The Solid State

Band Theory of Solids

You do not really understand something unless you can explain

it to your grandmother.A. Einstein

Typo in SyllabusNews FlashTypo in Syllabus

428 points gets you a C
429 points gets you a B
I already generously rounded the cutoff percentages
down! (429 is not quite 79.5%)
I have added two typo points to your grades.
The grades spreadsheet shows the grade you will get if
you do nothing for the rest of the semester.
Please make sure all earned bonus points are recorded!

10.6 Band Theory of Solids

What happens in crystalline solids when we bring
atoms so close together that their valence electrons
constitute a single system of electrons?

Band structure of diamond. http://home.att.net/~mopacmanual/node372.html

The energy levels of the

overlapping electron
shells are all slightly
altered.
The energy differences
are very small, but
enough so that a large
number of electrons can
be in close proximity and
still satisfy the Pauli
exclusion principle.
The result is the
formation of energy
bands, consisting of
many states close
together but slightly
split in energy.

The energy levels are so close together that for all

practical purposes we can consider bands as a
continuum of states, rather than discrete energy levels
as we have in isolated atoms (and in the core electrons
of atoms of metals).
A detailed analysis of energy bands shows that
there are as many separate energy levels in each
band as there are atoms in a crystal.*
Suppose there are N atoms in a crystal. Two
electrons can occupy each energy level (spin), so
there are 2N possible quantum states in each
band.
Lets consider sodium as an example. Sodium has a
single outer 3s electron.
*Kind of. I need to explain.

When you bring two

sodium atoms together,
the 3s energy level
splits into two separate
energy levels.
Things to note: 4
quantum states but only
2 electrons.
You could minimize electron energy by putting both 3s
electrons in the lower energy level, one spin up and the
other spin down.
There is an internuclear separation which minimizes
electron energy. If you bring the nuclei closer together,
energy increases.

When you bring five

sodium atoms together,
the 3s energy level
splits into five separate
energy levels.
The three new energy
levels fall in between
the two for 2 sodiums.
There are now 5 electrons occupying these energy
levels.
Ive suggested one possible minimum-energy
configuration. Notice how the sodium-sodium
internuclear distance must increase slightly.

When you bring N (some

big number) sodium
atoms together, the 3s
energy level splits into N
separate energy levels.
The result is an energy
band, containing N very
closely-spaced energy
levels.
There are now N electrons occupying this 3s band.
They go into the lowest energy levels they can find.
The shaded area represents available states, not filled
states. At the selected separation, these are the
available states.

Now lets take a closer look

at the energy levels in solid
sodium.
Remember, the 3s is the
outermost occupied level.
When sodium atoms are
brought within about 1 nm
of each other, the 3s levels
in the individual atoms
overlap enough to begin
the formation of the 3s
band.
The 3s band broadens as
the separation further
decreases.

3s band
begins to
form

Because only half the

states in the 3s band are
occupied, the electron
energy decreases as the
sodium-sodium separation
decreases below 1 nm.
At about 0.36 nm, two
things happen: the 3s
energy levels start to go up
(remember particle in
box?) and the 2p levels
start to form a band.
Further decrease in
interatomic separation
results in a net increase of
energy.

3s electron
energy is
minimized

What about the 3p and 4s

bands shown in the figure?
Dont worry about them
there are no electrons
available to occupy them!
Keep in mind, the bands do
exist, whether or not any
electrons are in them.
What about the 1s and 2s
energy levels, which are
not shown in the figure?
The sodium atoms do not
get close enough for them
to form bandsthey
remain as atomic states.

Figure 10-20 (the one on the last three slides) shows

energy levels as a function of interatomic separation.
Energy levels for an actual crystal structure also vary
with different directions in space.

http://cmt.dur.ac.uk/sjc/thesis/thesis/node39.html, band structure of silicon

As an aid to visualization, we often represent energy

bands like this (using sodium as an example):
This is highly schematic.
Real bands aren't boxes or
lines.
Sodium has a single 3s
electron, so the 3s energy
band contains twice as
many states as there are
electrons. The band is half
full.

