HSE — Valery Nakariakov — Solar Physics 1

Lecture 4: Magnetohydrodynamics (MHD),

MHD Equilibrium, MHD Waves

MHD describes large scale, slow dynamics of plasmas. More specifically,

we can apply MHD when

1. Characteristic time  ion gyroperiod and mean free path time,

2. Characteristic scale  ion gyroradius and mean free path length,
3. Plasma velocities are not relativistic.

In MHD, the plasma is considered as an electrically conducting fluid. Gov-

erning equations are equations of fluid dynamics and Maxwell’s equations.
A self-consistent set of MHD equations connects the plasma mass density
ρ, the plasma velocity V, the thermodynamic (also called gas or kinetic)
pressure P and the magnetic field B. In strict derivation of MHD, one
should neglect the motion of electrons and consider only heavy ions.
The 1-st equation is mass continuity
∂ρ
+ ∇(ρV) = 0, (1)
∂t
and it states that matter is neither created or destroyed.
The 2-nd is the equation of motion of an element of the fluid,
∂V
" #
ρ + (V∇)V = −∇P + j × B, (2)
∂t
also called the Euler equation. The vector j is the electric current density
which can be expressed through the magnetic field B. Mind that on the
lefthand side it is the total derivative, d/dt.
The 3-rd equation is the energy equation, which in the simplest adiabatic
case has the form
d P
!
= 0, (3)
dt ργ
where γ is the ratio of specific heats Cp /CV , and is normally taken as 5/3.
The temperature T of the plasma can be determined from the density ρ
and the thermodynamic pressure P , using the state equation (e.g. the ideal
HSE — Valery Nakariakov — Solar Physics 2

gas law). For example, in a pure hydrogen plasma, this equation is

kB
P =2 ρT, (4)
mp
where mp is the mass of a proton and kB is Boltzmann’s constant.
Now, let us derive the equation for the magnetic field using Maxwell’s
equations. Start with Ohm’s law,
j = σE0 , (5)
where σ is electrical conductivity (the physical quantity inverse to the resis-
tivity) and E0 is the electric field experienced by the plasma (fluid) element
in its rest frame. When the plasma is moving (with respect to the external
magnetic field) at the velocity V, applying the Lorentz transformation we
obtain
E0 = E + V × B. (6)

Now, Eq. (5) can be re-written as

1
j = E + V × B. (7)
σ
In the case of perfect conductivity, σ → ∞, we have
E = −V × B. (8)
Calculating the curl of the electric field E and using one of Maxwell’s
equation,
∂B
∇×E=− , (9)
∂t
we can exclude the electric field and obtain
∂B
= ∇ × (V × B), (10)
∂t
which is the 4-th MHD equation — the “induction equation”. In particular,
this equation describes the phenomenon of magnetic dynamo.

To close the set of MHD equations, we have to express the current density
j through the magnetic field B. Consider the other Maxwell’s equation,
1 ∂E
∇×B− = µ0 j (11)
c2 ∂t
HSE — Valery Nakariakov — Solar Physics 3

From Ohm’s law, we had E = −V × B. Consequently, we can estimate the

electric field as E ∼ V0 B, where V0 is a characteristic speed of the process.
Consider the ratio of two terms in Eq. (11):
1 ∂E
∇ × B and .
c2 ∂t
The first term is proportional to B/l0 , where l0 is a characteristic scale of
the process, the second to E/c2 t0 , where t0 is a characteristic time of the
process, V0 = l0 /t0 . When the process is not relativistic, V0  c, the first
term is very much greater than the second, and we have
1
j= ∇×B (12)
µ0
In addition, the magnetic field B must satisfy the condition ∇.B = 0.
HSE — Valery Nakariakov — Solar Physics 4

Thus, the closed set of MHD equations is

∂ρ
+ ∇(ρV) = 0, Mass Continuity Eq.,
∂t

d P
!
= 0, Energy Eq.,
d t ργ

dV 1
ρ = −∇P − B × (∇ × B), Euler’s Eq.,
dt µ0

∂B
= ∇ × (V × B), Induction Eq..
∂t

The equations are ideal, which means that all dissipative processes (finite
viscosity, electrical resistivity and thermal conductivity) were neglected.
Also, the magnetic field is subject to the condition
∇ · B = 0. (13)
HSE — Valery Nakariakov — Solar Physics 5

The Nobel Prize in Physics 1970 was given to Hannes Olof Gösta Alfvén
“for fundamental work and discoveries in magnetohydrodynamics with
fruitful applications in different parts of plasma physics”.
MHD is applicable from nanometre (10−9 m) scales in, e.g. physics of
semi-conductors, to galactic (1021 m) scales, e.g. galactic arms.
HSE — Valery Nakariakov — Solar Physics 6

Example: the applicability of MHD to the solar corona:

1. Speeds are much less than the speed of light.

(In the solar corona: V < a few thousand km/s).
2. Characteristic times are much longer than the gyroperiod and the
plasma period.
In the solar corona: fMHD < 1 Hz,
for fgyro = 1.52 × 103 × B(G) ≈ 1.52 × 104 Hz
−3
and fplasma = 9 × n1/2 8
e (m ) ≈ 2 × 10 Hz,
(for B = 10 G and ne = 5 × 1014 m−3 ).
3. Characteristic times are much longer than the collision times. Char-
acteristic spatial scales are larger than the mean free path length
7.2 × 107 T 2 (K)
λ  lii (m) ≈ .
n(m−3 )
For the typical conditions of the lower corona, lii ≈ 105 − 106 m.
4. Similar estimations should be made for the spatial scales, and the
conditions of applicability are well satisfied too.
HSE — Valery Nakariakov — Solar Physics 7

MHD Equilibrium

The static equilibrium conditions are:

∂
V = 0,
= 0. (14)
∂t
These conditions identically satisfy the continuity, energy and induction
equations.
From Euler’s equation we obtain the condition
1
−∇P − B × (∇ × B) = 0, (15)
µ0
which is called the equation of magnetostatics. This equation should be
supplemented with the condition ∇.B = 0.
Eq. (15) can be re-written (DYI) as
B2 
 
1
−∇ P +
 + (B.∇)B = 0. (16)
2µ0 µ0
The first term can be considered as the gradient of total pressure.
The total pressure consists of two terms, the gas (or thermodynamic) pres-
sure P , and the magnetic pressure B 2 /2µ0 .
The second term is magnetic tension. The force is directed anti-parallel to
the radius of the magnetic field line curvature.
HSE — Valery Nakariakov — Solar Physics 8

Plasma-β
Compare the terms in the magnetostatic equation,
1
−∇P + B × (∇ × B) = 0. (17)
µ0

We have that
P 1 B2
∇P ≈ and B × (∇ × B) ≈ , (18)
λ µ0 µ0 λ
where λ is a characteristic scale of the problem.
The ratio of the gas pressure gradient term and the Lorentz force is known
as the plasma-β,
gas pressure P
β≡ = 2 . (19)
magnetic pressure B /2µ0

Plasma-β can be estimated by the formula,

β = 3.5 × 10−21 n T B −2 , (20)
where n is in m−3 , T in K and B in G.
For example, in the solar corona, T = 106 K, n = 1014 m−3 , B = 10 G,
and β = 3.5 × 10−3 .
In photospheric magnetic flux tubes, T = 6 × 103 K, n = 1023 m−3 , B =
1000 G, and β = 2.
In the solar wind near the Earth’s orbit, T = 2 × 105 K, n = 107 m−3 ,
B = 6 × 10−5 G, and β = 2.
HSE — Valery Nakariakov — Solar Physics 9

Example: Sunspots.

