PYL 102

Monday, Sept. 23, 2024

Sommerfeld model
 Basic assumption of Drude model
1. Between collisions electrons move in straight line - effect of
electron-electron interaction is ignored (reasonably valid) – effect
of electron-ion in ignored (completely invalid).
2. Mean free time between collisions is  (probability of collision per
unit time is 1/; probability of having a collision in infinitesimal time
interval dt is dt/
3. Relaxation time is independent of position and velocity of electron.

4. Electrons achieve thermal equilibrium by collisions with lattice –

they emerge after collision at a random direction with speed
appropriate to the temperature of the region where collision
happened – the hotter the region; the higher the speed of the
emerging electrons.
Note: electrons (unlike gas molecules) are charged and are moving in the background of other
charged entities; electron densities also large (~1028/cm3 ); still Drude applied kinetic theory (valid for
neutral dilute classical gas) to metals.
Ohm’s Law

Collisions between electrons and

atoms in a conductor determines
the conductivity
 Electronic heat capacity
The Drude model predicts the electronic heat capacity to be the classical
“equipartition of energy” Cel = 3/2 nkB This is independent of temperature.

Experimentally, the low-temperature heat capacity of metals follows the

relationship CV = T + AT3 . The second term is obviously the phonon
(Debye) component, leading us to suspect that Cel = γT. Even at room
temperature, the electronic component of the heat capacity of metals is
much smaller than the Drude prediction. This is a severe failing of the
model.
Sommerfeld Model
1. Treats electrons using Fermi Dirac statistics.
2. Recognizes that their energies are discrete – treats them like a particle in a
box of constant energy.
3. Uses Pauli principle to distribute them in the available energy states.
 Improvement to the Drude Model

• Sommerfeld recognized we needed to utilize Pauli’s

exclusion principle. Electrons cannot all be in the
lowest energy state, since this would violate the
Pauli Principle.
 Sommerfeld still assumes the
free electron approximation

U(r)

U(r)

Neglect periodic potential & scattering

Reasonable for “simple metals” (Alkali Li,Na,K,Cs,Rb)

 Ground state of ideal electron gas
• These conduction electrons can be considered as moving independently in a
square well and the edges of well corresponds to the edges of the sample.
(ignores periodic potential from atoms)
• Electron confined in a cube of sides L, potential inside the cube is constant
(take it to be zero) – potential at boundaries. Assume non-interacting
electrons Ψ(𝑟1, 𝑟2,… 𝑟𝑁 )= Ψ 𝑟1 Ψ 𝑟2 … … . . Ψ(𝑟𝑁 )

 2 d 2 
U
− 2
+ U ( x) = 
 2m dx 
 is the total energy of the system. The square of the wave
function ( *) is equal to the probability of finding the
0 L particle at a particular region of space.

Cube V=L3 Electrons in crystals are moving through atomic potentials.

Sommerfield/Drude models ignore those; use flat average.
 U

 1 ik •r
0 L Eigenstates  (r ) = e
V
Possible Boundary conditions
 2k 2
1. Ψ(0)=0 and Ψ(L)=0 with eigenvalues  =
2. Ψ(x,y,z)= Ψ(x+L,y,z) 2m
Dispersion relation for e’s

n nx n y n z
where k= in 3D: kx = , ky = , kz =
L L L L
Where nx, ny and nz are integers
  
=1= e =e
ik x L ik y L ik z L
e
 For large N the filled states form a sphere in k-space – its radius is kF (this is called the Fermi wave-
4
vector) and volume 3 𝜋𝑘𝐹3 . This is the Fermi sphere. At T = 0

4 1 V  2m  2 3 22
𝑁 = 2 × 𝜋𝑘𝐹3 × N= 2  2 
EF
3 2𝜋 3 2   3
𝐿 2
 2N
2 3
EF =  3
2m  V 

ℏ2 𝑘 2
Total energy of the electronic system is E= 2× σ𝑘<𝑘𝐹
2𝑚

For large N; the values of k are arbitrarily close to each other – can treat as continuum:

𝑉 ℏ2 𝑘 2 𝑑𝑘 1 ℏ2 𝑘𝐹 5
E= 2 3‫׬‬ =
8𝜋 2𝑚 10 𝜋2 𝑚
𝐸 3
Average energy per particle is = 𝜀 In contrast to a classical gas, the degenerate quantum
𝑁 5 𝐹 mechanical electron gas has appreciable ground-state energy.
 4k 2 dkL3 k 2 dkL3
g (k )3 D dk = 2  =
8 3
2
2 is introduced for spin up and spin down conditions for a given k value.

−1 / 2
2mE  2mE  m
k=  dk =  2  dE
2    2
−1 / 2 3/ 2
2mE  2mE  m 1  2m 
Ng(E dE == 2 2  2 
( E )F3)DdE dE =   E 1/ 2 dE
     2 2 2   2 
The electronic density of states at E ≈ EF As only the electrons within ∼ kT of EF are able to take part in thermal
processes, only the density of electron states at the Fermi energy, g(EF), will be particularly important in many
calculations. ℏ2 2/3
In the free-electron approximation, the Fermi energy is given by 𝐸𝐹 = 2
3𝜋 𝑛
2𝑚
g(EF) ≡ dn/dEF ln(EF) = 2/3 ln(n) + constant

g(EF) = 3/2 (n/EF)

Fermi Energy for Sodium

Density=0.971 g/cc
Atomic Weight=22.99
N 6.02  10  0.971
23
=  10 6

V 22.99
= 2.543  10 28

ℏ2 2/3
Substituting we get 𝐸𝐹 =
2𝑚
2
3𝜋 𝑛

EF=3.16 eV
 Fermi-Dirac Distribution Function
Becomes a step function at T= 0.
f ( , T ) =
1
Low E: f ~ 1.
e  ( −  ) + 1 High E: f ~ 0.

 = chemical potential gives the probability of

occupation of a state of energy E
(T=0)=F Fermi energy

Right at the Fermi level: f = 1/2.

Turning to the finite T figure, f only varies significantly within kBT of µ. This has two implications.
1. As kT << EF, this implies that µ ≈ EF
2. Only electrons with energies within kT of µ, i.e. EF, will be able to contribute to thermal processes, transport etc..
Electrons further below µ will be unable to acquire sufficient thermal energy to be excited into empty states; states
more than ∼ kT above µ will be empty
 Electronic Heat Capacity
Width of shaded region ~ kT
Room temp T ~ 300K, TF ~ 104 K
→ Small width
→ Few electrons thermally excited
 0.01% @ room temp

How many electrons are excited thermally?

Shaded area  triangle. Area = (base)(height)/2
Number of excited electrons:  (g(F)/2)(kT)/2  g(F)(kT)/4
Excitation energy  kT (thermal)
2
Total thermal energy in electrons U(T): 𝑈 𝑇 𝛼 𝑘𝑇

𝑑𝑈
Electronic heat capacity 𝐶𝑒𝑙 =
𝑑𝑇

C~T
 -

1. Cel is proportional to T, as are experimental data

2. Cel is a factor ∼ kT/EF smaller than the classical (Drude) value, as are experimental data.