0% found this document useful (0 votes)
61 views32 pages

3400 - Lecture Notes - 2

Uploaded by

mohamedkaram
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
61 views32 pages

3400 - Lecture Notes - 2

Uploaded by

mohamedkaram
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 32

3.

LATTICE VIBRATIONS
• Atoms in lattice are not stationary even at T =
0K.
• They vibrate about particular equilibrium
positions at T = 0K ( zero-point energy).
• For T > 0K, vibration amplitude increases as
atoms gain thermal energy.

3.1 Sound Waves


• Sound waves travel through solids with typical
speeds ~ 5(km/s).
• At low frequencies (f < 1THz), λ~50Ao, one can
treat solid as continuum, ie. ignore discrete
nature.
At any frequency & in a given direction in
crystal, 3 sound waves possible:
One longitudinal and two transverse.

3.2 Continuum Model


Propagation of an elastic wave in a sample of long bar
follows the dispersion relation :
ω = vs q
Where vs is the wave velocity (sound velocity) which
equal in this case to
vs = Y / ρ
Y & ρ are Young’s modulus and density of the sample.
3.3 Discrete Lattice
As the wavelength decreases and q increases, the
atoms begin to scatter the wave, and hence to
decrease its velocity.
Thus, one expects the dispersion relation to be as
shown
ω Continuum

Discrete

0 q

A- Monatomic Chain
• Consider 1D chain of identical atoms each of mass
M, where atoms joined to nearest neighbors by
'spring' of force constant α, and spacing a apart.

• Look for wavelike solution:

u n = Ae i ( qX n −ω t )
Where Xn =na is the equilibrium position .
Substitute into eq. of motion get:


ω = 2
sin 2 12 qa
M
Hence, the Dispersion relation , which Determines how
wave vector q is related to ω,in 1D is :

ω = ω m sin 12 qa

Where ωm =(4α/M)1/2 .

• Note:

∗ ω is periodic in q & all values of ω contained in range:


-π/a < q < π/a.
* Range of q related to 1st Brillouin zone. Values of q
outside range have no physical meaning.
* N does not appear in dispersion relation, ie. equations
of motion of all atoms lead to the same algebraic
relation between ω and q.

1. Long wavelength limit:


As q →0, sinθ =θ, and then

ωma
ω= q
2
Or ωma
vs =
2
Hence,

α = Ya

2. short wavelength limit:


As q increases , the dispersion relation begins to
deviate from the straight line. It will saturate at q=π/a
with maximum frequency


ωm =
M

3. Phase & Group velosity :

ω
vp = defines the velocity of a pure wave with
q frequency ω and wave vector q.

∂ω
vg = defines the velocity of a wave pulse with
∂q average frequency ω and average wave vector q.

Thus, at q →0 , phase velocity = group velocity.


defines velocity of sound, which can be identified with
the appropriate Young’s modulus.
And, at q=π/a , group velocity is zero, ie. standing
waves.
B- Diatomic Chain
• Consider 1D chain of two different atoms of
masses M1 and M2, where the distance between two
neighbouring atoms is a.

Dispersion relation ,that determines how wave vector q


is related to ω, is

1 1 1 1 2 4 sin 2 (qa)
ω = α(
2
+ ) ±α ( + ) −
M1 M 2 M1 M 2 M 1M 2
3.5. Lattice Vibrations in 3D
• If we Extend properties of 1D chain to 3D
crystals:
1- For crystal lattice with unit cell containing only one
atom,
* 3 dispersion relations, one in each direction of
vibration.
* All three branches are acoustic.

2- Crystals with unit cell containing 2 different


atoms,
* 3 acoustic branches & 3 optical branches.
3- Generalize to crystals with s different atoms in unit
cell.
* 3 acoustic branches & 3(s - 1) optical
branches.

3.6 Phonons
• Define:
Phonons: quantum of vibrational (elastic) energy.
photons: quantum of electromagnetic energy.

* The energy of phonon is:


ε = hω
With a momentum equal to:

p = hq
* The number of phonons is given by Bose-Einstein
distribution:
1
n=
e hω / KT − 1
This number depends on temperature , at T=0 ⇒n=0.
At very high temperature,

KT
n =

Fact:
The number of phonons in a system is not conserved.

