MURI REVIEW - University Of Marylandspp.astro.umd.edu/SpaceWebProj/RESEARCH PAGE PDFs n pics/1...
Transcript of MURI REVIEW - University Of Marylandspp.astro.umd.edu/SpaceWebProj/RESEARCH PAGE PDFs n pics/1...
1
PARTICIPATING UNIVERSITIESUNIVERSITY OF MARYLAND, COLLEGE PARK
STANFORD UNIVERSITYUNIVERSITY OF CALIFORNIA, LOS ANGELES
DARTMOUTH COLLEGEVIRGINIA TECH
BOSTON COLLEGE
MURI REVIEWMarch 3, 2008
Dennis Papadopoulos
2
• DENNIS PAPADOPOULOS• ROALD SAGDEEV• GENNADY MILIKH• XI SHAO• NAIL GUMEROV• WALLY MANHEIMER• GLEN JOYCE• LEONID RUDAKOV• BENGT ELIASSON*• ANDREW DEMEKHOV*• OLEG POKHOTELOV*• ALEXEY KARAVAEV**• HIRA SCHROFF**• BIRU TESFAYE**• LUKE JOHNSON*** Visitor** Student
THE MURI TEAM - UMCP
3
HAARP DEMETER
DMSPCONJUGATE BUOYS
LAPD
WIDE RANGE OF CODES THAT COUPLE TO THE ABOVE EXPERIMENTS
RESOURCES
4γ
γω
/
/
ezz
ezz
Nvk
Nvk
Ω=−
Ω=− ω − kzvz = ±Ωp
kzvz ≈ Ωp
Parallel Wave Number
VLF
ULF
Relevant Wave Modes
5
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSU
EMICUM
VLF
ULF Alfven WavesUM
ProtonsPrecipitation Electrons
VLF/ULF
WIPPSU
6
Science Issues:The sheath surrounding an electric dipole antenna operating in a plasma has a significant effect on the tuning properties.
Terminal impedance characteristics vary with applied voltage.
Active tuning may be needed.
Stanford will use existing Antenna-In-Plasma (AIP) code to determine sheath effects on radiation process and validate using UCLA LAPD.
MURI Task Status:Stanford group has made a number of visits to the UCLA LAPD to assist in setting up the experiments necessary to validate the AIP code.
Preliminary impedance and pattern measurements have been obtained.
Stanford has begun simulating near-field properties of dipole antennas operating in conditions representative of LAPD environment.
AIP Code Validation (T. Chevalier)
7
Wave Injection at Low Latitudes
Investigate wave injection from Russian Alpha Navigation transmitter in Komsomolsk, Russia
Observe 1-hop signals at Conjugate point in Adelaide Australia
Results show growth and variation with geomagnetic conditions
Alpha Field Tests (M.Golkowski)
8
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFUM - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSU
EMICUM
VLF
ULF Alfven WavesUM
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
10
Why should one care? Necessary for VLF triggering
APPROACH: Compare plasmaspheric structures with electric fields in the ring current-plasmasphere overlap (Sub Auroral Polarization Streams & Ion Drifts)
Simulated equatorial density distributioncontaining a SAID generated trough
Simulated EUV image
Plasmaspheric ducts created by the SAPS Wave Structures (left) during the substorm ring current injection event from CRRES
RC ions
Plasmaspheric ducts formation (E. Mishin)
11
2D MODELING SHOWING S THAT TRANSIONOSPHERIC DUCTS WITH δn/n> .5 FORM IN 15 MINUTES WITH FULL HAARP F-REGION HEATING
Artificial Ducts Driven by F-Region Heating (G. Milikh)
13
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSU
EMICUM
VLF
ULF Alfven WavesUM
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
14
APPROACH
Compare the occurrence of VLF triggering from the Siple transmitter with the magnetic activity and background hiss and chorus emissions
Database: Siple June 1986 campaign
•Most favorable conditions for VLF triggering in the morning sector seem to be satisfied after weak/moderate substorms.
•The triggering occurred when broad-band hiss emissions were present and the pump frequency was above the top of the hiss band.
