Solid State Physics

FREE ELECTRON MODEL

A.H. Harker
Physics and Astronomy
UCL
6 The Free Electron Model
6.1 Basic Assumptions
In the free electron model, we assume that the valence electrons can
be treated as free, or at least moving in a region constant potential,
and non-interacting. We’ll examine the assumption of a constant po-
tential first, and try to justify the neglect of interactions later.

2
6.1.1 Constant Potential
Imagine stripping the valence electrons from the atoms, and arrang-
ing the resulting ion cores on the aatomic positions in the crystal.

Resulting potential – periodic array of Coulombic attractions.

3
From atomic theory, we are used to the idea that different electronic
functions must be orthogonal to each other (remember we used this
idea in discussing the short-range repulsive part of interatomic po-
tentials) i.e. if ψc(r) is a core function and ψv (r) is a valence function
Z
ψc(r)ψv (r)dr = 0.

Let’s see how orthogonality might be achieved for a slowly-varying

4
To achieve orthogonality:

we need high spatial frequency (large k) components in the wave.

and the constant potential assumption is a reasonable one. The net

6
7
So finally we assume that the attractive potential of the ion cores can
be represented by a flat-bottomed potential.

We go further, and assume that the potential is deep enough that we

8
6.1.2 Free Electron Fermi Gas
For the particle in a box with potential V
~2 2
− ∇ ψ + Vψ = E 0ψ,
2m
or, with a shift of origin for energy, E 0 − V → E,
~2 2
− ∇ ψ = Eψ,
2m
so that the wavefunctions have the form
1
ψk(r) = √ exp(ik.r),
V
where V is the volume of the material. These are travelling waves,
with energies
~2 k 2
Ek = ,
2m
dependent only on k = |k|. Note that Ek depends only on the magni-
tude of k, not its direction.

9
So we can use the result from section 4.4 (Lecture 10) that the number
of states with modulus of wavevector between k and k + dk is
V V
g(k)dk = 3 4πk dk = 2 k 2dk.
2
8π 2π
For electrons
2 2 ~ √
r
dE ~ k ~ 2mE
= = 2
=√ 2E.
dk m m ~ m
We also need to include a factor of 2 for spin up and spin down.
dk
g(E) = 2g(k)
dE
V 2 dk
= 2 2 k
2π dE √
V 2mE m
= 2 2 2
√
2π ~ ~ 2E
Vm √
= 2 3 2mE.
π ~
Note that as V increases, so does the density of states.

10
6.1.3 The Fermi Energy
Remember that the Fermi distribution function n(E)
1
n(E) =
exp((E − EF )/kBT ) + 1
at absolute zero is 1 up to the Fermi energy EF. Suppose the volume
V contains Ne electrons. ThenZ we know
∞
Ne = g(E)n(E)dE
Z0 E
F
= g(E)dE
0√
3 Z E √
V 2m F
= 2 3
E dE
π√ ~ 0
V 2m3 2EF3/2
=
π 2 ~3 3
so !2/3
~2 3π 2Ne
EF = .
2m V
11
We can define two related quantities:
• Fermi temperature, TF,
TF = EF/kB.
• Fermi wavevector, kF, the magnitude of the wavevector correspond-
ing to kF,
~2 kF 2
EF =
2m
so !1/3
3π 2Ne
kF = .
V

12
6.1.4 Orders of magnitude
For a typical solid, the interatomic spacing is about 2.5 × 10−10 m.
Assume each atom is in a cube with that dimension, and releases one
valence electron, giving an electron density Ne/V ≈ 6 × 1028 m−3.
Putting in the numbers, we find
• EF ≈ 9 × 10−19 J = 6 eV;
• TF ≈ 70, 000 K;
• kF ≈ 1.2×1010 m−1, comparable with the reciprocal lattice spacing
2.5 × 1010 m−1;
We can also estimate the electron velocity at the Fermi energy:
~kF
vF = ≈ 1.4 × 106m s−1,
m
which is fast, but not relativistic.
Z E
F 3
total energy of electrons = E g(E) dE = NeEF,
0 5
so that the average energy per electron is 35 EF.
13
6.1.5 The Fermi surface
In later sections we shall talk a good deal about the Fermi surface.
This is an constant-energy surface in reciprocal space (k-space) with
energy corresponding to the Fermi energy. For the free electron gas,
this is a sphere of radius kF.

14
6.2 Some simple properties of the free electron gas
6.2.1 Thermionic emission
If the work function φ is small enough, if the material is heated the
electrons may acquire enough thermal energy to escape the metal. A
small electric field is used to draw them away.

15
The current  
φ
J = BT 2 exp − ,
kB T
with a theoretical value
emkB2 6 A m−2 K−2.
B= = 1.2 × 10
π~2
Experimentally the exponential dependence is confirmed, with simi-
lar values for B.

16
6.2.2 Field emission
A large applied field alters the potential outside the metal enough to
allow electrons to tunnel out.

17
Very large fields are needed, but a sharp metal tip can give an image
which shows where the atoms are – fields vary across the atoms.

More detail from newer scanning probe microscopes.

18
6.2.3 Photoemission
A photon with energy greater than the work function can eject an
electron from the metal.

19
6.2.4 X-ray emission (Auger spectroscopy)
A high-energy electron incident on a metal may knock out an elec-
tron from a core state (almost unchanged from the atomic state). An
alectron from the band can fall into the empty state, emitting an x-
ray.

20
21
6.2.5 Contact potential
If two metals with different Fermi energies are brought into contact,
electrons will move so as to equalize the Fermi levels. As a result, one
becomes positively charged and the other negatively charged, creat-
ing a potential difference which prevents further electron flow.

22
Summary
• Justification for neglecting details of crystal potential.
• Fermi energy – several eV
• Fermi surface – comparable with size of Brillouin zone
• Simple properties

23