Magnetic Materials

Basic Concept Behind Magnetism

Circulating current sets up a magnetic
moment (
m
) perpendicular to the current
This results in a B field (magnetic field) that must
terminate back on itself (all magnets must have a
north & south, cant be alone like electrical charge)


Electrons in atoms are the same
concept-circulating charge that sets up a
magnetic moment and a magnetic field
The magnetic field is due to both the orbital angular momentum
and the spin
Only electrons in unfilled shells contribute to a net magnetization
(as those in full shells will cancel each other out)
I
A
e
e

orb
L
r
Fig. 8.4: An orbitting electron is equivalent to a magnetic
dipole moment
orb
.
Atomic Dipole Moment
Magnetic dipole moment: (orbital)
Electron current loop acts like a magnetic dipole with
dipole moment,
orb
= I.A = (-e/2m)L
where L is angular momentum of electron
Magnetic dipole moment: (spin)
Electron rotates about its axis with dipole moment

s
=(-e/m)S, where S is spin angular momentum of electron


Net magnetization is due to magnetic moments
from both forms of angular momentum
However, only electrons in un-filled shells
contribute to an overall magnetic moment
Current loop in an atom
A = r
2
I = e(/2)
L = mr
2

Fleming RH rule
e/2m = gyromagnetic ratio,
relates
m
to L (of q.mech)
g. ratio = e/m,
no classical deriv
n
Quantization of L
Atomic electron with L
interacts with B
z

L can never point in any
specific direction,
If it does so it violates
Uncertainity principle
Instead it is somewhere on
a cone in space so that its
projection is m
l
(h/2)
A.k.a. Space quantization
L
z
= m
l
(h/2), m
l
=0,1,2 ..

Ch6. Mod. Phys., Bieser
Larmor Frequency
The electron spin or proton spin will tend to
precess around the magnetic field with a
frequency traditionally called the Larmor
frequency. For a 1 Tesla magnetic field this
Larmor frequency would be

Quantization of S
In an applied magnetic field, m spin cannot
align with B because S is space quantized.
The torque that results cause the spin magnetic
moment to precess about B.
35
Spin Down
Spin Up S
z (along B
z
)
+ /2
S
S
0
3
2
m
s
= +
1
/
2
m
s
=
1
/
2
/2
3
2
Fig. 3.28: Spin angular momentum exhibits space
quantization. Its magnitude along z is quantized so that the
angle of S to the z-axis is also quantized.
2
1
2
3
= =
=

s Z
m S
S
Unfilled
s shell
e.g. H
A magnetic dipole moment (usually just called "magnetic moment") in a
magnetic field will have a potential energy related to its orientation with
respect to that field.











Note that the electron spin magnetic moment is opposite to the electron spin
while the proton spin magnetic moment is in the direction of the proton spin.
Bohr Magneton
Each spin magnetic moment (
s
) contributes a
average magnetic moment on the z axis in the
presence of a magnetic field



Potential energy (orientational) of the dipole
V = -
m
.B = -(-e/2m)L
z
B
z
= (eh/2m)Bm
l
= Bm
l


| = =

=
e e
s
e
z
Z
m
e
m
em
m
eS
2

0.927 x 10
-20
erg/gauss
9.3 x 10
-24
J m
2
/Wb
B=0
B0
m
l
=1
m
l
=0
m
l
=-1
Zeeman effect: Splitting of atomic level by magnetic field
E
E = B = h
Larmor

Orbital
B=0
B0
m
S
=1/2
m
S
=-1/2
E = 2B
Spin
Macroscopic Magnetism Terms
B
o
flux density in free space

o
permeability of vacuum
H applied field intensity(m.m.f)
M Magnetization (dipole moment /vol.)
B flux density in material
_
m
magnetic susceptibility=M/H
B =
o
(H+M)=
o
(1+_
m
)H=H
Permeability =
o

r


_
m
10
-5
local field ignored in dia, para,
but not in Ferro, similar to _
E
1 requires E
local

Molecular field theory
Where does M come from?

M that results from applied field is a function of material in the core
The material types can be divided into several main categories
Diamagnetism
Paramagnetism
Ferromagnetism
Antiferromagnetism
Ferrimagnetism
Recall
r
< 1
Types of Magnetism
Material (per cm
-3
)
Para - Al
Mn
W
+ 2.2 x 10
-5
+ 98
+ 36

Dia Cu
Au
Hg
Water
H
-1.0 x 10
-5
-3.6
-3.2
-9.0
-0.2 x 10
-8

_
m
(Fe) 10
+5

0<
r
<1
1<
r
<2
2<<
r

Diamagnetism
In atoms or ions with filled shells
with no permanent dipole moment
Change in orbital frequency of
electrons on application of field
Induced dipole opposes the field
model applicable to core electrons
and not to those in covalent bond
due to non-circular orbit.
All paramagnetic materials also
have diamagnetic contribution (-ve)
from filled shells, but may be
masked other magnetic effects.
-e
v
F
1

