Phy 462: Plasma Physics
Prasad Subramanian
[Link]@[Link]
Indian Institute of Science Education and Research (IISER), Pune
Subramanian Plasma Physics
Orbit theory
Motion of charged particles in electric/magnetic fields
Subramanian Plasma Physics
Orbit theory
Motion of charged particles in electric/magnetic fields
But we thought the whole point of a plasma was collective
effects as opposed to test-particle descriptions
Subramanian Plasma Physics
Orbit theory
Motion of charged particles in electric/magnetic fields
But we thought the whole point of a plasma was collective
effects as opposed to test-particle descriptions
Yes, but orbit theory is (very!) relevant for
Tenuous (i.e., low density) plasmas - e.g., some situations in
fusion plasmas, the Earth’s magnetosphere, the outer reaches
of the solar corona/solar wind
Subramanian Plasma Physics
Orbit theory
Motion of charged particles in electric/magnetic fields
But we thought the whole point of a plasma was collective
effects as opposed to test-particle descriptions
Yes, but orbit theory is (very!) relevant for
Tenuous (i.e., low density) plasmas - e.g., some situations in
fusion plasmas, the Earth’s magnetosphere, the outer reaches
of the solar corona/solar wind
Dynamics of high energy particles in tenuous plasmas (e.g.,
cosmic rays interacting with the Earth’s magnetosphere)
Subramanian Plasma Physics
Orbit theory
Motion of charged particles in electric/magnetic fields
But we thought the whole point of a plasma was collective
effects as opposed to test-particle descriptions
Yes, but orbit theory is (very!) relevant for
Tenuous (i.e., low density) plasmas - e.g., some situations in
fusion plasmas, the Earth’s magnetosphere, the outer reaches
of the solar corona/solar wind
Dynamics of high energy particles in tenuous plasmas (e.g.,
cosmic rays interacting with the Earth’s magnetosphere)
Orbit theory is relevant when the interaction of particles with
macroscopic (external) EM fields is more relevant than with
the microscopic/self-consistent fields associated with
individual particles.
Subramanian Plasma Physics
Orbit theory - outline
Particle motion in electric and magnetic fields
uniform/nonuniform, static, time-varying
Adiabatic invariants, particle drifts
applications to tokamaks and particle motion in the Earth’s
magnetic field
Chap 2 B & S, Chaps 3, 4, 5 Sturrock
Subramanian Plasma Physics
Motion in a static, uniform B field
A natural length - the particle Larmor (gyro) radius rL , and a
natural frequency - the gyrofrequency Ω
Subramanian Plasma Physics
Motion in a static, uniform B field
A natural length - the particle Larmor (gyro) radius rL , and a
natural frequency - the gyrofrequency Ω
Often, rL /L 1, 2π/ΩT 1; so they are natural ordering
parameters, one can expand motions perturbatively
Subramanian Plasma Physics
Motion in a static, uniform B field
A natural length - the particle Larmor (gyro) radius rL , and a
natural frequency - the gyrofrequency Ω
Often, rL /L 1, 2π/ΩT 1; so they are natural ordering
parameters, one can expand motions perturbatively
Average our rapid gyro motion and focus on motion of the
guiding center
Subramanian Plasma Physics
Motion in a static, uniform B field
