MICROELECTRONICS
ELE1055
Handout 7: Semiconductor Devices Part 1
p-n JUNCTIONS
Diode Applications and Importance
First used during World War II
Theory of operation developed in 1949, leading to invention of transistor
Most commonly used rectifier
Device isolation in integrated circuits
Special types: LED’s
lasers
photodiodes
solar cells
Diode Fabrication
A p-n junction diode is formed when, within a single-crystalline
semiconductor, the dopant type changes from p-type to n-type or vice-versa.
This can be achieved during the growth of the semiconductor by
changing the composition of the melt, or by changing the source gas in a
technique known as chemical vapour deposition.
However, far and away the commonest fabrication technique is
impurity diffusion after growth. Consider, for example, an n-type silicon
wafer into one side of which a high concentration of boron atoms is
introduced. Just as electrons and holes diffuse from areas of high
concentration to those of lower concentration, so the boron atoms will diffuse
into the interior of the wafer. The main difference is that high temperatures
( 1000oC) are required to enable the diffusion. When the wafer returns to
room temperature, the boron distribution is “frozen”.
The boron atoms can either be introduced directly in the diffusion furnace,
or by ion implantation prior to diffusion. Diodes can also be fabricated by
diffusing n-type dopants into p-type wafers.
For integrated circuits, selective masking is required. The process flow
below illustrates the technique.
1. Oxidize the silicon wafer
2. Apply a layer of photoresist (PR)
3. Expose PR through mask 1
4. Remove exposed PR in developer
5. Use HF to remove SiO2 in windows
6. Remove PR and diffuse boron through windows
7. Evaporate aluminium and pattern using mask 2
Mask 1 Mask 2
Diode Theory
1. Establishment of Equilibrium
In order to simplify our understanding of the p-n diode, we consider a
step junction i.e. a uniformly doped p-type region with an abrupt transition
to a uniformly doped n-type region. Although this is not a particularly
accurate picture for a diffused diode, it serves to illustrate the operation of a
p-n diode very well.
Before the junction is formed, the p-type material has a high
concentration of holes and the n-type material has a high concentration of
electrons. After junction formation, therefore, holes will diffuse from p-type
to n-type; electrons from n-type to p-type. One might expect this to continue
until hole and electron concentrations became uniform, and the diode was
annihilated.
However, when holes vacate the p-type material, negatively charged
acceptor ions occupy the “space” left. Similarly, positively charged donor
ions remain when electrons move out of the n-type material. These fixed ions
form what is called a space charge, and the part of the diode near the
junction, vacated by electrons and holes, is called the depletion region.
The fixed charges give rise to an electric field, as shown in the diagram,
which will introduce a drift current in the diode. Now we consider carrier
motion and current flow due to both diffusion and drift.
Carrier Motion Current Flow
Hole diffusion
Hole drift
Electron diffusion
Electron drift
The combined diffusion current flow is from left to right; that due to
drift from right to left. Equilibrium will be established when the net current
is zero i.e when diffusion and drift currents are equal.
2. Energy Band Diagram in Equilibrium
We now re-introduce a very important property of the Fermi level in
a semiconductor under equilibrium conditions. Because the Fermi level is
defined as the energy at which the probability of an electron state being
occupied is ½, it is representative of the average electron energy at any point.
In equilibrium, average electron energy cannot vary; otherwise electrons
will move from positions of high energy to those of low energy. This means
that, in equilibrium, the Fermi level is constant throughout a semiconductor.
This is just as true for a p-n junction as for a uniform semiconductor
sample. In equilibrium, therefore, the energy band diagrams both before and
after junction formation are as shown.
Referring to the second diagram, the energy bands outside the
depletion region are unperturbed, but inside the depletion region they must
“bend” in order to meet up.
3. Built-In Potential Barrier
Electron energy is given by:
E qV
where V is the electrostatic potential.
A graph of electrostatic potential for the conduction band (or valence
band) is, therefore, as follows.
V0, the potential difference between the two sides of the junction, is
known as the built-in potential barrier or contact potential. It is this built-in
potential barrier which gives rise to the forward turn-on voltage of a diode.
From the energy band diagram:
qV0 ( Eip EF ) ( EF Ein )
NA ND
qV0 kT ln.( ) kT ln.( )
ni ni
NA ND
qV0 kT ln.( 2
)
ni
kT N AND
V0 ln .( 2
)
q ni
4. Depletion Region Width
Once equilibrium is established in a diode, the electric field created by
the fixed impurity ions ensures that virtually no electrons and holes can exist
in the depletion region. It is fair to assume, therefore, that the charge density
within the depletion region is entirely due to the fixed impurities, and that
outside the depletion region there is charge neutrality. This is known as the
depletion approximation.