Sunspots appear as dark spots on the surface of the Sun. They typically
last for several days, although very large ones may live for several weeks.
Sunspots are magnetic regions on the Sun with magnetic field strengths
thousands of times stronger than the Earth’s magnetic field.
Consider a sunspot as a vertical magnetic flux tube. The magnetic field
B0 is vertical. The kinetic pressure is P0 and PE inside and outside, respec-
tively. The plasma temperature is T0 inside the sunspot and TE outside.
HSE — Valery Nakariakov — Solar Physics 10

TE B 0 T0 P0 TE

PE
PE

BE= 0
BE = 0

Surrounding Surrounding

Intensive
magnetic
flux tube

Sunspots are long-durational objects with no fast flows of plasma. So, it

is naturally to describe their structure in terms of magnetostatics. As the
magnetic field is not bent, the last term in Eq. (15), responsible for the
magnetic tension, is zero. The equilibrium condition becomes
B2 
 

∇ P + = 0, (21)
2µ0
This means that the total pressure must be equal inside and outside the
HSE — Valery Nakariakov — Solar Physics 11

sunspot,
B02
PE = P0 + . (22)
2µ0

Let us assume that the density of the plasmas inside and outside the
sunspot are equal, ρ0 = ρE . Now, we divide Eq. (22) by ρ0 ,

PE P0 B02
= + . (23)
ρE ρ0 2µ0 ρ0

Using the state equations,

kB kB
PE = 2 ρE TE , P0 = 2 ρ0 T0 , (24)
mi mi
we obtain from Eq. (23)
2kB 2kB B02
TE = T0 + . (25)
mi mi 2µ0 ρ0
This gives us
T0 B02 mi B02
=1− =1− (26)
TE 2µ0 2kB ρE TE 2µ0 PE
Thus, in a sunspot, TE > T0 . Indeed, temperatures in the dark centers
of sunspots drop to about 3700 K, compared to 5700 K for the surround-
ing photosphere. This is why sunspots are seen to be darker than the
surrounding.
HSE — Valery Nakariakov — Solar Physics 12

MHD Waves

Ideal MHD connects the magnetic field B, plasma velocity V, pressure P

and density ρ:
∂ρ
+ ∇(ρV) = 0, (27)
∂t
∂V 1
" #
ρ + (V∇)V = −∇P − B × (∇ × B), (28)
∂t µ0
∂B
= ∇ × (V × B), (29)
∂t
d P
!
= 0. (30)
dt ργ
Consider an equilibrium, described by the conditions
∂
= 0, V = 0, (31)
∂t
which gives us the magnetostatic equation
1
∇P0 + B0 × (∇ × B0 ) = 0. (32)
µ0
The simplest possible solution of the magnetostatic equation is a uniform
plasma:
P0 = const, B0 = const, (33)
and the equilibrium magnetic field B0 is straight.

Consider small perturbations of the equilibrium state:


B = B0 + B1 (r, t) 



V = 0 + V1 (r, t)



(34)
P = P0 + P1 (r, t) 



ρ = ρ0 + ρ1 (r, t)




Substitute these expressions into the MHD equations (27)–(30). Neglecting

terms which contain a product of two or more values with indices “1”, we
obtain the set of MHD equations, linearized near the equilibrium (33):
∂ρ1
+ ρ0 ∇V1 = 0, (35)
∂t
HSE — Valery Nakariakov — Solar Physics 13

∂V1 1
ρ0 = −∇P1 − B0 × (∇ × B1 ), (36)
∂t µ0
∂P1 γP0 ∂ρ1
− = 0, (37)
∂t ρ0 ∂t
∂B1
= ∇ × (V1 × B0 ), (38)
∂t
Let the equilibrium magnetic field B0 be in xz-plane,
B0 = B0 sin α ex + B0 cos α ez , (39)
where α is the angle between the magnetic field and the unit vector ez :

Consider plane waves, propagating along ez , so that all perturbed quan-

tities are proportional to exp(ikz − iωt). (This gives us ∂/∂t = −iω and
∇ = ik.) Projecting equations (35)–(38) onto the axes, we have
−iωρ1 + ikρ0 Vz1 = 0, (40)
ikB0 cos α
−iωρ0 Vx1 − Bx1 = 0, (41)
µ0
ikB0 cos α
−iωρ0 Vy1 − By1 = 0, (42)
µ0
ikB0 sin α
−iωρ0 Vz1 + ikP1 + Bx1 = 0, (43)
µ0
−iωBx1 + ikB0 sin αVz1 − ikB0 cos α Vx1 = 0, (44)
−iωBy1 − ikB0 cos α Vy1 = 0, (45)
−iωBz1 = 0, (46)
iωγP0
−iωP1 + ρ1 = 0. (47)
ρ0
HSE — Valery Nakariakov — Solar Physics 14

The set of equations (40)–(47) splits into two partial sub-sets. The first
one is formed by equations (42) and (45), describing By1 and Vy1 . The
consistency condition gives us
ω 2 − CA2 cos2 α k 2 = 0, (48)
where CA ≡ B0 /(µ0 ρ0 )1/2 is the Alfvén speed. This is dispersion relations
for Alfvén waves.

Main properties of Alfvén waves:

• they are transverse, V ⊥ k;

• Alfvén waves can be linearly polarised, elliptically polarised, or circu-
larly polarised;
• they are essentially incompressive: they do not modify the density of
the plasma, ∇.V = 0;
• their group speed is always parallel to the magnetic field, Vgroup k B0 ;
while the phase speed can be oblique to the field, Vphase may be 6k B0 ;
Vphase 6k Vgroup ;
• the absolute value of the group speed equals the Alfvén speed, CA .
HSE — Valery Nakariakov — Solar Physics 15

Lecture 5: MHD Waves (cont’d); Non-ideal

Effects; Magnetic Diffusion; Alfvén’s Theorem
The second partial set of equations is formed by equations (40), (41), (43),
(44) and (47) and describes variables Vx1 , Vz1 , Bx1 , P1 and ρ1 . The consis-
tency condition gives us
(ω 2 − CA2 cos2 α k 2 )(ω 2 − Cs2 k 2 ) − CA2 sin2 α ω 2 k 2 = 0, (49)
where Cs ≡ (γp0 /ρ0 )1/2 is the sound (or acoustic) speed.
Equation (49) is bi-quadratic with respect to ω and consequently has two
pairs of roots. They correspond to fast and slow magnetoacoustic waves.

Main properties of magnetoacoustic waves:

• they are longitudinal, Vkk;

• they are essentially compressive: they always perturb the density of
the plasma;
• the fast wave can propagate in the direction perpendicular to the
field at the speeds Vphase and Vgroup equal to the fast speed, VF ≡
 1/2
CA2 + Cs2 ;
• the fast wave cannot propagate along the field — if kkB0 the fast wave
becomes incompressive and degenerates to the Alfvén wave;
• in the β < 1 case, the slow wave propagates along the field at the speed
Cs and degenerates to the usual acoustic wave; the slow wave cannot
propagate across the field;
• in the slow wave the density and the absolute value of the magnetic
field are perturbed in anti-phase, while in the fast wave in phase.

Polar plots for phase speeds (ω/k) and group speeds (dω/dk) for β < 1:
HSE — Valery Nakariakov — Solar Physics 16
HSE — Valery Nakariakov — Solar Physics 17

Non-Ideal MHD Equations

We can account for non-ideal (e.g. dissipative) effects. In this case the set
of MHD equations become

∂V 1
ρ + ρ(V · ∇)V = −∇P − B × ∇ × B + F, (50)
∂t µ

∂B
= ∇ × (V × B) + η∇2 B, (51)
∂t

∂ρ
+ ∇ · (ρV) = 0, (52)
∂t

ργ d p
!
= −L. (53)
γ − 1 dt ργ
The parameter η is the magnetic diffusivity, connected with the electrical
conductivity σ,
η = 1/(µσ). (54)

The term F is an external force acting on a unit of volume of the plasma.