A natural length - the particle Larmor (gyro) radius rL , and a
natural frequency - the gyrofrequency Ω
Often, rL /L 1, 2π/ΩT 1; so they are natural ordering
parameters, one can expand motions perturbatively
Average our rapid gyro motion and focus on motion of the
guiding center
Concept of adiabatic constants of motion; e.g., the magnetic
moment
Subramanian Plasma Physics
Motion in a static, uniform B field
A natural length - the particle Larmor (gyro) radius rL , and a
natural frequency - the gyrofrequency Ω
Often, rL /L 1, 2π/ΩT 1; so they are natural ordering
parameters, one can expand motions perturbatively
Average our rapid gyro motion and focus on motion of the
guiding center
Concept of adiabatic constants of motion; e.g., the magnetic
moment
Radiation neglected, need to make sure that external fields are
dominant
Subramanian Plasma Physics
Particle gyrating in uniform B field
Subramanian Plasma Physics
Equation of motion
The usual
mj r̈j = ej [E(r, t) + ṙj × B(r, t)]
Consider uniform, constant B = Bẑ, E = 0
Subramanian Plasma Physics
Equation of motion
The usual
mj r̈j = ej [E(r, t) + ṙj × B(r, t)]
Consider uniform, constant B = Bẑ, E = 0
z̈ = 0, i.e., ż = v|| is constant
Subramanian Plasma Physics
Equation of motion
The usual
mj r̈j = ej [E(r, t) + ṙj × B(r, t)]
Consider uniform, constant B = Bẑ, E = 0
z̈ = 0, i.e., ż = v|| is constant
Also, ẍ = Ωẏ , ÿ = −Ωẋ, Ω ≡ eB/m
Also, mr̈ . ṙ = 0, → ⊥ and || kinetic energies are conserved
(no suprise there, magnetic fields do no work)
Subramanian Plasma Physics
Particle gyrating in uniform B field
Subramanian Plasma Physics
Equations for particle orbits
v⊥ v⊥
x= sin(Ωt + α) + x0 , y = cos(Ωt + α) + y0
Ω Ω
Clearly, motion in the ⊥ plane is a circle.
Subramanian Plasma Physics
Equations for particle orbits
v⊥ v⊥
x= sin(Ωt + α) + x0 , y = cos(Ωt + α) + y0
Ω Ω
Clearly, motion in the ⊥ plane is a circle.
Alternatively, recognize that B = (0, 0, B) and split
dv/dt = (e/mc)v × B into dvx /dt = (eB/mc)vy ,
dvy /dt = −(eB/mc)vx and dvz /dt = 0. This gives
q
vx = v⊥ cos(Ωt) , vy = −(e/|e|)v⊥ sin(Ωt) , v⊥ = vx2 + vy2 .
In other words, motion in the ⊥ plane is circular.
Subramanian Plasma Physics
Equations for particle orbits
v⊥ v⊥
x= sin(Ωt + α) + x0 , y = cos(Ωt + α) + y0
Ω Ω
Clearly, motion in the ⊥ plane is a circle.
Alternatively, recognize that B = (0, 0, B) and split
dv/dt = (e/mc)v × B into dvx /dt = (eB/mc)vy ,
dvy /dt = −(eB/mc)vx and dvz /dt = 0. This gives
q
vx = v⊥ cos(Ωt) , vy = −(e/|e|)v⊥ sin(Ωt) , v⊥ = vx2 + vy2 .
In other words, motion in the ⊥ plane is circular. How about the
pitch angle? Relation to instantaneous radius of curvature?
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv|| e
= E|| ,
dt m
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion;
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion; as regards ⊥ motion,
lets anticipate (Sturrock) that it can be split into a
time-dependent and a time-independent part: v⊥ = vd + vg (t);
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion; as regards ⊥ motion,
lets anticipate (Sturrock) that it can be split into a
time-dependent and a time-independent part: v⊥ = vd + vg (t);
the time-dependent part is the usual gyro motion and the
time-independent part is the (slow) drift.