-e
v
F
2

.
B
2

B
1

0
=> =eB
o
/2m
mv
2
/r =
(1/4
o
)(-e)Ze e(vxB)
mv
2
/r =
(1/4
o
)(-e)Ze
-e
v
-e
v
.
B
2

B
1

=0
=> =0
(per cm
-3
) (per cm
-3
)
He
Ne
Ar
Kr
Xe
F
-
Cl
-
Br
-
-1.9 x 10
-6
-7.6
-19
-29
-44
-9.4
-24.2
-34.5
Li
+
Na
+
K
+
Mg
+
Ca
2+
Sr
2+
0.7 x 10
-6
6.1
14.6
4.3
10.7
18.0
Rare-gas atoms and ions of filled shell structure
Langevin Theory of Diamagnetism
m
eB
o
2
e = e
B
m
r e
orb
|
|
.
|

\
|
= A
4
2 2
B
m
r e
orb
|
|
.
|

\
|
= A
6
2 2
( )
2
2
6
r NZ
m
e
o
|
|
.
|

\
|
= _
|
|
.
|

\
|
tc
= e
o
o
Ze e
r m
4
.
2
e e = e eBr r m r m
o
2 2
-e
v
F
1

-e
v
F
2

.
B
2

B
1

orb
0
mv
2
/r = (1/4
o
)(-e)Ze
-e
v
-e
v
.
B
2

B
1

orb
=0
=> =0
mv
2
/r = (1/4
o
)(-e)Ze e(vxB)
For B0
ignore B
2

I
electron
L

orb

Due to uncertainty principle electron orbits
around a spherical surface and not limited
to a circle. Hence r
2
is replaced by (2/3)r
2
to
projection taken prep
r
to field.
For N atoms/vol., each with Z electrons
|
|
.
|

\
|

A
= = _
o
orb
B
NZ
H
M
/
With N 10
29
m
-3
, Z=10, r=1A
o
=> 10
-5

In Absence of Magnetic Field
both dipoles cancel each other
On application of magnetic field the two dipole strengths differ
to give a resultant dipole moment that opposes the field

orb
=I.A = e
o
r
2
/2

Diamagnetic Susceptibility

dia
is independent of temperature, but depends on Z
Average on all
the occupied
orbitals in atom
3D orbit
R
2
xyz
= x
2
+y
2
+z
2

2D orbit
R
2
xy
= x
2
+y
2

Assuming
spherical
symmetry
R
2
xy
= (2/3)R
2
xyz

Langevin Theory of Paramagentism
Paramagnetic
Atoms have permanent dipole
It may be orbital or spin.
In absence of field, atomic dipole
point in random directions due to
thermal agitation and hence have
zero resultant field in all directions
dipoles tend to align with the
applied field to net magnetization
parallel to field.
Langevin Theory of Paramagentism
Pot. Energy of a magnetic dipole in magnetic field is
Component of avg. moment along dir
n
of field (say z)
Langevin Function


In most practical situations v << 1, with B = 0.1 Wb/m
2

and T = 300K, we can write L(v) (1/3)v
For N atoms/vol., Magnetization is

Langevin paramagnetic susceptibility
Curie Law
B V =
) (v L
z
=
v
v Coth v L
1
) ( ) ( =
v
1
L(v)
v/3
energy
thermal
nal orientatio
kT
B
v =

=
kT
B
z
3
2

=
kT
B N
N M
z
3
2

= =
kT
N
B
M
o
o
3 ) / (
2

=

= _
T
C
= _
With N 10
26
m
-3
, T=100K

=> 10
-5

k
N
C
o
3
2

=
Curie constt.

T
Curie Law

T
T
c

Curie Wiess Law
c
T T
C

= _
Sat
n
: dipoles
fully aligned
Langevins Function
# of particles whose axis falls within solid angle
d between cones and d is
kT B
o
Ce dn
/ cosu
=
}
e =
u
d e C n
vcos
u u t = e d d sin 2
u = = cos
o o
B B E
kT
B
v

=
}
t
u
u u t =
0
cos
sin 2 d e C n
v
Substituting cos=x - sin d = dx
) (
2
2
1
1
v v vx
e e
v
C
dx e C n

+

t
= t =
}
v
v
C
n sinh
4t
=
Each of the dn particles contributes a component of magnetic moment cos parallel to
the field, while by symmetry the component prep
r
to the field will cancel each other.
Magnetization of given mass of gas
} }
u u t u = u =
u
n
v
n
d Ce dn M
0
cos
0
sin 2 . . cos . cos
(

+
t =
(

t = t =

+
} 2
1
1
2
0
2 ) 1 ( 2 . 2
v
e e
v
e e
C vx
v
e
C dx e x C
v v v v vx
n
vx
) (
1
coth
sinh 2
cosh 2
sinh 4
2
v nL
v
v n
v
v
v
v
n
=
(

=
(

t
t
=
) (v L
n
M
= =
H
Langevin derived this formula for dipolar magnetization before its
electrical analog. His treatment was adapted to electrical case by Debye.
Average Magnetization
in the direction of field
Maxwell Boltzmann Statistics

d
kT E
e dn
/

Alternate method
Solid Sate Physics, Rogalski, and Palmer, Pg367
Shortcomings of Langevins theory
For intense fields and low temp. L(v)1,
but results at sat
n
do not match theory
Could not explain why particles have same
magnetic moment and not from - to +
On application of M-B stats, dia and para
susceptibility should cancel to give zero.

