PH Ysicsguide: Magnetostatics

ide
Magnetostatics
Sk Jahiruddin
gu Assistant Professor
ics
Sister Nibedita Govt. College, Kolkata
Author was the topper of IIT Bombay M.Sc Physics 2009-2011 batch
He ranked 007 in IIT JAM 2009 and 008 (JRF) in CSIR NET June
2011
ys
He has been teaching CSIR NET aspirants since 2012

Ph
1
©Sk Jahiruddin, 2020 Magnetostatics
Contents
ide
1 Magnetostatics 4
1.1 Steady current . . . . . . . . . . . . . . . . . 4
1.2 Force on a moving charge . . . . . . . . . . 6
1.3
gu 1.2.1 Force on current carrying wire . . . .
Electric current as a source of magnetic field

8
9
ics
1.3.1 Biot-Savart Law . . . . . . . . . . . . 9
1.3.2 Ampere’s Law . . . . . . . . . . . . . 11

ys
1.4 Surface Currents and Discontinuities . . . . 13
1.5 Magnetic Vector Potential . . . . . . . . . . 17

Ph
1.6 Magnetic Dipole . . . . . . . . . . . . . . . . 20
1.6.1 Torques and Forces on Magnetic Dipoles 28
1.7 Electric Current . . . . . . . . . . . . . . . . 28
1.7.1 Current Density . . . . . . . . . . . . 30
jahir@physicsguide.in 2 Physicsguide
1.7.2 Conservation of Charge . . . . . . . . 30
1.7.3 Ohm’s Law . . . . . . . . . . . . . . 31
ide
2 Magnetic Field in a medium 35
2.1 Bound Currents . . . . . . . . . . . . . . . . 37
2.2
gu Magnetic intensity . . . . . . . . . . . . . .
2.2.1 Linear magnetic material . . . . . . .

39
40
ics
2.3 Electromagnetic Boundary conditions . . . . 42
2.3.1 Magnetostatic Boundary condition in-

side a medium . . . . . . . . . . . . . 42
ys
2.3.2 Boundary condition in both electric

and magnetic field . . . . . . . . . . 42
Ph
2.3.3 Example: Field inside the gap . . . . 45
1 Magnetostatics
ide
Charges give rise to electric fields. Current give rise to mag-
netic fields. In this section, we will study the magnetic fields
induced by steady currents. This means that we are again
looking for time independent solutions to the Maxwell equa-
tions. We will also restrict to situations in which the charge
gu
density vanishes, so ρ = 0. We can then set E = 0 and focus
our attention only on the magnetic field. We’re left with
two Maxwell equations to solve:
ics
∇ × B = µ0 J (1.1)
and
∇·B=0 (1.2)
ys
If you fix the current density J , these equations have a

unique solution. Our goal in this section is to find it.
Ph
1.1 Steady current
Before we solve (1.1) and (1.2) let’s pause to think about the
kind of currents that we’re considering in this section. Be-
cause ρ = 0 , there can’t be any net charge. But, of course,
we still want charge to be moving! This means that we nec-

essarily have both positive and negative charges which bal-
ance out at all points in space. Nonetheless, these charges
ide
can move so there is a current even though there is no net
charge transport.
This may sound artificial, but in fact it’s exactly what
gu
happens in a typical wire. In that case, there is background
of positive charge due to the lattice of ions in the metal.
Meanwhile, the electrons are free to move. But they all

ics
move together so that at each point we still have ρ = 0.
The continuity equation, which captures the conservation
of electric charge, is
∂ρ
ys
+∇·J=0 (1.3)
∂t
Since the charge density is unchanging (and, indeed, van-
Ph
ishing), we have
∇·J=0 (1.4)
Mathematically, this is just saying that if a current flows
into some region of space, an equal current must flow out
to avoid the build up of charge. Note that this is consistent
with (1.1) since, for any vector field, ∇ · (∇ × B) = 0
1.2 Force on a moving charge
ide
The most basic form of the magnetic force is the force on
a charge q moving with velocity v, exerted by a magnetic
field B, which is
F = qv × B (1.5)
gu
The SI unit of magnetic field is the tesla (T). One tesla is
defined as one newton per ampere-meter. By (1.5) the force
on 1 C moving 1 m/s perpendicular to a field of 1 T is 1 N .
A magnetic field of 1 T is a rather strong field. The largest
ics
fields that can be produced by conventional electromagnets
are about 2 T. For high-field research in laboratories, fields
of 10 − 12 T , produced with superconducting magnets,
are used. The record for steady-state fields is about 30 T
ys
. At the surface of the Earth the magnetic field is about

