turbulence, shocks and particle accelerationhamish/stfc_ss15/horbury_turbulence.pdf · turbulence,...
TRANSCRIPT
Turbulence, shocks and particle acceleration
The solar wind as a plasma laboratory• Global morphology• Typical parameters• Measuring the solar wind
Turbulence• What is turbulence?• Hydrodynamic turbulence• Turbulence in collisionless plasmas• Open questions
Shocks– Conservation laws at shocks– Shocks in collisionless plasmas– Open questions
Particle acceleration– Acceleration at shocks– First order Fermi acceleration– Particle transport– Open questions
1
Tim Horbury, Imperial College London
Thanks to Chris Chen, Javier Pacheco
The solar wind as a plasma laboratory
• Collisionless, magnetised plasma• Continual, but variable, outflow from Sun’s
corona• Carries magnetic field, waves and turbulence
from corona• A collisionless plasma we can sample
directly
• Supersonic (super-Alfvénic, …)• Hot: >105 K• Rarefied: few per cm3 at Earth• Complex due to solar variability,
solar rotation, and in situ processes• Variable on all measured scales, from sub-second to centuries
2
SoHO/LASCO
Measuring the solar wind
• Spacecraft carry instruments to measure key plasma properties
• Bulk plasma parameters– Density, velocity, temperature
• Fields– Magnetic field, electric field
• Particle distribution functions– Essential for kinetic studies
• Energetic particles– Different energy ranges
• These measurements are hard – typically 1% accurate, but coverage in time, cadence, direction, energy, etc. is never as good as we would like
• Every measurement is wrong; the harder you push the data, the more you need to know about how the instruments work– Instrument teams love talking about their instruments; ask them
3
This is DSCOVR, launched 2015
Solar wind at 1 AU
• Number density: few per cm3
• Mostly protons and electrons, few percent helium, little bits of other things
• Bulk speed: few hundred km/s (~1keV for a proton)
• Proton thermal speed: ~ 50 km/s– Cold beam
• Plasma beta: ~1• Magnetic field: few nT• Alfven Mach number: a few
• Lots of exciting kinetic effects….
4
5
The real solar wind
Turbulence
6
da Vinci 1510
7
What is turbulence?
• Fluid phenomenon• Nonlinear energy transfer between
scales• Occurs when inertial forces dominate
viscous forces• Important in many engineering
problems
The Richardson cascade
Bigger whirls have little whirls,That feed on their velocity;
And little whirls have lesser whirls,And so on to viscosity.
Lewis Fry Richardson, 1920
Turbulence in space plasmas8
McComb 1995 RPP
9
The inertial range
• If energy input is steady, and far from dissipation scale, have a steady state Inertial range
• K41: k-5/3 spectrum
• We observe this in hydrodynamic fluids
• Inertial forces dominate viscosity• Reynolds number is scale dependent• At some scale, viscosity dominates
– Dissipation
Energy input Inertial range
Dissipation
Anisotropic MHD turbulence10
Turbulence in plasmas
Neutral fluids• Motion described by Navier-Stokes
equations Hydrodynamics• Energy transfer by velocity shearPlasmas• On sufficiently large scales, can treat
plasma as a fluid Magnetohydrodynamics• Multiple, finite amplitude waves can be
stable• Presence of a magnetic field
– Breaks isotropy– Key difference to neutral fluids
“The great power law in the sky”
• Measure interstellar density fluctuations using scintillations
• Consistent with Kolmogorov scaling over many orders of magnitude
Turbulence in space plasmas11
Turbulence in the solar wind12
Why study waves and turbulence in the solar wind?