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion; as regards ⊥ motion,
lets anticipate (Sturrock) that it can be split into a
time-dependent and a time-independent part: v⊥ = vd + vg (t);
the time-dependent part is the usual gyro motion and the
time-independent part is the (slow) drift. Substitute this
(assumed) form into the equation for (⊥) motion, separate out the
time-dependent and independent parts;
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion; as regards ⊥ motion,
lets anticipate (Sturrock) that it can be split into a
time-dependent and a time-independent part: v⊥ = vd + vg (t);
the time-dependent part is the usual gyro motion and the
time-independent part is the (slow) drift. Substitute this
(assumed) form into the equation for (⊥) motion, separate out the
time-dependent and independent parts; the time dependent part
yields the usual gyro motion dvg /dt = (e/mc)(vg × B),
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion; as regards ⊥ motion,
lets anticipate (Sturrock) that it can be split into a
time-dependent and a time-independent part: v⊥ = vd + vg (t);
the time-dependent part is the usual gyro motion and the
time-independent part is the (slow) drift. Substitute this
(assumed) form into the equation for (⊥) motion, separate out the
time-dependent and independent parts; the time dependent part
yields the usual gyro motion dvg /dt = (e/mc)(vg × B), but the
time-independent part yields the interesting E × B drift:
E⊥ + (1/c)vd × B = 0; crossing with B gives the drift velocity
Subramanian Plasma Physics
Motion in electric and magnetic fields
dv||
e dv⊥ e 1
= E|| , = E⊥ + v⊥ × B
dt m dt m c
Nothing too remarkable about the || motion; as regards ⊥ motion,
lets anticipate (Sturrock) that it can be split into a
time-dependent and a time-independent part: v⊥ = vd + vg (t);
the time-dependent part is the usual gyro motion and the
time-independent part is the (slow) drift. Substitute this
(assumed) form into the equation for (⊥) motion, separate out the
time-dependent and independent parts; the time dependent part
yields the usual gyro motion dvg /dt = (e/mc)(vg × B), but the
time-independent part yields the interesting E × B drift:
E⊥ + (1/c)vd × B = 0; crossing with B gives the drift velocity
E×B
vd = c
B2
Subramanian Plasma Physics
Electrons and protons drift in same direction in crossed
E/B fields
Subramanian Plasma Physics
Crossed E/B fields, when gyration speed < or > drift speed
Subramanian Plasma Physics
How about B and a non-EM field (e.g., gravity)?
dv/dt = g + (e/mc)v × B
Subramanian Plasma Physics
How about B and a non-EM field (e.g., gravity)?
dv/dt = g + (e/mc)v × B
By analogy with the electric field, the effective electric field is
Eeff = (m/e)g
Subramanian Plasma Physics
How about B and a non-EM field (e.g., gravity)?
dv/dt = g + (e/mc)v × B
By analogy with the electric field, the effective electric field is
Eeff = (m/e)g
so the drift velocity
mc g × B
vd =
e B2
depends upon the mass and charge!
Subramanian Plasma Physics
How about B and a non-EM field (e.g., gravity)?
dv/dt = g + (e/mc)v × B
By analogy with the electric field, the effective electric field is
Eeff = (m/e)g
so the drift velocity
mc g × B
vd =
e B2
depends upon the mass and charge!
There will be a net current J = ρ(g × B)/B 2
Subramanian Plasma Physics
How about B and a non-EM field (e.g., gravity)?
dv/dt = g + (e/mc)v × B
By analogy with the electric field, the effective electric field is
Eeff = (m/e)g
so the drift velocity
mc g × B
vd =
e B2
depends upon the mass and charge!
There will be a net current J = ρ(g × B)/B 2
Which species carries “most of” the current?
Subramanian Plasma Physics
Electrons and protons drift in opposite directions in
crossed g/B fields
Subramanian Plasma Physics
Inhomogeneous (but time-constant) B fields
