Parcours Ouvert

oP.2 EEE EXP4

CHAPTER I

INTRODUCTION TO SEMICONDUCTORS

Semiconductors are solid materials used in the manufacture of electronic

components. The most commonly used material is Silicon (Si). Other semiconductor
materials are also used but to a lesser extent: Germanium (Ge), Gallium Arsenide
(AsGa) and Indium Phosphide (InP).

I. THE SILICON CRYSTAL

I.1 The isolated atom

Silicon is the most widely used semiconductor in integrated circuits. Its

simplest physical model is that of Bohr (Figure 1):

- A positively charged nucleus (made up of protons and neutrons).

- Electrons placed in several orbits around the nucleus.

Figure 1: Silicon Bohr Model

The energy levels of the electrons are discrete and are associated with the notion of
an orbit or electron layer (K: 2 electrons; L: 8 electrons; M: 18 electrons etc.). The
electrons in the last layer are called valence electrons. They are used to make the
bonds between atoms to form the silicon crystal.

1 / 82
Silicon has 4 valence electrons (4 negative elementary charges), which means that,
for the model described above, the nucleus is positively charged to 4 positive
elementary charges (electrical neutrality).

I.2 The lattice atom

When atoms come together to form a solid, Pauli's exclusion principle

indicates that no more than two electrons in the solid can have the same energy.
When two atoms come together to form a covalent bond, the same Pauli's principle
would be violated if the energy levels of the two electrons in the bond were the
same. As a result, the energy levels of these electrons split into two different energy
levels. If the number of atoms increases, these two levels subdivide themselves into
other, very similar energy levels, which finally result in energy bands separated by a
gap.

Figure 2: Electron energy diagram of the networked electron

Depending on the interatomic distance in the network, very different material

properties are observed:

- for a metal, EG=0 . Electrons flow freely from the valence band to the
conduction band. The large number of electrons available (1022 to 1023 per cm3) allows
for high conduction at very low energy: metals are good conductors.

2 / 82
 - For a semiconductor, only a small amount of energy (EG ≈ 1 eV) is needed to
ionize an atom, i.e. to obtain a free or conduction electron. Average conduction is
observed at ambient temperatures, hence the name semiconductor.

- for an insulator, the energy required to ionize an atom is greater: EG ≥ 3 eV

(diamond: 7 eV). There is practically no electronic conduction, the body is an
insulator.

Figure 3: Electron energy diagram of the networked electron

The following figure (Figure 4) shows the bonds between silicon atoms forming the
silicon crystal. It is the pooling of 2 valence electrons that makes a bond. There must
be 4 of them around each silicon nucleus. Under the effect of thermal agitation, a
few electrons and free holes are generated.

Under the effect of an additional energy input to the valence electrons (energy
above an E0 threshold or photoelectric threshold), the latter can leave their orbit and
become free; thus leaving the valence band and passing into the conduction band.

3 / 82
 Figure 4: Silicon crystal

They become free electrons or conduction electrons since they can give rise to
an electric current if they are in motion (Figure 5).

Figure 5: Displacement of electrons and holes (T>0 K)

The electrons, having left the valence band, give rise to electron gaps called free
holes (equivalent to free electrons, but with a positive charge), for reasons of
electrical neutrality. This is called the generation of electron-hole pairs.

Conversely, when a free electron encounters a hole, it can fill it. This is the
phenomenon of recombination of electron-hole pairs.

The average time between generation and recombination is called the carrier

4 / 82
lifetime. For pure silicon, it is about one second, but it decreases when the crystal
is impure or has defects.

An electric current consists of a movement of free electrons and a movement of free

holes.

Pure material is said to be intrinsic because it is only defined by its own

characteristics.

I.3 Fermi level

The Fermi level represents the maximum energy of the states occupied by the
electrons at absolute zero. Its value depends on the number of valence electrons and
the density of accessible states and is therefore characteristic of the material. In
semiconductors, the Fermi level is in the gap. This means that at 0 K, the valence
band is full while the conduction band is empty. When the temperature rises, some
electrons are able to exceed the Fermi level and thus reach the conduction band.

The probability of occupation of an energy level E at a temperature T by an

electron is given by the Fermi-Dirac statistic.

With : Boltzmann constant =1.38 10-23 J.K-1=8.6 10-5 eV.K-1

Absolute temperature (Kelvin)

Energy level (J or eV)

: Fermi level, characteristic of the material. It is located around the middle of the
band gap for a pure semiconductor.

