Magnetic Reconnection

Transcription

1 Magnetic Reconnection J. Egedal In collaboration with Joe Olson, Cary Forest and the MPDX team UW-Madison, WI Les Houches, March, 2015

2 Madison Plasma Dynamo experiment 2

3 Key new hardware for TREX Cylindrical Insert: Houses internal coils to drive reconnection Asymmetric, symmetric reconnection and strong guide field reconnection To be pulsed at a 10 s rep rate Hugely flexible facility User Facility 3

4 Magnetic Reconnection A change in magnetic topology in the presence of a plasma Consider a small perturbation Plasma carrying a current Magnetic fields j

5 Magnetic Reconnection A change in magnetic topology in the presence of a plasma Consider a small perturbation

6 Magnetic Reconnection A change in magnetic topology in the presence of a plasma Consider a small perturbation Nearly all the initial magnetic energy is converted into: 1. thermal energy 2. kinetic energy on fast electrons and ions 3. kinetic energy of large scale flows

7 Coronal Mass Ejections The most powerful explosions in our solar system Can power the US consumption of electricity for 10 million years

8 Outline More Pretty Pictures Models for Reconnection Role of Pressure Tensor Electron Heating TREX, the Terrestrial Reconnection EXperiment ( The Reconnection EXperiment) Conclusions

9 Coronal Mass Ejections Movie from NASA s Solar Dynamics Observatory (SDO)

10 Outstanding Problems Arcade Arcade as seen from above Heating 3D effects Trigger

11 Space Weather The Solar Wind affects the Earth s environment

12 The Earth s Magnetic Shield During Before reconnection

13 Magnetic Storms

14 Aurora Borealis October 26 th, 2011, Nantucket Island, Massachusetts, USA

15 Aurora Borealis October 26 th, 2011, Kola Peninsula, Russia

16 Carrington Flare (1859, Sep 1, am 11:18) Richard Carrington (England) first observed a solar flare in White flare for 5 minutes. Very bright aura appeared next day in many places on Earth including Cuba, the Bahamas, Jamaica, El Salvador and Hawaii. Largest magnetic storm in recent 200 years (> 1000 nt). Telegraph systems all over Europe and North America failed, in some cases even shocking telegraph operators. Telegraph pylons threw sparks and telegraph paper spontaneously caught Fire. (Loomis 1861)

17 Magnetic storm and aurora on March 13, that lead to Quebeck blackout (for 6 million people) Magnetic storm ~ 540 nt, Solar flare X4.6. A Carrington Flare today billion dollars of damage

18 Occurrence frequency of flares? Nanoflares 1000 in 1 year 100 in 1 year 10 in 1 year 1 in 1 year 1 in 10 years 1 in 100 years 1 in 1000 years 1 in years Solar flares New Kepler data from 83,000 stars for 120 days Largest solar flare C M X X10 X1000 X100000

19 The Magnetosphere as a Laboratory

20 MMS Successfully Launched! Magnetospheric Multiscale Missing March, 12, 2015

21 Electromagnetism 101 Faraday s law: EMF Area Faraday s law for a conducting ring: EMF=0. db dt The magnetic flux through the ring is trapped This also holds if the ring is made of plasma plasma frozen in condition

22 Magnetic Topology Constant in Ideal Plasma Ideal Plasma E ' E v B 0 Plasma and B frozen together B B Ideal MHD: E B=0, Excellent for 99.9% of all plasmas, 99.9% of the time.

23 Reconnection: A Long Standing Problem

24 Reconnection: A Long Standing Problem Simplest model for reconnection: E + v B = j [Sweet-Parker (1957)] t X E X j X

25 Reconnection: A Long Standing Problem Simplest model for reconnection: E + v B = j [Sweet-Parker (1957)] L Outflow speed: (Alfven speed) Sweet-Parker: L >> : t sp t R t A 2 0L L v A Unfavorable for fast reconnection Two months for a coronal mass ejections

26 Things Good to Know: CGL Model Consider flux tube with reflecting boundaries: A 0 A Conserved quantities: p = n B p 0 n 0 B 0 l 0 l J = p 2 = n l 0 p 0 n 0 l 2 = n3 2 B 0 3 n 0 B 2 p n B 3 2 p nb Flux= A B # Partic = n A l l = n 0 l 0 n B B 0 CGL-scalings [Chew, Goldberger, Low, 1956]

27 Things Good to Know: Whistlers

28 Plasma Kinetic Description The collisionless Vlasov equation: + Maxwell s eqs. Vlasov-Maxwell system of equations Can be solved numerically (PIC-codes)

29 Fluid Formulation (Conservation Laws) mass: momentum: Sweet-Parker energy: Isotropic (scalar) pressure is the standard closure! p = n T Add Maxwell s eqs to complete the fluid model...

30 Kinetic Simulation Isotropic pressure Two-Fluid Simulation GEM challenge (Hall reconnection) E + v B = (j B)/ne [Birn, Drake, et al. (2001)] Out of plane current Aspect ratio: 1 / 10 v in ~ v A / 10 Particle In Cell (PIC) simulation, By W. Daughton

31 Hall Effect documented in Simulations and in MRX

32 Kinetic Simulation Isotropic pressure Two-Fluid Simulation vs. Kinetic Out of plane current Particle In Cell (PIC) simulation,

33 Electron Trapped by, B g = 0.4 e / T e E E E E

34 Pressure Anisotropy WIND Spacecraft Observations in Magnetotail, 60R E Measurements within the ion diffusion region reveal: Strong anisotropy in f e p > p 34