Effect on the Earth• Can trigger reconnection, substorms,
aurorae, …Understanding solar processes• Signature of coronal heating, etc.Application to other plasmas• Astrophysics: particle propagation• Dense plasmas: transportTurbulence as a universal phenomenon• Comparison with hydrodynamics
D. Vier/SoHO/Hubble/Dimotakis et al
Turbulence13
Interpreting spacecraft measurements
• In the solar wind (usually),VA ~50 km/s, VSW >~300 km/s
• Therefore,VSW>>VA
• Taylor’s hypothesis: time series can be considered a spatial sample• We can convert spacecraft frequency f into a plasma frame wavenumber k:
k = 2f / VSW
• Almost always valid in the solar wind• Makes analysis much easier• Not valid in, e.g. magnetosheath, upper corona
Turbulence14
Interpreting spacecraft measurements
• Solar wind flows radially away from Sun, over spacecraft• Time series is a one dimensional spatial sample through the plasma• Measure variations along one flow line
Flow
Turbulence15
Alfvén wavesField-parallel Alfvén wave:• B and V variations anti-
correlatedField-anti-parallel Alfvén
wave:• B and V variations
correlated
• See this very clearly in the solar wind
• Most common in high speed wind
16
Average magnetic field sunwardPositive correlationPropagating anti-parallel to fieldPropagating away from Sun in plasma frame
Propagation direction of Alfvén waves
• Waves are usually propagating away from the Sun
Average magnetic field anti-sunwardNegative correlationPropagating parallel to fieldPropagating away from Sun in plasma frame
17
Dominance of outward-propagating waves
• Solar wind accelerates as it leaves the corona
• Alfvén speed decreases as field magnitude drops
• Alfvén critical point: equal speed (~10-20 solar radii)
• Above critical point, all waves carried outward
Therefore,• Outward-propagating low
frequency waves generated in corona!
Distance from Sun
Speed Solar wind speed
Alfvén speed
Critical point
Turbulence18
Active turbulent cascade in fast wind
• Bavassano et al (1982)• Fast wind: “knee” in
spectrum• Spectrum steepens
further from the Sun• Evidence of energy
transfer between scales: turbulent cascade
after Bavassano et al 1982
MHD turbulence
19
Turbulence20
Field-aligned anisotropy
• Power levels tend to be perpendicular to local magnetic field direction
• anisotropy
• Dots: local minimum variance direction
• Track large scale changes in field direction
• Small scale turbulence “rides” on the back of large scale waves
Anisotropic MHD turbulence21
Anisotropy of energy transfer
Neutral fluid• No preferred direction
isotropyPlasmas• Magnetic field breaks symmetry
anisotropy
• Shebalin (1983): power tends to move perpendicular to magnetic field in wavevector space
‘hydro-like’ region
Magnetic fieldk||
k
Critical balance• MHD turbulence has two timescales
– Alfven timescale: time for two wavepackets to decorrelate
– Nonlinear timescale: time for eddies to decay and transfer energy
• “Weak” MHD: Alfven time is faster, 3/2 spectrum (Kraichnan, Iroshnikov)
• “Strong” MHD: Nonlinear time is faster, 5/3 isotropic spectrum
• “Critical balance” (Goldreich and Sridhar, 1995): Two timecales balance on every scale
• Predicts 5/3 field-perpendicular spectrum, 2 field-parallel
• Consistent with observations…
Turbulence in space plasmas22
after Wicks et al. 2010 MNRASafter Wicks et al. 2010 MNRAS
k||-2
k⊥-5/3isotropic
anisotropic
inertial range
Anisotropic MHD turbulence23
Anisotropic energy transfer
Eddies• On average, tend to become smaller
perpendicular to field• Results in long, fine structures along the
magnetic field
k||
k
Wavevectors• Energy tends to
move perpendicular to magnetic field
Hydrodynamics
MHD
Turbulence beyond MHD
• Collisionless plasmas have no real viscosity
• What happens at the proton gyroscale?
• We observe a spectrum at scales below the ion gyroscale
• It seems that there is an additional cascade to smaller scales
• Seems to be kinetic Alfven waves but this is still an open question
24
Alex
andr
ova
et a
l. 20
09 P
RL
How
es e
t al.
