Now let's consider how the energy levels are arranged. If the
atoms had no interactions, e.g., if they were very far apart,
then the atomic energy levels would not be shifted but would
have a high degree of degeneracy, i.e., there would be N levels
at each of the atomic energy eigenvalues. Let's look at one energy where there
are N degenerate levels. As we turn on the interactions
between the atoms, e.g., by bringing the atoms closer together,
this degeneracy is lifted and the levels spread out
in energy. The maximum spread will be of the order of the interaction energy.
(This maximum spread in energy is called the bandwidth.)
Since there will be N levels packed into this range,
the separation between levels is quite small. As
,
the energy level separation goes to zero. We call this spread of levels
a ``band.'' The energy difference between the highest and lowest levels
is the ``bandwidth.''
It's often useful to plot the energy E versus wavevector k of the
band. Let's start by considering a free electron gas. In this case
the system is isotropic, homogeneous, and has complete translational
symmetry. So the momentum is a good quantum number.
For simplicity let's consider a one dimensional system. The energy
is given by
![]() |
(1) |
Now let's consider a crystal with a periodic array of atoms. Now we
only have discrete translational symmetry. Let a be the distance
between 2 atoms in the lattice. a is called the lattice constant.
For simplicity, let's consider a one dimensional lattice of length L=Nawhere N is the number of atoms in the lattice.
As we saw in lecture 1 when we considered the case of periodic
boundary conditions, we found that the allowed wavevectors
were
![]() |
(3) |
![]() |
(4) |
The magnitude of the gap is of the same order of magnitude as the periodic potential (V(G)). So we have 2 bands which are separated by a gap:
This is the repeated zone scheme. Since translating by G doesn't add anything new, we usually just draw:
The range from
to
is
called the first Brillouin zone.
So far we have been considering a lattice that is a periodic array of atoms. But rather than just one atom on each lattice site, we could have something more complicated, like 2 atoms or a molecule or several atoms. This unit which is repeated periodically is called a ``unit cell.'' A unit cell can be just one atom or several atoms. If the unit cell is a molecule which retains some measure of its individual identity in the solid, then we have a molecular solid. Each unit cell contributes exactly one independent value of k to each energy band. Taking into account the 2 spin orientations of the electron, there are 2N independent levels in each energy band, where N is the number of unit cells in the crystal.
For simplicity, let's consider a one dimensional lattice of length L=Nawhere N is the number of atoms in the lattice.
As we saw in lecture 1 when we considered the case of periodic
boundary conditions, we found that the allowed wavevectors
were
![]() |
(7) |
![]() |
(8) |
![]() |
(9) |
![]() |
(10) |
Just as the lattice looks the same if you translate by a lattice vector
or any of its multiples, so in k-space, things look the same
if you translate by a reciprocal lattice vector G.
Let's go back to the band picture. If each unit cell contributes 1 valence electron to a band, then the band will be half full, the Fermi energy will lie in the band, and the system will be metallic. A solid with a partially filled band is called a metal. In a metal electrons can flow and carry current because electrons in filled states below the Fermi energy can easily jump to empty states above the Fermi energy. The energy difference between the filled and empty states can easily be supplied by the applied electric field and thermal excitation.
Experimentally the way to tell the difference between a metal and an insulator is by measuring electrical resistance. The difference between a good conductor and a good insulator is striking. The electrical resistivity of a pure metal may be as low as 10-10 ohm-cm at a temperature of 1 K (ignoring the possibility of superconductivity). The resistance of a good insulator may be as high as 1022ohm-cm. This range of 1032 may be the widest of any common property of solids.
In a metal, the resistivity increases linearly with increasing temperature because the electrons scatter from phonons (lattice vibrations), and the number of phonons increases with temperature.
Photons can also be used to excite electrons from the valence band into the conduction band. When electrons make transitions from the conduction band into the valence band and recombine with holes, photons can be given off. Semiconductor lasers take advantage of this.