Based on Boltzmann's approximation,

If >> (at ) then the expression of the probability

is approximated by :

5 / 82
 - In the conduction band:

- In the valence band :

Figure 6: Probability of occupation of an energy level by an electron

We can thus determine the probability of obtaining a free electron in the conduction
band :

This expression shows the very great influence on the number of available carriers,
the width of the band gap of the semiconductor and its temperature.

I.4 Intrinsic electron and hole densities

The effect of temperature is that the energy input is proportional to the

absolute temperature (in degrees Kelvin). This thermal energy creates an intrinsic

density of free electrons equal to an intrinsic density of free holes .

The numbers of free electrons or holes per cm3 of silicon are statistical averages
resulting from the combination of the two phenomena of electron-hole pair
generation and recombination.

6 / 82
At room temperature (T=300 K),

This density of electrons and free holes is very low compared to the density of atoms
in a silicon crystal, which is 5 1022 cm-3. This is why an intrinsic (undoped) silicon
crystal is a very poor conductor of electricity.

As the temperature rises, these intrinsic electron and hole densities increase.

II. N-DOPING OF SILICON

II.1 Description

Doping consists of introducing into the semiconductor crystal atoms of

another nature, called impurities, which locally take the place of a semiconductor
atom. A crystal thus doped is called extrinsic, as opposed to pure crystal, called
intrinsic.

This silicon doping operation is carried out at a high temperature of about 1000°C by
epitaxy, ion implantation and/or diffusion.

A silicon crystal has an atom density of 5,1022 cm-3. When this crystal is doped,
impurities are introduced into it at a density which can vary between 1014 and 1019 cm-
3. More, it would degenerate the silicon crystal which would lose its semiconductor
qualities.

Indeed, contrary to an intrinsic or extrinsic semiconductor, the Fermi level of a

degenerated material is no longer in the forbidden band but in one of the allowed
bands (conduction or valence).

The purpose of doping is to create either a large number of free electrons in

the crystal (N-type semiconductor) or a large number of free holes (P-type
semiconductor).

In an extrinsic semiconductor, the number of P carriers (holes) will no longer be

equal to the number of N carriers (free electrons).

For N-doping, phosphorus is most often used as an impurity. Phosphorus has 5

valence electrons, which makes it interesting for an N-type doping of Si.

7 / 82
 Figure 7: Potential dopants

Figure 8: Bohr Model of Phosphorus

Later on, phosphorus will be used as a doping agent. The latter is a donor-type
impurity, because it gives free electrons to the crystal in the following way:

Figure 9: Phosphorus-doped silicon

When a phosphorus atom replaces a silicon atom, 4 of its valence electrons

participate in the bonds with the 4 neighbouring atoms of the silicon crystal. The 5th
electron is in surplus and becomes easily free. Its activation energy is very low (≈ 0.01
eV); it is therefore free at room temperature.

The level of Fermi is also altered by doping. The higher the N-type doping, the
closer the level is to the conduction band.

8 / 82
 Figure 10: Energy diagram of an N-doped material

II.2 Density of electrons and free holes

Each phosphorus atom introduced into the silicon crystal will result in the
creation of a free electron. The density of free electrons is equal to the density of
impurities of the donor type. A phosphorus atom that releases 1 electron becomes
positively charged to maintain electrical neutrality. This is called a positively ionized
impurity. This positive ion cannot give rise to an electric current since it is fixed in the
crystal.

The density of positively ionized impurities is equal to the density of impurities of the
donor type.

These last 2 considerations show that the silicon crystal always remains electrically
neutral: there are as many negatively charged free electrons as positively charged
but fixed impurities.

By defining:

The density of free electrons (per cm3) in an N-type semiconductor,

The density of holes in an N-type semiconductor,

the density of the donor atoms introduced into the crystal.

We can write, room temperature:

9 / 82
In addition, it is demonstrated in solid-state physics that, for a given temperature
and at equilibrium, the product of the density of free electrons by the density of
holes is a constant, regardless of the doping:

(Mass Action Law)

Nc and NV represent the effective state densities in the conduction and valence
bands. This shows the dependence of the intrinsic density on temperature and
bandgap width.

Example: Calculation of minority holder density

A 27 oC, ft = ni = 1.45 1010 cm-3

Hence,

ni2 = ft ni = 2.10 1020cm-6

For ND = 1018 cm-3, we have:

nn = ND = 1018 cm-3

cm-3

and thus

In an N-type semiconductor, the free electrons are called the majority carriers, while
the holes are called the minority carriers.