Interaction of phonons with other excitation (or with


themselves) governed by conservation laws.
3.6. Density of States
ƒ In calculating physical properties of solids, the no.
of modes in a given frequency (or energy, or q-
space) range is required.
1D density of States
ƒ A 'better' way of obtaining this is to apply the
periodic boundary condition:
u(x=L) = u(x=0)
eiqL =1
qL = n2π, ie. q = n2π/L

Thus, 'length' occupied by one mode in q-space is


2π/L . Hence the No. of modes in the range dq
(i.e. the density of states in q-space) is:
L/2π dq

ƒ To obtain density of states in frequency space,


need dispersion relation. Since there is the same
No. of states in g(q)dq & g(ω)dω, ie.
L
g (ω )d (ω ) = g (q)d (q ) = dq

L dq
or g (ω ) =
2π dω

Since the modes lying in( –ve) q-region must be included


as well, then:
L 1
g (ω ) = (1)
π dω / dq
Note the relation between the density of states
and the group velocity.
* At long wavelength limit:

ω = vsq, ie. dω = vsdq


and hence,
L 1
g (ω ) = (2)
π vs

* At Dispersive region:

ω = ω m sin 12 qa
or 2 ω
q= sin −1 ( )
a ωm

Differentiating:
dq 2 1
= ( 2 ) -1/2

dω a ω m − ω 2
Hence density of states in ω-space is:
2L 1
g (ω ) = ( 2 ) -1/2
(3)
aπ ω m − ω 2
3D density of States
Volume occupied by one mode in q-space is:
(2π/L)3 = (8π3)/V
Thus, No. of modes contained in a small spherical shell
in q-space is: 4πq2dq / 8π3/V

Hence, density of states in ω-space is

V 2 dq
g (ω ) = 4πq
8π 3 dω

V 2 dq
or g (ω ) = q (4)
2π 2

This is the density of state if we assume that each
value of q associated with one single mode. But, in fact
for each q there are 3 modes (1 longitudinal & 2
transverse ), then

3V 2 dq
g (ω ) = q
2π 2
dω (5)
For non-dispersive limit;
3V ω 2 1
g (ω ) = ( )
2π vs vs
2

or, 3V ω 2
g (ω ) = 2 3
2π vs
3.7. Specific Heat of Lattice
• General definition of specific heat is;

ΔQ
C=
ΔT
Where ΔQ is the heat required to increase the
temperature of one mole by ΔT.

• At constant volume all internal energy E converts to


heat, i.e. ΔQ= ΔE , then one can define the specific
heat at constant volume as;

⎛ ∂E ⎞
Cv = ⎜ ⎟ (1)
⎝ ∂T ⎠ v

Cv
Experimentally, it was found that
there is a relation between Cv & T
as is shown in the figure.
T
Theoretically, three different approaches were used to
determine Cv:
A- Classical Theory:
• Vibration of lattice may be treated as that of a
harmonic oscillator.
• Thus, the average energy for 1D oscillator is

ε = kT
Therefore, the total thermal energy per mole in 3D is
E = 3N AkT = 3RT
where NA & R are Avogadro’s no. and the universal gas
constant respectively. Substituting into (1), we find

C v = 3R (2)
This relation predicts a constant value to Cv whatever
the temperature is. Although it is in agreement with
experiment at high T, it fails totally at low T.

B- Einstein model:

* Considering the atoms as independent oscillators,


vibrating with a common frequency.
* Treating the oscillator quantum mechanically,
where the average energy for 1D oscillator is


ε =
e hω / kT − 1

• At high T , ε = kT (and the classic value is reached)

• At low T ,total energy is;

hω E
E = 3N
e hωE / kT − 1

Where the common frequency of the oscillators, ωE , is


known as Einstein frequency.
• Substituting into (1), we find
⎛ hω E ⎞ e hω / kT
2

C v = 3R⎜ ⎟
(
⎝ kT ⎠ e hω E / kT − 1
2
)
Introducing Einstein temperature, θ E = hω E / k , the
above expression can be reduced to:

⎛θE ⎞ eθ E / T
2

C v = 3R⎜ ⎟ (3)
(
⎝ T ⎠ eθ E / T − 1
2
)
This formula is in a good agreement with the
experimental results as it shown in the following figure:

• At very low temperature, T<< θE , and hence,

⎛θ ⎞
2

C v = 3R⎜ E ⎟ e −θ E / T
⎝T ⎠

Which indicates that Cv approaches zero exponentially.