3s
2p
2s
1s

At T=0 the band is filled exactly halfway up, and the

Fermi level, F, is right in the middle of the band.
Sodium is a metal because an applied field can easily
give energy to and accelerate an electron.

Magnesium has two 3s

electrons. You expect the 3s
3p
band to be full, the 3p band to
be empty, with a forbidden
gap in between. Magnesium
should be an insulator. (Why?)
3s
But magnesium is a metal
(actually, a semimetal).
The 3p and 3s bands overlap.
There are many empty states
nearby into which electrons
can be accelerated.

3p
3s

Materials which have bands

either completely full or
completely empty are
insulators (unless band
overlap occurs, as was the
case for magnesium).

3p

3s

In a carbon atom, the 2p shell contains 2 electrons.

There are 6 available states, so one would expect the
2p band to be 1/3 full* and carbon to be a conductor.
But carbon is an **insulator. Why?
*2N 2p electrons in a crystal with N atoms. 6N 2p states, when you
include spin.
**If the diamond in your diamond ring conducts electricity, its time to
take it back!

Figure 10.23 shows

energy bands in carbon
(and silicon) as a
function of interatomic
separation.
At large separation,
there is a filled 2s band
and a 1/3 filled 2p band.

But electron energy can be lowered if the carbon-carbon

separation is reduced.
There is a range of carbon-carbon separations for which
the 2s and 2p bands overlap and form a hybrid band
containing 8N states (Beiser calls them levels).

But the minimum total

electron energy occurs
at this carbon carbon
separation.
At this separation there
is a valence band
containing 4N quantum
states and occupied by
4N electrons.
The filled band is separated by a large gap from the
empty conduction band. The gap is 6 eVremember,
kT is about 0.025 eV at room temperature. The gap is
too large for ordinary electric fields to move an electron
into the conduction band. Carbon is an insulator.

Silicon has a similar

band structure. The
forbidden gap is about 1
eV.
The probability of a
single electron being
excited across the gap
is small, proportional to
exp(-Egap/kT).
However, there are enough 3s+3p electrons in silicon
that some of them might make it into the conduction
band. We need to consider such a special case.
On page 355 Beiser says there are as many levels in a band (N) as there
are atoms in a crystal. In figure 10.23 the caption implies there are 2N
levels in a 2s band. This is inconsistent. Lets call a level an energy
level and a state a quantum state, including spin. Then figure 10.23
should talk about 2N states in a 2s band.

A semiconductor is a material which has a filled (at

T=0) valence band separated by a small gap from an
empty (at T=0) conduction band.

A semiconductor at
T= 0 K.

conduction
band

valence
band
Because the gap is small, at room temperature there
will be a few electrons in the conduction band. These
electrons can be accelerated by an applied electric
field.

A semiconductor at room
temperature.
Although e(-Egap/kT) is small,
Ne(-Egap/kT) can be significant.
The smaller the gap, the
more charge carriers in the
conduction band.

CB
F
VB

Thus, at very low temperatures, silicon is an insulator,

but at room temperature, it is a weak conductor
(intermediate between conductor and insulator, hence
semiconductor).
A rough rule of thumb: a band gap of less than 3 eV
gives rise to a semiconductor.

It is possible to engineer properties of semiconductors

to make them more suitable for use in devices.
We do this by doping them with impurities.
Heres an example: arsenic impurities in silicon.
As has an outer electron configuration of 4s2 4p3. After
it shares 4 of its outer electrons with neighboring
silicons, the remaining electron is very loosely bound.
Homework problems 10.21 and 10.22
(assigned although listed under section
10.8) show that kT at room temperature
is more than enough energy to ionize
the fifth outer electron.