10 / 82
 Figure 11: Variation with temperature

The figure above shows the variation with temperature for 3 semiconductor

materials: Germanium ( =0.66 eV), Silicon ( =1.12 eV) and Gallium Arsenide (
=1.42 eV).

III. P-DOPING OF SILICON

III.1 Description

For a P-doping, Boron is most often used as an impurity. Boron has 3 valence
electrons, which makes it interesting for a P-type doping of Si.

Figure 12: Potential dopants

11 / 82
Boron is an acceptor-type impurity because its presence in the crystal corresponds
to as many holes capable of accepting an electron to saturate the covalent bond.
The semiconductor doped with an acceptor type impurity is said to be p-type
because of the large number of holes thus created.

Figure 13: Bohr model of boron

As far as the Fermi level is concerned. The higher the P-type doping, the closer it
is to the valence band.

Figure 14: Boron-doped silicon

Figure 15: Energy diagram of a P-doped material

12 / 82
 III.2 Density of electrons and free holes

A free hole is generated by an impurity of the "acceptor" type. To each free

hole corresponds a fixed negatively ionised impurity. This ensures the electrical
neutrality of the crystal.

By defining:

The density of free holes in a P-type semiconductor,

The density of free electrons in a P-type semiconductor,

The density of the acceptor-type atoms introduced into the crystal.

We can write:

In a P-type semiconductor, the holes are the majority carriers and the free electrons
are the minority.

IV. MOVEMENT OF ELECTRONS AND HOLES

In a semiconductor, the current is bipolar in nature, as it is based on the

simultaneous movement of 2 types of free carriers, electrons and holes. In addition,
the displacement of each type of carrier can be caused by 2 different phenomena
that can coexist:

- The conduction or drift that results from the action of an electric field on the
carriers.

- Diffusion that results from non-uniform carrier density in the crystal.

In total, the current in a semiconductor can therefore have 4 components

(conduction and diffusion of electrons and holes).

IV.1 Conduction current

13 / 82
 This is the action of an electric field E on the carriers P and N. The electric field

causes them to move at an average speed which depends on the amplitude of the
electric field. The proportionality factor that relates the speed to the E field is called
mobility μ.

(in cm2/Vs)

It can be seen from the figure below (Figure 16) that the velocities are only
proportional to the electric field at low field strengths. From a certain value of the
107
electric field, the velocity of the carriers tends towards a limit of about cm/s for
pure materials.

Figure 16: Carrier velocity as a function of electric field

The figure below (Figure 17) shows the variation of mobilities as a function of doping
for 3 semiconductor materials. It can be noted that the mobility of holes is always
lower than that of electrons. This is due to the differences in effective masses
(different energy levels) of the two types of carriers, with the holes moving in the
valence band at lower energy.

14 / 82
 Figure 17: Carrier mobility as a function of doping

Mobilities and are complex functions depending on many parameters:

material, temperature, doping, electric field...

Mobility depends mainly on it:

- The type of carriers, the mobility of the electrons is two to three times higher
than that of the holes.

µn = 600 cm2/Vs. This average value depends on the device.

µp = 300 cm2/Vs, This average value depends on the device.

- The density of the dopants in the crystal and the electric field,

- Temperature.

A charge displacement is equivalent to a conduction current: the holes move

in the direction of the E field, the electrons in the opposite direction.

The conduction current density due to the holes is equal to the charge flow per
unit time in a unit section.

15 / 82
 q: (q= 1.6 10-19 C ) is the elementary charge of the electron or hole in absolute
value.

Replacing it with its value, it comes:

The same goes for electrons:

We note that the two current densities are of the same sign, because the electrons
with a negative charge -q flow in the opposite direction of the field (the 2 signs -
compensate each other).

The total conduction current density applies:

or else:

with it:

σ is the conductivity of the silicon rod. It depends on the density of electrons and
holes and therefore on the doping. The resistivity is determined directly with the
conductivity.

(Ω.cm)

The resistivity of the material can thus be determined by doping control.

The figure below (Figure 13) shows the variation in resistivity with the doping for P
and N doped silicon at 300K.

16 / 82
 Figure 18: Silicon resistivity as a function of doping at 300 K

IV.2 Broadcast current

As we have seen previously, if we consider a semiconductor in thermal

equilibrium, a process of electron-hole pair generation by thermal agitation is
constantly occurring. The opposite process, recombination, sees an electron
recombining with a hole.

