(iii) Theoretical Geomagnetism Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.0 Introduction
In earlier lectures we have established that: (i) Earths outer core consists of liquid metal (mostly Fe). (ii) The majority of the geomagnetic field originates in the core. (iii) This magnetic field changes on timescales of 1 up to 108 yrs. It seems likely that it is the motion of liquid metal in the core that both generates the geomagnetic field and causes its time variations. But, how does this happen in detail? To answer we need to understand the physics of electromagnetism applied to moving conductors..........

Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.1 Induction by a moving conductor 8.2 Maxwells equations of Electromagnetism 8.3 Approximations and moving reference frames 8.4 The Induction Equation 8.5 Summary Appendix: Ohms law from microscopic considerations

8.1 Induction by a moving conductor

To begin, let us consider what happens when an electrical conductor is pulled through a magnetic field:

3 effects result: (1) An electrical current is induced in the conductor.

(From Davidson, 2001)

(2) This current causes a magnetic field that adds to the original field, such that the conductor appears to drag the field along with it. (3) The combined magnetic field interacts with the current resulting in a Lorentz force that acts on conductor, opposing its motion.

Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.1 Induction by a moving conductor 8.2 Maxwells equations of Electromagnetism 8.3 Approximations and moving reference frames 8.4 The Induction Equation 8.5 Summary Appendix Ohms law from microscopic considerations

8.2.1 Basic Assumptions

Our mathematical description of electromagnetism will be in terms of macroscopic, continuum, physical quantities measured in SI units:

B(r, , , t)

Magnetic Flux Density (often called Magnetic Field) in Tesla (T). Electric Field in Volts per metre (Vm-1). Electric Current Density in Amperes per cubic metre (Am-3) Electric Charge Density in Coulombs per cubic metre (Cm-3)

E(r, , , t)
J (r, , , t)
e (r, , , t)

i.e. It is assumed that these fields are well defined and smoothly varying everywhere in space and individual particles are ignored. (N.B. Good approximation if considering length scales much larger than the size and mean free path length of the particles in the medium.)

8.2.2 Maxwells equations: An Overview

The relationship between magnetic fields, electric fields and electrical currents are encapsulated in just 4 simple equations:

E =

e
Gausss Law: E- fields produced by charge density

B =0
B E = t

Lack of monopole sources of B-field so B-field is solenoidal.

Faradays Law of Induction: E-field induced by changing B-field

E B = 0 J + 0 0 t

Ampere-Maxwell law: B-field produced by currents or by changing E-field(displacement currents)

Note, this is the form of Maxwells equations for materials with no permanent magnetization or electric polarization and where
0
= electrical permittivity of free space = 8.85 x10 -12 C2 N-1 m-2 10-7 N A-2

0 = magnetic permeability of free space = 4! x

and

( 0 0 )1/2 = c = 3 108 ms1

8.2.3 Maxwells equations: Gausss law of electrostatics

Consider the Gausss law

E =

e
0

If there is no time-varying magnetic field, then by Faradays Law

E =0
2

and

E=

The electric potential is then defined by Poissons equation,

e
0

with solutions of the form:

(r) =

1 4 0

e (r ) 3 , d r |r r |

Thus and hence E at any point can be obtained by integration over all the sources of electric field (the charge density). Considering a single charge at the origin and noting that the force on another charge is F = qE, gives Coulombs law of electrostatics.

8.2.4 Maxwells equations: Divergence-free Magnetic fields

Consider the Maxwell equation,

B =0
This expresses the observational fact that there are no point sources/sinks of magnetic field (i.e. no magnetic monopoles) Therefore: (i) Magnetic field lines can never end, but are always closed (ii) Same number of field lines as enter must leave a volume The continuity equation for incompressible fluids ( u = 0) takes the same form, so the same amount of fluid as enters a volume must leave a volume.

8.2.5 Maxwells equations: Faradays Law of EM Induction

The differential form of Faradays Law may be integrated over a surface S perpendicular to the magnetic field,

(
S

E) dS =
S

B dS t

Then using Stokes Theorem the LHS can be re-written as a line integral round the curve C surrounding S such that,

E dl =
C

d dt

where =
S

B dS

Therefore, the change in magnetic flux through S is equal to the induced electric field integrated round C.