Permanent magnetic moment of a
given atom is not freely rotating but
restricted to a finite set of orientations
relative to applied field.
Let total angular momentum quantum
number of each atom be J = L+S
Relation between angular momentum
and magnetic moment of each atom
depends on the relative proportions of
spin and orbital angular momentum. It
is contained in Lande splitting factor g


Magnetic dipole moment is given as
35
Spin Down
Spin Up S
z (along B
z
)
+ /2
S
S
0
3
2
m
s
= +
1
/
2
m
s
=
1
/
2
/2
3
2
Fig. 3.28: Spin angular momentum exhibits space
quantization. Its magnitude along z is quantized so that the
angle of S to the z-axis is also quantized.
Quantum Theory of Paramagnetism
) 1 ( 2
) 1 ( ) 1 ( ) 1 (
1
+
+ + + +
+ =
J J
L L S S J J
g
L=L
i
S=S
i

Summation over
incomplete shell
2 1
)] 1 ( [ J
2
+ | =
|
.
|

\
|
= J J g
m
e
g
J
J = h[J(J+1)]
1/2

where
S
L
J

J
=cos
is not collinear with J
but is tilted towards spin
because of its higher
gyromagnetic ratio.

It thus precesses
around J.

Only component along
J is taken as prep
r
part
cancels
External magnetic field splits each energy
level into 2J+1 levels.
Component of each magnetic moment in the
dir
n
of field is
Energy values of each magnetic moment in
field B is
Use M-B stat. to obtain average magnetic
moment per atom as
J m
gm
J
| =

+
=

+
=

=
J
J m
kT B
J
J m
kT B
m
J
J
m
J
J
m
J
e
e
/
/
m
j
= J, -(J-1), .. 0, .. J
J m
Bm g B
J
| =
)
`

|
|
.
|

\
|

| =

|
|
.
|

\
|
| = | =

| =

+ +
=

+
=

+
=
+
=

+
=

x
x J
Jx
J
J m
x m
J
J m
x m
J
J m
x m
J
J
J m
x m
J
J m
x m
J
e
e
e
dx
d
g e
dx
d
g
e
e m
g
e
e m
g
J
J
J
J
J
J
J
J
J
J
1
1
ln ln
) (
) 1 2 (
) (
) (
Let x = Bg/kT be ratio of magnetic
interaction energy to thermal energy
) ( B
2
coth
2
1
2
1 2
coth
2
1 2
) 2 / sinh(
2 / ) 1 2 sinh[
ln Jx gJ
J
Jx
J
Jx
J
J
J
J
gJ
x
x J
dx
d
g
J
| =
)
`

|
.
|

\
|

|
.
|

\
| + +
| =
)
`

|
|
.
|

\
| +
| =
B
J
(Jx) is known as Brillouin function
Fig 9.2 pg366, Rogalski, Palmer
for J large,
B
J
(Jx) L(Jx)
For N atoms/vol., Paramagnetic
susceptibility is

Each atom was assumed in state
with the same J quantum number.
This assumption is valid only in
weak field. For x<<1 i.e. for high
temperatures, we get Curie law



For x>>1 i.e. for low temperatures,
we have

) ( B
) / (
Jx
B
gJ N
B
N
H
M
J
o
o
o o
p
|
=

= = _
x
x
x
1
3
coth + ~ or Jx
J
J
Jx
J
|
.
|

\
|
+
~
3
1
) ( B
kT
g B J
B
gJ N
o
o
o
p
|
|
.
|

\
|
+ |
= _
3
1
T
C
kT
N
g J J
kT
N
eff
o o
p
=

= | +

= _
2 2 2
3
] ) 1 ( [
3
k
N
C
eff o
3
2

=
| = | + = p J J g
eff
) 1 (
Curie const.
1 coth ~ x or
1 ) ( B ~ Jx
J
o
o
p
B
gJ N |
= _
eff o o
N gJ N M ~ | =
Magnetization M
o
is a constant for a given N and J and
represents a sat
n
value, where every atom contributes its
max. possible magnetic moment
p

is eff. number of Bohr
magnetons

for the atom
Unfilled 4f shell screened by filled 5p, 5d shells, not effected by
other crystal ions, free ion behavior, L & S are strongly coupled.
Incomplete outermost 3d shell. Strongly coupled to neighboring
ions. Orbital motion quenched. Only S contributes.
Quantum Theory of Paramagnetism
Atomic magnetic moment in terms
of total angular momentum J