Small perturbations: B(r) ≈ B(r0 ) + (δr . ∇)B|r0
Subramanian Plasma Physics
Inhomogeneous (but time-constant) B fields
Small perturbations: B(r) ≈ B(r0 ) + (δr . ∇)B|r0
δB is small over a Larmor radius
Subramanian Plasma Physics
Inhomogeneous (but time-constant) B fields
Small perturbations: B(r) ≈ B(r0 ) + (δr . ∇)B|r0
δB is small over a Larmor radius
One can average over the fast gyromotions and concentrate
on the guiding center motion
dΩ
Ω(y ) ' Ω(y0 ) + (y − y0 ) |y =y0
dy
Subramanian Plasma Physics
Guiding center
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Subramanian Plasma Physics
Curvature drift
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ,˙ where ξ ≡ x + iy (convince
yourself)
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ,˙ where ξ ≡ x + iy (convince
yourself)
ξ¨ + iΩξ˙ = 0;
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ,˙ where ξ ≡ x + iy (convince
yourself)
0
ξ¨ + iΩξ˙ = 0; using Ω(y ) = Ω0 + (y − y0 )Ω0 , we get
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (y − y0 )ξ˙
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ,˙ where ξ ≡ x + iy (convince
yourself)
0
ξ¨ + iΩξ˙ = 0; using Ω(y ) = Ω0 + (y − y0 )Ω0 , we get
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (y − y0 )ξ˙
The spirit of the approximation is that y and Ω0 on the RHS
can be replaced by their uniform B field values
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ,˙ where ξ ≡ x + iy (convince
yourself)
0
ξ¨ + iΩξ˙ = 0; using Ω(y ) = Ω0 + (y − y0 )Ω0 , we get
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (y − y0 )ξ˙
The spirit of the approximation is that y and Ω0 on the RHS
can be replaced by their uniform B field values
˙ = ξ(0)exp(−iΩ
ξ(t) ˙ 0 t)
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ, ˙ where ξ ≡ x + iy (convince
yourself)
0
ξ¨ + iΩξ˙ = 0; using Ω(y ) = Ω0 + (y − y0 )Ω0 , we get
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (y − y0 )ξ˙
The spirit of the approximation is that y and Ω0 on the RHS
can be replaced by their uniform B field values
˙ = ξ(0)exp(−iΩ
ξ(t) ˙ ˙
0 t) with ξ(0) = v⊥ exp(−iα)
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ, ˙ where ξ ≡ x + iy (convince
yourself)
0
ξ¨ + iΩξ˙ = 0; using Ω(y ) = Ω0 + (y − y0 )Ω0 , we get
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (y − y0 )ξ˙
The spirit of the approximation is that y and Ω0 on the RHS
can be replaced by their uniform B field values
˙ = ξ(0)exp(−iΩ
ξ(t) ˙ ˙
0 t) with ξ(0) = v⊥ exp(−iα)
which gives
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (v⊥2 /Ω0 )cos(Ω0 t + α) exp(−iΩ0 t + α)
Integrating once gives
Subramanian Plasma Physics
B fields with transverse gradient - the ∇B drift
Recall, ẍ = Ω(y )ẏ , ÿ = Ω(x)ẋ
Alternatively, ξ¨ = −iΩ(y )ξ, ˙ where ξ ≡ x + iy (convince
yourself)
0
ξ¨ + iΩξ˙ = 0; using Ω(y ) = Ω0 + (y − y0 )Ω0 , we get
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (y − y0 )ξ˙
The spirit of the approximation is that y and Ω0 on the RHS
can be replaced by their uniform B field values
˙ = ξ(0)exp(−iΩ
ξ(t) ˙ ˙
0 t) with ξ(0) = v⊥ exp(−iα)
which gives
0
ξ¨ + iΩ0 ξ˙ = −iΩ0 (v⊥2 /Ω0 )cos(Ω0 t + α) exp(−iΩ0 t + α)
Integrating once gives
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt+α)− 1−cos2(Ωt+α)−2sin(Ωt+α)sinα
2Ω20
with a similar expression for ẏ (t) with all the time-dependent
sins and coss interchanged (but the constant term 1 is
absent).
Subramanian Plasma Physics
Grad B drift
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt +α)− 1−cos2(Ωt +α)−2sin(Ωt +α)sinα
2Ω20
Subramanian Plasma Physics
Grad B drift
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt +α)− 1−cos2(Ωt +α)−2sin(Ωt +α)sinα
2Ω20
Its only the constant term that matters; the gyrofrequency (Ω0 )
and twice-gyrofrequency (2Ω0 ) can be averaged out; they don’t
matter to us.