8.2.6 Maxwells equations: Ampere-Maxwell Law

Assuming that changes in the electric field are slow then, Integrating this over a surface S perpendicular to the B-field, Then using Stokes theorem on the LHS gives us Amperes law: If we instead take the curl of Poisson equation:

B = 0 J
(
S

B) dS = 0
S

J dS

B dl = 0 I
C

where

I=
S

J dS

B = 0 J we obtain a vectorial
2

B = 0

The solution of this equation is known as the Biot-Savart Law, 0 0 J (r ) (r r ) 3 J (r ) 3 d r B(r) = d r = 4 |r r | 4 |r r |3

Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.1 Induction by a moving conductor 8.2 Maxwells equations of Electromagnetism 8.3 Approximations and moving reference frames 8.5 The Induction Equation 8.6 Summary Appendix: Ohms law from microscopic considerations

8.3.1 Approximations: Neglect of the displacement current

Consider the relative magnitude of the 2nd term on the RHS and the term on the LHS of the Ampere-Maxwell equation:

| 0 0 E/t| | B|

0 0 E/

B/l

From similar scale analysis of terms in the Faradays Law equation we have, | E| | B/t| so that E/l B/ therefore

1 l2 | 0 0 E/t| 2 2 | B| c

8.3.1 Approximations: Neglect of the displacement current

l, are the characteristic length and time scales associated with EM field changes we wish to study. For Earths core, the largest lengthscales are ~ 3481km and shortest timescales ~1 yr

1 l2 | 0 0 E/t| 2 2 | B| c

0.1ms1 3 108 ms1

1020

-19

Therefore can safely neglect the displacement current term: Physically this represents filtering out EM waves from the system.

8.3.2 Moving reference frames: The Lorentz transformations

The observed electric and magnetic field depend on the frame of reference in which they are measured (Einstein, 1905). Let E and B be the fields in one frame of reference and E , B be those measured in a second inertial frame moving at velocity u relative to the first. The transformation btw the frames is:

E = (E + u B) + (1 )

(u E)u u2 (u B)u B = (B + 0 0 (u E)) + (1 ) u2

where = (1 u2 /c2 )1/2

8.3.2 Moving reference frames: The Lorentz transformations

E = (E + u B) + (1 ) (u E)u u2 (u B)u B = (B + 0 0 (u E)) + (1 ) u2

Assuming motions in Earths core are non-relativistic i.e. u2 << c2 2 2 1/2 then = (1 u /c ) ~ 1 and the transformations simplify to

E =E+uB

and B = B

Thus moving to a reference frame travelling along with the core fluid: (i) the magnetic field is unchanged, (ii) the electric field is modified by effect of magnetic field on moving charged particles.

8.3.3 Ohms law in a moving reference frame

Ohms law is an empirical law stating the experimentally observed relation between the electric fields and electric current density. For stationary conductors it takes the form:

J = E

where

= electrical conductivity ( 1 m1 )

When considering a moving electrical conductor, the effective electric field in the frame moving with the conductor must be used:

J = E = (E + u B)
Again, note, this a non-relativistic approximation ( u2 << c2 ). For Earths core (predominantly liquid Fe at high P,T) the electrical conductivity is thought to be large ~ 0.5x106 1 m1

For moving conductors (where u << c ) and considering slow changes in the EM fields ( (l/ )2 /c2 << 1 ) then the evolution of the magnetic fields and electric currents are specified by:

8.3.4 Summary of Equations describing Electrodynamics of 2the Core 2

B =0

B t B = 0 J E =
J = (E + u B)
N.B.1: Only time-derivative of the magnetic field remains. N.B.2: Neglect of the displacement current means we no longer need to consider the Gausss electrostatic eqn (decoupled). Often referred to as: the magnetohydrodynamic (MHD) approximation of electrodynamics

Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.1 Induction by a moving conductor 8.2 Maxwells equations of Electromagnetism 8.3 Approximations and moving reference frames 8.4 The Induction Equation 8.5 Summary Appendix: Ohms law from macroscopic considerations

8.5.1 The Magnetic Induction Equation

Recall the equations governing electrodynamics under the MHD approximation:

B =0

(1) (2) (3) (4)

B t B = 0 J E =
J = (E + u B)
Substituting from (4) into (3) gives

1 ( 0

B) = E + u B

(5)

Take the curl of this and using the magnetic diffusivity =

1 0

B) =

E+

(u B) (6)

8.5.1 The Magnetic Induction Equation

Next, substituting from (2) into (6) and rearranging,

B = t

(u B)

B)

(7)

If = constant then, we can use a standard vector calculus identity together with (1) to re-write the last term as
( B) = ( B) = ( ( B)
2

B) =

Substituting this into (7) we arrive at the Magnetic Induction eqn:

B = t

(u B) +

(8)

Thus, under the MHD approximation, if we know the motion of the conductor u and the present magnetic field, we can calculate how the field evolves in time. This single equation describes the electrodynamics of Earths core !