•Consistent with the step-like background electron distribution.
Moderately-strong hiss creates a sharp increase (step) on the electron distribution function (Trakhtengerts & Co).
dBWavelet on the top of the hiss band is strongly amplified
Negative Energy Wave
TOWARD PREDICTING VLF TRIGGERING(E. Mishin & A. Gibby )
15
New simulation code describes dynamics of propagation and amplification of 1 kHz (20 μscec pulse) along entire L=4.2 line with spatial resolution of 1 km.
How to maximize injected whistler amplitude
Propagation/Amplification (A. Streltsov)
16
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSU
EMICUM
VLF
ULF Alfven WavesUM
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
17
Personnel:
Martin Lampe, NRL Dennis Papadopoulos, U MSteve Slinker, NRL Glenn Joyce, U MGuru Ganguli, NRL Wally Manheimer, U M
Major accomplishments:
• HEMPIC code development (quasineutral + eliminates c, ωp timescales ⇒ very fast)
• Theory and simulation of ducting (with Anatoly Streltsov)
• Theory and simulation of whistler growth in homogeneous systems
• Nonlinear amplification of wave packets propagating along the earth's dipole field
- Wave growth driven by resonant electrons propagating toward equator- Large frequency shifts (triggering of fallers)- Extensive simulation studies, theory in progress
Whistler Amplification and Stimulated EmissionNRL / Maryland collaboration
18
V||
V⊥
V|| 0
Electron distribution
cold electrons
fast electrons
V⊥ 0
HEMPIC SIMULATIONS OF WHISTLER INSTABILITY
•Realistic non-uniform geomagnetic field •Single-frequency wave packet (bounded in space) initiated at t=0•Fast electron distribution: ring distribution of constants of the motion v2 and v
2/ B0(z)•Inflow of fresh electrons from the boundary of the simulation domain
20
Dashed line indicates the angular frequency ω=1200 sec–1 that is initiated at t=0
WAVE FREQUENCY EVOLUTION
21
• For a ring distribution and a single frequency wave, electrons are resonant at two discrete locations, one on each side of the equator.Wave propagates to right, resonant electrons propagate (at a much larger velocity) to left.
• Wave initially grows at each of these resonant points.
• Resonant electrons are strongly phase-bunched at each resonant point.As these bunched electrons propagate to the left, they drive waves, at lower frequency than the initial triggering wave.
• Electrons propagating toward the equator lose energy and drive wave growth.Electrons propagating away from the equator gain energy and damp the waves.
• Resonant electrons remain resonant for a long time, because the wave frequency adjusts to the changing magnetic field, so as to maintain resonance.However very few electrons are phase trapped. The unstable waves are triggered by resonant untrapped electrons.
• We think we understand the main features and are working on a quantitative theory.
AN OUTLINE OF THE PHYSICS
22
Physics:
• More realistic thermal, anisotropic and loss cone electron distributions
• Can a large-amplitude injected wave grossly modify the electron phase-space distribution so as to trigger rapid growth of new waves?
• Instability of obliquely-propagating waves and 2-D mode spectra
• Instability of ducted whistlers in 2-D
Code development:
• Will need to parallelize the HEMPIC code (very easy)
• Further code development may be necessary. Can use Δt >> 1/Ω and Δx >> 1/λ by:
- Expanding fields in linear normal modes. Very efficient if not too many modes needed.
- For the particle kinetics, writing eqs for the deviation from unperturbed gyro motion. Using the mode expansion, these eqs can be integrated with long time steps.
WHISTLER INSTABILITY STUDIES: NEXT STEPS
23
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSU
EMICUM
VLF
ULF Alfven WavesUM
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
24
• Observations of LEP events using DEMETER and worldwide VLF observations. Burst mode observations over active thunderstorms (U. Inan)
• WIPP Code Issues (P. Kulkarni)– What is the precipitation induced by ground-based VLF
transmitters?– What factor affects most strongly induced precipitation: Source location,
operating frequency or radiated power
• Tentative Results: – The NWC transmitter in Australia induces the most >100 keV precipitation of the existing ground-based VLF sources
– Source location, much more than operating frequency orradiated power, determines energetic electron precipitation
Electron loss in theinner magnetosphere
25
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
EMICUM
VLF
ULF
ProtonsPrecipitation Electrons
VLF/ULF
LHWSU
Alfven WavesUM
WIPPSU
26
• Only one (peak near L=1.5 ) Proton Belt.• Stably Trapped for centuries !!