In an equilibrium semiconductor the two processes balance each other so that the
free carrier concentration remains constant.

If additional energy is supplied to the surface of a material to generate carriers, the

material will become unbalanced. The quantity of majority carriers will not change
too much while the quantity of minority carriers will vary a lot. It will therefore have
diffusion of these excess charges from the surface to the volume. They will
recombine to tend towards their equilibrium density in the volume.

Suppose we illuminate an N-doped silicon rod with a source of sufficient intensity

to create electron-hole pairs. The holes are the minority carriers and their density in

the material is then (hole density at equilibrium).

If the illumination is removed, these additional holes recombine with the high
electron population of the N-doped silicon and their population decreases as a

17 / 82
function of x according to :

with : Diffusion length of the holes (it is about 10 µm for silicon).

The reasoning is similar for a P-doped material, where is the scattering length of
the minority electrons.

The diffusion current density J is proportional to the slope (the derivative) of the
carrier density, the elementary charge q and a proportionality factor D called the
diffusion constant.

The diffusion current density of the holes is given by :

N-doped silicon
P-doped silicon

Figure 19: Electron scattering in silicon

The electron scattering current density (A/cm²) given by :

This gives the total diffusion current density:

18 / 82
 The diffusion length and diffusion constant are used to determine the lifetime
of excess carriers :

For an N-doped material:

For a P-doped material:

is called the excess carrier life. For silicon, depending on the quality of
the material.

If there is both an electric field and a concentration gradient, the currents are
added :

IV.3 Relationship between mobility and diffusion constant

The mobility μ and the diffusion constant D are not independent. They are
linked by a relationship demonstrated in solid state physics and known as the
Einstein relationship:

(Einstein's Relation) and

With K: Boltzmann constant

T: temperature in degrees Kelvin

q: elementary charge of the electron.

UT = 26 mV at 27°C ambient temperature or 300°K

19 / 82
The relationship that often recurs in the study of semiconductors to the
dimension of a voltage. It is called the thermodynamic potential UT.

CHAPTER II

RADIATION-SEMICONDUCTOR INTERACTION

I. Photon-electron interaction

In the study of radiation-matter interactions, especially when there is an energy

exchange, the corpuscular representation of the radiation is more suitable than the
wave aspect.
The energy of a photon is given by :

or

Let's now consider the electron in the semiconductor. Due to the symmetry
properties of the crystal lattice, the energy levels of the bands are not equal in all
directions: there are privileged conduction axes. Semiconductors where the

20 / 82
direction corresponding to the maximum energy of the valence band and the
direction corresponding to the minimum energy of the conduction band coincide
are said to have a direct gap (GaAs, GaSb), the others are said to have an indirect gap
(Si, Ge, GaP).

Fig. 1: Nature of the gap in a semiconductor

The interaction of the radiation with the electrons in the semiconductor manifests
itself in three distinct processes.

- A photon can induce the jump of an electron from an occupied state of the valence
band to a free state of the conduction band, this is the fundamental absorption. This
process is used in radiation sensors.

- An electron in the conduction band can spontaneously fall back to an empty state
in the valence band with emission of a photon, this is the spontaneous emission. It
occurs without a preferred direction; it is isotropic. It is spontaneous in the sense that
it is impossible to exert an action that modifies it. This process is used in radiation
emitters such as light-emitting diodes.

- A photon present in the semiconductor can induce the transition of an electron

from the conduction band to an empty state in the valence band, with the emission
of a second photon of the same energy, this is the stimulated emission. The photon
produced by this de-excitation is inseparable from the incident photon, it is
incorporated into the incident wave, which it reinforces. Moreover, it clings in phase
to the incident photon and is emitted in the same direction, the stimulated emission
is not isotropic. This process is used in semiconductor lasers.

21 / 82
 Fig. 2: Different processes of electron-photon interaction

Fig. 3: Transitions between the extrema of the valence and conduction bands

In a direct gap semiconductor, the electronic transitions between the extrema of

the valence and conduction bands are vertical, and therefore radiative. In an indirect
gap semiconductor, the electronic transitions between the extrema are oblique and
therefore non-radiative.

However, in an indirect gap semiconductor, electrons can be excited vertically, i.e.

optically, from the top of the valence band to the central minimum of the
conduction band. The electrons excited in this way are then thermalized in the
minimum of the conduction band and can participate in the conduction
phenomena.