8.5.2 Magnetic Reynolds Number

B = t (u B) +
2

Assume that the velocity field has a characteristic magnitude U Assume that the magnetic field has a characteristic magnitude Assume that the lengthscale over which both fields change is

Then the ratio of the magnitudes of the terms on the RHS will be:

UB/L UL (u B)| = Rm | 2 B| B/L2

Rm

is known as the magnetic Reynolds number.

For global UL = UL0 3 104 3.481 106 0.5 106 4 107 motions in : Rm = 650 Earths core

8.5.3 Perfect Conductivity: Frozen Flux

Consider the case in which Rm = so the second term on the right hand side is negligible (i.e. perfect conductivity), then

B = t

(u B)

(9)

Using the standard vector calculus relation,

(u B) = (B )u (u )B since u=0 and B = 0.

So,
B + (u t )B = (B )u or DB = (B Dt )u

(10)

Advection of Magnetic Field along with flow

Stretching of magnetic field by shear of flow

8.5.3 Perfect Conductivity: Frozen Flux

Consider a short line element dl drawn into a fluid at some instant. It subsequently moves with the fluid, so the rate of change of dl is u(r + dl) u(r) where r and r + dl are the position vectors at the ends of dl. The equation of evolution of dl is then:

D(dl) = u(r + dl) u(r) = (dl Dt

)u

But this is identical to (10) therefore when Rm = , the magnetic field evolves just like the line element, moving along with the fluid. The magnetic field lines can be thought of as being frozen into the fluid (i.e. fluid elements lying on a field line at some instant must continue to lie on the fluid elements at all later times).

8.5.3 Perfect Conductivity: Frozen Flux

In the Frozen Flux approximation, we schematically imagine flow sweeping the magnetic field along with it, moving and stretching it.

(From Baumjohann and Treumann, 1997)

8.5.3 Perfect Conductivity: Frozen Flux

In the limit of a perfect conductor , there are also strong consequences for changes in the flux through material contours. e.g. Consider a surface S bounded by a closed material contour C that moves along with the fluid.
d dt B dS =
S S

B dS + t

B u dl =
C S

B dS t

u B dl
C

=
S

B t

(u B) dS = 0

i.e. The total flux enclosed by a material surface cannot change with time even if the shape of the evolves : FLUX IS FROZEN !
t=t1 t=t2

The limit Rm is often called the frozen flux approximation.

8.5.4 Magnetic diffusion

Next consider the opposite limit in when Rm = 0 so the first term on the RHS is negligible (i.e.no flow), then

B = t

This is a classic vector diffusion equation. Simple dimensional analysis yields, L2 B B = 2 so D = D L Thus magnetic field features with larger spatial gradients (smaller L ) will diffuse faster for an particular magnetic diffusivity . For global scale fields in Earths core :
D = (3.481 106 )2 0.5 106 4 107 7.6 1012 s

240, 000 yrs

( More precise calculations give ~ 30,000yrs for dipole part of geomagnetic field)

8.5.4 Magnetic Diffusion

Magnetic diffusion can also allow magnetic field structures to merge as well as decay, for example

(From Baumjohann and Treumannn, 1997)

8.5.4 Combination of Advection and Diffusion

First, in cartesian co-ordinates (x,y,z), consider an initially constant magnetic field B 0 = (0, 0, B0 ) only in the z direction. This is acted on by a constant flow in the x direction which has a shear in the z direction u = (u(z), 0, 0) Consider 1st only advection and use: .
(u B) = (B )u (u
B0

)B

since

u=0

and

B = 0.

e.g.

Bz = 0 => Bz = B0 for all t t By = 0 => By = 0 for all t t u(z) Bx = B0 t z

Bx = B0 du t dz

Integrating the last w.r.t time and using Bx =0 at t=0 gives,

So this simple shear causes the field to grow linearly with time in the direction of flow (will be important for dynamo action.... ).