26 years
Proton (> 80 MeV) Belt Dynamics (1979-2005)(from NOAA 5-14 POES Satellites)
27
Over the south Atlantic, the inner proton belt is closest to the surfaceProtons in this region are the largest radiation source for LEO satellites
South Atlantic Anomaly
28
• Proton remediation can be done much more slowly– Years instead of weeks to remove particles– Particles take decades or centuries to
return• Because it can be done more slowly, it
may be cheaper and easier• Proton remediation would have
immediate operational impact
Compared to HANE Remediation
30
L=1.5
<b> pT
days
1000
Lifetime vs. <b>
Input power required for <b>=25 pT at few Hz at L=1.5 is 600 Watts per ΔL/Lª.1
/
/p z z
z A
k v
k Vω
Ω =
=
SAW Proton Precipitation (X. Shao)
31
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSU
EMICUM
VLF
ULF Alfven WavesUM
ProtonsPrecipitation Electrons
VLF/ULF RMFSU - UCLA
HAARP F-RegionUM
WIPPSU
32
2 22 2 3
21
2 2
2
/
1( | |) ( )
for
As a result 1/ / beforereaching resonance (1/ 0)
z z e
pe pj
je j
j
z e z
z
k v
k c
k c
k vk
γ
ω ωωω ω ω ω
ωω
γ
=
− = Ω
= − −+ Ω − Ω
→ ∞ → Ω
→ Ω
→
∑
At resonance I can have reflection or absorption or mode conversion or tunneling. What is R in our case?
Electron Precipitation by ULF waves (X. Shao)
33
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSUVLF
ULF
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
EMICUM
Alfven WavesUM
34
Step 1.Step 2.
USE ENERGY (30 GJ/Ton) STORED IN RELEASING A LARGE AMOUNT OF LOW IONIZATION POTENTIAL GAS (e.g. Li) AT ORBITAL VELOCITY TO GENERATE THE RESONANT WAVES – GANGULI ET AL (2007)
RELEASEPHOTO IONIZATION
• CONVERSION EFICIENCY FROM FREE ENERGY TO RESONANT SPECTRAL ENERGY
• SATURATION LEVEL OF PRIMARY ALFVEN ION CYCLOTRON INSTABILITY – VT
NOVEL WAVE INJECTION CONCEPTSNEUTRAL GAS INJECTION
36
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSUVLF
ULF
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
EMICUM
Alfven WavesUM
37
RMF driven by mechanically rotating permanent magnet
RMF driven by two orthogonal phase-
delayed current loops
Oblique Rotator
The RMF can be generated either by a pair of poly-phase coils, superconducting or else, or rotating permanent magnet.