2011
PR
L
E and B spectrum in the kinetic regime
• MHD: E=-VxB
• Not so on kinetic scales
• Evidence for kinetic Alfven waves?• Bale, 2005
• See also: Galtier, Hall MHD
Turbulence in space plasmas25
Collisionless plasma turbulence: open questionsMHD scales• How does the magnetic field affect the MHD cascade?• How are compressive fluctuations related to the cascade?• How does the balance of Alfvenic fluctuations affect the cascade?
Kinetic scales• What is the nature of the cascade below the proton gyroscale?• Is energy removed at this scale?• What happens at the electron gyroscale?
How universal is turbulence between neutral fluids and plasmas?
26
Shocks
27
What is a shock?
• Occurs when velocity change is larger than wave speed
• Acts to decelerate and deflect supersonic flow around an obstacle
• Compresses and heats the fluid
• Fluid phenomenon• But, shock transition is kinetic, not
fluid
Shocks and waves: information speed
• Body moving through fluid• Waves propagate upstream:
send information to deflect flow
• If body moves faster than waves, something else must propagate information – shock
• Fundamentally nonlinear• Related to wave mode
• Mach number: ratio of inflow speed to sound speed
• Plasmas: 3 wave speeds, so 3 Mach numbers
How do shocks form?
30
• Consider a sound wave in an Ideal gas (adiabatic)• Compressional wave – perturbation in pressure P
where
• Small P • No significant change in cs within wave• Wave profile does not evolve
• Large P• Wave steepens• Shock forms• Mach number
• shock strength
Pcs vp CC
sflow cVM
What happens at a shock?
31
• Consider ideal “thin” shock
• Propagating/standing shock?• Work in rest frame of shock
• Mass IN = Mass OUT• Momentum IN =Momentum OUT• Energy IN = Energy OUT
• Internal distribution of energy given by• Equation of state
• Easy for Ideal gases nkTPV kTP
Shocks: fluid view
• Take a fluid approximation– Mass density – Flow velocity v– Pressure p (isotropic)– Assume static, 1D shock– Assume perfect gas– Assume fluid reacts adiabatically
32
Rankine-Hugoniot relations
33
Hydrodynamic shocks: kinetic view
• Where does the kinetic energy go?– Heating
• Collisions rapidly thermalise the particles
• Shock thickness around the collisional mean free path
34
Shocks Space Physics 2009
Astrophysical shocks• Supernovae• Jets• Bowshocks• Blast waves
Shocks in the solar system
36
• Planetary bow shocks
• CME and CIR shocks
• Termination shock
• Heliospheric bow shock (maybe not?)
Plasma shocks
• More complex than hydrodynamic shocks– Multiple wave modes possible
• Alfven mode shock: no compression, “discontinuity”• Fast mode shock: compressive• Slow mode shock: compressive
• Presence of magnetic field• Interactions via charges and fields – not just collisions
• Collisionless shocks possible– Particles can travel far upstream– Must more complex behaviour– Energy partitioning
37
MHD Rankine-Hugoniot relations
38
0
0
0
tntn
n
n
uBu
B
u
B
0
02
012
0
0
22
00
22
tn
tn
n
nnn
Bu
Bpu
BBupuu
Bu
Bu
Mass conservation
Continuity of Bn
Continuity of Et
Energy conservation
Momentum conservation || n
Momentum conservation n
• R-H relations true for all discontinuities, not just shocks• If inflow velocity → subsonic;; solutions revert to MHD waves: fast, slow, Alfvén
Fast and Slow Mode Shocks
• MHD shocks retain characteristics of linear wave modes• Alfven mode
– not compressive so no shock (in isotropic plasma)– rotational discontinuity
• Fast mode– density and |B| correlated– so B turns away from normal– occurs more commonly
• Slow mode– density and |B| anti-correlated– so B turns towards normal– occurs less commonly