For example, if we take into account the gravity and the viscosity,

1
" #
2
F = −ρ g + νρ ∇ V + ∇(∇ · V) , (55)
3
where g is the gravity acceleration and ν is the coefficient of kinematic
viscosity (assumed uniform).
Incompressible limit. Consider the situation when ρ = ρ0 = const. Then,
from the continuity Eq.,
∂ρ0
0= = −∇ · (ρ0 V) = 0, (56)
∂t

ρ0 ∇ · (V) = 0, ∇ · (V) = 0 (57)

Motions which satisfy this condition (e.g. Alfvén waves) are incompress-
ible.
HSE — Valery Nakariakov — Solar Physics 18

The ratio of specific heats γ is usually about 5/3. In some cases, isothermal
(T = const) processes can be considered, with γ = 1.
The righthand side of equation (53) contains the energy loss/gain function
L, discussed later. When L = 0, the equation reduces to the adiabatic
equation.
In addition, the electric current density j, the electric field E and the
temperature T can be determined from the equations:
j = ∇ × B/µ, (58)
E = −V × B + j/σ, (59)
kB
P = ρT, (60)
m
where kB is the Boltzmann constant and m is the mean particle mass.
In equation (60), the expression kB /m = R̃/µ̃, where R̃ and µ̃ = m/mp
are the gas constant and the mean atomic weight (the average mass per
particle in units of the proton mass), is often used.
If there are only protons and electrons, ne = np ,
ne me + mp np mp np
m= ≈ = 0.5mp , (61)
ne + np ne + np
c.f. Eq. (4).
For example, in the solar corona the presence of He (alpha-particles) and
other elements (in addition to H) makes m/mp = µ̃ ≈ 0.6.
The total number of particles per unit volume
np + ne ≈ 2ne
n= . (62)
np + ne + nother ≈ 1.9ne

Consequently,
ρ = np mp + ne me + nother mother ≈ ne mp (63)

The adiabatic equation ((53) with L = 0) can also be taken in several

different forms, e.g.
dp γp dρ
− =0 (64)
dt ρ dt
(Exercise: Derive it from equation (53))
HSE — Valery Nakariakov — Solar Physics 19

Non-adiabatic effects in the energy equations

The right hand side of equation (53) is the energy loss/gain function,
L = ∇ · q + Lr − j 2 /σ − H, (65)
where

q is the heat flux due to the thermal conduction, q = −κ∇T , with κ being
the thermal conductivity.
Lr is the radiation function, in the optically thick plasma of the solar
interior it is Lr = −κr ∇2 T , with κr being the coefficient of radiative
conductivity;
j 2 /σ is the ohmic dissipation; and
H represents the sum of all the other heating sources.

In rarified and magnetised plasmas,

q = −κ̂ ∇T, (66)
where κ̂ is the thermal conduction tensor. In this case
∇ · q = ∇|| · (κ|| ∇|| T ) + ∇⊥ · (κ⊥ ∇⊥ T ). (67)

Thermal conduction along the field is primarily by electrons,

κ|| = 10−11 T 5/2 Wm−1 K−1 . (68)

Conduction perpendicular to the field is mainly by protons, and

2
κ⊥ −31 n
= 2 × 10 , (69)
κ|| T 3B 2
where the field is in teslas.
In strongly magnetised plasmas, thermal conduction across the magnetic
field is dramatically depressed.
Estimate the ratio for the coronal parameters, n = 1015 m−3 , T = 106 K
and B = 10 G = 10−3 T.
In the optically thin part of the solar atmosphere (where T ≥ 2 × 104 K)
(what are the parts of the atmosphere where this condition is fulfilled?)
the radiation function takes the form
Lr = ne nH Q(T ) ≈ n2e Q(T ), (70)
HSE — Valery Nakariakov — Solar Physics 20

where ne is the electron concentration and nH is the hydrogen concentra-

tion, (ne ≈ nH ) and Q(T ) is a function of temperature T . Often, the
function can be approximated as χT α , where χ and α are constant:

When the plasma pressure P remains constant (isobaric processes), a con-

venient alternative form of equation (53) is
dT
= −L,
ρcp (71)
dt
where cp is specific heat at constant pressure,
γ kB
cp = .
γ−1m

4. Energetics
In MHD, three different types of energy are considered:

• internal energy ↔ entropy,

• EM energy ↔ Poynting flux,
• mechanical energy ↔ kinetic energy

increase
= heat flux − radiation + heat sources
in entropy

inflow of EM electrical a rise in magnetic

= +
energy E × H energy E · j energy B 2 /2µ

d V 2 
 

ρ = −V · ∇P + V · j × B + V · F
dt 2
HSE — Valery Nakariakov — Solar Physics 21

Consequences of the Induction Equation

Consider the induction equation with the diffusive term

∂B
= ∇ × (V × B) + η∇2 B (72)
∂t
Compare the right hand side terms.
Let the plasma have the typical speed V0 and the length scale l0 , then

V0 B
“convective term”: ∇ × (V × B) ≈ , (73)
l0

ηB
“diffusive term”: η∇2 B = . (74)
l02

Their ratio is
V0 B l02 l0 V0
= = Rm . (75)
l0 ηB η

This dimensionless parameter Rm is called the magnetic Reynolds number.

HSE — Valery Nakariakov — Solar Physics 22

1. Diffusive Limit
If Rm  1, the convective term can be neglected with respect to the
diffusive term, and the induction equation becomes
∂B
= η∇2 B. (76)
∂t

This is the diffusion equation. It implies that field variations on a length

scale l0 are destroyed over a diffusion time scale,
τd = l02 /η. (77)

The smaller the length-scale, the faster the magnetic field diffuses away.
In a fully-ionised plasma,
τd ≈ 10−9 l02 T 3/2 , (78)
where the length scale is in m and the temperature in in K.
E.g., in the solar corona, T = 106 K, and the typical length scale is 1 Mm=
106 m, thus
τd ≈ 10−9 1012 109 = 1012 s = 30, 000 years (!!) (79)

Since solar flares represent a release of magnetic energy over a time-scale

of 100 or 1000 s, it seems that a length-scale as small as
τd = 102 ≈ 10−9 l02 109 → l0 ≈ 10 − 100 m (80)

Example: Consider the diffusion of a unidirectional magnetic field B =

B(x, t)ey with the initial step-function profile:

 +B0 , x > 0
B(x, 0) =  (81)
−B0 , x < 0
(“a current sheet”, e.g. in helmet streamers).
HSE — Valery Nakariakov — Solar Physics 23

(Image of the corona taken during a solar eclipse. Several helmet streamers
are well seen.)
Suppose the field remains unidirectional. Then Eq. (76) becomes
∂B ∂ 2B
=η 2. (82)
∂t ∂x

The PDE should be supplemented by initial conditions (208) and the

boundary conditions, e.g.
B(±∞, t) = ±B0 . (83)

The solution which satisfies the boundary conditions is

B(x, t) = B0 erf(ξ), (84)
where ξ = x/(4ηt)1/2 , and
2 Z ξ
erf(ξ) = exp(−u2 ) du, (85)
π 1/2 0
is the error function.
HSE — Valery Nakariakov — Solar Physics 24

(Here, there are three curves shown, corresponding to t = 0, t = 1 and

t = 2. Mark the curves in Figure with the appropriate value of t.)
The gradient of the magnetic field causes the current:
1 1 dB
j= ∇×B ⇒ jz = (86)
µ µ dx
What happens with the other components of the current density? Draw the
structure of the current in Figure for different times.
The width of the current sheet behaves like l = 4(ηt)1/2 . Notice that the
field density at large distances remains constant in time. The field lines
in the sheet are not moving outwards, since those at large distances are
unaffected. Rather, the field in the sheet is diffusing away, and so it is
being annihilated. (The magnetic energy is being converted into heat by
ohmic dissipation).
HSE — Valery Nakariakov — Solar Physics 25

2. Perfectly Conducting Limit

When Rm  1, the induction equation reduces to
∂B
= ∇ × (V × B) (87)
∂t

In this large magnetic Reynolds number limit, the frozen-flux (Alfvén’s)

theorem holds:
In a perfectly conducting plasma, magnetic field lines behave as
if they move with the plasma.
In other words:
The total amount of magnetic flux passing through any closed
circuit moving with the local fluid velocity is constant in time
Proof: We show the above statement by proving that the time rate of
change of the magnetic flux through such a circuit is zero.
Consider a closed curve C bounding a surface S which is moving with the
plasma.