But careful measurements show that it does that
slower.
Although the fact that (Cv → 0 as T → 0 ) agrees with
experiments, the manner in which this is approached is
not.

C- Debye model:

* Treating the atoms not as independent oscillators,


but as coupled oscillators vibrating collectively as sound
wave follows the dispersion relation:
ω = vsq

• Total energy in such system is ,


E = ∫ ε (ω ) g (ω ) dω
where g(ω) is the density of state, and hence the
integration is taken over all allowed frequencies.
Or;
3V hω
2 3 ∫
E ω 2
hω / kT

2π v s e −1

Debye stated the integral limit : (from 0 to ωD), where


ωD is Debye frequency wich given by

ω D = v s (6π 2 n )
1/ 3

Specific heat Cv then can be found to be,


3θ /T
⎛T ⎞ D
x 4e x
C v = 9 R⎜⎜
⎝θD
⎟⎟

∫ (e
0
x
−1 )
2
dx

With x = hω / kT and θ D = hω D / k

Here θD is Debye temperature.

At very low temperature, T<< θD , and hence

3
12π 4 ⎛ T ⎞
Cv = R⎜⎜ ⎟⎟
5 ⎝θD ⎠
4. FREE ELECTRON THEORY OF METALS
Classical theory : Drude & Lorentz (based on free-
electron model) gave explanations for conductivity and
optical properties in metal.
Quantum theory: Summerfield included Pauli's
exclusion principle into classical theory.

4.1 Conduction electrons


• A valence electron in a free atom, becomes a
conduction electron in a solid.
• The core electrons do not contribute anything to
the electric current.
• Hence,

No. of conduction electrons = No. of atoms x atomic valence

ρm N A
N = Zv
M′

Where M`is the molar mass of the element (gm/mole).


4.2 The Free-Electron Gas
• Assumptions:
a. Electrons free to move in a background of
uniform (+ve) charge provided by ions (jellium
model), ie. electrons behave like a gas.
b. No interactions between electrons, because of:

1- Pauli exclusion.
2- To minimize the system energy, electrons tend to
stay away of each other.

The two above reasons lead to the hole concept:

Each electron is surrounded by a spherical region which


is lacking of other electrons Known as a hole.

4.3 Electrical Conductivity

• Apply external electric field (ε) on a wire has a


length L and cross section A, we find:

I V Lρ
J= ε= and R=
A L A
Where J is the current density, and ρ is the electrical
resistivity =1/σ (σ is the electrical conductivity).

• According to Ohm’s Law:

J = σε (1)

• When external electric field is applied two main


things happen:
1. Coulomb force causes electrons to accelerate, ie. ,
where vd is the drift velocity .
2. Scattering (with other electrons, ions, impurities
etc) produces a decay of vd back to zero. This
retardation is represented by: -vd/τ, where τ is the
relaxation time.
Thus, adding up total force on electron get:

dvd v
m∗ = −eε − m∗ d
dt τ
In steady state, dv/dt=0, and hence


vd = − ∗
ε
m

But since J represents the amount of charge crossing a


unit area per unit time, then

Ne 2τ (2)
J = (− Ne)vd = ∗
ε
m

Comparing (1) with (2) leads to;

Ne 2τ
σ= ∗
m

Since τ is the time between two following collisions, it


can be expressed as;
l
τ= ( ~10-14 s in metals )
vr

Where l is distance between two following collisions


(electron’s mean free path) and vr is the random velocity.