On application of magnetic field,
Zeeman splitting takes place


If N were density of atoms in a
state, then Magnetization M is


Solving the equations, let x = gB/kT
J
m
e
g
|
.
|

\
|
=
2
g Lande factor: its value depends upon relative
orientations of spin and orbital angular momenta
E
pot
= -
m
.B = gBm
j
B=0
B0
m
j
=1/2
m
j
=-1/2
E = gB
Parallel
to field
Opposite
M=g(N
parallel
-

N
opposite
)
N
opp
/N
ll
=e
-E/kT

From MB stats, we have
Component
parallel to field
N = N
opp
+N
ll

) tanh(x Ng
e e
e e
Ng M
x x
x x
| =
+

| =

x
M
Ng
At low fields, M is proportional to B at higher fields M
begins to saturate when all dipoles are aligned with field
- Conclusion same as that from classical treatment
Magnetism in metals
Langevins theory says
paramagnetic
1/T, but in paramagnetic
metals it is independent of T

Pauli (1927) demonstrated that this is due
paramagnetism of free electrons (e-gas).

Reason: metals follow Fermi-Dirac stats
instead of M-B stats used by Langevin.

In absence of field, spins of free electrons
are randomly oriented with zero magnetization

Electrons divide into two equal halves with spins aligned
or opposite to the direction in which field is to be applied.

When field is applied, energy of electrons with spins
aligned to field is lowered, and raised for opposite spins.

Situation being energetically unstable, so some electrons
near Fermi level switch to lower energy of aligned spins

This leads to net magnetization due relative increase in
population of aligned spin.
B=0
B0
m
s
=1/2
m
s
=-1/2
E = 2B
Aligned
to field
Opposite
E
pot
= -
m
.B = Bm
s
B=0
E
F

B
E
F

} 2
# electrons (N
eff
) which change their
direction =density of states x E
Increase in net magnetization
Pauli spin susceptibility of e-gas



p
is independent of T, as temperature
has little effect on g(E
F
)
In terms of classical susceptibility
With T
F
30,000K,
p
is 10
-2
as
observed experimentally.
For transition metals with g(E
F
) large

p
is also exceptionally high
Diamagnetic contribution is only 1/3
p


B E g N
F eff
| = ) (
2
1
Only half of total
density of states flip
| ~ 2
eff
N M
) (
) / (
2
F o
o
p
E g
B
M
| =

~ _
2 / 1
2 / 3
*
2
2
2
1
) (
F F
E
m
E g
|
|
.
|

\
|
t
=

3 / 2 2
*
) 3 (
2
N
m
E
F
t
|
.
|

\
|
=

From free
electron theory
of metals
F
F
E
N
E g
2
3
) ( =
F
o p
E
N
2
3
2
| ~ _
F
p
T
T
_ ~ _
2
3
as E
F
= kT
F

2 2 2
] ) 1 ( [
3
|

= | +

= _
kT
N
g J J
kT
N
o o
For l=0, J=s, => g=2. Also for J=1/2
B
E
F

} 2
Ferromagnetic
Posses magnetization even without the
presence of an applied field
Exists up to T
C
, Curie temperature, beyond
which it is paramagnetic
Only certain materials are ferromagnetic
criteria 1 is that there is an unfilled shell
criteria 2 is that there is a positive exchange energy
Fig 8.15: In a magnetized region of a ferromagnetic material such as iron all the magnetic moments are spontaneously
aligned in the same direction There is a strong magnetization vector M even in the absence of an applied field
Molecular field Theory
Ferromagnetics are spontaneously
magnetized which produces internal field for
more magnetization
Magnetizing field from this magnetization is,
where is Weiss Const. 10
4
For J=1/2 we can express magnetization as
Transcendental eqn., is solved graphically


For T < T
F
(some

critical value), curves
interact to yield a sol
n
i.e. finite spontaneous
magnetization or ferromagnetism.

We find T
F
Curie temp. at tangent tanh(x) x


) B/kT g tanh( | | = Ng M
M
x
T >T
f
T =T
f

T <T
f
, M x
M tanh(x)
0
H
W
= M, or B =
o
M
) M/kT g tanh( | | =
o
Ng M
) tanh(x Ng M | = x
g
kT
M
o
|
=
2
) ( |
=
g N
kT
o
f
With T
f
= 1000K, N=10
29
m
-3
=> 10
-4

Weiss (1907)
Something like E
local

Refer to last slide for tanh relation
Curie Weiss law of Ferromagnetism
Max. Magnetization M(0) is

Normalized Saturated Magnetization
v/s normalized Temperature is
Total field H
tot
is sum of applied field
H and molecular field H
w

For small H


Curie-Weiss law
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Iron
M
sat
(T)
M
sat
(0)
T / T
C