I.2 Semiconductors adapted to optoelectronics

The choice of a semiconductor for the realization of an optoelectronic

component depends on the use of this component. For example, the sensitivity

22 / 82
curve of the eye conditions the choice of materials for display systems, the solar
spectrum that of energy converters.

The human eye does not have the same vision during the day and at night. It is
centered in the green-yellow in day vision and in the blue-green in night vision.

Fig. 4: Sensitivity spectrum of the eye (photopic=day vision, scotopic=night

vision)

The radiation emitted by the sun corresponds to that of a black body at a

temperature of 6000°C. The intensity of radiation above the atmosphere is 1.35
kW/m2 , with a spectrum centred around = 480 nm. At the ground surface the
power density is only 0.9 kW/m2 , due to absorption mainly by ozone, water and
carbon dioxide.

23 / 82
 Fig. 5: Solar spectrum (on the ground = AM0, above the atmosphere = AM1)

Fig. 6: Semiconductor materials used in optoelectronics as a function of

wavelength

II. PHOTODETECTION

II.1 Photoexcitation

The excess of carriers is created with the help of radiation. Let the flux of
incident photons of E energy. If is the reflection coefficient of the semiconductor

24 / 82
for radiation of energy E, the flux of photons entering the semiconductor, we
have :

Photons are absorbed as they propagate through the material. Material is defined
as the relative change in radiation density per unit length.

Thus, the flux of photons of energy E on the abscissa x inside the semiconductor
is written as a function of the incident flux.

The reflection coefficient varies little with the energy for radiation whose energy
is close to the gap of the material on the other hand its value is very sensitive to the
angle of incidence of the radiation. It is minimum at normal incidence where it is
given by

n is the refractive index of the material and its relative dielectric constant. For all
semiconductors n varies between 3 and 4 so that R varies between 0.25 and 0.35.
Thus under the best conditions, 30% of the radiation is reflected by the
semiconductor.

If is zero, the radiation of energy E passes through the material without

attenuation, the material is transparent to this radiation (case if E<Eg).

II.2 Distribution of photocarriers in a semiconductor

When the semiconductor is photoexcited, the excess carriers created diffuse and
recombine. The evolution of this out-of-equilibrium regime is governed by the

25 / 82
continuity equations in which the generation rates and are the optical
generation rates.

where represents the photon flux in the semiconductor immediately

below the surface.

The electric field has two components, the one resulting from possible polarization (

) and the one resulting from the internal field resulting from the difference in
mobility of electrons and holes ( ).

At the instant t=0 when the semiconductor is photoexcited, the carrier densities
(electrons and holes) are equal. At the instant t≠0, this is no longer the case because
the electrons diffuse faster than the holes. The result is the existence of a space
charge zone and an internal electric field. This field is such that it accelerates the
slower carriers and brakes the faster ones.

The resolution of the previous system of equations can only be achieved by using a
third equation which is none other than the Poisson equation.

Solving these equations is quite complex. Following several simplifications, we

obtain for a material of type n, in a low excitation regime (n>>p) :

26 / 82
Also in a material of type p

In each case, the equation governs the evolution of the minority carrier density, the
majority carrier density being unaffected by the radiation.

Consider a semiconductor sample of thickness d subjected to no external voltage (

), and uniformly illuminated on one of its faces by a permanent radiation of energy

greater than the gap. Since the regime is stationary, the diffusion equation is
independent of time.

so

Fig. 7: Semiconductor illumination

If the diffusion length is entered, , the equation is reduced to :

The solution of this differential equation is the sum of the general solution of the
equation without second members and a particular solution of the equation with
second members.

27 / 82
The constants A and B are determined by the boundary conditions, which are given
by the surface recombination velocities in and :

and

The writing of the general solution is a bit heavy, we will only present the two
borderline cases of the thick sample and the thin sample.

● Thick sample ( and )

Excess carriers do not reach the rear side, which can therefore be rejected ad
infinitum. When , , , which leads to . The constant A is then
determined by the speed of recombination at the surface .
You get:

and

so

Fig. 8: Distribution curve of the photocarriers in a thick semiconductor

28 / 82
 ● Thin sample ( and )
Assuming negligible recombination rates. As a result, the second member of the
equation is reduced to a constant.