Using these terms (σ) can be rewritten as

Ne 2l
σ= ∗
m vr

In metals :
N ~1029 m-3 & vr ~106 m.s-1 ⇒ σ ~ 107 Ω-1.m-1
In semiconductors:
N ~1020 m-3 & vr ~104 m.s-1 ⇒ σ ~ 1 Ω-1.m-1

The mean free path, l , for electrons in metals may be


estimated via: vrτ = l . This give l ~ 10-8 m ~ 102 Ao, which
is 20 times larger than the interatomic spacing.
i.e., electrons can move freely in a perfect periodic
structure.
scattering occurs when periodicity is interrupted by
phonons or impurities.
4.4 Temperature dependence of Electrical Resistivity
In metals, the temperature dependence of ρ is
determined by τ, according to the relation:

m∗ 1
ρ= 2
Ne τ

But total scattering rate τ given by addition of two


individual scattering rates:
1 1 1
= +
τ τ ph τ i
Which known as Matthiesens' Rule.
Therefore;

m∗ 1 m∗ 1
ρ = ρi + ρ ph = + 2
Ne τ i Ne τ ph
2

Impurity Scattering ⇒ ρi , the residual resistivity, which


is independent of T.
It is determined by distance between impurities.
Electron-phonon scattering⇒ ρph , the ideal resistivity,
which is dependent of T.
* At very low temperature, scattering by phonons is
negligible. So τph ⇒∞ & ρph ⇒0, and hence ρ = ρi
* At high temperature, no. of phonons increases as T.
Hence scattering increases as T .
τ proportional to 1/T (true for T > ¼θD).
4.5. Heat Capacity of Conduction Electrons
Based on Drude-Lorentz model, the average
energy of electrons per mole is :
E = 32 N AkT = 32 RT
Hence, the electrons heat capacity is
Ce = 32 R
The total heat capacity in metals should then be;
C = C ph + Ce
This means that at high temperature
C=4.5 R ~ 9 cal/mole K
But, experiments show that C~3R , and Ce is
smaller than the classical value by a factor of
10-2.

Quantum explanation
* Pauli's exclusion principle implies that only 2
electrons of opposite spin orientation are allowed
in each state.
* At T = 0K, electrons occupy all energy levels
from the ground state upward. The energy of
the highest filled state is known as the Fermi
energy.
* At T = 0K, electrons distribution function is

f(E)= { 0,
1,
E > EF
E<E
F
Thermal Distribution
* For T > 0, electrons with energy ~ EF can gain
thermal energy and occupy energy states >EF.
* In many cases, it is the electrons near EF that
dominate the properties of the metal.
* For T > 0, electrons distribution function is

1
f (E) =
e ( E - E F ) / KT + 1
i.e. Fermi-Dirac distribution.
* Using this distribution function, one can find
energy of electron gas, hence the heat capacity
of the electrons.
ƒ electron per mole as
N A ( kT / E F )

ƒ Which leads to an approximate thermal energy

N A ( kT )2
E=
EF

ƒ Hence, the electrons heat capacity is,


kT
Ce = 2 R
EF
T
Or
Ce = 2 R
TF

With EF = kTF, where TF is Fermi temperture.


* The exact value of electrons heat capacity is
then,
π 2 kT
Ce = R
2 EF
4.6. Fermi Surface
Because of the random motion, the energy of an
electron is mainly kinetic. Hence:
E = 12 m∗vr2

• In the velocity space, all conduction electrons


confined in a sphere of radius vF which is called
Fermi velocity. Energy at the surface of this
sphere is; vy
• Fermi Surface

E F = 12 m∗vF2 v
vx
vF
vF is very large ~ 10 m.s .
6 -1

• Fermi surface (FS) is temperature


Independent, and so the Fermi velocity.

• The value of EF is determined by the electron


concentration N according to :

EF =
h2
2m ∗
(3π 2 N )
2/3
~ 5 eV

4.7. Electrical Conductivity (with respect to FS):

Ne 2lF
σ= ∗
m vF
Although the formula of σ
is the same as the classical
case, the actual picture is
different.