Fig. 8.21: Normalized saturated magnetization vs. reduced
temperature T/T
C
where T
C
is the Curie temperature (1043K).
| = Ng M
0 T
|
|
.
|

\
|
=
f
T
T
M
M
tanh
) 0 (
Universal curve for all materials
M H H H H
W total
+ = + =
) ( ) 0 ( M H
kT
g
M M
o
+
|
=
=> tanh(x) x
) ( M H
T
T
M
f
+
|
|
.
|

\
|

=
f f
f
T T
C
T T
T
H
M

=
|
|
.
|

\
|

|
|
.
|

\
|

=
) (
1
2
) (
2
|
=
|
|
.
|

\
|

=
g N
T
C
o
f
Origin of Weiss Molecular field
Large Mol. field explained as exchange interaction between
electron spins
Strength of interaction is represented by exchange constant J
It was expt
ly
found to depend on ratio of inter-atomic spacing r
and radius of unfilled shell r
d

J is related to Weiss constant , J 0.1eV
Dipole-dipole interaction is too small to account for strong
interaction.
Heisenberg proposed that this strong interaction is because of
spin dependent force due to Paulis exclusion principle.
While being inside the strong coulomb force electron spins can
be either parallel or antiparallel.
When they are parallel, Pauli exclusion principle requires that
they be far apart. If they are antiparallel their may come close to
have overlapping wave functions but energy rises due to strong
Coulomb repulsion
This explains why parallel spin is favored, but there are other
effects that may still have antiparallel spins.
Heisenberg 1928)
r/r
d
3 for
ferromagnetism
Slater (1930)
Element r/r
d

Fe 3.26
Co 3.64
Ni 3.94
Cr 2.60
Mn 2.94
Gd 3.10
Exchange Energy
E
ex
=-2J
e
S
1
S
2
S
1
and S
2
are spins of electrons
J
e
is negative for most materials
So the exchange energy is -ve
(minimum) if the spins are misaligned
For Fe, Co, Ni
J
e
is positive
So the exchange energy E
ex
is -ve
(minimum) if the spins are aligned
Fe, Co, Ni most common examples of
ferromagnets
Domains
Experimental results
show that freshly smelted
(solidified) iron below
Curie temperature does
not act like magnet
Schematic illustration of magnetic domains in the
grains of an unmagnetized polycrystalline iron
sample. Very small grains have single domains.
10
9
10
15
atoms
Anisotropic energy
There are easy
directions: those most
likely for the atoms in
the domain to align to
For cubic crystal
Magnetocrystalline anisotropy in a single
iron crystal. M vs. H depends on the crystal
on the crystal direction and is easiest along
[100] and hardest along [111]
2 2 2
2
2 2 2 2 2 2
1
) ( | o + o + | + | o = K K E
an
K
1
, K
2
are constants, ,, are direction cosines
For Ni the [111] is preferred over [100]
Pg 285, Shankaran, Kumar, Prasad
Why do domains form?
If material is magnetized, there will be
magnetic field lines in the surrounding
area.
Magnetized bar of ferromagnet in
which there is only one domain and
hence an external magnetic field.
These field lines, just like electric field
represent potential energy
(magnetostatic energy).
Formation of multiple domains with
opposite magetization reduces the
external field
Why dont domains keep subdividing?
Exchange energy is less for single domain but has high magnetostatic energy,
Multiple domains has less magnetostatic energy, but high exchange energy.
Domain or Bloch walls
To reduce high energy boundaries, domain
walls occur over many atoms that slowly
misalign to minimize the exchange force energy
created
In a Bloch wall the neighboring spin magnetic moments rotate
gradually and it takes several hundred atomic spacings to rotate
the magnetic moment by 180.
Potential energy of a domain wall
depends on the exchange and anisotropy
energies
Domains in applied field
Atoms systematically
align with the field
Amount of magnetization
(M) for a given applied
field depends on how
easy it is to align the
domains in that direction
for that particular material
Energy is lost to the
lattice as domain walls
are moved
This makes process
irreversible
A hysteresis loop forms
Hysteresis Loop
Demagnetization
Increase magnetization
Atoms align to applied magnetic field
Magnetization increases from 0 to Sat
n

(M
sat
)