0.5 × 10−4 T . Very large and very small fields are impor-
tant in astronomy; e. g.108 T near pulsars, and 10−10 T in
Ph
interstellar space in the galaxy.
The direction of the magnetic force on q is sideways, i.e.,

perpendicular to v and to B , as shown in Fig. 8.1 for a
positive charge. Therefore, the magnetic force does no work
on q :
dW = F · dx = F · vdt = 0 (1.6)
ide
The magnetic force affects the direction of motion of q, but
not its kinetic energy. Because the magnetic force does no
work, it is not possible to define a potential energy function
for the magnetic force. The magnetic force is velocity de-
pendent, which is quite different from the other fundamental
gu
forces we encounter in physics.
ics
ys
Figure 1.1: The magnetic force. The force on a moving

Ph
charge is sideways, perpendicular to both v and B. The

figure shows F and the circular trajectory for a positive
charge in a uniform field.
Equation (1.5) may be taken as the definition of the

magnetic field. From the measurement of the force on a
test charge q, e.g. by observing its deflection for a known

velocity, the defined quantity B could in principle be de-
duced from (1.5)
ide
When both electric and magnetic field is present then
the force equation is known as lorentz force equation
F = q(E + v × B) (1.7)
1.2.1
guForce on current carrying wire
ics
A current-carrying wire in a magnetic field experiences a
force, from the magnetic force on the individual moving
charges that make the current. Electric current is motion of
ys
charge, and (1.5) is the force on a moving charge. Suppose

the current consists of particles with charge q and linear
density n1 , moving with mean velocity v . Then the net
Ph
force dF on a small segment d` of the wire is
dF = (n1 d`) qv × B (1.8)
because n1 d` is the number of moving charges in d`. The

current I in the wire is qn1 v, and v is parallel to d`, so the
total force on the wire is
Z
F= Id` × B (1.9)
wire
ide
1.3 Electric current as a source of mag-
netic field
gu
Where does a magnetic field come from? What is its source?
The most familiar source is a permanent magnet, i.e., a
piece of magnetized iron or other ferromagnetic material.
ics
But at a more basic level the source of B is to be found
in electric current. We will study magnetic materials, and
see that the magnetic and field of a ferromagnet comes from
properties of atomic electrons - their spin and orbital motion
ys
- which, although not classical currents, do involve dynamics

of charged particles. However, now we are concerned with
B from macroscopic steady currents.
Ph
1.3.1 Biot-Savart Law
The magnetic field of a steady line current is given by the

Biot-Savart law: The field dB at a point P, due to an in-
finitesimal current element Id` at a point P 0 . Figure below

shows the geometry. For this elemental case the Biot-Savart
law is
ide
gu
ics
Figure 1.2: Elemental form of the Biot-Savart law. Id` is
the source of magnetic field, and dB is the resulting field at
P.
ys
µ0 Id` × r̂
dB = (1.10)
4π r2
Ph
Here r is the distance between P 0 (the source point) and P

(the field point), r̂ is the unit vector in the direction from
P 0 to P, and r = rr̂. Note that the Biot-Savart law is an
inverse-square law, like Coulomb’s law, but the direction of
the magnetic field is azimuthal, around the axis of Id`.
Do few examples from Griffiths.
1.3.2 Ampere’s Law
ide
The integral of the current density over the surface S is the
same thing as the total current I that passes through S.
Ampère’s law in integral form then reads
I
B · dr = µ0 I (1.11)
gu C
For the problems where cylindrical symmetry is there, it is

easy to apply Ampere’s law to calculate Magnetic field.
ics
Consider an infinite, straight wire carrying current I.
We’ll take it to point in the ẑ direction. The symmetry
of the problem is jumping up and down telling us that we
need to use cylindrical polar coordinates, (r, ϕ, z), where
ys
p
r = x2 + y 2 is the radial distance away from the wire.
We take the open surface S to lie in the x − y plane,

Ph
centered on the wire. For the line integral in (1.11) to give

something that doesn’t vanish, it’s clear that the magnetic
field has to have some component that lies along the cir-
cumference of the disc.
ide
gu
ics
Figure 1.3: Magnetic field of long straight wire
But, by the symmetry of the problem, that’s actually

the only component that B can have: it must be of the
ys
form B = B(r)ϕ̂ . (If this was a bit too quick, we’ll derive
this more carefully below). Any magnetic field of this form
automatically satisfies the second Maxwell equation ∇·B =
Ph
0. We need only worry about Ampère’s law which tells us

I Z 2π
B · dr = B(r) rdϕ = 2πrB(r) = µ0 I
C 0
We see that the strength of the magnetic field is

µ0 I
B= ϕ̂ (1.12)
2πr
The magnetic field circles the wire using the “right-hand

rule”: stick the thumb of your right hand in the direction
of the current and your fingers curl in the direction of the
ide
magnetic field.
1.4 Surface Currents and Discontinuities
gu
Consider the flat plane lying at z = 0 with a surface current
density that we’ll call K . Note that K is the current per
ics
unit length, as opposed to J which is the current per unit
area. You can think of the surface current as a bunch of
wires, all lying parallel to each other.
We’ll take the current to lie in the x-direction: K = K x̂

ys
as shown below.
Ph
Figure 1.4: Surface Current
From our previous result, we know that the B field
should curl around the current in the right-handed sense.