Lorntz Model Summerfield model


No. of electrons
responsible about N N(vd /vF)
conductivity
Velocity of those vd vF
electrons

4.8 Hall Effect in Metals

• Applying a magnetic field B normally on a wire


carries a current of density Jx will produce an
electric field normal to both Jx and B known as
Hall field εH.
• In steady state;
Force of Hall field = Lorentz force
-e εH = -e vx B
ε H = vx B
Or,

1
εH = − JxB
Ne
Since Jx = N(-e)vx ,where N is the electron density.
The proportionality constant, i.e. ε H / J x B ,is known
as Hall constant RH, therefore,

1
RH = −
Ne

4.9 Failure of Free-Electron Model


1- The model suggests that electrical conductivity
is proportional to electron concentration, but it
is found that the divalent metals (Zn, Cd,..) and
even trivalent metals (Al, In) are less
conductive than the monovalent metals such as
(Cu,Na,Ag,..) .
2- Some metals show positive Hall constant (Zn,
Cd,..).
3- Fermi surface was found to be often non-
spherical in shape.
5.1. Energy Bands in Solids

* Solving Schrodiger equation for a free atom gives a


series of discrete energy levels.

* When two atoms come together to form a


molecule, each of the atomic levels will split into two
closely spaced levels.

* The amount of splitting depends strongly on:


1- interatomic distance; the closer the atoms,
the greater the splitting.
2- atomic orbital; the higher the energy, the
greater the splitting.
This is an eigenvalue equation which leads to many
solutions. Therefore, for each value of k there are a
large number of solutions (i.e. a set of discrete
energies E1,k, E2,k, ….).

* Each band covers a certain range of energy,


extending from the lowest to the highest value it
takes when plotted in k-space.
* Within each band the electron states can be
classified according to their momentum (p=ħk).

3-The crystal potential

V(r) is the crystal potential “seen” by the electron.


The most important property of this potential is its
periodicity. This potential is composed of two parts;
ionic & electronic:

V (r ) = Vi (r ) + Ve (r ) (4)
* When N atoms come together to form a solid,
each of the atomic levels will split into N very closely
spaced sublevels; i.e. energy band.

* The energy spectrum in a solid is composed of a


set energy bands. Regions separating these bands
are energy gaps, i.e. regions of forbidden energy.

• The wave function describing the electronic


states in a band, known as the crystal orbital,
can extend throughout the solid, unlike the
atomic orbital which is localized.

5.2. The Bloch Theorem

The wavelike properties of an electron in a


crystalline solid can be described using the following
form of Schrodinger equation:

⎡ h2 2 ⎤ (1)

⎢ 2m ∇ + V ( r ) ⎥ψ (r ) = Eψ (r )
⎣ ⎦

Where V(r) is the crystal potential and ψ(r) & E are


the state function and energy of this electron.
1-The Bloch Function
The solution of (1) for a periodic potential is given
by
ψ (r ) = f (r )u (r ) (2)

Where both the function u(r) and the probability


density ψ(r) 2 must have the same translational
symmetry as the lattice , i.e,

u (r + R) = u (r )
and
f (r + R) = f (r )
2 2

The only function that satisfies this is eikr, hence (2)


can be written as
ψ k (r ) = eik .r uk (r ) (3)

Which known as Bloch function.


Note:
*This state function has the form of a traveling
plane wave so it propagates through the crystal like a
free particle.
*This function represents the crystal orbital of the
solid.
2-The Energy Bands
Substituting from (3) into (1) gives
⎡ h2 ⎤
⎢− (∇ + ik ) + V (r )⎥uk (r ) = Ek uk (r )
2

⎣ 2m ⎦
The first term represents the interaction with the
ion cores, and given by

Vi (r ) = ∑ vi (r − R j )
j

Since electron-electron interaction is very weak and


has some difficulties to be calculated, the second
term in (4) can be dropped out. The major effect of
this interaction is to screen the ions from other
electrons and hence to weaken the elecrton-ion
interaction.
Thus,

V (r ) = Vi (r ) = ∑ vs (r − R j )
j

Where vs(r-Rj) is the potential of the screened ion


located at Rj.

5.3. Number of States in the Band


Since each Bloch function ψ(r)n,k defines a specific
electron state, let see how many states there are in
each band.
For 1D case: ψ k ( x) = eik . xuk ( x)
Applying the periodic boundary condition on this
function shows that the allowed number of k is ;

k =n
L
For the first zone, the number of states is
2π 2π L
( ) /( ) = = N
a L a

Where N is the number of unit cells. Hence, we end


with these two important facts:

1- In each band the number of states (orbitals)


inside the first zone is equal to the number of unit
cells in the crystal N.
2- In each band the maximum number of electrons is
equal to 2N .

You might also like