The magnetic flux of the field B through the elementary area A is

B · n dA,
where dA is a differential of area enclosed within the circuit and n is the
unit vector normal to A.
The flux through a circuit may change if
HSE — Valery Nakariakov — Solar Physics 26

• either the field strength at a point enclosed by the circuit changes

• or the motion of the boundary results in change in the amount of the
field enclosed.

The first type of change is given by

∂B
· n dA. (88)
∂t
The total change is given by integrating over the entire surface S.
The second type of change can be visualised by imagining a piece dl of
boundary moving at velocity V past a magnetic field B. The change in
the amount of flux enclosed within the area bounded by the curve due to
the motion of dl during time t is
B · (V × dl)t. (89)
In the following, consider the change of the flux during a unit time, t = 1.
Using the vector identity
a · (b × c) = (a × b) · c,
we can rewrite Eq. (89) as
B · (V × dl) = (B × V) · dl = −(V × B) · dl. (90)
The total change in flux through the circuit due to motion of the contour
C, is obtained by integrating the above expression around C.
Combining both the effects, the total change of the flux through a circuit:
d Z Z ∂B I
B · n dA = · ndA − (V × B) · dl. (91)
dt S S ∂t C

We can then use the Stokes theorem,

Z I
(∇ × Q) · n dA = Q · dl,
S C
and get I Z
(V × B) · dl = [∇ × (V × B)] · n dA. (92)
C S

Consequently, Eq. (91) becomes

d Z ∂B
Z " #
B · n dA = − ∇ × (V × B) · n dA, (93)
dt S S ∂t
HSE — Valery Nakariakov — Solar Physics 27

and, together with the induction equation, it gives

d Z
B · n dA = 0. (94)
dt S

Thus, the magnetic flux passing through the circuit C is constant.

Consequently, the magnetic field lines are frozen into the plasma: plasma
can move freely along field lines, but, in motion perpendicular to them,
either the field lines are dragged with the plasma or the field lines push
the plasma.
Alfvén’s theorem prohibits reconnection of magnetic field lines.
Consequently, the magnetic field lines are frozen into the plasma: plasma
can move freely along field lines, but, in motion perpendicular to them,
either the field lines are dragged with the plasma or the field lines push
the plasma.
Using the identity
∇ × (a × b) = a(∇ · b) − b(∇ · a) + (b · ∇) a − (a · ∇) b
we may rewrite the induction equation as
dB
= (B · ∇)V − B(∇ · V), (95)
dt
with the use of ∇B = 0. Also, notice that there is the total derivative on
the LHS of (95).
Consider a magnetic flux tube: it is the volume enclosed by the set of field
lines which intersect a simple closed curve.
For example, sunspots, coronal loops, erupting prominence, magnetic ele-
ments:
HSE — Valery Nakariakov — Solar Physics 28

Eq. (95) implies that changes in the magnetic field following the motion
are produced when a flux tube is stretched, sheared or expanded.
According to the first term:

• an accelerating motion along the field amplifies the field,

• a shearing motion normal to the field makes the field change direction
by increasing the field component along the flow direction.

According to the second term:

• an expansion (∇ · V > 0) decreases the field,

• a compression (∇ · V < 0) amplifies the field.
HSE — Valery Nakariakov — Solar Physics 29

Lecture 6: Hydrostatic Equilibrium; Thermal

equilibrium; Parker’s Solar Wind; Coronal Loop
Equilibria
Hydrostatic Pressure Balance

The magnetohydrostatic equilibrium condition is

0 = −∇P + j × B + ρg, (96)

coupled with
∇ · B = 0, (97)
µj = ∇ × B, (98)
ρRT
P = , (99)
µ̃
and T satisfies an energy equation.
Before investigating any specific phenomena we need to consider the basic
pressure balance when the magnetic field does not exert any force.
Consider the simple case of a uniform vertical magnetic field. For simplicity
we assume that the temperature is known.
Thus,

B = B0 ẑ, g = −gẑ.

Hence, j = 0 and there is no Lorentz force.

In addition, the pressure is P = P (z) and (181) becomes
dP g µ̃ P (z)
= −ρ(z)g = − P (z) = − , (100)
dz RT (z) Λ(z)
where
RT (z)
Λ(z) = , (101)
µ̃g
is the pressure scale height.
HSE — Valery Nakariakov — Solar Physics 30

Eq. (100) is a separable, first order ordinary differential equation so that

dP 1
=− dz,
P Λ(z)
and integrating gives
log P = −n(z) + log P (0),
where
1 Z z
n(z) =du,
0 Λ(u)

is the ‘integrated number’ of scale heights between the arbitrary level at

which the pressure is P (0) and the height z. Therefore,

P (z) = P (0) exp [−n(z)]. (102)

If the atmosphere is isothermal so that both T and Λ are constant, then
(102) gives
P (z) = P (0) exp (−z/Λ), ρ(z) = ρ(0) exp [−z/Λ], (103)
so that the pressure decreases exponentially on a typical length scale given
by the pressure scale height Λ:

(Here, three curves are shown, corresponding to Λ = 1, Λ = 2 and Λ = 3.

Mark the curves in the figure with the appropriate value of Λ.)
Consider typical values of the pressure scale height. Taking the solar grav-
itational constant as g = 274 ms−2 and R = 8.3 × 103 J K−1 mol−1 then Λ
takes the following values:
HSE — Valery Nakariakov — Solar Physics 31

1. In the photosphere T = 6, 000 K and µ̃ = 1.3 so that

RT 8.3 × 103 × 6 × 103
Λ= = = 140 km.
µ̃g 1.3 × 274

2. In the corona T > 106 K and µ̃ = 0.6 giving

RT 8.3 × 103 T
Λ= = ≈ 50.5T m.
µ̃g 0.5 × 274
Thus, the scale height can be estimated as
Λ/Mm ≈ 50T /MK.
E.g., the scale height of the corona observed by TRACE-171 Å (the
temperature is 1 MK) is 50 Mm.
(This figure is comparable to the size of a loop).
HSE — Valery Nakariakov — Solar Physics 32

Example: Density stratification in a polar plume.

Polar plumes are cool, dense, linear, magnetically open structures that
arise from predominantly unipolar magnetic footpoints in the solar
polar coronal holes.

The solid line shows the hydrostatic solution for T = 106 K.

3. In the Earth’s atmosphere T = 300 K, g = 9.81 ms−1 , µ̃ = 29 in air
and so
8.3 × 300
Λ= = 8.7 km.
29 × 9.81
HSE — Valery Nakariakov — Solar Physics 33

Note that the height of Mount Everest is about 8.8 km. Thus, the air
pressure at the summit of Everest is about 1/e = 0.37 that of the air
pressure at sea level.
HSE — Valery Nakariakov — Solar Physics 34

Hydrostatic equilibrium at larger heights

(Image of the solar corona taken with a coronagraph).