Subramanian Plasma Physics
Grad B drift
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt +α)− 1−cos2(Ωt +α)−2sin(Ωt +α)sinα
2Ω20
Its only the constant term that matters; the gyrofrequency (Ω0 )
and twice-gyrofrequency (2Ω0 ) can be averaged out; they don’t
matter to us. So the average is
2 0
v⊥ Ω0
hvi = − , 0 , v||
2Ω20
Subramanian Plasma Physics
Grad B drift
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt +α)− 1−cos2(Ωt +α)−2sin(Ωt +α)sinα
2Ω20
Its only the constant term that matters; the gyrofrequency (Ω0 )
and twice-gyrofrequency (2Ω0 ) can be averaged out; they don’t
matter to us. So the average is
2 0
v⊥ Ω0
hvi = − , 0 , v||
2Ω20
So this correponds to the grad B drift
W⊥ (B × ∇)B
vG =
eB 3
Subramanian Plasma Physics
Grad B drift
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt +α)− 1−cos2(Ωt +α)−2sin(Ωt +α)sinα
2Ω20
Its only the constant term that matters; the gyrofrequency (Ω0 )
and twice-gyrofrequency (2Ω0 ) can be averaged out; they don’t
matter to us. So the average is
2 0
v⊥ Ω0
hvi = − , 0 , v||
2Ω20
So this correponds to the grad B drift
W⊥ (B × ∇)B
vG =
eB 3
The drift is ⊥ to both B and ∇B;
Subramanian Plasma Physics
Grad B drift
0
Ω0 v⊥2
ẋ(t) = v⊥ cos(Ωt +α)− 1−cos2(Ωt +α)−2sin(Ωt +α)sinα
2Ω20
Its only the constant term that matters; the gyrofrequency (Ω0 )
and twice-gyrofrequency (2Ω0 ) can be averaged out; they don’t
matter to us. So the average is
2 0
v⊥ Ω0
hvi = − , 0 , v||
2Ω20
So this correponds to the grad B drift
W⊥ (B × ∇)B
vG =
eB 3
The drift is ⊥ to both B and ∇B; will this lead to a net current,
though?
Subramanian Plasma Physics
How good is the guiding center approximation?
Subramanian Plasma Physics
Curvature drift
Curvature force
F= mv||2 Rc /Rc2 ; draw analogy with vD = (F × B)/eB 2 to get
Subramanian Plasma Physics
Curvature drift
Curvature force
F= mv||2 Rc /Rc2 ; draw analogy with vD = (F × B)/eB 2 to get
mv||2 Rc × B
vC =
eB 2 Rc2
Subramanian Plasma Physics
Points to ponder
How about gradients in density? (∇n drifts)
Can (guiding center) particle drifts always be identified with a
bulk current?
Subramanian Plasma Physics
Points to ponder
How about gradients in density? (∇n drifts)
Can (guiding center) particle drifts always be identified with a
bulk current? Remember, a macroscopic current density is
constructed using a (bulk) density and a (bulk)
speed/velocity;
Subramanian Plasma Physics
Points to ponder
How about gradients in density? (∇n drifts)
Can (guiding center) particle drifts always be identified with a
bulk current? Remember, a macroscopic current density is
constructed using a (bulk) density and a (bulk)
speed/velocity; both bulk quantities, which involve averaging
over a particle distribution;
Subramanian Plasma Physics
Points to ponder
How about gradients in density? (∇n drifts)
Can (guiding center) particle drifts always be identified with a
bulk current? Remember, a macroscopic current density is
constructed using a (bulk) density and a (bulk)
speed/velocity; both bulk quantities, which involve averaging
over a particle distribution; not quite the same as an
individual particle, even if one is averaging over fast timescales
for an individual particle to get the guiding center motion.
Subramanian Plasma Physics
Longitudinal gradients in B
B varies in the same direction as v|| ; the gyroradius gets tighter as
the particle progresses
Subramanian Plasma Physics
Longitudinal gradients in B
Subramanian Plasma Physics
The magnetic mirror effect
In cylindrical coordinates, ∇ . B = 0 is
1 ∂ ∂Bz
(rBr ) + =0
r ∂r ∂z
Subramanian Plasma Physics
The magnetic mirror effect
In cylindrical coordinates, ∇ . B = 0 is
1 ∂ ∂Bz
(rBr ) + =0
r ∂r ∂z
(neglecting radial variation of ∂Bz /∂z).