A. Karavaev, X. Shao, N. Gumerov, G. Joyce + UCLA
Ways to Create RMF
Mech to em energy
38
Two independent coils:4 turns eachOperation frequencies 50<f<500 kHzCurrent magnitude 100 – 300 A
Device used to create RMF in LAPD
Setup for the Experiments
39
z
xy
B0
B0
I0cos(ωt)-I0cos(ωt)
B0
I0cos(ωt)-I0cos(ωt)
ω = Ω/10“Dipoles”
“Rotating Field”
z
xI0sin(ωt)
-I0sin(ωt)
I0cos(ωt)- I0cos(ωt)
2 wires
4 wires
Case 1
Case 2
Case 3
Configurations to Compare Theoretical and Computational Results
40
]~ˆ~ˆ)[(ˆ),( θθbeberrSeBtrB rrozo +−+=r
)sin(~)cos(~
tbb
tbbrω
ω
θ =
=
zz etrbrrStrE ˆ)cos())((),( 0 ωω−=
Axial Screen Current Jz
))(/1()/( 2 BJneJnemErrvv
×+= ν
nebJJnemE rz /)/( 2 ><+>=< θθ ν
Hall Term
Pondermotive force
Induces Azimuthal Eθ field
)()( rneVrJ θθ −=
• Modifies background
magnetic field•Dependence
on n• Hall term allows field
penetration into the plasma•Oscillating
collisionless skin depth
RMF Basics
41
Over 450 access portsComputer controlled data acquisition systemMicrowave interferometersLaser induced fluorescenceDC Magnetic field: 0.05 – 4 kG, variable on axisHighly ionized plasmas up to n≈5x1012 cm-3
Plasma column up to 2000Rci across diameterHighly reproducible plasma with 1 Hz operation frequency
Large Plasma Device (LAPD), University of California, Los Angeles
Setup for the Experiments
42
z
x
y
B0
Lx
Typical Settings
Domain Size:Lx = 80 (c/ωpc) ~ 12.5 λLz = 160 (c/ωpc) ~ 25 λ
Grid: Nx = 128, Nz = 256
Time step: ht = 1/(10Ω)
Max time: tmax = 100/Ω
Lz
HEMPIC Algorithm:
1). 3rd order Predictor-Corrector in time;2). FFT-based solver for elliptic equations;3). FFT-based spatial differentiation/convolution to compute the nonlinear terms;For a serial CPU code running on PC1 timestep takes about 0.7 seconds(1000 time steps ~ 12 min). (For 3D problem on grid 64x64x128 1 timestep takes about 45 seconds (1000 time steps ~ 12.5 hours).
Our goal is to reduce this time by orders of magnitude to enable computation of larger 2D and 3D problems for reasonable time. To enable this we will use Graphical Processing Units (GPU), which realize massively parallel computing in the scale of PC and algorithmic improvements (advancing in time in the Fourier space, use of Adams-type methods, finite-difference approximations for nonlinear terms, etc.).
Computational Model for 2D/3D Whistlers
43
050
100150200250300350400450500550
Bz, G
auss
-150
-100
-50
0
50
100
150
Cur
rent
, Am
ps
Background magnetic field50 Gauss200 Gauss
Two nearly identical currents with phase shift π/2Frequency f=292 kHz (Ωi<ω<<Ωe<<ωpe)
n≈7.0x1011 cm-3
n≈4.5x1010 cm-3
Source position
p27p31
p32p33 p35
p36p34
Input currents
Ambient magnetic field along the chamber
Setup for the Experiments
44
One loop experiment Two loops experiment
( , where T is the oscillating period)
One-loop: Jz current just oscillates around 0 with frequency ω.
Two-loop: Jz always has non-vanishing amplitude and rotates with the magnetic field about z-axis.
The graphs are in the same scale.
2 1 4t t T− =
Current Jz along ambient magnetic field at t1 and t2
Two-Loop Antenna vs. One-Loop Antenna
45
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
Mag
netic
Fie
ld, G
auss
Bx
Bx
(Bx2+By
2)1/2
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
Mag
netic
Fie
ld, G
auss
Bx
Bx
(Bx2+By
2)1/2
One-Loop antenna
Two-Loop antenna
Time dependence of magnetic field at the central point of xy-plane (Port 33)
One loop antenna
The magnitude of magnetic field oscillates (reaching 0) at twice the rotation frequency
Two loops antennaWhile x and ycomponents of magnetic field oscillate, the magnitude of magnetic field was kept at 0.2-0.4 G. Magnetic field vector rotates around z-axes.
t1 t2
Two-Loop Antenna vs. Single-Loop Antenna in Magnetized Plasma
46
Modeling shows that the pertiurbed magnetic field depends on n
That is δB~n1/2. Validated by experiment.
Can we improve antenna efficiency by injecting plasma ?