Determination of the density stratification on larger scale, e.g. in coronal
holes, requires taking into account the effects of the spherical geometry
and the change of the gravitational acceleration with height,
GM
g(r) = (104)
r2
where r is the radial coordinate.
The magnetic field is assumed to be strictly radial,
2
B0 R
B= . (105)
r2

In the following we consider spherically symmetric isothermal (T = const)

atmosphere.
Again, there is no the Lorenz force. The magnetostatic equation is similar
to (100),
dP (r)
= −ρ(r)g(r), (106)
dr
HSE — Valery Nakariakov — Solar Physics 35

which, with the use of the state equation,

ρRT
P =
µ̃
can be rewritten as
dρ(r) R2 1
= − 2 ρ(r). (107)
dr r Λ
Here, the scale height Λ was defined by substituting the value of the gravi-
2
tational acceleration g(R ) = GM /R at the solar surface into Eq. (101).
ODE (107) is separable,
Z dρ Z R2

=− dr (108)
ρ Λr2
with the solution
R2
 

ρ(r) = C exp    . (109)

Λr

Determining the constant C from the condition ρ(R ) = ρ0 , we obtain

R R
" !#
ρ(r) = ρ0 exp −1 . (110)
Λ r

This solution coincides well with the observationally determined empirical

dependence. E.g., the profile of plasma concentration in polar coronal holes
determined with SPARTAN 201-01 (from Fisher & Guhathakurta 1995):
HSE — Valery Nakariakov — Solar Physics 36

An empirical model was constructed by Esser et al. (1999),

2.494 × 106 1.034 × 107 3.711 × 108
ne = + + , (111)
r3.76 r9.64 r16.86
which corresponds to the theoretical dependence reasonably well.
Notice that Eq. (110) gives infinite density at r → ∞, so can be applied
at the heights below 5–6 R only. At larger distances from the Sun steady
flows of plasma (the solar wind) must be accounted for.
HSE — Valery Nakariakov — Solar Physics 37

Parker’s solar wind model

The corona cannot remain in static equilibrium but is continually expand-
ing. The continual expansion is called the solar wind.
Assume that the expanding plasma of the solar wind is isothermal and
steady.
The governing equations can be obtained from the MHD equations setting
∂/∂t = 0:
∇ · (ρV) = 0, (112)
ρ(V · ∇)V = −∇P + ρg, (113)
P = ρRT, (114)
and
T = T0 . (115)

Also, we restrict our attention to the spherically symmetric solution. The

velocity V is taken as purely radial, V = ver and the gravitational accel-
eration g = ger obeys the inverse square law,
GM
g=− 2 . (116)
r
The temperature and, consequently, the sound speed
Cs2 = P/ρ, (117)
are constant.
We are interested, for simplicity, in the dependence on the r coordinate
only. Thus, the expressions for the differential operators in the spherical
coordinates are
da 1 d  2 
∇a = , ∇·A= 2 r Ar .
dr r dr
In the spherical geometry, the governing equations describing and the ra-
dially symmetric values are

dv d p GM ρ
ρv =− − , (118)
dr dr r2

d 2
(r ρv) = 0 ⇒ r2 ρv = const. (119)
dr
HSE — Valery Nakariakov — Solar Physics 38

Substituting (117) into (118) we exclude the pressure from the equations,
dv d ρ GM ρ
ρv = −Cs2 − , (120)
dr dr r2
or
dv 1 d ρ GM
v = −Cs2 − 2 . (121)
dr ρdr r

To exclude ρ, we use (119),

d 2 d dρ
(r ρv) = ρ (r2 v) + r2 v = 0, (122)
dr dr dr
and obtain
1 dρ 1 d 2
=− 2 (r v). (123)
ρ dr r v dr

Now, Eq. (121) becomes

dv Cs2 d 2 GM
v = 2 (r v) − 2 . (124)
dr r vdr r

Rewriting this equation, we obtain

C2 d v 2Cs2 GM
 
v − s  = − 2 , (125)
v dr r r
and, then
C2 d v C2
 
v − s  = 2 2s (r − rc ) , (126)
v dr r
where rc = GM /(2Cs2 ) is the critical radius showing the position where
the wind speed reaches the sound speed, v = Cs .
This is a separable ODE, which can readily be integrated,
C2 C2
 
Z Z
v − s  dv = 2 2s (r − rc ) dr, (127)
v r
giving the solution
!2 !2
v v r rc
!
− log = 4 log + 4 + C. (128)
Cs Cs rc r

The constant of integration C can be determined from boundary condi-

tions, and it determines the specific solution. Several types of solution are
present in the figure:
HSE — Valery Nakariakov — Solar Physics 39

Types I and II are double valued (two values of the velocity at the same
distance), and are non-physical.
Types III has supersonic speeds at the Sun which are not observed.
Types IV seem also be physically possible. (The “solar breeze” solutions).
The unique solution of type V passes through the critical point (r = rc , v =
Cs ) and is given by C = −3. It can be obtained from the general solution
(128) by putting the coordinates of the critical point. This is the “solar
wind” solution (Parker, 1958). It was discovered by Soviet Luna-2, Luna-3
and Venera-1 probes in 1959.
Let us estimate the critical radius rc . For a typical coronal sound speed of
about 105 m/s, and the critical radius is
GM
rc = 2
≈ 6 × 109 m ≈ 9 − 10R . (129)
2Cs

At the Earth’s orbit, the solar wind speed can be obtained by substituting
r = 214R to Eq. (128), which gives v = 310 km/s.
For the radial flow, the rotation of the Sun makes the solar magnetic field
twist up into a spiral.
Suppose the magnetic field is inclined at an angle φ to the radial solar wind
velocity:
HSE — Valery Nakariakov — Solar Physics 40

The component of the vector V perpendicular to the vector B, v sin φ,

must equal the speed of the field line in that direction, because the field is
frozen in the plasma. But, the field is dragged by the solar rotation with
the angular frequency Ω. The normal component of the speed of the field
line is Ω(r − R ). Consequently,
v sin φ = Ω(r − R ) cos φ, (130)
which gives us
Ω(r − R )
tan φ = (131)
v
Taking v ≈ 310 km/s and calculating the frequency of the equatorial ro-
tation period, which is about 26 days, Ω = 2π/(26 × 24 × 60 × 60) ≈
2.8 × 10−6 rad/s, we obtain that near the Earth’s orbit, r ≈ 214R , the
angle is about 45◦ .
HSE — Valery Nakariakov — Solar Physics 41

In-situ observations have established that there are actually two component
in the solar wind,

• relatively low-speed streams (v < 350 km/s) - the “slow solar wind” and
• high-speed streams (v up to 800 km/s) - the “fast wind”.

The slow wind is denser and carries greater flux of particles. The presence
of the fast wind has been observed at higher solar latitudes:
HSE — Valery Nakariakov — Solar Physics 42
HSE — Valery Nakariakov — Solar Physics 43

Realtime monitoring of the solar wind near the Earth’s orbit:

http://www.swpc.noaa.gov/products/ace-real-time-solar-wind
HSE — Valery Nakariakov — Solar Physics 44

Coronal Loop Equilibria

1. Static Equilibria
Consider a semi-circular loop of constant cross-section with the curvature
radius RL . The plane of the loop is perpendicular to the solar surface. In
the following, we neglect 2D effects such as the loop curvature and twisting,
and transversal structuring. Consequently, we can consider the loop as a
straight cylinder, confined between two planes representing the footpoints:

Also, we assume that the gravitational acceleration does not change much
on the height of the loop and so is constant.
If s measures distance along magnetic field lines (along the z-direction in
the sketch above), the component of the magnetostatic equation
−∇P + j × B + ρg = 0, (132)
parallel to B is
dP
= −gρ. (133)
ds
To connect the pressure P and the density ρ we use the state equation
kB T
P = ρ. (134)
m
HSE — Valery Nakariakov — Solar Physics 45

We restrict our attention to consideration of the isothermal loops with the

stationary temperature T = const.
Assuming that the gravitational acceleration g is vertical and is of constant
magnitude g0 (= 274 m/s2 ), its projection to the loop gradient is
g(s) = g0 cos θ(s) (135)
where the angle θ measures the local inclination of the loop to the horizon.
In terms of the distance along the loop, the angle can be rewritten as
θ = s/RL , (136)
and, consequently,
s
!
g(s) = g0 cos . (137)
RL