Subramanian Plasma Physics
The magnetic mirror effect
In cylindrical coordinates, ∇ . B = 0 is
1 ∂ ∂Bz
(rBr ) + =0
r ∂r ∂z
(neglecting radial variation of ∂Bz /∂z). This gives
Z r
0 ∂Bz 0 1 ∂Bz
rBr = − r dr = − r 2
0 ∂z 2 ∂z
We are concerned here with the longitudinal force Fz on the
particle;
Subramanian Plasma Physics
The magnetic mirror effect
In cylindrical coordinates, ∇ . B = 0 is
1 ∂ ∂Bz
(rBr ) + =0
r ∂r ∂z
(neglecting radial variation of ∂Bz /∂z). This gives
Z r
0 ∂Bz 0 1 ∂Bz
rBr = − r dr = − r 2
0 ∂z 2 ∂z
We are concerned here with the longitudinal force Fz on the
particle; Fz = −qvφ Br /c = |q|v⊥ Br /c.
Subramanian Plasma Physics
The magnetic mirror effect
In cylindrical coordinates, ∇ . B = 0 is
1 ∂ ∂Bz
(rBr ) + =0
r ∂r ∂z
(neglecting radial variation of ∂Bz /∂z). This gives
Z r
0 ∂Bz 0 1 ∂Bz
rBr = − r dr = − r 2
0 ∂z 2 ∂z
We are concerned here with the longitudinal force Fz on the
particle; Fz = −qvφ Br /c = |q|v⊥ Br /c. Using r = rL = v⊥ /Ω and
(1/2)mv⊥2 = W⊥ , we have
W⊥ ∂Bz
Fz = − = −µ . ∇B
Bz ∂z
As the field gradient flattens out, the longitudinal force decreases
→ mirroring
Subramanian Plasma Physics
Mirroring - another way
Use conservation of magnetic moment
µ ≡ W⊥ (0)/Bz (0) = W⊥ (z)/Bz (z) or (Ω/2πc)|q|πrL2 (like current
x area) and conservation of total (|| and ⊥) energy:
Subramanian Plasma Physics
Mirroring - another way
Use conservation of magnetic moment
µ ≡ W⊥ (0)/Bz (0) = W⊥ (z)/Bz (z) or (Ω/2πc)|q|πrL2 (like current
x area) and conservation of total (|| and ⊥) energy:
W (0) = W|| (0) + W⊥ (0) = W|| (z) + W⊥ (z), to get
Bz (z)
W|| (z) = W|| (0) − W⊥ (0) −1
Bz (0)
Subramanian Plasma Physics
Mirroring - another way
Use conservation of magnetic moment
µ ≡ W⊥ (0)/Bz (0) = W⊥ (z)/Bz (z) or (Ω/2πc)|q|πrL2 (like current
x area) and conservation of total (|| and ⊥) energy:
W (0) = W|| (0) + W⊥ (0) = W|| (z) + W⊥ (z), to get
Bz (z)
W|| (z) = W|| (0) − W⊥ (0) −1
Bz (0)
As the particle moves into a stronger B field, the || component of
velocity could vanish at some point, and it could actually get
reflected.
Subramanian Plasma Physics
Mirroring - another way
Use conservation of magnetic moment
µ ≡ W⊥ (0)/Bz (0) = W⊥ (z)/Bz (z) or (Ω/2πc)|q|πrL2 (like current
x area) and conservation of total (|| and ⊥) energy:
W (0) = W|| (0) + W⊥ (0) = W|| (z) + W⊥ (z), to get
Bz (z)
W|| (z) = W|| (0) − W⊥ (0) −1
Bz (0)
As the particle moves into a stronger B field, the || component of
velocity could vanish at some point, and it could actually get
reflected. Where it gets reflected (and in fact, if it does at all)
depends upon W|| (0)/W⊥ (0) (i.e., the pitch angle)
Subramanian Plasma Physics
Mirrors in practice
Subramanian Plasma Physics
Loss cones
v⊥ (0) = v (0)sinαm , where αm is the “critical” angle below
which particles will not be reflected at the point of maximum
magnetic field Bm .
Subramanian Plasma Physics
Loss cones
v⊥ (0) = v (0)sinαm , where αm is the “critical” angle below
which particles will not be reflected at the point of maximum
magnetic field Bm .