If magnetic field removed, there is a
remainance (M
r
) left in the material
Coercive force (H
C
) is the magnetic
field in the opposite dir
n
required to
remove this magnetic field
This however isnt how to demagnetize
If you remove applied field, the
magnetization in the material will go to
some non zero value
To demagnetize, keep cycling through
smaller and smaller hysteresis cycle.
Antiferromagnetism
Individual atoms bond as to give no
magnetic moment even in the presence
of a field due to the crystal structure
Exists only below Neel Temperature: T
N
Discovered in 1938 in MnO by Bizette, Squire, and Tsai
Theory by Neel and Bitter, and later by Van Vlack
Explained by two interpenetrating simple
cubic sublattices with antiparallel spins
Since Mn has ve J
e
hence it
prefers antiparallel spins for
nearest neighbors (corner and
body center) and so on.
Two interpenetrating simple cubic
sublattices with antiparallel spins.
Let Internal field H
A
for A ions and H
B

for B ions be given as

Total magnetization



To find T
Neel
, i.e. M
A
, M
B
finite at H=0
For T>T
N
we get trivial

sol
n
M
A
=M
B
=0
T = T
N
, for non-trivial sol
n
for M
A
,M
B
=>
T
N
=

T
C
, expt
ly
T
N
<T
C

Susceptibility of poly-crystalline form is
found be averaging over all dir
n

B A
M H H =
A B
M H H =
Magnetization of A is opposite to B,

is internal field const, represents
strength of exchange interaction
|
.
|

\
|
|
| =
W
o
H
kT
Ng M
g
tanh ) (
) (
) (
) (
2 2
A
o
B B
o
A
M H
kT
g N
M M H
kT
g N
M
|
=
|
=
for T large,
assume
tanh(x) x
) 2 (
) (
2
M H
kT
g N
M M M
o
B A

|
= + =
) ( M H
T
T
M
f
+
|
|
.
|

\
|

=
f f
f
T T
C
T T
T
H
M

=
|
|
.
|

\
|

|
|
.
|

\
|

=
) (
1
) 2 ( M H
T
T
M
f

|
|
.
|

\
|

=
+
= = _
C T
C
H
M 2
As before
Similarly, if we
assume M & H
are parallel
k
g N
C
o
2
) ( |
= = C T
C
where C is double of that for ferro
0 0 =

+ =

+
B A A B
M
T
C
M M
T
C
M
1 =

N
T
C
as two sublattice model is over simplified

_ + _ = _
3
2
3
1
// pc
Quantities
//
and

below T
N
refer to susceptibilities for the field parallel and prep
r
to spontaneous spin dir
n
-T
C

1/3
1/3
1/3
0
0
//
= _
= _

??
Ferrimagnetism
Results in a net magnetization even when there
is no applied field (similar to ferromagnetic).
It comes from opposite magnetizations in crystal
structure of differing magnitudes resulting in a
net permanent magnetization in one direction
Fig. 1.2: Illustration of magnetic ordering in a ferrimagnetic crystal. All Aatoms have their spins aligned in one
direction and all Batoms have their spins aligned in the opposite direction. As the magnetic moment of an A
atom is greater than that of a Batom, there is net magnetization, M, in the crystal.
e.g. Magnetite (lodestone) Fe
3
O
4
Generalized formula Fe
2
MO
4

M is a divalent metal ion, e.g.
Cu, Pb, Mg, Mn, Co, Ni or Fe
Tetrahedral site(M
2+
), each with
4 neighboring O atoms
Octahedral sites(Fe
3+
), each with
6 neighboring O atoms
At low T, Tetra spin aligned along
cube edge, Octa opposite to it.
Understood as two interlaced sublattices.
Exchange energy minimum when spins
on different sublattices are antiparallel.
Spins on same sublattice aligned becuz
of exchange interaction with other
sublattice.
Good insulators, minimize losses
due to eddy currents
Comparison
Type Susceptibility Temp. dependence Examples
Dia
Complete
shells
Small, -ve
-ve
Large ve
Independent
Varies with mag. Field T< 20K
Exists below critical temp only
Organic materials, light elements, Noble gases
Alkali earths, Bismuth
Superconductors
Para
Incomplete
shells
Small +ve
Large +ve
Independent
= C/T
Alkali metals, transition metals
Rare earths
Ferro
Incomplete
shells,
strong
exchange
energy
Very large +ve T> , = C/(T- )
T< , = Complex
Some transition and rare earth elements
Antiferro Small +ve T> T
N
, = C/(T+ )
T< T
N
, T
Salts transition elements
Ferri Very large +ve T> T
N
, = C/(T )
T< T
N
, = Complex
Ferrites
Spin waves
In 1-D array of N magnetic atoms, each with
spin S,
Each spin interacts in its neighborhood via
Heisenberg exchange interaction.
System is in ground state with spins aligned
First excited state is not a flip spin, instead
A small disturbance in alignment makes spins
precess around magnetization vector M
o
.
Spin precessional motion is correlated
through exchange forces and propagates
as a wave
Collective behavior is much like lattice
vibrations
Helps to predict low temperature behavior
M
o
z
Ferromagnetic resonance
Dispersion relation for Spin waves
Effective field at atom p is
Rate of change of mom
n
= Torque on S
p

Then eqn of motion is
With spins only along z-axis, sol
n
is likely
to be a traveling spin wave.
With periodic boundary conditions we
expect quantization of k, Magnon
On substituting the sol
n
we get
dispersion relation
Total exchange energy is
Spin wave energy

Using dispersion relation
) S S )( / 2 (
1 1 +
+ | =
p p e p
g J B
dt
d
B
p
p p
S
=
) S S ( S
2
1 1 +
+ =
p p p
e
p
J
dt
dS