Using (134) and (137), we can rewrite the magnetostatic equation (133) as
kB T dρ s
!
= −g0 cos ρ. (138)
m ds RL

Equation (138) is the 1st order linear ODE, and we can write the solution
as
0
   
Z s g m s
0
ρ = ρ0 exp − cos   ds0  , (139)
0 kB T RL
where ρ0 is the density at s = 0 (the loop footpoint).
Evaluating the integral, we obtain
g0 mRL s
" !#
ρ = ρ0 exp − sin , (140)
kB T RL

or, introducing Λ = kB T /gm as the scale height (the same as it was intro-
duced in the case of the vertical magnetic field),
RL s
" !#
ρ = ρ0 exp − sin . (141)
Λ RL

Distribution of density along the loop, the loop length is normalised to the
loop radius:
HSE — Valery Nakariakov — Solar Physics 46

(Here, three curves are shown, corresponding to Λ/RL = 0.5, Λ/RL = 1

and Λ/RL = 1.5. Mark the curves in Figure with the appropriate value of
Λ.)
Interestingly,
s
!
RL sin = z, (142)
RL
where z is the distance of a loop point from the surface (the height), and,
consequently,

z
" #
ρ = ρ0 exp − . (143)
Λ
Thus, the plasma inside a coronal loop has the same stratification as it is
in a non-structured atmosphere - c.f. the communicating water tubes.

2. Static Energy Balance Models

HSE — Valery Nakariakov — Solar Physics 47

Thermal equilibrium
In more realistic models, the hydrostatic equilibrium should be supple-
mented with the thermal equilibrium between thermal conduction, radia-
tion and heating (see Eq. (65)):
d dT
!
κ0 T 5/2 = χn2e T −1/2 − H, (144)
ds ds
where s is the coordinate along the magnetic field.

In particular, for short coronal loops, with the major radius shorter than
the scale height of the stratification, (RL < Λ) the loop pressure P (s) can
be taken to be constant,
P (s) = P0 (145)
and, consequently, from the state equation, the density is
ne (s) = P0 /2kB T (s). (146)

Assuming that all three terms in Eq. (144) are of the same order, we get,
comparing the terms on RHS of Eq. (144),
P02 χT −5/2
H≈ . (147)
4kB2
and the LHS and RHS terms,
κ0 T 7/2 P02 χT −5/2
≈ , (148)
RL2 4kB2
the scaling law:
T ∝ (P0 RL )1/3 . (149)
HSE — Valery Nakariakov — Solar Physics 48

This scaling law is called RTV (after Rosner, Tucker & Vaiana) was, as
the zero-order approximation, confirmed observationally in the soft X-ray
and EUV band:

3. Steady Flows in Loops

There are steady flows observed in coronal loops (usually, subsonic, below
the sound speed.

(Right: Intensity map of Fe xii 195Å for an active region taken with
Yohkoh/EIS Left: Doppler velocity map. Blue (Red) indicates that the
plasma moves toward (away) us.)
In particular, if there is a pressure difference between the two footpoints,
siphon flows are generated.
Consider steady (∂/∂t = 0) flow along a semicircular loop of uniform cross-
section (the same model as above).
HSE — Valery Nakariakov — Solar Physics 49

The MHD equations reduce to

d
(ρV ) = 0, (150)
ds
dV dP
ρV =− − ρg0 cos(s/RL ), (151)
ds ds
d P
!
= 0, (152)
ds ργ
where V is the component of the bulk velocity vector, tangential to the
loop.
Eqs. (150)-(152) can be combined into one ODE
C 2 dV
 
V − s  = −g0 cos(s/RL ), (153)
V ds
where the sound speed, Cs2 = γp/ρ is introduced.
The ODE is separable, and so can easily be integrated,
C2
 
Z Z
V − s  dV = −g cos(s/RL )ds, (154)
0
V
which gives us
V2 s
− Cs2 log V + C = −gRL sin . (155)
2 RL
Applying the boundary condition
V (s = 0) = V0 , (156)
where V0 is the initial speed of the flow, we obtain the solution
V 2 − V02 V0 s
+ Cs2 log = −gRL sin . (157)
2 V RL
HSE — Valery Nakariakov — Solar Physics 50

Eq. (153) has a critical point at the loop apex (top, or summit) s = πRL /2,
where the flow can become sonic (V = Cs ). The starting velocity V0∗ for
flow to pass through the critical point can be determined from (157) by
putting V = Cs and s = πRL /2:
Cs2 − V0∗ 2 2 V0∗
+ Cs log = −gRL . (158)
2 Cs

For initial speeds slower than V0∗ , the flow is subsonic and symmetric about
the loop apex, and for V0 > V0∗ , the results have no physical meaning.
4. Total Pressure Balance Across the Loop
In the direction across the loop axis, physical quantities inside and outside
the loop are connected with each other.
Consider the loop as a magnetic cylinder parallel to the surface of the Sun:
HSE — Valery Nakariakov — Solar Physics 51

The perpendicular component of magnetostatic equation (132) is

B2
  

−∇⊥ P + (j × B)⊥ = −∇⊥ P + (B · ∇)B/µ − ∇   =

2µ ⊥
B2 
 

−∇⊥ P − ∇⊥  = 0, (159)
2µ

Consequently, the total pressure balance must be kept across the loop,
B02 Be2
P0 + = Pe + . (160)
2µ 2µ
HSE — Valery Nakariakov — Solar Physics 52

Lecture 7: Potential and Force-Free Fields;

Prominences; Magnetic Reconnection
Potential and Force-Free Fields
Magnetic Field Lines
If the magnetic field B = (Bx , By , Bz ) is known as a function of position,
then the magnetic lines of force, called the magnetic field lines, are defined
by
dx dy dz ds
= = = . (161)
Bx By Bz B
The solution to (161), a system of ordinary differential equations, defines a
curve in three dimensional space that is the field line. In parametric form,
in terms of the parameter s, the field lines satisfy
dx Bx dy By dz Bz
= , = , = , (162)
ds B ds B ds B
where the parameter s is the distance along the field line.
Example: Consider the field given by B = B0 (y/a, x/a, 0), where B0 and
a are constants, calculate the equations of the field lines. Using (161), the
field lines are given by
dx dy
= , ⇒ xdx = ydy,
(y/a) (x/a)
and so
x2 − y 2 = ±c2 = constant.
Therefore the field lines are hyperbolae.
HSE — Valery Nakariakov — Solar Physics 53

y
2

–4 –2 0 2 x 4

–2

–4

This is a neutral point or an X-point.

HSE — Valery Nakariakov — Solar Physics 54

Potential Fields
If β  1, we may also neglect the gas pressure (with respect to the mag-
netic pressure!) magnetostatic equation (181) reduces to the low β plasma
approximation
j×B=0 (163)
and the magnetic field is called force-free.
A simple solution to (163) is given by assuming that the current density
j is identically zero so that the magnetic field is potential. Thus, the field
must satisfy the conditions
1
j = ∇ × B = 0, (164)
µ0
and
∇ · B = 0, (165)

The most general solution to (164) is

B = ∇φ, (166)
where φ is the scalar magnetic potential.
Substituting (166) into condition (165), we get
∂ 2φ ∂ 2φ ∂ 2φ
2
∇ φ= + + = 0, (167)
∂x2 ∂y 2 ∂z 2
which is Laplace’s equation. It is commonly used for the determination of
the basic geometry of the magnetic field. Eq. (167).
Separable solutions to (167):
We solve (167) in the two dimensional plane x − y where x and y are the
horizontal and vertical coordinates respectively, subject to the boundary
conditions
φ(x, 0) = F (x), φ(0, y) = φ(l, y) = 0, φ → 0 as y → ∞. (168)

Set φ = X(x)Y (y) and substitute into (167). This gives

X 00 Y + XY 00 = 0,
and rearranging we obtain
X 00 Y 00
=− = constant = −k 2 .
X Y
HSE — Valery Nakariakov — Solar Physics 55

This is equivalent to an equation for X(x) and one for Y (y). Hence,
Y 00 = k 2 Y ⇒ Y (y) = ae−ky + beky .
and
X 00 = −k 2 X, ⇒ X(x) = c sin kx + d cos kx.