The electrons from the As+ ions are then free to move
throughout the crystal.

The arsenic impurity creates a donor impurity level.

Because it takes only a few meV to ionize the As and
place the resulting electron in the conduction band, the
donor impurity level sits just below the conduction band.
A few donor impurities can
produce many electrons in
the conduction band and
give rise to significant
conductivity.
The Fermi level lies
somewhere between the
top of the valence band and
the highest-energy electron
in the conduction band.

CB
F

VB

These semiconductors are n-type, because conduction

is by
n egative electrons.

Heres another example: gallium impurities in silicon.

Ga has an outer electron configuration of 4s2 4p1. It
wants to borrow 3 electrons from its 4 silicon nearest
neighbors.
After a Ga borrows 3 electrons,
it is easy for the 4th silicon to
convince the Ga to borrow
one more electron, creating a Gaion.
The gallium impurity creates an
acceptor impurity level.

The lattice depicted here

is highly schematic, not
realistic, and should be
used as a visual aid only!

Because the acceptor ionization energy is very small,

the acceptor levels sit just above the valence band (a
small fraction of an eV).
Electrons from the valence
band can easily get to the
acceptor bound states. That
leaves holes in the valence
band.
The holes represent states
into which other electrons in
the valence band can move.
Thus, electrons can easily
move around in the presence
of an applied field.

CB

F
VB
This is called a p-type
semiconductor because
positive holes are the
charge carriers.

Alternatively, we can look at the holes, say they move

around, and say that conduction is due to holes.

Again, it turns out that there don't have to be very

many acceptor atoms around to result in a significant
number of holes in the valence band.
What's all this fuss about holes. It seems like a hole is
really just an electron missing from the valence band
and sitting in an acceptor state.
Why don't we just talk about conduction by these
electrons moving from state to state, instead of
worrying about holes?
Answer: holes really are more than just missing
electrons. For one thing, electrons in this case are
"stuck" on acceptor atoms, but the holes are free to
move about in the valence band.
For another thing, we can dope semiconductors so that
there are excesses of holes and electrons. A "hole"
really is more than just a missing electron which is
somewhere else.

Optical properties of solids are closely related to band

structures.
Visible photons have energies ranging from about 1 eV
to about 3 eV.
Metals can absorb visible photons
because there are many empty
states for electrons to move to.
In diamond, the valence band is full and
the conduction band is empty.
A 3 eV photon cannot excite an electron
across the 6 eV band gap. Diamond
cannot absorb visible photons. Diamonds
are transparent.

Shouldnt all insulators be transparent?

Yes. They are opaque because they contain impurities
and irregularities which scatter visible light.
Insulators are transparent to high-energy UV
(ultraviolet) photons.

Now, skipping ahead briefly

Section 10.8 is a lengthy section which attempts to

give a better justification of energy bands.
The section is not required, but does contain a few key
points:
The periodicity of a crystal lattice leads naturally
(through the solution of Schrdingers equation)
to allowed and forbidden energy bands for
electrons.
Electrons cannot exist in forbidden energy states
because there are no solutions to Schrdingers
equation there.
Electrons do not (in general) scatter from a
perfect lattice, but they do interact with the
lattice.

Sometimes the lattice holds the free electrons

in place. These electrons act like they have an
unusually high mass (meffective > melectron).
Sometimes the electrons act like they have an
unusually small mass (meffective < melectron).
Sometimes the electrons even have a negative
effective mass.
Huh?
Sure! Push on an electron in one direction. If the
push gives it enough energy to Bragg scatter off
the lattice (the exception to the no scattering
rule), the electron goes in the opposite direction.
Negative mass!

Ceramics majors would benefit from studying section

10.8. The only material you are required to know (for
exam or quiz purposes) is that given above.
One more thing the band structure I showed earlier:

Notice the gap that extends in all directions through

the crystal.

Lets move on to lecture 34 and see how far we get