But, with an infinite number of wires, this can only mean
that B is oriented along the y direction. In fact, from the
ide
symmetry of the problem, it must look like
gu
Figure 1.5: Magnetic field due to a Surface Current
ics
with B pointing in the −ŷ direction when z > 0 and in
the +ŷ direction when z < 0 we write
ys
B = −B(z)ŷ
with B(z) = −B(−z). We invoke Ampère’s law using the

Ph
following open surface:
ide
Figure 1.6:
gu
with length L in the y direction and extending to ±z.
We have
ics
I
B · dr = LB(z) − LB(−z) = 2LB(z) = µ0 KL
C
so we find that the magnetic field is constant above an infi-

ys
nite plane of surface current

µ0 K
B(z) = z>0
2
Ph
This is rather similar to the case of the electric field in the

presence of an infinite plane of surface charge.
The analogy with electrostatics continues. The magnetic

field is not continuous across a plane of surface current. We
have
B z → 0+ − B z → 0− = µ0 K

In fact, this is a general result that holds for any surface

current K. We can prove this statement by using the same
curve that we used in the Figure above and shrinking it until
ide
it barely touches the surface on both sides. If the normal to
the surface is n̂ and B± denotes the magnetic field on either
side of the surface, then
gu n̂ × B|+ − n̂ × B|− = µ0 K
Meanwhile, the magnetic field normal to the surface is

(1.13)
ics
continuous. We can write the boundary conditions as
⊥ ⊥
Babove = Bbelow (1.14)
and
ys
k k
Babove − Bbelow = µ0 K (1.15)
When we looked at electric fields, we saw that the nor-

Ph
mal component was discontinuous in the presence of surface

charge while the tangential component is continuous.
For magnetic fields, it’s the other way around: the tan-
gential component is discontinuous in the presence of surface
currents.
1.5 Magnetic Vector Potential
ide
At this point it is useful to compare the principles of mag-
netostatics and electrostatics, given in the table below.
Electrostatics Magnetostatics
∇ × B = µ0 J ∇ · E = ρ/0
gu ∇·B=0
B curls around I
∇×E=0
E diverges from q
ics
Note this difference: The electric field has scalar sources,
but the magnetic field has vector sources. Electric charge,
i.e., a point source of E from which E diverges, is a com-
ys
mon property of elementary particles. Motion of charged

particles-current-is the source of B, around which B curls.
Ph
Magnetic charge, i.e., a point source from which B di-

verges in the rest frame of the particle, apparently does
not exist in nature, or at least it has not been observed.
Such a hypothetical charge is called a magnetic monopole.
Dirac showed that it is possible to construct a consistent
quantum theory with both electric charges and magnetic
monopoles. However, the fundamental magnetic charge g
and electric charge e would necessarily be quantized, and

satisfy the relation eg = n/2, where n is an integer. Many
experimenters have searched for magnetic monopoles, but
ide
so far the results are negative. Some speCUlative theories
of high-energy physics, such as grand unified field theories,
predict the existence of very massive magnetic monopoles,
too massive to be produced at current high-energy acceler-
ators,
gu
but which might have been produced in the big bang.
Searching for magnetic monopoles continues to be an inter-
ics
esting experimental challenge
If magnetic monopoles do not exist, then the equation

∇ · B = 0 is a universal equation of magnetism. Whether
ys
magnetic monopoles exist or not, the source of the magnetic

fields we encounter in physics are not point magnetic charges
but rather currents of electric charge, corresponding to the
Ph
source equation ∇ × B = µ0 J.
We found in electrostatics that it is useful to introduce

a scalar potential V (x) for the electrostatic field, such that
E = −∇V. This guarantees that ∇ × E = 0. In an analo-
gous way we may introduce a vector potential A(x) for the
magnetic field, such that
ide
B=∇×A (1.16)
This guarantees that ∇ · B = 0
However, (1.16) does not uniquely determine A for a
gu
given magnetic field B : If f (x) is an arbitrary scalar func-
tion, then A + ∇f has the same curl as A (namely B)
because ∇ × ∇f is identically 0. Therefore we may impose
ics
a condition on A called a gauge condition, to remove this
ambiguity. The Coulomb gauge condition, which we will
use in magnetostatics, is
ys
∇·A=0 (1.17)
Ph
Taking (1.16) and (1.17) together is still not enough