8.5.4 Combination of Advection and Diffusion

But have ignored diffusion, if this were present the field growth would be slower since the x-component of the induction equation is:

2 Bx u Bx = B0 + t z z 2
Diffusion will become more important as the field gradients grow, until eventually a steady state is reach where advection and diffusion are in balance:

2 Bx u B0 = z z 2
Integrating twice w.r.t. z the steady state profile for Bx can be found:
z

Bx =
0

B0 u(z )dz + cz + d

So, even if diffusion is not initially important in the induction equation, it will often be a vital ingredient of the saturated state.

Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.1 Induction by a moving conductor 8.2 Maxwells equations of Electromagnetism 8.3 Approximations and moving reference frames 8.4 The Induction Equation 8.5 Summary Appendix: Ohms law from microscopic considerations

8.6 Summary: self-assessment questions

(1) Can you state and describe the physical meaning of each of Maxwells equations? (2) Can you derive the magnetic Induction equation? (3) Do you understand the consequences of the frozen flux approximation to the Induction equation? (4) Can you estimate the magnetic diffusion time for a particular system? Next time: Dynamo Theory

References
- Baumjohann, W.A. and Treumann, R.A., (1997) Basic Space Plasma Physics, Imperial College Press. (Detailed derivation of Generalised Ohms law, Section 7.3). - Davidson, P.A., (2001) An introduction to magnetohydrodynamics, Cambridge University Press. (Good for physical understanding, especially Chapters 1,2,4). - Jackson, J.D., (1998) Classical Electrodynamics, John Wiley and Sons Inc. (A wealth of detail on electromagnetism in general). - Roberts, P.H., (2007) Theory of the geodynamo. In Treatise on Geophysics, Vol 5 Geomagnetism, Ed. M. Kono, Chapter 8.03, pp.67-102. (especially section 8.03.2)

Lecture 8: Fundamentals of Electromagnetism and the Induction equation

8.1 Induction by a moving conductor 8.2 Maxwells equations of Electromagnetism 8.3 Approximations and moving reference frames 8.4 The Induction Equation 8.5 Summary Appendix: Ohms law from microscopic considerations

Ohms Law: A microscopic derivation

In previous section we ignored all microscopic effects and stated Ohms law as an empirical result. But, it can also be derived from microscopic considerations of motions of electrons and ions in a conductor.
The equations of conservation of momentum density for the electrons and ions are (e.g. Baumjohann and Treumann, 1997 p.140):

(ne v e ) + t

(ne v e v e ) =

1 me

Pe

ne e R (E + v e B) + me me

(ni v i ) + t
Temporal variation of flux density

(ni v i v i ) =
NL momentum density flux

1 mi

Pi +

ni e R (E + v i B) + mi mi
EM Lorentz force Collisions btw ions and electrons

Fluid Pressure tensor

ns ms

= number density of species = mass of species = bulk velocity of species

vs

Ohms Law: A microscopic derivation

Multiply the electron eqn by mi and the ion equation by me and subtract the equations. Then assuming me /mi << 1 and ne ni n (quasi-neutral) and defining the macroscopic current as J = e(ni v i ne v e ) : me J = P e + ne(E + v e B) R e t Note the dependence of the current density only on the electron fluid pressure, electron velocity and electron-ion collisions. Again assuming me /mi << 1 and ne ni n (quasi-neutral) then the bulk fluid velocity is approximately that of the ions so,
vi u and ve u J ne

This allows the Lorentz force on the electrons to be split into 2 terms, one involving the bulk velocity and one the current density: me J = P e + ne(E + u B) J B R e t

Ohms Law: A microscopic derivation

Next, we note that we can write the collision term as J ne2 R = nme c (v i v e ) = ne where = me c c is the frequency of collisions. where Using this leads to an expression of the Generalised Ohms law:

J 1 me J + = ne2 t ne
When

P e + (E + u B)

1 (J B) ne

pe << 1 can neglect pressure gradient on RHS. ne U BLp me << 1 can neglect current changes on LHS. When ne2
When

B << 1 ne

can neglect Lorentz (Hall) term on RHS.

These are all true in Earths core so finally we are left with,

J = (E + u B)