) sin(
) cos(
t kpa A S
t kpa A S
py
px
e =
e =
a is separation between
adjacent spins,
2 2
A S S
pz
=
Na n k / 2t =
-N n <+N
)) cos( 1 ( 4 ka S J
z e
= e

=
+
+ =
N
p p p
e
J
E
1 p
1 1
) S S ( S
2
2
1

to void p,p-1 duplication

z S y S x S S
pz y p px p
+ + =
| | ) 2 / ( sin 2
2 2 2
ka A S
NJ
E
e
=

2 2
2
A
S
N
S
NJ
E
z
e
e
+ =

Energy for aligned spins

+ Energy for spin waves

k
/a
8J
e
S
z
/
0
p p
g S | =
High T sets up higher modes of Spin
waves and hence loss of magnetization
Electron Paramagnetic Resonance
On applications of static field induced magnetization is
obtained after sufficient time has elapsed for the system to
reach its equilibrium.
Application of alternating magnetic field helps us to
determine the nature of state of dipoles, interaction between
dipoles, and interaction between dipoles and lattice.
We will deal with paramagnetic system where interaction
between dipoles is weak.
We expect a paramagnetic resonance when frequency of
external field is equal to Larmor frequency.
When a magnetic field B is applied, the magnetization
vector M moves as
When B is a constant field, M simply precesses around B
with Larmor frequency
When field is itself varying, we expect complicated motion
B
m
e
dt
d

|
.
|

\
|
=

2
= g(e/2m) is
gyromagnetic ratio
B M
dt
dM
=
B has two parts (i) large static component
B
o
along z-axis, (ii) small alternating
transverse component b in xy-plane
b is small, so precession only about z-axis
with Larmor frequency
o
Similarly M will also have two parts, (i)
longitudinal M
z
, and transverse m (xy plane)
Substituting them in eqn of motion

b B + = z B
o
o o
B = e
m M + = z M
z
0 ) ( ) ( ) ( ~ = = =
x y y x
z
o x x z
y
y z o y
x
b m b m
dt
dM
B m b M
dt
dm
b M B m
dt
dm
Becuz of
second
order small
terms
B M
dt
dM
=
y x o
b b z B + + = B
y x z
m m z M + + = M
o o
B = e
With M
z
assumed to be constant, we have

Complete sol
n
depends on nature of b
Case-1: b is plane polarized signal
Sol
n
will also be of similar nature
On substitution, we get

Susceptibility is a tensor
=> m not in same direction as b but lags
behind

=> , at =
o
=> EPR
b + = z B
o
B
o o
B = e
m M + = z M
z
0 ) ( ) ( ) ( ~ = = =
x y y x
z
o x x z
y
y z o y
x
b m b m
dt
dM
B m b M
dt
dm
b M B m
dt
dm
t i
e
e
=
o
b b
t i
e m
e
=
o
m
y o x
o
z
y y x o
o
z
x
b b i
M
m b i b
M
m e e
e e

e e
e e

= +

= ( ; ) (
2 2 2 2
t i
x
e m m
e
o
=
t i
e b
e
o x
b =
2 2
2 2
/
/
e e
e
_

_
e e
e
_

= = =

= = =
o
z o
yx
o y
x
xy
o
z o o
yy
o x
x
xx
M
i
b
m
M
b
m
(

(
(

=
(

y
x
yy yx
xy xx
y
x
b
b
m
m
_ _
_ _
Case-2: Assume circular polarization

We get circular susceptibility



anticlock
exhibits resonance at
o
as
Larmor
is
anticlockwise

x y o y o x
ib b t b b t b b = e = e = sin cos
x y o y o x
ib b t b b t b b = e = e = sin cos
e + e

= _
e e

= _
o
z o
clock
o
z o
anticlock
M M
x y
im m =
x y
im m =
On applying static magnetic field dipoles
align, in doing so they lose some
magnetic energy.
This can occur if dipoles are coupled to
environment (lattice)
Rewriting eq
n
of motion as Bloch Eqns
starting at some arbitrary value M
z
will
approach M
o
after spin-lattice relaxation
time
1
At High T, interaction is more =>
1
small
starting at some arbitrary value m will
approach m=0 after spin-spin relaxation
time
2
Density => Dipole-dipole interaction
=>
2