0.0
0.5
1.0
1.5
2.0
Mag
netic
Fie
ld, G
auss
(Bx2+By
2)1/2 n=7.0x1011cm-3
(Bx2+By
2)1/2 n=4.5x1010cm-3
Time dependence of magnetic field at central point of xy-plane (Port 33) for two–loop antenna in plasma with different plasma densities
1
2
4.1BB
≈ 1
2
3.94nn
≈From experiment: and - good agreement with models
Experiment 2N= 7x1011 cm-3
Experiment 1N=4.5x1010 cm-3
Dependence of Induced Magnetic Field on number density
47
ω = Ω/10
z
xy
Ey
B0
Ey
Numerical (Ωt = 100)
zx
x
zTheory predicts that the result has reflection asymmetry about the axis.
RMF Modeling
48
B0
I0cos(ωt)-I0cos(ωt)
Numerical (Ωt = 100)
Analytical Integral (stationary)
ω = Ω/10
Imaginary Part
Real Partz
xy
Ey
Ey
zx
Theory predicts a maximum on the axis.
Case 1
49
ω = Ω/10
z
xy Ey
B0
I0cos(ωt)-I0cos(ωt)
Numerical (Ωt = 100)
Ey
x
z
zx
Theory predicts a null on the axis.
Case 2
50
Magnetic field configuration for some moment of time for ambient magnetic field 50 Gauss
P27 – 2.9 m P31 – 1.6 m P32 – 1.3 m P33 – 1 m p35
Experiments Results
53
n=4.5x1010cm-3
Decay rate across ambient magnetic field lines (port 33)
Decay rate along ambient magnetic field lines
p35
p36 p33 p32 p31
p27
n=7.x1011cm-3 N=4.5x1010 cm-3
N= 7x1011 cm-3
Total magnetic field across ambient magnetic filed line
Spatial Decay Rates
Hall term allows enhanced penetration
55
n0
Bz0
Ω(t)
Constant Ω Modulated Ω(t)
Ω ΩBφ Bφ
Alfven Wave Generation with Rotating Magnetic Field
(MHD Simulation)
56
n = n0 n = 10n0
n
n0
Bz0
Bφ/Bz0
Z/λ 0
Z/λ 0
r/λ0 r/λ0Bφ/Bz0
Ω(t)
Bφ/Bz0
Vφ/VA0
n = n0
n = 10n0
n = n0
Poynting Fluxes: S( n = 10 n0) ~ 2äS( n = n0)
Alfven Wave Generation with Modulated RotatingMagnetic Field (MHD Simulation)
57
• Demonstrated concepts of new type RMF-based antenna/active device for space applications .• Differences between two-loop and classic one-loop antenna
• Two-loop antenna produces RMF and drives non-vanishing current along ambient magnetic field.
• Induced magnetic field by RMF is proportional to plasma frequency (ω<<ωp). That is δB~n1/2.• Spatial decay rate for the perturbation propagating across magnetic field much larger than c/ωe• Second-Order analysis shows excitation of azimuthal current perpendicular to the ambient magnetic field. This current helps perturbation penetrate deep into surrounding plasma.
Conclusion
58
1. Developing 3D semi-analytical model and 3D EMHD code to model RMF-induced whistler-mode wave propagation along magnetic field lines. Use the model to guide and explain experiments.
2. Conducting experiments with various parameters (e.g. plasma density, background magnetic field, and driving current magnitude), tilted rotating magnetic field, locally increased density by laser pulse or electron beam
3. Conducting experiments with finer spatial resolution along ambient magnetic field to investigate second-order phenomenon.
4. Use simulation to investigate parameter regime with perturbed magnetic field larger than ambient magnetic field.
5. Study energetic particles interaction with RMF
Future Work
59
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSUVLF
ULF
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
EMICUM
Alfven WavesUM
60
VLF/ELF/ULFInjected into Waveguide
BB
WhistlersSAW
Msonic Msonic
Electrojet Current
F-region ModulatedHeating
DiamagneticCurrent
J =B × ∇δpB2 exp(iωt)
F-region generation does not require ejet current. Can be located anywhere and operated continuously. Drives Msonic wave while ejet modulation drives SAW
G. Milikh, H. Schroff UM; D.Piddyachiy, U.Inan SU; C.Chang, T.Wallace BAE; M. Parrot CNRS
Cash et al. 2006
)( hVn A
R Δ≈
πω
HAARP ULF Generation
61
.25 Hz .5 Hz Natural lines
HAARP
IAR
At Juneau wave evanescent 1/R2
ULF Signals at Gakona and Juneau
67
UH heating
Bill Bristow UAL
B. Eliasson
K. Papadopoulos
Can we drive whistlers with F-region heating?
