events were once thought to be rare, however, it is now clear [ Reames, 1993; Reames et al. 1994] that the events are observed at 1 AU at a rate of about 100 events/yr during solar maximum. Since they come from a restricted longitude interval as seen in Figure 2, 20 or less when we allow for field-line motion with solar wind speed, the total number of events on the solar disk must be 1000/yr at solar maximum. The number of hard X-ray bursts, H flares and type III bursts vary from 4000/yr to 10000/yr [see Reames 1993]. Even at the present level of instrument sensitivity, the He-rich events account for a significant fraction of solar flares; this fraction might increase when instruments with 100 times this sensitivity are flown aboard the WIND spacecraft. Meanwhile, Cane et al. [1988] found a total of 235 proton events in 20 years, or 20/yr at solar maximum. In this case there is no correction for longitude since the events come from the visible disk and even from far behind the west limb. The rate of CMEs is 500/yr, however, many are too slow to form the strong shock that is necessary for particle acceleration.

A summary of the properties of gradual and impulsive events is given in Table 1. These observations have led to a new paradigm: in most of the large proton events a CME-driven shock wave accelerates the particles from the ambient plasma of the corona and solar wind as it propagates over a large region of space and time. The particles that are actually accelerated in impulsive flares have unusual He-rich, Fe-rich, and electron-rich abundances that were probably produced by wave-particle interactions induced by the streaming electrons in the flare plasma [ Temerin and Roth 1992; Miller and Viñas 1993].