1. SEP Acceleration Mechanisms Dennis K. Haggerty and Edmond C. Roelof Johns Hopkins U./Applied Physics Lab. ACE/SOHO/STEREO/Wind Workshop Kennebunkport, ME (8-10 June)

2. The use of Spikes, Pulses, and Ramps in the ACE/EPAM near- relativistic electron events The temporal profile of electron events observed at 1 AU is a continuum, from Gaussian spikes through long-lived ramp events. To use these classifications quantitatively we have assigned a value “v” to each electron event, between zero and 2.5: 0.0 < v < 0.5 represent spikes. 0.5 < v < 1.5 represent pulses. 1.5 < v < 2.5 represent ramps. We sort 204 near-relativistic beam-like electron events based on their assigned temporal value into groups of spikes, pulses and ramps. We then examine the intensity, spectra, and timing association with solar electromagnetic emissions for each type of event.

3. Spike and Pulse events often come as part of a sequence of events Spikes and pulses often occur (mixed) in sequences of electron events that are individually (< 18 h) apart. Although such a sequence may culminate in a Ramp event, most Ramp events are isolated in time. Example of an event sequence of > 5 spike/pulse events. The general profile is not expected from a CME-driven shock 1 2 3 ? ? 4 5

4. Classes of beam-like solar electron events: Implications for acceleration and injection When these beam-like electron events are classified according to their intensity- time profiles (spikes, pulses, ramps), the different classes have different distributions of: Peak intensity Spectral index CME velocity Injection delays after solar electromagnetic emission (SXRs, typeiii radio bursts).

5. Association with 2-5 MeV protons We compared the peak intensity of the energetic electron beams with the prompt peak intensity of the energetic protons measured by EPAM. There is a very good correlation in peak intensity between the peak intensity of the near-relativistic electron beams and prompt (SEP) peak in the energetic proton events. We get this good correlation because we restricted the study to beams and the prompt component (other studies mix together non-beams, the ESP part of the event, etc…

6. Association with near-relativistic protons We have analyzed the rise phase of 19 solar energetic particle events (SEPs) using simultaneous data from: EPAM/DE4:(175-315 keV) electrons SOHO/ERNE:(~100 MeV) protons We compare the onsets and the relative timing of these large SEP events By choosing electron beam events that we know are magnetically well-connected to the injection site, we believe it likely that the protons are also beam-like. Hence these proton events are also likely to have nearly scatter-free propagation, so that their rise history measured at 1 AU is essentially their injection history near the Sun.

7. Almost all values fall in same ranges (~3) 3<(a/  )<10electrons 2<(a/  )<8protons Highly unlikely by chance because individual values of a and  vary over ranges ~10. Implication: Common acceleration/release mechanism (on average over all events) Ratio of turnover-to-rise times

8. Conclusions (1) Non-beam events (~500/700) can come from any longitude or can cluster about active region longitudes Beam events Spikes and Pulses (smaller events)— Occur in sequences Tend to map to open field lines near flaring active regions Ramps (larger events)— Isolated (or at culmination of sequences) Can map near or relatively far from flaring active regions Consistent with acceleration by CME-driven shocks

9. Conclusions (2) Spike and Pulse events have properties consistent with a more localized and sporadic acceleration mechanism. Shorter delays with respect to solar electromagnetic emissions Poor correlation with CME-driven shocks and type-II events Smaller intensity and softer spectra Fairly good agreement with coronal models of open fields near the associated active region They also tend to occur in sequences. Reconnection events are known to recur at the same general site (same active region near coronal holes with open field lines). It is difficult to produce a nearly symmetric Spike (decay time only a little longer than the rise time) by a moving shock. Acceleration by magnetic reconnection is a more likely candidate. Injection into the heliosphere based on access to open coronal fields.

10. Conclusions (3) Ramp events (electrons and protons) have properties consistent with acceleration by shocks driven by large CMEs: Significant delays with respect to prompt solar electromagnetic emissions (e.g. HXR, microwave, type-III) Higher intensity and harder spectra Delays with respect to CME launch times (implied injection until CME develops a strong coronal shock). Correlated prompt-peak intensity (the larger events so most were ramps) Similar rise-to-maximum history functions Their long duration (and continuing rise while the anisotropy is still significant) is easily explainable by continual injection/release from an outgoing CME-driven shock. On the other hand, such smooth and continued release is difficult to explain by an explosive or sporadic process like reconnection.

11. Summary Although the associations favor CME-driven shock acceleration for the long- lived Ramp events, and an explosive process like magnetic reconnection for the shorter-lived Spike and Pulse events, the classification of 204 beam-like electron events yields a continuum of injection profiles from Pulse to Spike to Ramp. The rise-to-maximum functions (of time) also form a continuum. Obviously both shocks and reconnection can contribute to particle acceleration in a particular event. Violent reconnection would be expected to produce shocks, and CME launches require magnetic reconnection. It is reasonable to think of two acceleration process (shocks and reconnection), in which the former dominates for Ramps and the latter for Spikes and Pulses.

12. END