Collisionless shocks: magnetic field angle
Quasi-perpendicular• Ions gyrate back into
shock• Sharp transition
Quasi-parallel• Ions travel upstream• Unstable: generate
waves• Extended, messy shock
Shock
Magnetic field
Ions gyrate into shock
θBN largequasi-perpendicular
Shock
Magnetic field
Ions escape
θBN smallquasi-parallel
Quasi-perpendicular shocks (Bn~90°)
41
VSWVMS
BMS
BSW
ESW
ESh
dcs
dsh df~0.68 Vsw/ci
jsh
je jf
|B|
t
Overshoot Ramp Foot
Reflected ionsGyrating ions Figure after Baumjohann and Treumann. 1997(not to scale)
Waves
42
Different types of Q shock: the importance of Mach number
• Low Mach number: 1- 3 e.g. Interplanetary shocks: – Sharp transition; little overshoot– Sometimes see clear dispersive
whistler leading shock ramp
• Moderate to High Mach number:. 3-10 e.g.Terrestrial bow shock:– Foot; Ramp; Overshoot;
Downstream wave decay– Ion reflection plays essential
role
43
Shock variability• Simulations show non-stationary
structure– Important for particle heating?– Reformation of the ramp– Rippling of the shock front
Lembege and Savioni, 1992
Lembege et al, 2004
44
The Earth’s bow shock• Curved geometry
– Shock orientation (Bn) varies• Electrons and ions reflected/escape to
form foreshock– Morphology depends on particle
energy, cross-field drift speed– Not steady state but time varying
• Upstream particles generate waves– ULF wave foreshock
• ULF waves convect back towards the shock – Potential to modify shock
geometry– Seen in modulation of reflected ion
populations• Rankine-Hugoniot only true if include
foreshock From Treumann and Scholer, 2001
45
Generation of ion beams Kucharek et al, Ann. Geophys., 2004
• Ion beams originate in the shock ramp – they do not leak from the magnetosheath• The density of the ion beam is higher for high Mach number shocks• Gyrating ions and ion beams come from same population
– Specular reflection not sufficient for ions to escape • scattering/multiple shock interactions?
– Predicted by simulations [Burgess, Ann. Geophys., 1987]
The termination shock
• Very large radius of curvature• Most similar to astrophysical
shocks?• How different is this to
planetary bowshocks?• The Voyager spacecraft have
now sampled this shock
46
Termination shock
• ~MeV ions rise by a factor of ~10 for months before shock crossing
• Solar wind slows down• Mass loading by energetic particles
– A “cosmic ray modulated shock”?• Probably more similar to many
astrophysical shocks
Florinski et al., 2009
47
Energy partitioning
• Where does the energy go?• Proton heating
– Temperature anisotropy?• Electron heating
– Temperature anisotropy?• Supra-thermal particles• Heavy species: helium, …
• Can depend on θBN, Mach number, …
48
Particle acceleration
What are energetic particles?
• Particles with kinetic energies significantly above that of the bulk plasma• They must somehow be accelerated out of the thermal population
Sources• Galaxy
– Supernova shocks, etc.• Sun
– Solar flares• Interplanetary shocks• Pick-up ions
– Interstellar neutrals
• Note: we only detect energetic particles because the interplanetary medium and interstellar medium are collisionless
50
Galactic cosmic rays
51
http://www.physics.utah.edu/~whanlon/spectrum.html
• We measure fast ions and electrons up to relativistic energies
• These are “cosmic rays”• Flux is modulated by solar cycle
Energetic particles in the heliosphere
52
• Also see energetic “tails” on bulk plasma distributions
• Particles are accelerated within the solar system– “Solar energetic particles”
• Highest energy events (rare) reach ~GeVand are seen as cosmic rays– “ground level events”
• Where does this particle acceleration happen?
