Lecture 14

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Lecture #14

Program:
1. Introduction
2. Magnetization and it’s relation to free energy
3. Measuring the magnetization – the force exerted by a spatially varying magnetic
field.
4. Atomic susceptibility
5. Key equation of atomic magnetization
6. Larmor Diamagnetism, susceptibility of insulators filled shells
7. Calculating the total angular momenta of atoms in ground states: Hund’s rules
8. Paramagnetism: susceptibility of insulators with partially filled shells
9. Curie’s Law: T – dependence of paramagnets
10. Susceptibly of metals: Pauli Paramagnetism

Handouts:
1. Payne et al Iterative minimization techniques for ab-initio total-energy
calculations.
2. Ground states of ions with partially filled f and d shells as constructed from
Hund’s rules (page 652 out of Solid State Physics Ashcroft and Mermin).

1
Introduction:
Today we will discuss the magnetic properties of materials, Fe is probably the most well
known magnetic material. Other materials exhibit appreciable magnetic properties such
as Ni, Co and some alloys such as Ni-Fe and Ni-Co. This type of magnetism is called
Ferromagnetism we will learn about it in the next lecture. It turns out that all substances
exhibit magnetic properties albeit these might be exceedingly weak.
Materials that exhibit this weak magnetic response can be classified into two: those that
are attracted by the magnetic field (like Aluminum are called paramagnetic) and those
that are repelled (these are called diamagnetic one example is Bismuth)

Magnetism is an entirely QM phenomenon strictly classical system at equilibrium will


not display any magnetic moment even in the presence of a magnetic field.
The magnetic moment of a free atom has three principle sources:
1. intrinsic spin angular momentum (nucleus and electronic 1/1000 ratio)
2. orbital angular momentum
3. field induced orbital angular momentum
The first two effects lead to paramagnetic behavior whereas the third leads to
diamagnetic contributions.

We will focus on atomic contributions (non-interacting particles) neglecting electron-


electron interactions these should allow us to explain the origin of diamagnetism and
paramagnetism. Then we will analyze the effects of interactions and in particular the
phenomenon of magnetic ordering which leads to ferromagnetism and
antiferromagnetism.

The Magnetization and it’s relation to the total energy

2
The definition of the magnetization density for a QM system of volume V at T=0 in a
uniform magnetic field is defined as:
r
r r 1 ∂E0 H ( )
( )
M H =−
V ∂H
r

r
( )
where E0 H is the ground state energy of the system in the presence of the field H

which is the field acting on the individual magnetic moments.


At temperatures other than 0 the system at thermal equilibrium will be in an excited state
(or a superposition of excited states) the average magnetization density is given by:
En

r ∑M e kBT

( )
n
M H ,T = n
En

∑e
n
k BT

r
r r 1 ∂En H ( )
( )
Mn H = −
V ∂H
r

Which can be seen as a weighted average of the magnetization corresponding to the


individual excited states.

In terms of a thermodynamic quantities one can define the partition function and the
corresponding Helmholtz free energy F:
F En
− −
e k BT
= ∑e k BT

En

r ∑M e k BT

1 ∂F
( )
n
M H ,T = n
=−
E
− n V ∂H
∑e n
k BT

Measuring the magnetization


The magnetization can be measured by recording the force exerted on a specimen by an
inhomogenous field that is varying slowly over the length of the sample, the
corresponding change in free energy would be:

3
∂F ∂H ∂H
F ( H ( x + dx ) ) − F ( H ( x ) ) = dx = −VM dx
∂H ∂x ∂x
f
→ =M
∂H
∂x

Definition of magnetic susceptibility


r
µ 0 ∂M 1 ∂2 F
χ= r =−
∂H V ∂H 2
χ < 0 → diamagnetic
χ > 0 → paramagnetic

Calculation of atomic susceptibility:

We would like to now explore the atomic origin of the magnetic susceptibility

1. As we saw in the previous lecture, in the presence of a uniform magnetic field the
Hamiltonian of an atom is modified in the following way:
1 ˆ
( ( )) ( )
2
Hˆ = P − qAˆ Rˆ + Vˆ Rˆ
2me
r
in particular the effects of a magnetic field H are incorporated via a vector potential
A related to the magnetic field by:
r r r
H =∇× A
The simplest and commonly referred to case is that of a uniform magnetic field which
can be shown to have a vector potential equal to
r 1r r
A= − r×H
2
r
In this case H is a constant and therefore all operators commute with it which allow
us to simplify the form of the Hamiltonian:
1
∑ i ( ( )) ( )
2
Hˆ = ˆ
P − ˆ Rˆ + Vˆ Rˆ
qA
2me i
r2
Pˆi 2 µ r q 2
H
=∑ + Vˆ − ∑ B
Lˆi ⋅ H + ∑ Rˆ⊥2i
i 2me
14243
i h i 8me
paramagnetic