to make A(x) unique, because adding a constant does not
change either ∇ × A or ∇ · A. But imposing an appropriate
boundary condition, such as requiring A → 0 at infinity,
makes A unique.
Example: Consider the uniform field B = B0 k . A
vector potential function, satisfying (1.16) for the uniform

field and the Coulomb gauge condition (1.17) is
ide
1 1
A(x) = B × x = B0 (−y î + xĵ)
2 2
An example of uniform B is the field inside a solenoid, so
for B = B0 k̂ we may picture a long, tightly and uniformly
gu
wound solenoid, whose axis is the z axis. Note for this case
that A(x) is parallel to the surface currents (azimuthal) and
that A(x) curls around the B field.
ics
The analytic expression of vector potential in vector no-
tation is Z 0
µ0 3 0 J (x )
A(x) = dx (1.18)
4π V |x − x0 |
ys
1.6 Magnetic Dipole

Ph
Vector potential, while in the general coordinate as shown

below figure, can be expanded as
ide
gu
Figure 1.7: Multipole expansion of vector potential
As in the multipole expansion of V, we call the first term

ics
(which goes like 1/r) the monopole term, the second (which
goes like 1/r2 ) dipole, the third quadrupole, and so on.
ys
I I
µ0 I 1 1
A(r) = dI + 2 r0 cos αdl0
0
4π r r
I (1.19)
1 2 3 1
Ph
+ 3 (r0 ) cos2 α − dl0 + · · ·

r 2 2
The magnetic monopole term is always zero
The dipole term

I I
µ0 I µ 0 I
Adip (r) = 2
r0 cos αdl0 = 2
(r̂ · r0 ) dl0 (1.20)
4πr 4πr
We define the dipole moment

Z
m ≡ I da = Ia (1.21)
ide
Then the vector potential for the the dipole term
µ0 m × r̂
Adip (r) = (1.22)
gu 4π r2
Magnetic field of a magnetic dipole can be written as

µ0 1
ics
Bdip (r) = [3(m · r̂)r̂ − m] (1.23)
4π r3
The magnetic field of a (perfect) dipole is easiest to cal-

ys
culate if we put m at the origin and let it point in the z

-direction see figure below.
Ph
ide
gu
Figure 1.8: Calculation of magnetic field due to a magnetic
dipole
ics
µ0 m × r̂
According to Equation Adip (r) = the potential
4π r2
at point (r, θ, φ) is
ys
µ0 m sin θ
Adip (r) = φ̂ (1.24)
4π r2
Ph
Hence the magnetic field

µ0 m
Bdip (r) = ∇ × A = (2 cos θr̂ + sin θθ̂) (1.25)
4πr3
Example: Find the magnetic dipole moment of the

book type loop shown in Figure below. All sides have length
w, and it carries a current I.
ide
gu Figure 1.9:
Solution: This wire could be considered the superposi-

ics
tion of two plane square loops The ”extra” sides (AB) can-
cel when the two are put together, since the currents flow
in opposite directions. The net magnetic dipole moment is
ys
m = Iw2 ŷ + Iw2 ẑ
√
its magnitude is 2Iw2 , and it points along the 45◦ line
z=y
Ph
Example: A circular loop of wire, with radius R, lies in

the xy plane (centered at the origin) and carries a current
I running counterclockwise as viewed from the positive z
axis.
(a) What is its magnetic dipole moment?
(b) What is the (approximate) magnetic field at points far

from the origin?
(c) Show that, for points on the z axis, your answer is con-
ide
µ0 I R2
sistent with the exact field B = 3/2 when z R.
2 (R +z 2 )
2
Solution: (a) m = Ia = IπR2 ẑ
gu
(b) B u µ0 IπR2
4π r3 (2 cos θr̂
BuB≈
+ sin θθ̂)
(c) On the z axis, θ = 0, r = z, r̂ = ẑ( for z > 0). So

µ0 IR2
ẑ
ics
2z 3
for z < 0, θ = π, r̂ = −ẑ, so the field is the same, with |z|3
in place of z 3
ys

µ0 I R2
The original ans B = 3/2 reduces (for z
2 (R2 +z 2 )
R) to B ≈ µ0 IR2 /2|z|3 . So the ans are same.
Ph
Example: (a) A disk of radius R, carrying a uniform

surface charge σ, is rotating at constant angular velocity ω
. Find its magnetic dipole moment.
(b) A spherical shell of radius R, carrying a uniform

surface charge σ, is set spinning at angular velocity ω. Find
the magnetic dipole moment of the spinning spherical shell.
Also find the magnetic vector potential of the spinning

shell.
ide
Solution: (a) For a ring, m = Iπr2 . So the current due
to a ring with radius extended to small amount r → r + dr
is (see figure (a) below )
gu
ics
ys
Figure 1.10:
Ph
I → σvdr = σωrdr
So the magnetic dipole moment