Now applying the boundary conditions, (168) gives b = d = 0 and

nπ
sin kl = 0, ⇒k= ,
l
where n is integer. The full solution to φ is obtained by summing over all
the possible solutions. Defining Ak = ac, we obtain
Ak sin kxe−ky ,
X
φ(x, y) = (169)
k

where
X
F (x) = Ak sin kx, (170)
k
and
nπ
k= . (171)
l
As a simple example assume that F (x) is given by only one Fourier com-
ponent, namely F (x) = sin πx/l. This implies that all the coefficients in
the summation in (169) are zero except for the first one which is unity.
Thus, the potential solution is simply
πx
φ(x, y) = sin e−πy/l . (172)
l

Now that φ is known we may calculate the components of the magnetic

field by using B = ∇φ. In this way,
∂φ πx
Bx = = B0 cos e−πy/l , (173)
∂x l

∂φ πx
By = = −B0 sin e−πy/l (174)
∂y l
HSE — Valery Nakariakov — Solar Physics 56

Coronal Arcades:
In the photosphere magnetograms show that there are regions of opposite
polarity that are separated by a magnetic polarity inversion line. When
the magnetic field joins the opposite polarities across this inversion line the
field forms a “coronal arcade”. These arcades of magnetic field lines are
clearly seen in the soft x-ray images of the Sun.

Example: Potential extrapolation of the coronal field

It is impossible to measure directly the magnetic field in the corona. Mea-
surements of the magnetic field at the photosphere provide us with a bound-
ary condition for the Laplace equation. Solving it, e.g., in the spherical
geometry we get the geometry of the magnetic field in the corona, e.g.:
HSE — Valery Nakariakov — Solar Physics 57

However, potential fields do not have electric currents that are necessary
for plasma heating and impulsive energy releases, e.g. flares and coronal
mass ejections (CME).
Force-Free Fields
Again, we assume that λ  Λ and β  1 and we again have the force-free
field equation (163). If the magnetic field is not potential (|j| 6= 0) then
the general solution is that the current must be parallel to the magnetic
field. Thus,
µ0 j = αB, ⇒ ∇ × B = αB, (175)
for some scalar function α which may be a function of position and time.
Property of α: The scalar function α(r) is not completely arbitrary since
B must satisfy the conditions:

• ∇ · B = 0 and
• the vector identity ∇ · (∇ × B) = 0.

So using (175) we obtain

∇ · (∇ × B) = ∇ · (αB)
= α∇ · B + B · ∇α.
HSE — Valery Nakariakov — Solar Physics 58

Hence,
B · ∇α = 0, (176)
so that α is constant along each field line, although it may vary from field
line to field line. If α = 0, then the magnetic field reduces to the potential
case already considered.
If α is constant everywhere then

∇ × B = αB ⇒ ∇ × (∇ × B) = ∇ × (αB) = α∇ × B = α2 B.

However, ∇ × ∇ × B = ∇(∇ · B) − ∇2 B and so

−∇2 B = α2 B. (177)
This is a Helmholtz equation.
If α is a function of position, i.e. α(r), then we have
∇ × (∇ × B) = ∇ × (αB) = α∇ × B + ∇α × B
= α2 B + ∇α × B
Hence, we get two coupled equations for B and α, namely
∇2 B + α2 B = B × ∇α, (178)
and
B · ∇α = 0. (179)
They are usually solved numerically.
Example: Nonlinear Force-Free (NLFF) extrapolation of photospheric
sources:
HSE — Valery Nakariakov — Solar Physics 59

Example: Comparison of potential and NLFF extrapolations

HSE — Valery Nakariakov — Solar Physics 60

Prominences
Support of Prominences

Prominences (chromospheric filaments) are observed to situate above the

photospheric “neutral line”: the boundary between the regions of positive
and negative magnetic polarities:

The simplest model of the internal structure of a prominence was suggested

HSE — Valery Nakariakov — Solar Physics 61

by Kippenhahn and Schlüter (1957). The temperature is assumed to be a

constant T = T0 .
Since the width of a prominence is much shorter than the height and length,
we assume that inside the prominence we may neglect variations in the
vertical direction and only consider variations in the horizontal direction
across the prominence. Thus,

B = (Bx0 , By0 , Bz (x)) , p = p(x), ρ = ρ(x), (180)

where Bx0 and By0 are constants. These function must satisfy the magne-
tostatic equations:
0 = −∇p + j × B + ρg, (181)
coupled with
∇ · B = 0, (182)
µj = ∇ × B, (183)
ρRT
P = , (184)
µ̃
and T satisfies an energy equation.
The horizontal and vertical components of the force balance equation (181)
are

dP Bz dBz
=− , (185)
dx µ dx

Bx0 dBz
= ρg. (186)
µ dx

Since the temperature is constant we may use the gas law, (184) to elimi-
nate the density in favour of the pressure to obtain from (186)
Bx0 dBz P
= , (187)
µ dx Λ
where the pressure scale height, Λ = RT0 /µ̃g. Solving (185) we get

Bz2
P+ = constant. (188)
2µ
HSE — Valery Nakariakov — Solar Physics 62

To determine the constant we apply the boundary conditions that the

pressure and density tend to zero as we move away from the prominence
and that Bz tends to a constant value Bz0 . Thus,

P →0 as |x| → ∞ (189)
|Bz | → Bz0 as |x| → ∞ (190)
(191)

Thus,
1  2
Bz0 − Bz2 .

P = (192)
2µ

Substituting (192) into (187) gives

Bx0 dBz 1  2
Bz0 − Bz2 ,

=
µ dx 2µΛ
Z dBz x
2 = + constant,
Bz0 − Bz2 2ΛBx0
1 −1 Bz x
!
tanh = + C,
Bz0 Bz0 2ΛBx0
Bz0 x
!
Bz = Bz0 tanh +C .
2Bx0 Λ

From symmetry at x = 0 we must have Bz (0) = 0 and this gives C = 0.

Therefore,

Bz0 x
!
Bz = Bz0 tanh . (193)
2Bx0 Λ
and the pressure from (192) is
2
Bz0 Bz0 x
!
P = sech2 . (194)
2µ 2Bx0 Λ

Since the temperature is constant the density is given by the gas law as
2
µ̃ Bz0 Bz0 x
!
ρ= sech2 . (195)
RT0 2µ 2Bx0 Λ
The profiles of Bz and ρ are shown in figures:
HSE — Valery Nakariakov — Solar Physics 63

Equation of the field lines for the Kippenhahn–Schlüter prominence model

is
dx dz
= ,
Bx Bz

Bz0 Bz0 x
Z !
⇒ tanh dx = z + c,
Bx0 2Bx0 Λ
and so integrating we obtain
Bz0 x
( !)
2Λ log cosh = z + c.
2Bx0 Λ
HSE — Valery Nakariakov — Solar Physics 64

Note how the magnetic field lines are bent and the magnetic tension force
opposes the force due to gravity. In addition, the magnetic pressure is
higher away from the centre of the prominence and so there is a magnetic
pressure acting towards the centre that compresses the plasma and opposes
the outward pressure gradient.
Thus, the key element of the magnetic configuration that supports a promi-
nence is the magnetic dip.
HSE — Valery Nakariakov — Solar Physics 65

Another possible magnetic configuration with a magnetic dip is the “inverse

polarity”:

Sometimes, under the prominence there could be a magnetic x-point.