68
Field Experiment Lab Experiment Theory Data Analysis
Propagation
Space Based Ground BasedAIP Code Validation
SU - UCLA
RMFSU - UCLA
Neutral Gas InjectionVT
Amplification
VLF
ULF
VLF
VLF
Radiation/InjectionAlpha Field Tests
SU
HAARP F-RegionUM
RMF - HEDUM
Natural DuctsBC
Artificial DuctsUM
ASE Triggering - Siple DataBC - SU - DC
ASE Triggering ModelingNRL - UM
LEPSU
LHWSUVLF
ULF
WIPPSU
ProtonsPrecipitation Electrons
VLF/ULF
EMICUM
Alfven WavesUM
69
<b>≈25 pT
V ≈ 2x1020 m3
W=<b>2/2μ ≈ 3x10-16J/m3
Total Energy ≈ 60 kJ
Tconf ≈100-200 sec
Power ≈300-600 Watts
reflection
R ≈.90-.95
L≈1.5
Pitch Angle Diffusion Coefficient for Protons at 1 Hz
.1-3. Hz
Need1-3 Hz to resonate with 30-100 MeV protons at L=1.5
<b> pT
days
1000
Lifetime vs. <b>
<b> = 1 pT
Scale D at <b> = 1pTto get lifetime at larger <b>
Pick lifetime of 103 days and getpower needed
Energetic Proton Removal
70
60 KM
75 KM
120 KM
SAW A
M MP=IL
HED
For HMDP≈300 (M/1010 A-m2)2 Watts
Ground Based Antenna Choice
VED
• VED minimal injection
For VMDP≈(300/4) (M/1010 A-m 2)2 (δ/75 km)2 Watts
What about HED ?
VMD HMD
71
<b>75 km
120 km
A≈1010 m2
BR
To inject 600 W we require <b>≈20-25 pT at 75 km, the bottom of the magnetized ionosphere.
RMF
Rotating Magnetic FieldA superconducting magnet rotating at ULF frequency has an image in phase and increases its power by a factor of 4. This innovative antenna design can give the required 600 W with a magnetic moment of 1010 A-m2
Innovative Antenna Design
72
• Rotating superconducting magnets are useful for frequencies of up to 10 Hz
• They are compact sources of large moments and can be used in arrays• Example design:
– Superconducting coil 5 m high x 5 m wide x 5 m long– 25 m2 area– 100 Amps DC current– 4 x 104 turns– M = 108 A-m2
• In LTS wire ($2/kA-m) this could be a few million dollars including Dewar, He refrigeration, and rotation (HTS wire is still $50/kA-m, Cu now ~$100/kA-m)
Superconducting RMF
73
What about Tesla Technology HED Revisited
l
leff
Figure 1: Antenna with return current
Δ
( / 2 ) ln[ / ( )](8 / ) ln[ / ( )]
L l l lR l l l
μ π δ δπ σ δ δ
= Δ += Δ +
2 2l δ<
σ mho/m .1 Hz 1 Hz 10 Hz
10-4 150 km 50 km 16 km
10-3 50 km 17 km 5 km
10-2 17 km 5 km 1.7 km
Resistive impedance dominates if
Traditional HED antennae operate in the reactive impedance region
74
Point HED Design
~Δ
Figure 2: Plan view of HED antenna
B ≈ .4(P /MW )1/ 2(l /10km)5 / 2(σ /10−3)1/ 2nT
PULF ≈≈ 5(P /MW )(l /10km)5(σ /10−3)kW
1 km
76
Total energy = E= Volumex(b2/2μ)=20 (b/20 pΤ)2 kJPower required =P=E/T= 300 (b/20 pT)2 (60 sec/T) Watts
60 KM
75 KM
120 KM
SAW A
M MP=IL
(5/75)2 4
P≈300 (M/1010 A-m2) Watts
Energy – Power Requirements
77
Key Points• First observation of Magnetosonic waves in
the Pc1 range generated by modulated ionospheric heating using the HAARP heater
• Msonic waves generated by modulated collisionless F-region electron heating and are independent of the presence of elctrojet currents
• Detection by the Demeter satellite flying over HAARP indicate ULF power in excess of 5 kW