Z R
m= πr2 σωrdr = πσωR4 /4
0
(b) See figure (b) above. The total charge on the shaded
ring is dq = σ(2πR sin θ)Rdθ. The time for one revolution
is dt = 2π/ω .
ide
So the current in the ring is I = dq 2
dt = σωR sin θdθ. The
area of the ring is π(R sin θ)2 , so the magnetic moment of
the ring is dm = σωR2 sin θdθ πR2 sin2 θ, and the total

gu
dipole moment of the shell is
m = σωπR4
Z π
sin3 θdθ = (4/3)σωπR4
0
ics
Hence
4π
m= σωR4 ẑ
3
Dipole term of the vector potential is

ys
µ0 m × r̂ µ0 m sin θ
Adip (r) = = φ̂
4π r2 4π r2
Ph
In this problem
µ0 4π 4 sin θ µ0 σωR4 sin θ

Adip = σωR 2 φ̂ = φ̂
4π 3 r 3 r2
1.6.1 Torques and Forces on Magnetic Dipoles
ide
Torque on a magnetic dipole of dipole moment m while
placed in magnetic field B
N=m×B (1.26)
gu
Force on a magnetic dipole is zero when placed in a uniform
field. But when placed in a non-uniform field then there is
a net force acts
F = ∇(m · B) (1.27)
ics
Potential energy of a magnetic dipole while placed in a mag-
netic field is
U = −m · B (1.28)
ys
1.7 Electric Current

Ph
The conceptually simplest example of an electric current

is the current in a thin conducting wire. In an ideal one-
dimensional wire the current I is defined as the net charge
passing a point P per unit time. In a real wire I is defined
as the charge per unit time passing through a cross section
of the wire at P .
dQ
I= (1.29)
dt
ide
The unit of current is the ampere (A), which is the basic
electric unit in the SI system. If the wire carries a current
of 1 A at point P , then 1 C of net charge passes P each
second.
gu
If the current in a wire is due to charges q moving with
mean velocity v, and the charges have linear density nL (=
number of charge carriers per unit length) , then
ics
I = qnL v (1.30)
In n is the charge carrier density per unit volume and A is
ys
the surface area of the current carrying wire, q is the amount

of charge each particle carries, then the current is
I = qnAv (1.31)
Ph
What is the current for an orbiting electron? Well the

charge is e, velocity is v and the radius of atom is taken to
be R. Then the electron crosses a point once in one time
period. So the current is
e
I=
T
As T is 2πR/v , the current

ev
I= (1.32)
2πR
ide
You will need this expression many times
1.7.1 Current Density
gu
If the current in a volume of space is due to charges q with
volume number density n ( = number of charge carriers per
unit volume) moving with mean velocity v , then the current
ics
density is
J = qnv (1.33)
Note that the units of qnv are A/m2 . Current density is
ys
the flux of electric charge. In general flux is equal to density

times velocity.
Ph
1.7.2 Conservation of Charge
The net charge of an isolated system is constant. But charge

is not only conserved overall in a system, it is also conserved
point by point throughout the system. This local conserva-
tion of charge is described mathematically by the continuity
equation
∂ρ
∇·J=− (1.34)
∂t
ide
Here ρ(x, t) is the volume charge density (= charge per unit
volume), and J(x, t) is the volume current density (= cur-
rent per unit area). Equation (1.34) is universally true, for
arbitrary time dependence.
gu
Integral form of the continuity equation,
I
S
J · dA = −
d
Z
dt V
ρd3 x (1.35)
ics
states that the rate of charge passing outward through
the closed surface S is equal to the rate of decrease of charge
in the enclosed volume V . Equation (1.34), or equivalently
ys
(1.35), is a basic equation of electrodynamics, expressing

local conservation of charge.
Ph
1.7.3 Ohm’s Law
How is the current related to the potential gradient in a con-

ductor? If two terminal points on a conductor are held at
a constant potential difference V , e.g., by connecting them
to the electrodes of a battery, then in equilibrium a steady
current flows through the conductor. Let I be the total cur-