ACE/SWICS
Solar energetic particles
• Energy in SEPs can reach 10% of the total energy of a coronal magnetic ejection
• Cant be ignored as part of the event
• SEPs flood the heliosphere
53
Mewaldt et al., 2007
Impulsive vs gradual events
54
•Protons •Protons
Parent solar activity PROTONS
Impulsive event
Gradual event
shock
Acceleration at shocks in the solar wind
• Interplanetary shocks can accelerate particles locally
• In some CMEs, around 10% of the total kinetic energy is in suprathermal particles
• Particles can also be accelerated nearer the Sun, and in flares
• Bothmer et al., GRL, 1995
55
Acceleration at shocks: cartoon
• Particles bounce of scattering centres either side of the shock• These scattering centres are converging: particles accelerate• Scattering centres are “turbulence”
56
Marc Pulupa
Fermi acceleration
First order Fermi (“diffusive shock acceleration”)• Converging scattering centres around shock• Elastic scatters result in energy gain• Steady state spectrum is a power law function of shock compression ratio• Spectral index of 2 for a strong shock (close to observed)
Second order Fermi• Particles scatter off randomly moving objects (e.g.”gas clouds”)• More likely to hit one coming towards you, so on average gain energy• Statistical• Spectrum quite similar to first order Fermi
• Widely regarded as broadly correct mechanism• Injection problem: how do particles get out of the thermal population?
57
Hot ions upstream of the bow shock
First order Fermi-acceleration:
• Particle history:1. Crosses shock2. Scatters off density irregularities
(waves, turbulence) in sheath3. Moves upstream4. Scatters again… from pulsations…?5. Sees upstream and downstream
irregularities are converging at uu-ud
6. Gains energy
58
Away from shock
shock
uu ud
ion
Shock normal angle: perpendicular is faster
59
Giacalone
Particle transport
• Particles travel from acceleration site to spacecraft
• See dispersion signatures from events occurring near the Sun
• Scattering can occur along the route
60
Klecker and Kallenbach, 2003
Particel dropouts: hints about transport
61
Mazur et al., 2000, APJ Ruffolo et al., 2003, APJ Lett.
Particle acceleration: open questions
Injection problem• How do we get particles out of the thermal population?
Where are SEPs accelerated?• Flares? Near-Sun shocks? Interplanetary shocks?• Need to go close to the Sun within one scattering mean free path
How representative are planetary bowshock to astrophysical shocks?• Curvature, Mach number, …
Is the termination shock our first astrophysical shock?
62
Summary
Turbulence• Importance of the magnetic field• The kinetic cascade
Shocks• Partitioning of energy between species• Importance of curvature
Particle acceleration• Injection problem• Relating solar system to astrophysical shocks
63
References: turbulence
• Books– Landau & Lifshitz (1959) “Fluid Mechanics”– Frisch (1995) “Turbulence”– Biskamp (2003) “Magnetohydrodynamic Turbulence”
• Review Papers– Schekochihin & Cowley (2007) “Turbulence and Magnetic Fields in
Astrophysical Plasmas”– Alexandrova et al. (2013) Space Sci. Rev. “Solar Wind Turbulence and the
Role of Ion Instabilities”– Bruno & Carbone (2013) Living Rev. Sol. Phys. “The Solar Wind as a
Turbulence Laboratory”
64
References: shoks
• Books– Landau & Lifshitz (1959) “Fluid Mechanics”– Burgess (1995) “Collisionless Shocks” in Kivelson & Russell “Introduction
to Space Physics”
• Review Articles– Treumann (2009) Astron. Astrophys. Rev. “Fundamentals of collisionless
shocks for astrophysical application, 1. Non-relativistic shocks” – Krasnoselskikh et al. (2013) Space Sci. Rev. “The dynamic
Quasiperpendicular Shock: Cluster Discoveries”– Burgess & Scholer (2013) Space Sci. Rev. “Microphysics of Quasi-parallel
Shocks in Collisionless Plasmas”
65