A possible scenario of the magnetic dip formation:
HSE — Valery Nakariakov — Solar Physics 66

2. Thermodynamics of Prominences

3.1 Radiatively-Driven Thermal Instability

Consider a plasma which is in equilibrium with temperature T0 and density
ρ0 , under a balance between

• mechanical heating of amount H = hρ per unit volume (where h is

constant, and
• optically thin radiation of amount Lr = χρ2 T α (where χ and α are
constant).

Also, we assume that the thermal conduction and the ohmic dissipation
are negligible.
So, the energy loss function is
L = Lr − H = χρ20 T0α − hρ0 = 0, (196)
(It is equal to zero because of the equilibrium condition).
For a perturbation at constant pressure, the energy equation becomes
∂T
cp = h − χρT α , (197)
∂t
Also, we can express the density through the temperature to close the
equation,
mp0
ρ= . (198)
kb T
So, Equation (22) of Lecture 2 becomes
T α−1 
 
∂T α
cp = χρ0 T0 1 − α−1 . (199)
∂t T0

Thus, if α < 1, a small decrease in temperature (T < T0 ) makes the right-

hand side negative, so that ∂T /∂t and the perturbation continues: this
is the thermal instability. If the plasma at coronal temperatures (α < 1)
locally cools, the radiation increases and the plasma cools further. This
runaway process continues until the temperature reaches the value when
α > 1 (about 104 K), and a new equilibrium is reached.
α > 1: α < 1:
HSE — Valery Nakariakov — Solar Physics 67

1 1

0.8 0.8

0.6 0.6
T/T_0 T/T_0
0.4 0.4

0.2 0.2

0 1 2 t/t_rad 3 4 5 0 1 2 t/t_rad 3 4 5

The time-scale (growth-time) of the instability is

cp
τrad = . (200)
χρ0 T0α−1

The growth-time (s) for thermal instability in a plasma of number density

n0 (m−3 ) and temperature T0 (K):

n0 T0
105 5 × 105 106 2 × 106 107
1014 440 2200 3.2 × 104 1.3 × 105 3.2 × 106
1015 44 220 3.2 × 103 1.3 × 104 3.2 × 105
1016 4.4 22 320 1.3 × 103 3.2 × 104

The thermal instability can be prevented by the efficient heat conduction

(along magnetic field lines).
Taking into account this effect, Eq. (197) becomes

∂T 1
= h − χρT α + ∇ · (κ|| ∇T ),
cp (201)
∂t ρ0
where κ|| is the coefficient of thermal conduction parallel to the field,

κ|| = κ0 T 5/2 . (202)

HSE — Valery Nakariakov — Solar Physics 68

Thus, if the length of a field line is L, the thermal conduction time is

L2 ρ 0 cp
τcon = 5/2
. (203)
κ0 T0

When the length L is so small that

τcon < τrad , (204)
the plasma is stable. This condition can be rewritten as
 1/2
7/2−α
κ0 T0
L < Lmax = 
. (205)
χρ20
 

Thus, when the length of magnetic field lines (say, in a coronal loop) ex-
ceeds the threshold value Lmax , the plasma filling the loop becomes ther-
mally unstable and cools down until a new equilibrium is reached, with
cooler temperature and, consequently, higher density:
condensation takes place.
HSE — Valery Nakariakov — Solar Physics 69

Magnetic reconnection
If sharp gradients (small characteristic spatial scales) are allowed to form
then diffusion processes can become important, Rm  1 (locally). This
diffusion can lead to the violation of Alfvén’s theorem, and allow changes
in the more global topology of the field.

Consider a current sheet. In its very vicinity the resistivity can be taken
as finite.

The plasma diffuses into the current layer at some relatively small inflow
velocity Vin . (More specifically: there is the total pressure balance across
the current sheet; in the vicinity of the current sheet there are large gradi-
ents of the field and hence diffusion; the total pressure outside the current
sheet is getting higher, resulting into a pressure gradient forces, moving
field lines toward the current sheet from the top and bottom).
In the current sheet the oppositely directed magnetic field lines get recon-
nected, resulting into the magnetic tension forces in the horizontal direc-
tion. These forces drive the frozen-in plasma — the sling shot effect.
HSE — Valery Nakariakov — Solar Physics 70

The plasma is accelerated along the layer (in the sketch in the horizontal
direction), and eventually expelled from its ends at some relatively large
velocity Vout .
It is the Sweet–Parker stationary reconnection.
• Conservation of mass: mass flux in equals mass flux out (in the incom-
pressible limit):
2LVin ≈ 2δVout
• Conservation of energy (magnetic energy flux in equals kinetic energy
flux out, neglecting heating etc.):
B2  2
   
ρVout
2LVin  ≈ 2δVout  
2µ 2
Here we took that in the same time t the side “L” moves the distance
dL ≈ Vin t, while the side δ moves the distance dδ ≈ Vout t. Thus on the
LHS the volume is 2LdL, while on the RHS it is 2δdδ
Combining these two equations shows that the outflow scales with the
upstream Alfvén speed:
B
Vout ≈ CA = . (206)
(µρ)1/2

The timescale for propagation of the outflow across the system, τout ≈
L/Vout , is much larger than the diffusion timescale, τd ≈ L2 /η, see the
discussion of the diffusion equation (77). The ratio of these time scales,
τd LVout
≈ =S (207)
τout η
— the Lundqvist number. This number determines the Sweet–Parker re-
connection rate, MSP = S −1/2 .
More general estimation of the energy conversion in Sweet–Parker recon-
nection:

• The input energy is the energy stored in the magnetic field.

• Change of B because of reconnection generates steep gradients of B,
hence increase in ∇ × B. It leads to the increase in the current density
j.
HSE — Valery Nakariakov — Solar Physics 71

• As the diffusivity is not negligible in the reconnection region (in the

vicinity of the current sheet), the current is subject to Ohmic dissipa-
tion, hence increase in internal energy of the plasma.
• Also, the slingshot effect generates bulk flows of plasma, hence increase
in its kinetic energy.
• The electric field E = −V × B accelerates plasma particles: non-
thermal high energy particles.

There is however a problem: for typical parameters of the corona, the

characteristic time of energy release by magnetic reconnection is about a
few tens of days, as the Lundqvist number is very large, S = 1012 . This
is too long to explain dynamical phenomena (e.g. flares and CME) in the
solar atmosphere. The problem of “fast reconnection” is one of the key
problems of modern solar and space plasma physics. Possible solutions:
anomalous resistivity, non-MHD processes...
For example, Petschek Reconnection: a much larger reconnection rate
would be possible if the diffusion region were much shorter, LP  L —
reconnection of an X-point rather than a current sheet. The inflow and
outflow are separated by slow mode shocks.
HSE — Valery Nakariakov — Solar Physics 72

Acceleration of the outflows occurs outside the diffusion region, at the

standing slow shocks (which are also current sheets).
The rate of Petchek reconnection is MPet = π/8log S.
Another option: anomalous resistivity that arises through a number of
(non-MHD) ways of scattering particles and retarding the flow of current,
e.g.:
Wave-particle interactions in the diffusion region;
Turbulent/stochastic motions in the diffusion region;
Effects of various possible instabilities in the diffusion region.
For example, numerical experiments on the onset of anomalous resistivity:

In this case, the resistivity of the plasma is


 ηclassical , |j| < jcritical ,
η= (208)
ηanomalous , |j| > jcritical
Thus, the anomalous resistivity may occur in the regions of steep gradients
of the magnetic field, e.g. near current sheets.
Plasmoid instability: fragmentation of the reconnecting current sheet in a
number of magnetic flux ropes (“plasmoids”).
HSE — Valery Nakariakov — Solar Physics 73

It may increase the volume in which the reconnection occur, and hence the
amount of the converted energy.