B M
dt
dM
=
2
,
,
,
) (
t
=
y x
y x
y x
m
B M
dt
m d
1
) (
t

=
o z
z
z
M M
B M
dt
dM
Bloch Equations
@ liquid N
2
T,
1
10
-6
s

2
10
-10
s

to setup coherence
Transverse Relaxation time
The time
2
is known as the transverse relaxation time,
or, more descriptively, as the spinspin relaxation time.
It arises because neighboring dipoles, coupled via the
familiar magnetic dipole-dipole interaction, attempt to
break up any initial coherence between the directions of
the individual transverse moments.
The time
2
is usually very short, often of the order of 10
-
10
s, and is independent of temperature.
It does, however, depend upon concentration of the
magnetic atoms; the larger the concentration the closer
the dipoles, which leads to a strong interaction and
consequently a shorter relaxation time
2
.
Relaxation Process
How does this great disparity between
1
and
2
affect our picture of
the magnetization process, and why does such a disparity exist in the
first place?
Let us begin with the first question. Suppose that there are only three
dipoles, which were originally in complete alignment with each other
at the instant t=0 (a).
A static field B
0
in the z-direction is applied, after which we observe
the subsequent precessional motion of the dipoles.
Since
2
<<
1
, the first thing to take place is that m0 (b). The
phases between the individual moments have been quickly
reshuffled to yield a vanishing transverse magnetization. For this
reason, time
2
is referred to as the dephasing time
After the dephasing, the moment begin to spiral toward the direction
of field, resulting in an increased magnetization in that direction (c)
after a time
1
.
Energy dissipation
The essential reason for the disparity in the magnitudes of

2
and
1
is that longitudinal relaxation process involves a
dissipation of energy, whereas transverse does not.
Note that transverse relaxation requires no tilt toward B
0
,
and hence no energy exchange.
When the moments try to tilt toward B
0
they do so only if
they can release some of their energy and the faster they
can do this the shorter is
1
.
However, conditions for exchanging large amounts of
energy are rather stringent in magnetic interactions. It is
this difficulty in disposing of the energy quickly that makes
the moments magnetize slowly.
Re()= ,
Im()=
EPR experiment
To rf supply and circuit for
measuring inductance and
losses Expt
l
. setup for EPR measurement
Field few kilogauss,
o
W
range,
For ease Fix , sweep B, then
repeat for +
Re() measured from change in
inductive current and Im() from
energy absorption
Used as principal tool
to analyze matter
If
1
is more,
population inversion possible
(W) MASER is converse of EPR
Spin Resonance Spectrometer
Nuclear magnetic resonance
Nuclear analog of EPR, involves nuclear
spin
Nucleon spin is vector sum of constituents
Nucleon spin quantum number
Application of magnetic field causes
precessional motion, leading to NMR
Q.Mech views NMR as transition between
Nuclear Zeeman sublevels with E = g
n

n
B
o

When external ac field is applied aborption
occurs when = E (resonance condition)
Transition between nonadjacent is forbidden
by selection rules
I = I
I g I
2
n
| =
|
|
.
|

\
|
=
p
n n
M
e
g
g
n
is nuclear g-factor 1
M
p
proton mass

n
nuclear Bohr magneton
2000 times smaller to
o
p
n
o
B
M
e g
|
|
.
|

\
|
= e = e
2
1-10 MHz for
B
o
=0.5Wb/m
2
Comfortable
range, accuracy
Expt
l
setup similar to EPR, measurements made from
absorption spectra
Used to study microscopic constitution of matter,
nuclear data, relaxation times, interaction of nucleus
with matter in solids.
Cooling by Adiabatic demagnetization
To create temperatures 0.01
o
K
Placed sample in field of 1Wb/m
2

Spins align losing energy to lattice, => sample heats up
Cool sample by liquid He bath.
Isolate sample and remove applied field.
Dipoles randomize absorbing energy from lattice
Sample cools down (adiabatic demagnetization)
Nuclear Magnetic Resonance
When the nuclear magnetic moment associated with a
nuclear spin is placed in an external magnetic field, the
different spin states are given different magnetic
potential energies. In the presence of the static magnetic
field which produces a small amount of spin polarization,
a radio frequency signal of the proper frequency can
induce a transition between spin states. This "spin flip"
places some of the spins in their higher energy state. If
the radio frequency signal is then switched off, the
relaxation of the spins back to the lower state produces a
measurable amount of RF signal at the resonant
frequency associated with the spin flip. This process is
called Nuclear Magnetic Resonance (NMR).
The precession of the proton spin in the magnetic field is
the interaction which is used in proton NMR. As a
practical technique, a sample containing protons
(hydrogen nuclei) is placed in a strong magnetic field to
produce partial polarization of the protons. A strong RF
field is also imposed on the sample to excite some of the
nuclear spins into their higher energy state. When this
strong RF signal is switched off, the spins tend to return
to their lower state, producing a small amount of
radiation at the Larmor frequency associated with that
field. The emission of radiation is associated with the
"spin relaxation" of the protons from their excited state. It
induces a radio frequency signal in a detector coil which
is amplified to display the NMR signal.
Since the Larmor frequency of
the detected signal is
proportional to the applied
magnetic field, changing the
magnitude of that field
produces a different detected
frequency. Placing a magnetic
field gradient across a sample
allows you to locate the source
of the proton NMR signal in the
sample. This is used to great
advantage in the medical
imaging process known as
Magnetic Resonance Imaging.