Flare particle acceleration in the interaction of twisted coronal flux ropes

Transcription

1 Flare particle acceleration in the interaction of twisted coronal flux ropes James Threlfall, A. W. Hood (University of St Andrews) P. K. Browning (University of

2 Motivation Su et al., Nat. Phys. (213) Clear evidence of restructuring of magnetic fields here (and most flares). Tangled/twisted coronal fields "reconnect" to relax to lower energy state. Released energy: heat, bulk plasma motion and accelerates particles.

3 3D Magnetic Reconnection Reconnection historically studied using 2D steady state models: limitations and properties reasonably well known and understood In 3D: No fundamental restriction on where reconnection occurs. Z Necessary and sufficient condition for reconnection: E ds 6= (e.g. Schindler et al. 1988; Hesse and Schindler 1988) "Cut and paste" 2D field line picture no longer holds: 3D reconnection happens continuously and continually within finite volume.

4 Our Approach: Test Particles In uniform B-field, particles gyrate orbit field lines with Larmor/gyro-radius: r g = mv? eb Averaging over gyro-motion (GCA: guiding centre approximation) reduces complexity (provided environment unchanged during orbit). Typically leads to fast parallel motion (particularly when some component of E-field parallel to B) and slower perpendicular drifts. Downside: collisions and back-reactions upon global fields omitted Use relativistic form of GCA ODEs (Northrop, 1963)

5 Blatant plug! Try it yourselves: Relativistic and non-relativistic GCA schemes, 4th/5th order Runge Kutta scheme - only needs E and B. Assumes well separated spatial and temporal scales Adapted to take input from analytical fields or various MHD codes.

6 What configurations to probe? *shameless self-promotion warning* Isolated topological features - separators (Threlfall et al. A&A, 215,216a) Non-flaring Active Region model (MHD) (Threlfall et al. A&A, 216b) Non-topological coronal reconnection model (Threlfall et al., Solar Physics, 217) Multi-thread avalanche energy release (MHD) (Threlfall et al., 218, A&A, 611, A4)

7 Multi-thread cascade Becomes kink unstable Tam et al. (215), Hood et al. (216) First demo of single coronal loop thread destabilising neighbouring threads, leading to a cascade MHD Energy release in discrete bursts (nanoflares?) Study (up to) 23 threads More details in paper(s)! stable

8 Multi-thread cascade j If this is a "nanoflare storm", how do particles respond?

9 Our Study A B First step: Study particle behaviour in two loop config. Cases: 1. Loop A does not destabilise Loop B 2. Loop A triggers disruption in Loop B

10 Case 1 - single loop disruption Single loop benchmark case Blue loop initially kink unstable Green initially marginally stable Resistivity: ηbkg + ηanom (j>jcrit) Particles: randomise positions, pitch angles and Maxwellian energies purple = current > jcrit

11 Case 1: Pre-onset ηanom only 2 1 upper boundary final positions little/no acceleration. Pre-onset energy distributions: y (Mm) 2 ηanom ηanom+ηbkg -1 y (Mm) x (Mm) lower boundary final positions f(e) -4-6 Initial Maxwellian Initial Energy Distrib. Electrons: final, inc. bkg res. Electrons: final, no bkg res. Protons: final, inc. bkg res. Protons: final, no bkg res log 1 (Energy) [ev] x (Mm)

12 Case 1: Kink Instability Onset ηanom only Thin beams of accelerated particles at top and bottom boundaries. More kev-mev orbits when including ηbkg 2 f(e) -4-6 Initial Energy Distrib. Electrons: final, inc. bkg res. Electrons: final, no bkg res. Protons: final, inc. bkg res. Protons: final, no bkg res. Helical currents >jcrit log 1 (Energy) [ev]

13 Case 1: Fragmented current sheets dissipate ηanom only Broader regions of accelerated orbits Good agreement with previous works (e.g. Gordovskyy et al. 211,12) Energy dists well-matched with and without ηbkg: 2 f(e) -4-6 Initial Energy Distrib. Electrons: final, inc. bkg res. Electrons: final, no bkg res. Protons: final, inc. bkg res. Protons: final, no bkg res Thin current sheets rapidly dissipate to sub-critical levels log 1 (Energy) [ev]

14 Case 2 How do things change when a second loop becomes destabilised?

15 Differences between cases? Case 1: Case 2: Secondary disruption Key differences: Orientation of initial helical instability ηbkg= Insert particles at multiple stages (blue arrows) de (% of total energy) Change in Energy (relative to initial state) magnetic energy kinetic energy thermal energy left loop kink instability onset right loop destabilises time (t/τ A )

16 Case 2: kink instability onset upper boundary final positions 1 y (Mm) x (Mm) lower boundary final positions 1 y (Mm) Acceleration signatures ONLY in left hand loop core x (Mm)

17 Case 2: intermediate phase upper boundary final positions current >jcrit dissipates 1 y (Mm) x (Mm) lower boundary final positions 1 sheath field distorted y (Mm) -1 Few MeV orbits observed x (Mm)

18 Case 2: Secondary disruption upper boundary final positions Fragmented currents >jcrit 1 y (Mm) x (Mm) lower boundary final positions 1 y (Mm) Acceleration signatures spread THROUGHOUT boundaries x (Mm)

19 Case 2: Energy Distributions Proton and electron energisation nearly identical. Distributions follow reconnection rate. Hard-soft-hard pattern for two loops. Not obvious what happens when more loops included. f(e) Final electron spectra orbits initialised at t=τ A (p= -3.69) t=9τ A (p= -1.37) t=115τ A (p= -1.29) t=19τ A (p= -1.83) t=235τ A (p= -1.34) t=27τ A (p= -1.47) log 1 (energy) [ev]

20 Summary Threlfall, Hood & Browning, 218, A&A, 611, A4 3D reconnection fundamentally different to 2D. Guiding centre theory not new: careful application to 3D magnetic reconnection configurations is! Parallel electric field is crucial and sometimes overlooked! Multi-thread MHD loop cascade/eruption (2 loops): Orbits in single loop destabilisation agree with Gordovskyy et al. (211,212) Secondary disruption can be triggered by orientation of helical instability. Energised orbit final positions fill volume of both loops during second eruption Energy distribs repeatedly harden then soften, matching reconnection