A graph of charge density across the junction is as follows:
To preserve overall charge neutrality, the total negative and
positive charges must be equal.
Charge = Charge Density × Volume
qN A Ax p qN D Axn
where A is the cross-sectional area.
N A x p N D xn (1)
We now use Poisson’s Equation, which in one form states that:
where is the permittivity of the semiconductor, is the electric field and
is the charge density.
Using the depletion approximation, Poisson’s Equation implies that
for -xp < x < 0, (2)
and for 0 < x < xn ,
and because the charge density is zero outside the depletion region, the
electric field must be zero there also.
Therefore, a graph of electric field across the junction is as follows.
Using equation (2),
by considering the slope of the left-hand side of the graph.
(3)
[We could equally well have proved that
by considering the right-hand side of the graph, but this is already apparent
by substituting equation (1) in equation (3).]
We also know that electric field is related to potential by the equation:
or
The built-in potential of a p-n junction is, therefore, given by:
The right-hand side of this expression is simply the area under the graph of
electric field within the depletion region. This is a negative area as it lies
entirely below the x-axis.
or (4)
using equation (3).
Let W be the total depletion region width
i.e.
using equation (1).
(5)
Substituting equation (5) in equation (4):
In many cases, the doping on one side of the junction is much
heavier than on the other side e.g. shallow diffused junctions. This is known
as a one-sided step junction.
If ND >> NA ,
If NA >> ND ,
It is, therefore, the lower of the two doping levels which determines the
depletion region width, and from equation (1) it can be seen that the
depletion region will be mostly on the low-doped side of the junction.
5. Forward and Reverse Bias
Having studied the p-n diode under equilibrium, we now investigate
the effect of applying bias.
Forward bias for a p-n diode means a positive voltage applied to the
p-side. The graph of electrostatic potential at equilibrium was:
so a positive voltage at the p-side will reduce the potential barrier by an
amount equal to the forward bias (Vf). The energy band bending and
depletion region width will also be reduced.
The opposite effect is observed for reverse bias.
EQUILIBRIUM FORWARD BIAS REVERSE BIAS
In summary,
Equilibrium
1. Electrostatic potential barrier V0
2. Drift and diffusion currents equal and opposite
Forward Bias
1. Electrostatic potential barrier V0 - Vf
2. Diffusion current increases exponentially
3. Drift current remains small
Reverse Bias
1. Electrostatic potential barrier V0 + Vr
2. Diffusion current decreases exponentially
3. Drift current now larger than diffusion current
N.B. Changes in the potential barrier have little effect on drift current
because the inbuilt slope is already carrying all available carriers across the
junction.
6. Diode Equation
A simplified derivation of the diode equation follows from the previous
discussion.
The net forward current is given by:
I I diff I drift
The diffusion current is exponentially dependent on bias voltage.
qV
i.e. I diff I 0 e kT
where I0 is a constant.
qV
I I0e kT
Idrift
In equilibrium (i.e. V = 0) , I = 0
I0 Idrift 0
I drift I0
qV
i.e. I I 0 (e kT
1)
The I/V characteristic highlights the diode’s rectifying properties.
Current conduction occurs for forward bias, but in reverse bias only a small
voltage-independent current flows. This is often referred to as the reverse
saturation current.
7. Reverse Breakdown
At a critical value of reverse bias, the reverse current of a p-n diode
suddenly increases sharply - a phenomenon known as reverse breakdown.
The modified I/V characteristic is as follows:
Reverse breakdown is not inherently destructive, as long as the
current is limited by the external circuit.
There are two mechanisms which can cause reverse breakdown.
Zener Breakdown: As the reverse bias applied to a p-n diode increases, so
does the electric field in the depletion region, which is at its maximum at the
junction itself. At a critical value of electric field, electrons can be torn from
their host atoms, and accelerated to the n-side of the junction. Because of the
large number of atoms, the current can increase rapidly. This process is
known as field ionization or Zener breakdown.
Avalanche Breakdown: The increased electric field has another effect. The
small number of electrons comprising the reverse saturation current are
accelerated to high velocities. When they then collide with atoms in the
crystal lattice, additional electrons can be knocked out. Both the original
electrons and the new ones are then accelerated by the electric field. The
number of electrons multiplies rapidly. The process is known as impact
ionization or avalanche breakdown.