78
The Fundamentals• For Pc1 frequencies (.1-7 Hz) the ionosphere behaves as:
• A resonator for Shear Alfven (SA) waves, confined along the B lines with an almost vertical structure at high latitudes • A waveguide for Magnetosonic (MS) waves propagating isotropically and ducted horizontally over long distances
Bo Reflection due to gradη at one to two thousand km
Partial reflection at E-region
Standing SA wave
Ducted MS wave
79
SA Waves – Ionospheric Alfven Resonator (IAR)
Cash et al. 2006
k
E
B
b
SvgSA wave is guided along the B field
Reflections create standing wave structure
Notice
b·B=0
Natural SA waves
80
MS (Compressional) Waves Alfvenic Duct
k
Bb
Svg E
D=εΕ
VA
Alfvenic Duct
Notice b parallel to B
Isotropic Mode
AIC instability drives SA waves
SA waves mode converted at the boundary of duct and propagate laterally
as MS waves over large distances
Natural generation and ducting of MS waves
81
ULF Generation by Ejet Modulation
• Ejet modulation cannot drive b field parallel to ambient B. This type of modulation can create only SA waves. The waves cannot propagate laterally since they are evanescent in the Earth-Ionosphere Waveguide and do not couple to the Alfvenic Duct
• SA waves can be detected: (a) In the near zone below the heated spot and (b) By satellites over-flying the heated spot but confined to the magnetic flux tube that spans the heated spot.
D/E region heating + Electrojet
Evanescent in EI Waveguide
SA waves do not Excite Ionospheric Duct
82
ULF Signals at Juneau
• 28 April, 2007 UTC 05:01:00 – 05:05:45• Detected 1 Hz & 3 Hz peaks• Amplitudes at 1 Hz: 0.28 pT NS; 0.23 pT EW
2~ 1/b R EVANESCENT
85
F-Region Msonic ULF GenerationCollisionless upper hybrid F-region modulated heating results in Δp exp(iωt) that drives a b exp(iωt) with having a component parallel to B (a msonic mode). The wave propagates isotropically but is reflected at the D/E region and is much weaker on the ground. It can be measured by satellites or at large lateral distances (skip zone)
MHD Simulation
Vr
b/Bºβ (Δp/p)ºβ
86
Example of F-Region Msonic Generation Detected by the Demeter satellite
1OO nT
O-mode at 4.4 MHz HAARP at 3.5 MW modulated at .1 Hz between 6:47:30 and 6:59:30 UT
No ejet
No D/E region Demeter pass
No ULF detection on the ground - .1 Hz detection at Demeter between 6:51:30 and 6:53:00
89
Proton Energy
<δB> at L = 1.5
Life Time
50 MeV 20 pT 820 days
100 MeV 20 pT 620 days
200 MeV 20 pT 650 days
Equilibrium Distribution Function g(α0) Vs. Pitch Angle α0
)/)(exp()( 220 δωωωω −−∝f
f0 = 6 Hz, δf = 0.7 f0
Life Time and Equilibrium Equatorial Particle Flux Distribution
90
Assumptions: 1). Stationary ions; 2). Negligible displacement current.
Computational model (Lampe et al. JCP 214(2006) 284-298):
Evolution equation:
Elliptic equations for each time step:
(Electron plasma frequency) (Electron cyclotron frequency)
Basic Equations for Quasineutral Cold Fluid Electron Plasma Simulations