2. The interaction of the field with the individual electron spin is given by:

4

Hˆ = − 0 B Sˆz H
1 h 3
424
magnetic moment

Sˆz = ∑ szi
i

h
szi = σ i
2

g 0 ; 2.0023

Combining both contributions leads to the identification of those terms in the


Hamiltonian which depend on the magnetic field:
µ r r q2 r 2
(
Hˆ magnetic = − B L + g 0 Sˆ ⋅ H +
h 8m
)
H ∑ ( Ri2 )
i

qh  Joule 
µB = ≈ −9.27 × 10−24 
2me  Tesla 

Since the relative magnitude of the magnetic energy levels are very small compared to
excitation energies as we discussed in the previous lecture we can compute the changes to
the atomic levels using a perturbation approach:
Hˆ = Hˆ natomic + Hˆ nmagnetic
1424 3
perturbation

where,
Hˆ natomic n = En n
The corrections to the energy eigenvalues are given by,
En = Enatomic + Enmagnetic
2
n Hˆ magnetic n′
Enmagnetic = n Hˆ magnetic n + ∑
n ≠ n′ Enatomic − Enatomic

The key equation of magnetism


By substituting the form of the magnetic Hamiltonian into the perturbation expression for
the magnetic energy we obtain:
µ r
( )
Enmagnetic = − B n Lˆ + g 0 Sˆ n ⋅ H
h
µB r r
( )
2

n L + g 0 Sˆ ⋅ H n′
h q2 2
+∑ + H n ∑ ( xi2 + yi2 ) n
n ≠ n′
atomic
En − En′
atomic
1444424444
8m i
3
 q2 
O  H 2 ma02 
m 
 

5
This equation is used as a basis for theories that explain the magnetic susceptibility of
atoms, ions and molecules regardless of whether they have filled or partially filled outer
shells. This equation is also used to describe solids which can be described as a collection
of individual ions such as ionic or molecular solids in such cases the susceptibility is
calculated ion by ion. The first term is the dominant term and is on the order of unity
which means that if you apply a filed of 10 4 gauss the correction is about 10 −4 eV .
The last term has is typically smaller than the first by approximately 5 orders of
magnitude

6
Larmor Diamagnetism, susceptibility of insulators with all shells filled
In an atom such as He, Ne, Ca which has all the electronic shells filled the ground state
will be of spin and orbital angular momentum of 0 (follows from the spherical
symmetry),
Jˆ 0 = Lˆ 0 = Sˆ 0 = 0
therefore only the last term in the equation for the energy shift will be non-zero.
q2 2
E0magnetic = H 0 ∑ ( xi2 + yi2 ) 0
8m i
The susceptibility of a solid composed of N such ions is given by:
N ∂ 2 E0magnetic q2 N
χ =− = − 0 ∑ ri 2 0
V ∂H 2
6m V i
Comments:
1. This is known as the Larmor diamagnetic susceptibility (or the Langevin
susceptibility) where diamagnetic refers to χ < 0 which means that the induced
moment is opposite to the field.
2. Alkali-halide salts such as NaCl exhibit Larmor diamagnetism where the
susceptibility of the solid can be taken to be the sum of the independent
susceptibility for the negative and positive ions.
3. Diamagnetic susceptibilities are on the order of 10−5 which means that the
magnetization is very small compared to the field.
4. The trend of diamagnetic susceptibility in the periodic table is:

Li + -0.7 He -1.9
Na + -6.1 Ne -7.2
K+ -14.6 Ar -19.7
Cs + -35.1 Kr -28

Atoms with partially filled shells:

Hund’s Rules
The Hund rules apply to the occupation of levels by electrons in a given shell at ground
state,
1. The total electron spin, S is maximized which means that S z = ∑ msi is
i
maximized consistent with pauli’s exclusion principle.
2. The maximum value of the orbital angular momentum L = ∑ mi is assumed,
i
consistent with S.
3. The value of the total angular momentum is
 L − S shell is less than half full
J =
 L + S shell is more than half full

7
QM Theory of paramagnetism
Electronic paramagnetism can be found in: atoms, molecules and lattice defects
containing an odd number of electrons (free sodium atoms, gaseous nitric acid, free
atoms and ions with partially filled inner shells such as transition metals, rare earth and
actinide elements.
A few compounds with even number of electrons including molecular oxygen and
metals.
One can distinguish between two cases:
1. J=0 the ground state is non-degenerate. For symmetry reasons associated with J=0
states the linear term in the energy shift is 0.