R
rent at either point, i.e., the integrated flux J·dA through
a surface inside the conductor surrounding the point. It is
ide
found empirically that formany cases the current and po-
tential difference are proportional,
V = IR (1.36)
gu
The constant of proportionality R is called the resistance
of the conductor. The SI unit of resistance is the ohm (Ω),
defined by 1Ω = 1V/A . The reciprocal 1/R is called the
ics
conductance, and its unit is Ω−1 , or siemens (S) . Ohm’s law
holds to a very good approximation for many conductors.
However, it is not a universal principle, there are examples
where it does not hold.
ys
The resistance R of a sample of matter is a function of

the geometry (size and shape) of the sample, and of the ma-
Ph
terial composition. For example, the resistance of a uniform

cylinder, of length L and cross section A, is proportional to
L and inversely proportional to A,
R = ρL/A (1.37)
The parameter ρ (not to be confused with charge density
ρ(x)! is an intrinsic property of the material called the re-

sistivity.
ide
We may also write a local form of Ohm’s law, which is
a more basic equation.
J(x) = σE(x) (1.38)
gu
Again, σ is the conductivity.
Example Two long coaxial metal cylinders (radii a and

ics
b) are separated by material of conductivity σ as shown in
figure below . If they are maintained at a potential difference
V, what current flows from one to the other, in a length L?
ys
Ph
Figure 1.11:
Solution: The field between the cylinders is

λ
E= ŝ
ide
2π0 s
where λ is the charge per unit length on the inner cylinder.
The current is
Z Z
σ
I = J · da = σ E · da = λL
gu 0
The integral is over any surface enclosing the inner cylinder.
The potential difference between the cylinders is
ics
Z a
λ b
V =− E · dl = ln
b 2π0 a
2πσL
I= V
ys
ln(b/a)
So the resistance of the system is
ln(b/a)
Ph
R=
2πσL
2 Magnetic Field in a medium
ide
Electric fields are created by charges; magnetic fields are
created by currents. We learned in our first course that the
simplest way to characterise any localised current distribu-
tion is through a magnetic dipole moment m. For example,
a current I moving in a planar loop of area A with normal
gu
n̂ has magnetic dipole moment,
m = IAn̂
ics
Magnetic vector potential and magnetic field due to a ma-
gentic dipole are

µ0 m × r µ0 3(m · r̂)r̂ − m
A(r) = ⇒ B(r) = (2.1)
ys
4π r3 4π r3
Current loops, and their associated dipole moments, al-

Ph
ready exist inside materials. They arise through two mech-

anisms:
Electrons orbiting the nucleus carry angular momentum

and act as magnetic dipole moments.
Electrons carry an intrinsic spin. This is purely a quan-
tum mechanical effect. This too contributes to the mag-

netic dipole moment.
ide
We define the magnetisation M to be the average mag-
netic dipole moment per unit volume.
In most (but not all) materials, if there is no applied
gu
magnetic field then the different atomic dipoles all point
in random directions. This means that, after averaging,
hmi = 0 when B = 0. However, when a magnetic field is
applied, the dipoles line up. The magnetisation typically
ics
takes the form M ∝ B. We’re going to use a slightly strange
notation for the proportionality constant. (It’s historical
but, as we’ll see, it turns out to simplify a later equation)
ys
1 χm
M= B (2.2)
µ0 1 + χm
where χm is the magnetic susceptibility. The magnetic prop-
Ph
erties of materials fall into three different categories. The

first two are dictated by the sign of χm :
Diamagnetism: −1 < χm < 0. The magnetisation of

diamagnetic materials points in the opposite direction
to the applied magnetic field. Most metals are diamag-
netic, including copper and gold. Most non-metallic ma-

terials are also diamagnetic, including, importantly, wa-
ter with χm ≈ −10−5 . This means, famously, that frogs
ide
are also diamagnetic. Superconductors can be thought
of as ”perfect” diamagnets with χm = −1.
Paramagnetism: χm > 0. In paramagnets, the magneti-
gu
sation points in the same direction as the field. There
are a number of paramagnetic metals, including Tung-
sten, Cesium and Aluminium.
ics
Ferromagnetism: M 6= 0 when B = 0. Materials with
this property are what you usually call ”magnets”. They’re
the things stuck to your fridge. The direction of B is
from the south pole to the north. Only a few elements
ys
are ferromagnetic. The most familiar is Iron. Nickel

and Cobalt are other examples.
Ph
2.1 Bound Currents
When a material becomes magnetised (at least in an anisotropic

way), there will necessarily be regions in which there is a
current. This is called the bound current.
Let’s first give an intuitive picture for where these bound

currents appear from. Consider a bunch of equal magnetic
dipoles arranged uniformly on a plane like the left picture
ide
: The currents in the interior region cancel out and we’re
left only with a surface current around the edge. We know
that this is surface current a K. We’ll follow this notation
and call the surface current arising from a constant, internal
gu
magnetisation Kbound .
ics
ys
Figure 2.1: Origin of bound surface and volume current