Semiconductors whose primary source of carriers comes from the direct excitation of electrons from the valence band to the conduction band are called intrinsic semiconductors. Most of the electrical current carriers in extrinsic semiconductors come from impurities. These impurities produce states in the band gap which can supply electrons to the conduction band or holes to the valence band. Most electronic devices use extrinsic semiconductors that have been subjected to selective doping.
If a semiconductor has primarily donor impurities, we call it an n-type semiconductor because it has primarily negatively charged carriers. If a semiconductor has primarily acceptor impurities, we call it a p-type semiconductor because it has primarily positively charged carriers.
Let's suppose that initially the barrier between the p and n doped
semiconductors is infinitely high. The chemical potential will
be higher in the n-type semiconductor than in the p-type semiconductor.
Now imagine that we remove the barrier. Electrons will flow over to the p-side, and holes to the n-side until the chemical potentials are the same.
As soon as a small charge transfer by diffusion has taken place, there is left behind on the p-side an excess of - ionized acceptor atoms and on the n-side an excess of + ionized donor atoms. This double layer of charge creates an electric field directed from n to p that inhibits further diffusion and maintains the separation of the two carrier types. We can draw the potential seen by the electrons. The potential drop will be at the interface because the ionized donors and acceptors attract each other.
Electrons would rather go downhill than uphill. If we apply a voltage
across the junction that increases the size of the drop, we encourage
electrons to flow from the p-side to the n-side. This is called
reverse bias. But they already want to do this, so it doesn't make
much difference in the current. If we really crank up the
reverse bias, we get what is called ``breakdown'' and electrons
avalanche from the p-side to the n-side.
If we apply voltage in the other direction, the electrons
are less reluctant to go from the n-side to the p-side.
(the conductivity
where V is the barrier height.) This is
called forward bias. This asymmetry in the preference
of the direction of the
current is how a diode works. A diode allows current to go
one way but not the other way. Increasing current flows as the
forward bias increases but not much current flows when reverse
bias is applied.
A bipolar transistor is a current amplifier. In normal operation the emitter
to base junction is forward biased, and the collector to base
junction is reverse biased. Consider the electrons coming into
the base from the emitter due to the forward biased emitter-base
junction. For a thin enough base section, these carriers sweep through the
base layer, cross the base-collector junction, and contribute
to the collector current. The essential action is the emitting of
carriers from the emitter region and the collection of practically all
of these carriers by the collector. Let's denote this current Iec.
A small hole current from the base region also flows across the
emitter junction. We will denote this hole current Ibe.
This adds to the electron current from the emitter to the base.
By proper design of the impurity concentrations and base layer width, the
ratio
Iec/Ibe can be made very large (
).
If the input current is taken to be the small hole current Ibe,
and the output current is taken to be the large emitter-collector
current Iec, a significant current gain is thus achieved.
The positive gate voltage attracts electrons to the interface. By adjusting the magnitude of the gate voltage Vg, we can adjust the charge density at the interface between the semiconductor and the insulator. This is like a capacitor where Q=CVg. The current flows between the source and the drain. Since current is I=dQ/dt, we can adjust the amount of current by adjusting the gate voltage.
The potential seen by the electrons is lower near the interface between the semiconductor and the oxide layer than deep inside the semiconductor. We can describe this lower potential by ``band bending.''
(When one puts this 2DEG in a large magnetic field perpendicular to the plane of the interface, one gets the quantum Hall effect. The discovery of the integer quantum Hall effect by Klaus von Klitzing won him the Nobel prize in 1985. The 1998 Nobel prize was for the fractional quantum Hall effect which was discovered by Daniel Tsui and Horst Stormer and explained by Robert Laughlin. In the fractional quantum Hall effect the 2DEG becomes a quantum fluid with fractionally charged excitations which have charges like e/3. The smaller the fraction, the larger the applied field.)