From the constancy of the magnetic moment,
µ ≡ v⊥ (0)2 /B(0) = v02 /Bm
Subramanian Plasma Physics
Loss cones
v⊥ (0) = v (0)sinαm , where αm is the “critical” angle below
which particles will not be reflected at the point of maximum
magnetic field Bm .
From the constancy of the magnetic moment,
µ ≡ v⊥ (0)2 /B(0) = v02 /Bm (since, at the point of maximum
magnetic field, the velocity is all perpendicular; nothing
parallel, and we have also used conservation of total kinetic
energy)
we get sin2 αm = B(0)/Bm
Subramanian Plasma Physics
Loss cones
v⊥ (0) = v (0)sinαm , where αm is the “critical” angle below
which particles will not be reflected at the point of maximum
magnetic field Bm .
From the constancy of the magnetic moment,
µ ≡ v⊥ (0)2 /B(0) = v02 /Bm (since, at the point of maximum
magnetic field, the velocity is all perpendicular; nothing
parallel, and we have also used conservation of total kinetic
energy)
we get sin2 αm = B(0)/Bm
So particles with pitch angles < αm “leak through” the
mirror, and are lost from the system - the loss cone
Subramanian Plasma Physics
Loss cones
Subramanian Plasma Physics
Trapped particles in a tokamak - “banana” orbits
Subramanian Plasma Physics
Trapped particles in the Earth’s magnetic field - the van
Allen belts
Subramanian Plasma Physics
van Allen belts - mirroring, ∇B drifts, etc.
Enhanced energetic particle population poses credible danger to
satellites
Subramanian Plasma Physics
Draining the van Allen belts (!)
[Link]
Subramanian Plasma Physics
Cosmic ray Forbush decreases
Subramanian Plasma Physics
Cosmic ray Forbush decreases
Subramanian Plasma Physics
Forbush decrease example - GRAPES-3 data
Subramanian Plasma Physics
Loss cone example - precursors to Forbush decreases
Forbush decreases - decreases in flux of high energy (galactic)
cosmic rays due to “umbrella” of enhanced magnetic field
carried by Earth-directed coronal mass ejections (from the
Sun)
Subramanian Plasma Physics
Loss cone example - precursors to Forbush decreases
Forbush decreases - decreases in flux of high energy (galactic)
cosmic rays due to “umbrella” of enhanced magnetic field
carried by Earth-directed coronal mass ejections (from the
Sun)
We often have precursors coming to us in advance of the main
Forbush decrease;
Subramanian Plasma Physics
Loss cone example - precursors to Forbush decreases
Forbush decreases - decreases in flux of high energy (galactic)
cosmic rays due to “umbrella” of enhanced magnetic field
carried by Earth-directed coronal mass ejections (from the
Sun)
We often have precursors coming to us in advance of the main
Forbush decrease; very important, giving us advance
information regarding the impending geomagnetic storm
Subramanian Plasma Physics
Loss cone example - precursors to Forbush decreases
Forbush decreases - decreases in flux of high energy (galactic)
cosmic rays due to “umbrella” of enhanced magnetic field
carried by Earth-directed coronal mass ejections (from the
Sun)
We often have precursors coming to us in advance of the main
Forbush decrease; very important, giving us advance
information regarding the impending geomagnetic storm
These precursors are loss cones escaping from the B field
enhancements
Subramanian Plasma Physics
Loss cone example - precursors to Forbush decreases
Forbush decreases - decreases in flux of high energy (galactic)
cosmic rays due to “umbrella” of enhanced magnetic field
carried by Earth-directed coronal mass ejections (from the
Sun)
We often have precursors coming to us in advance of the main
Forbush decrease; very important, giving us advance
information regarding the impending geomagnetic storm
These precursors are loss cones escaping from the B field
enhancements
Lots of exciting solar-terrestrial research opportunities using
cosmic ray data from the GRAPES-3 muon telescope at Ooty
(operated by the TIFR)
Subramanian Plasma Physics
Forbush decrease precursors - several factors
Subramanian Plasma Physics
Loss cone anisotropy example
Subramanian Plasma Physics