Ph
Now consider instead a situation where the dipoles are

arranged on a plane, but have different sizes. We’ll put the
big ones to the left and the small ones to the right, like In
this case, the currents in the interior no longer cancel. As
we can see from the right side of the picture picture, they go
into the page. since M is out of the page, and we’ve arranged
things so that M varies from left to right, this suggests that
J bound ∼ ∇ × M.
Bound surface and volume currents are
ide
Kbound = M × n̂ (2.3)
and
Jbound = ∇ × M (2.4)
gu
Note that the bound current is a steady current, in the
sense that it obeys ∇ · J bound = 0 .
ics
2.2 Magnetic intensity
Recall that Amp‘ere’s law describes the magnetic field gen-

ys
erated by static currents. We’ve now learned that, in a

material, there can be two contributions to a current: the
bound current J bound that we’ve discussed above, and the
Ph
current J free from freely flowing electrons that we were im-

plicitly talking. Amp‘ere’s law does not distinguish between
these two currents; the magnetic field receives contributions
from both.
∇ × B = µ0 (J free + J bound )
(2.5)
= µ0 J free + µ0 ∇ × M
We define the Magnetic Intensity or magnetising field, H as

1
H= B−M (2.6)
ide
µ0
This obeys
∇ × H = J free (2.7)
and in integral form
gu I
H · dl = Ifenc (2.8)
We see that the field H plays a similar role to the electric

ics
displacement D; the effect of the bound currents have been
absorbed into H, so that only the free currents contribute.
Note, however, that we can’t quite forget about B entirely,
since it obeys ∇ · B = 0 . In contrast, we don’t necessarily
ys
have ”∇ · H = 000 . Rather annoyingly, in a number of books

H is called the magnetic field and B is called the magnetic
induction. But this is stupid terminology so we won’t use
Ph
it.
2.2.1 Linear magnetic material
For most of the magnetic material the magnetization vector

and magnetic intensity are proportional. They are called
linear magnetic material
M = χm H (2.9)
ide
The constant of proportionality χm is called the magnetic
susceptibility; it is a dimensionless quantity that varies from
one substance to another - positive for paramagnets and
negative for diamagnets.
gu
From the definition of H
B = µ0 (H + M) = µ0 (1 + χm ) H
ics
or
B = µH (2.10)
where
ys
µ ≡ µ0 (1 + χm ) (2.11)
µ is called the permeability of the material.
Ph
Do some example from Griffiths.
2.3 Electromagnetic Boundary conditions
ide
2.3.1 Magnetostatic Boundary condition inside a
medium
Boundary conditions inside the magnetic medium.
and gu k
⊥
Babove
k
⊥
− Bbelow =0
Habove − Hbelow = Kf × n̂
(2.12)
(2.13)
ics
this can also be written as in terms of B
k k
Babove − Bbelow = µ0 (K × n̂) (2.14)
ys
2.3.2 Boundary condition in both electric and mag-

netic field
Ph
In the presence of surface charge, the electric field normal to

the surface is discontinuous, while the electric field tangent
to the surface is continuous. For magnetic fields, it’s the
other way around: in the presence of a surface current, the
magnetic field normal to the surface is continuous while the
magnetic field tangent to the surface is discontinuous.
What happens with dielectrics? Now we have two op-

tions of the electric field, E and D, and two options for the
magnetic field, B and H . They can’t both be continu-
ide
ous because they’re related by D = E and B = µH and
we’ll be interested in situation where ( and possibly µ) are
different on either side. Nonetheless, we can use the same
kind of computations that we saw previously to derive the
gu
boundary conditions. Roughly, we get one boundary condi-
tion from each of the Maxwell equations.
ics
ys
Figure 2.2: Boundary conditions of electric field

Ph
For example, consider the Gaussian pillbox shown in the

left-hand figure above. Integrating the Maxwell equation
∇ · D = ρ free tells us that the normal component of D is
discontinuous in the presence of surface charge,
n̂ · (D2 − D1 ) = σ (2.15)
where n̂ is the normal component pointing from 1 into 2.

Here σ refers only to the free surface charge. It does not
include any bound charges. Similarly, integrating ∇ · B =
ide
0 over the same Gaussian pillbox tells us that the normal
component of the magnetic field is continuous.
n̂ · (B2 − B1 ) = 0 (2.16)
gu
To determine the tangential components, we integrate
the appropriate field around the loop shown in the right-
ics
hand figure above. By Stoke’s theorem, this is going to
be equal to the integral of the curl of the field over the
bounding surface. This tells us what the appropriate field
is: it’s whatever appears in the Maxwell equations with a
ys
curl. So if we integrate E around the loop, we get the result
n̂ × (E2 − E1 ) = 0 (2.17)
Ph
Meanwhile, integrating H around the loop tells us the dis-

continuity condition for the magnetic field
n̂ × (H2 − H1 ) = K (2.18)
where K is the surface current.
2.3.3 Example: Field inside the gap
ide
Here is a very interesting problem. A C-magnet is shown in
Figure . All dimensions are in cm. The relative permeability
of the soft Fe yoke is 3000 . If a current I = 1 amp is to
produce a field of about 100 gauss in the gap, how many
turns of wire are required?
gu
ics
ys
Ph
Figure 2.3: Magnetic field in the gap
Solution: You must know the boundary condition of

the magnetic field. Consider the surfaces which act like the
cross section of the magnet (red colored in the image). The
normal component of B is continuous