35 Wind Spacecraft Observations Measurements within the ion diffusion region reveal: Strong anisotropy in f e p > p E A 1 E A 2 Simulations: Ion density near uniform Quasi neutrality: n e n i A 2 A 1 : Parallel compression by E Trapped electrons dominate, Zero heat transport CGL: p n 3 /B 2, p nb Jan Egedal

36 Electrons in an Expanding Flux Tube Trapped electron Magnetic moment: m 2B mirror force: v 2 Passing Trapped Trapped Passing

37 Electrons in an Expanding Flux Tube Trapped: Passing: Drift kinetic equation: J. Egedal et al., JGR (2009) J. Egedal et al., POP (2013)

38 EoS Implemented in Two-Fluid Code Moments over kinetic model yields fluid closure with anisotropic pressure, EoS, [Le et al., PRL 2009] EoS implemented by O Ohia using the HiFi framework developed in part by VS Lukin Boltzmann CGL Standard two-fluid equations Anisotropic pressure model Jan Egedal Les Houches, March,

39 Kinetic Simulation Anisotropic pressure Isotropic pressure New EoS Implemented in Two-Fluid Code Ohia et al., PRL, 2012 Out of plane current Jan Egedal MPPC meeting, Princeton, March,

40 Threshold for Guide-Field Reconnection Kinetic simulation results at m i /m e = 1836, [A Le et al., PRL 2013]

41 Regimes of the Electron Diffusion Region J y in PIC simulations at m i /m e = 1836 B g /B r (4) (3) (2) (1) x/d 5 i B g /B r Regimes of reconnection vs. m i /m e and B g 0 (4) (3) (2) (1) m i /m e Jan Egedal Unexplored regime of reconnection, relevant to the MMS mission Les Houches, March, 2015 [Le et al., PRL, 2013] 41

42 Role of Collisions for Pressure Anisotropy B g /B 0 =0.28 m i /m e = 1836 Collisionless if ei 0.1v A > di or Jan Egedal Les Houches, March,

43 Role of Collisions: Condition for anisotropy S > 10 4 L d i TREX Daughton W and Roytershteyn V, (2012) Space Science Reviews 172, Ji H and Daughton W, (2011) Phys. Plasmas 18, Jan Egedal Les Houches, March, 2015

44 EoS for anti-parallel reconnection? The electrons are magnetized in the inflow region: z/d e z/d e z/d e x/d e Jan Egedal Les Houches, March, 2015

45 x/d Electron distributions in the layer [J. Ng et al., PRL2011] z/d e z/d e

46 EoS for anti-parallel reconnection? The electrons are magnetized in the inflow region: z/d e z/d e Momentum balance: With CGL: z/d e x/d e 46 Jan Egedal Les Houches, March, 2015

47 Estimates of in the Magnetotail Magnetotail data, 18 events

48 Breakdown of Parallel Adiabaticity Magnetotail data, 18 events v A 1 v A 2 Expect non-adiabatic behavior if v e > v te or Including role of E for driving flows (kinetic model) 0.02

49 Simulation with e ~ d i long, 180 billion particles stays high in the exhaust electron heating?

50 Simulation with e ~ d i long, 180 billion particles!

51 Strong Double Layer Instability Observed

52 confines electrons, further energized by E Trapped region with pitch angle scattering Heated by v d E, Flat-top or

53 Generation of Super-Thermals log 10 ( # of electrons with energies above 1.7mc 2 ) log 10 [ f (E) ] Continuity Eq. for electrons with v 1 < v < v 2 : Steady state:

54 Model applicable to solar flares? Ohm s law: Before reconnection: p = nt e e ~ T e log(n/n 0 ) During reconnection: e 100T e yielding bulk energization Loop top confinement by E explored by [Li, Drake, et al., 2012,2013,2014] Superthermal tail generated by v d E log 10 [ f (E) ] Advantages: Role of free-streaming losses addressed The model is simple, and works with one or more X-lines

55 Log 10 (P /P EOS ) Log 10 (P /P 0 ) Trigger of Reconnection could be Related? New Giant Simulation with B g =1:

56 Requirements for Experiment Large normalized size of experiment: L di ~ 10 (high n, large L) Low collisionality to allow p >> p : ei 0.1vA > di (low n, high T e, high B) Low electron pressure: e < 0.05 (low n, T e, high B) Manageable loop voltage: 0.1v A B rec (2 R) < 5kV (high n, low B) Variable guide field: B g = 0 4B rec Symmetric inflows Experimental window available in Hydrogen or Helium plasma with n ~ m -3, T e ~ 15 ev, B rec ~ 15 mt, L ~ 2 m 56

57 Asymmetric reconnection in TREX Simple configuration using the HH-coils plus two internal coils Low plasma This will be the first configuration to be implemented High plasma Jan Egedal

58 Experimental Setup Joseph Olson

59 Reaching Target Plasma Helium plasma 60G Helmholtz field -400 V cathode bias 59

60 Preliminary Results ~44.5 μs after second pulse Jan Egedal 60

61 Strong Guide-field Reconnection Strong guide-field in TREX Pulsed operation of magnetic coils B g = 0.25T at 1m 3 min between pulses Optimize configuration for maximal length of reconnection layer to explore turbulent reconnection

62 Strong Guide-field Reconnection Pulsed operation of magnetic coils 3 min between pulses B g = 0.25T at 1m FLARE Jan Egedal ρ s = (2m i T e ) 1/2 /eb = ion sound Larmor Radius

63 Conclusion Reconnection is still purely understood, but the pressure tensor is important The construction of the new Terrestrial Reconnection EXperiment is well under way TREX provides huge flexibility in available configurations, and the insert will allow for fast turn-around. MPDX/TREX in the process of becoming a user facility Jan Egedal