Mechanisms for particle heating in flares J. F. Drake University of Maryland J. T. Dahlin University of Maryland M. Swisdak University of Maryland C. Haggerty University of Delaware M. A. Shay University of Delaware T. D. Phan U. C. Berkeley
RHESSI occulted flare observations 30-50keV 17GHz Krucker et al 2010 Observations of a December 31, 2007, occulted flare A large fraction of electrons in the flaring region are part of the energetic component (10keV to several MeV) The pressure of the energetic electrons approaches that of the magnetic field A significant fraction of electrons in the flaring region undergo acceleration
Numbers related to the RHESSI December 31 2007 occulted flare observations Characteristic size of x-ray source L ~ 10 4 km Hard x-ray decay time ~ 40s Alfven speed (B ~ 50G, n ~ 10 9 cm -3) c A ~ 3x10 3 km/s Super hot electrons T h ~ 10keV Thermal transit time t th ~ 0.1 s Mean-free-path L mfp ~ 10 6 km >> L collisionless Reconnection time scales for island with scale w ~ 10 3 km t r ~ 1.5s Plenty of time for reconnection energy release Electron transport suppression required Islands, double-layers, mirroring or?
Main Points Electrons and ions in flares are dominantly accelerated by Fermi reflection in contracting magnetic field lines Two-step acceleration mechanism Electron and ion acceleration in a single reconnection exhaust drives bulk heating up to around 10 kev Producing an energetic electron and ion component up to around an MeV or beyond requires multi-island reconnection
The critical role of the reconnection exhaust The change in topology for reconnection to proceed takes place in the dissipation region This is not where significant magnetic energy is released Energy release primarily takes place downstream of the x- line where newly-reconnected field lines relax their tension Not at small spatial scales
Basic mechanisms for particle energy gain during reconnection In the guiding center limit Curvature drift Slingshot term (Fermi reflection) increases the parallel energy Grad B drift dε dt = qv E + q v c E + µ B t + q v B E v c = v 2 Ω b ( b b) Betatron acceleration increases perpendicular energy µ conservation 2 v B = v 2Ω b B B
Electron heating during reconnection 2-D PIC simulations of electron-proton system with a weak and strong guide fields (0.2 and 1.0 times the reconnection field) 819.2d i x 409.6d i Compare all of the heating mechanisms Dahlin et al 14
Electron heating mechanisms: weak guide field Slingshot term dominates (Fermi reflection) Parallel electric field term small a surprise Grad B term is an energy sink Electrons entering the exhaust where B is low lose energy because µ is conserved. Dahlin et al 2014
Electron heating mechanisms: strong guide field Fermi and parallel electric field term dominate Longer current layers where E 0 with a guide field
Acceleration mechanism for highest energy electrons Fermi reflection dominates energy gain for highest energy electrons Where v c ~ v 2 dε dt ~ qv E + q v c E Recent simulations of pair and relativistic reconnection also see the dominance of Fermi reflection (Guo et al 14, Sironi and Spitkovsky 14) E V c
The spatial distribution of electron heating Tradition paradigm based on fluid turbulence that dissipation takes place at small scales is wrong in weakly collisional systems Phys. Plasmas 21 Electron heating rate from Fermi Electrons gain energy in the entire reconnection exhaust
Electron-ion Energy Partition: the magnetosphere Where does the released magnetic energy go? Available magnetic energy per particle from Poynting flux W 0 = 1 2 B up n up 4π = m 2 ic Aup Magnetopause enthalpy flux observations (Phan et al 13, 14) ΔW ΔW e = 5 2 ΔT = 0.043W i = 5 e 0 2 ΔT = 0.33W ΔW i 0 flow = 0.5W 0 Parallel heating exceeds perpendicular heating Magnetotail observations (Eastwood et al 13) Ions carry most of the released magnetic energy Electrons and Poynting flux are smaller
Ion heating mechanism: single x-line Ion energy gain from Fermi reflection leads to large parallel heating of ions Measured throughout the magnetosphere For C A ~ 2000km/s have T ~ 25keV Measured scaling of ion temperature consistent with Fermi reflection (Phan et al 2014) 2 ΔT i ~ 0.13m i c A Hoshino et al 98 Gosling et al 05 Phan et al 07 Smaller than expected. Why? C A ΔT i = 1 3 m c 2 i A exhaust RD RD
Electron heating mechanism: single x-line Magnetosphere observations ΔT e = 0.017m i c A 2 Fermi reflection of electrons Single pass yields energy increment ~ m e v 0 c A is too small to explain observations How do the electrons gain so much energy?
tion parameters are given in Br 1.0 1.0 1.0 2.236 2.236 2.236 1.0 2.236 0.447 1.0 1.0 1.0 n Te Ti 0.2 0.25 0.25 0.2 2.25 0.25 0.2 0.0625 0.3125 0.2 2.25 0.25 0.2 2.25 The 0.25 development of a large scale potential boosts electron 0.2 0.25 1.25 heating 0.2 0.0625 0.3125 0.2 1.25 A1.25 large-scale potential develops to keep hot electrons in the exhaust 0.2 0.05 from 0.05 escaping upstream (Egedal et al 08) 0.2 0.25 0.25 0.2 0.0833 0.25 0.2 0.25 0.0833" % A large scale potential controls the relative heating of electrons and ions n Δϕ ~ Te ln $$ exhaust '' nup & # mass rameters. Electron to ion grid spacing x and upstream plane) magnetic field Br, deneratures Te and Ti. The potential holds in electrons and enables them to undergo multiple Fermi reflections a double current sheet[16]. ion is used to initiate recons evolved until reconnection d then for analysis purposes Haggerty et al e simulation data is time ave steps, which is typically on 1 ma wave periods ωpe. effects of the parallel poten, the average ion heating is 2105
The potential suppresses ion heating In the frame of the exhaust ions move inward at C A Ion velocity is reduced by the potential to V d 1 2 m V 2 i d = 1 2 m C 2 i A eϕ C A shock V d V d exhaust shock
Scaling of electron and ion heating The total energy gain of electron plus ions has a universal scaling The partition of energy going to electrons and ions can vary Red triangles have large upstream T e Larger potential Single x-line reconnection splits released energy between electrons, ions and bulk flow For B ~ 50G, with n ~ 10 9 cm -3, obtain T hot ~ 15keV
SPP observations Reconnection at 10R s will differ significantly from at 1AU Low β compared with 1AU Reconnection with a strong guide field? Not well explored at 1AU Direct measurement of reconnection exhausts from coronal and chromospheric reconnection? Solar wind observations have revealed that reconnection exhausts extend long distances Very exciting possibilities for exploring the role of nano-flares in coronal heating
Conclusions Fermi reflection dominates electron energy gain during magnetic reconnection in flares Strong anisotropy with T >> T Single x-line reconnection Electrons undergo multiple Fermi reflections in a reconnection exhaust confined by a large-scale potential The same potential reduces ion energy gain The temperatures increases to ~ 10keV Multi-island reconnection Equal-size merging islands increase electron energy by a factor of two After N mergers, energy gain by 2 N The relativistic electrons in ~ 10s
Electron spectra reveal strong anisotropy With a guide field the dominant acceleration mechanisms accelerate electrons parallel to the local magnetic field Fermi reflection and E Extreme anisotropy in the spectrum of energetic electrons More than a factor of 10 3 solid parallel dashed - perp t
Electron heating: dependence on the guide field Fermi reflection dominates for weak guide field E dominates for strong guide field B r /B g