⊥ ⊥
Babove − Bbelow =0
ide
And there are no parallel component of Magnetic field (re-
member that in a solenoid the field is along the axis only).
So, the magnetic intensity in the gap is H = µB0 , while that
inside the magnet is H = µ0Bµr , where µr is the relative per-
gu
meability of the iron. Ampere’s circuital law
I
H · dI = N I
ics
applied to the closed loop passing through inside the magnet
and covering the whole magnet
B B
d+ (4l − d) = N I
ys
µ0 µ0 µr
hence
B 1
N= d + (4l − d)
Ph
µ0 I µr
now apply the given data - l = 20 cm = 0.2 m , d = 2 cm
= 0.02 m, and B = 100 gauss = 100 × 10−4 Tesla.
100 × 10−4

0.2 × 4 − 0.02
N= 0.02 +
4π × 10−7 × 1 3000
= 161 turns

PH Ysicsguide: Magnetostatics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PH Ysicsguide: Magnetostatics

Uploaded by

Copyright:

Available Formats

ide

He has been teaching CSIR NET aspirants since 2012

1.1 Steady current . . . . . . . . . . . . . . . . . 4

1.2 Force on a moving charge . . . . . . . . . . 6

Electric current as a source of magnetic field

1.3.2 Ampere’s Law . . . . . . . . . . . . . 11

1.4 Surface Currents and Discontinuities . . . . 13

1.5 Magnetic Vector Potential . . . . . . . . . . 17

1.6 Magnetic Dipole . . . . . . . . . . . . . . . . 20

1.6.1 Torques and Forces on Magnetic Dipoles 28

1.7 Electric Current . . . . . . . . . . . . . . . . 28

1.7.1 Current Density . . . . . . . . . . . . 30

1.7.2 Conservation of Charge . . . . . . . . 30

1.7.3 Ohm’s Law . . . . . . . . . . . . . . 31

2.1 Bound Currents . . . . . . . . . . . . . . . . 37

2.2.1 Linear magnetic material . . . . . . .

2.3.1 Magnetostatic Boundary condition in-

2.3.2 Boundary condition in both electric

2.3.3 Example: Field inside the gap . . . . 45

If you fix the current density J , these equations have a

1.1 Steady current

we still want charge to be moving! This means that we nec-

This may sound artificial, but in fact it’s exactly what

Meanwhile, the electrons are free to move. But they all

1.2 Force on a moving charge

. At the surface of the Earth the magnetic field is about

interstellar space in the galaxy.

The direction of the magnetic force on q is sideways, i.e.,

Figure 1.1: The magnetic force. The force on a moving

charge is sideways, perpendicular to both v and B. The

Equation (1.5) may be taken as the definition of the

test charge q, e.g. by observing its deflection for a known

charge, and (1.5) is the force on a moving charge. Suppose

force dF on a small segment d` of the wire is

dF = (n1 d`) qv × B (1.8)

because n1 d` is the number of moving charges in d`. The

- which, although not classical currents, do involve dynamics

1.3.1 Biot-Savart Law

The magnetic field of a steady line current is given by the

finitesimal current element Id` at a point P 0 . Figure below

Here r is the distance between P 0 (the source point) and P

Do few examples from Griffiths.

1.3.2 Ampere’s Law

For the problems where cylindrical symmetry is there, it is

We take the open surface S to lie in the x − y plane,

centered on the wire. For the line integral in (1.11) to give

But, by the symmetry of the problem, that’s actually

0. We need only worry about Ampère’s law which tells us

We see that the strength of the magnetic field is

The magnetic field circles the wire using the “right-hand

1.4 Surface Currents and Discontinuities

We’ll take the current to lie in the x-direction: K = K x̂

Figure 1.4: Surface Current

From our previous result, we know that the B field

should curl around the current in the right-handed sense.

with B(z) = −B(−z). We invoke Ampère’s law using the

following open surface:

so we find that the magnetic field is constant above an infi-

nite plane of surface current

This is rather similar to the case of the electric field in the

The analogy with electrostatics continues. The magnetic

In fact, this is a general result that holds for any surface

Meanwhile, the magnetic field normal to the surface is

Electrons orbiting the nucleus carry angular momentum

Electrons carry an intrinsic spin. This is purely a quan-

Diamagnetism: −1 < χm < 0. The magnetisation of

Paramagnetism: χm > 0. In paramagnets, the magneti-