µ r
( )
E0magnetic = − B 0 Lˆ + g 0 Sˆ 0 ⋅ H
h
1444 424444 3
=0

µB r
( )
2

0 Lˆ + g 0 Sˆ ⋅ H n′
h q2 2
−∑ + H 0 ∑ ( xi2 + yi2 ) 0
n Enatomic − E0atomic 1444424444
8m i
3
 q2 
O  H 2 ma02 
m 
 
Larmor Diamagnetism

When the solid has N/V atoms per unit volume the susceptibility is given by
N ∂ 2 E0magnetic
χ =−
V ∂H 2
We have both Larmor diamagnetism and a term which goes like

µB
( )
2

0 Lˆz + g0 Sˆz n′
h
−2
Enatomic − E0atomic
which causes the magnetic moment to align with the field – Van Vleck Paramagnetism.

2. J is not equal to 0 the first term in the energy shift is dominant and the ground
state is 2J+1 degenerate in the zero field.
Wigner Eckart Theorem states that (valid for cases where kT<<energy splitting
between the ground state and first excited state resulting from the application of the
field):
Lˆ + g 0 Sˆ = g ( JLS ) Jˆ
under these conditions the contribution of the magnetic field to the Hamiltonian is
just:
µ r
Hˆ = − g ( J , L, S ) B Jˆ ⋅ H
14243 h
144244
Lande g factor
3
µˆ − magnetic moment

8
which means that the interaction energy is proportional to the total angular momentum J.

Magnetization of a set of identical ions of angular momentum J – Curie’s Law


Let us now consider a system where only the lowest 2J+1 states are thermally excited.
We define the magnetic moment of an atom or ion in free space:
µ r
µˆ = g ( JLS ) B J
h
The Hamiltonian of the system in a magnetic field can be written as:
r r m gµ
Hˆ = − µ ⋅ H = − J B H
h
where mJ is the azimuthal quantum number and has values − J , − J + 1....J
The free energy is given by (assuming that only the lowest 2J+1 levels are excited)
 1  1
β g µB H  J +  β g µB H  J + 
−  2  2
β g µ B HmJ
J
e h
−e h
e −β F
= ∑
mJ =− J
e h
=

β gµB H β g µB H
e h 2
−e h 2

Curies Law
The expression for the magnetization is
r N ∂F N g µB J  β gµB H 
M =− =− BJ  − 
V ∂H V h  h 
where the Brillouin function BJ ( x ) is given by ,
2J +1 2J + 1 1 x
BJ ( x ) = coth x− coth
2J 2J 2J 2J
in an atom with total angular momentum J there will be 2J+1 equally spaced energy
levels, the magnetization is given by,
NgJ µ B
M =− BJ ( x )
h
1 gJ µ B B
x=−
h k BT
2J +1 2J +1 1 x
BJ ( x ) = co h x− coth
2J 2J 2J 2J

9
Let’s examine the two temperature limits

1. Low temperature limit


gµB H
− >> 1
hk BT
As T approaches 0 for fixed H BJ → 1
N gµB
lim M = − J
T →0 V h
each atom is completely aligned by the field mJ = J
2. High temperature limit
gµ H 1 x x3
For − B = 1 → cot x = + − + ... and the susceptibility is given by,
hk BT x 3 45
∂M N J ( J + 1) g µ B C
2 2

χ= ≅ =
∂H V h 2 3k BT T
where C is known as the Curie constant

10
The inverse dependence of the susceptibility on the temperature is known as Curie’s Law,
it describes systems where the “permanent moments” are aligned by the field in the same
direction and are disordered by thermal energy.
χ paramagnetic = 10−2 − 10−3
χ diamagnetic = 10−5

Curies Law in Solids


N J ( J + 1) g µ B
2 2

χ=
V h 2 3k BT
This expression works well for crystalline rare earths (La, Nd, Ps, Er ect.) because their
partially filled f shell is not the outermost shell and does not interact strongly with the
crystal lattice. In transition metal ions Curie’s law is obeyed but one needs to calculate
the prefactor by setting J=S (L=0). This happens because the d shell is the outermost and
is therefore subject to substantial interactions with the lattice this leads to the so called
lattice splitting as opposed to spin orbit splitting which underlies Hund’s third rule. It is
necessary for transition metals to modify Hund’s third rule and approximately set J=S

11

You might also like