The only solution for A and B is:

is constant throughout the semiconductor. If recombination rates are not

negligible, the density of photocarriers decreases near the surfaces.

Fig. 9: Distribution curve of the light carriers in a thin semiconductor

CHAPTER III
Different structures

29 / 82
 THE PN JUNCTION

I. JUNCTION PN AT EQUILIBRIUM

I.1 Reminder of the properties of an N-doped semiconductor

An N-doped silicon crystal is richer in free electrons the higher the dopant
concentration. These electrons form a gas of mobile negative elementary charges.
The dopants are positively ionized. These positive charges are fixed and cannot
create an electric current.

Main features :

● Many free electrons: N D=nn

● As many positively ionised and fixed impurities: ND.

ni2
● Very few free holes: pn=
nn

● Electrical neutrality.

I.2 Reminder of the properties of a P-doped semiconductor

A P-doped silicon crystal is richer in free holes the higher the concentration of
dopants. These holes form a gas of mobile positive elementary charges. The dopants
are negatively ionized. These negative charges are fixed and cannot create an
electric current.

Main features :

● Numerous free holes: N A = p p.

● As many negatively ionised and fixed impurities: NA.

2
ni
● Very few free electrons: n p =
pp

● Electrical neutrality.

I.3 Construction of the PN junction

30 / 82
 A junction is made when a type N material and a type P material meet. As soon
as the junction is created, an equilibrium is created by the succession of the
following phenomena:

a. Broadcast

The free electrons on the N side are much more numerous than on the P side, so
there is a strong gradient at the junction, which leads to a strong diffusion of the
electrons from the N side to the P side. The holes on the P side are much more
numerous than on the N side, so there is a strong gradient at the junction, which
leads to a strong diffusion of the holes from the P side to the N side.

The current of diffusion is therefore due to the majority holders.

b. Recombination

The holes, which arrive on the N side, meet a large number of free electrons, with
which they recombine. The electrons, arriving from the P side, meet a large number
of holes, with which they recombine. In the vicinity of the junction, most of the
electrons and holes have recombined with each other and the crystal is no longer
electrically neutral:

- On the P side, only the negative ionised charges remain. Their density is -
qNA.

- On the N side, only the positive ionised charges remain. Their density is
+qND.

An electric field E is created around the junction due to the ionised charges. This
electric field is directed from the N side to the P side.

c. Conduction

The action of the field E on the free carriers is such that the electrons are pushed by
a force F = - qE in the opposite direction to that of the field. It drives the minority
holes from N towards P and the minority electrons from P towards N. The
conduction (or drift) current due to the presence of this electric field is therefore a
minority carrier current.

The electrons are thus pushed back to the N-side, the conduction current. The
holes are likewise repelled on the P side by this conduction.

31 / 82
 d. Équilibre

Equilibrium is reached when the diffusion tendency (due to the carrier gradient)
is exactly balanced by the conduction tendency (due to the electric field) which acts
in the opposite direction. The central area where the electric field exists is almost
free of free carriers. This is called the depletion zone or space charge zone (SZC). Its
width is of the order of a few microns.

Figure 1: PN junction at equilibrium

I.4 Charge Density, Electric Field and Equilibrium Potential

The PN junction consists of 2 neutral zones on the P and N sides (zero charge
density) and a depletion zone negatively charged on the P side and positively
charged on the N side. The charge density of the depletion zone is equal to the
dopant density times the elementary charge q: -qNA on the P side and +qND on the
N side.

The PN junction remains electrically neutral overall, and therefore the dimensions of
the depletion zone must satisfy the relationship:

Figure 2: Load density in a PN junction

32 / 82
The electric field E in the PN junction and the potential ( ) are obtained by
integrating the charge density according to the FISH relationship:

Gauss's law

with: Load density according to x,

Vacuum Permittivity ( F.m-1)

Relative Permittivity of Semiconductor ( )

● On the P side (WP < x < 0), the ionized atoms have lost a hole :

so

A first integration gives,

One second gives,

potential in x=−W p

● On the N side (0<x< WN), the ionized atoms have lost an electron:

33 / 82
A first integration gives,

One second gives,

potential in x=−W n

The electric field is directed according to x and given by .

The modulus of the electric field, which varies linearly with distance, is maximum for
, i.e. in the plane of the junction:

Figure 3: Electric field in a PN junction

By posing :

With : Total Space Load Zone

We get it,

34 / 82
The potential is the electrostatic potential and represents the height of the
potential barrier of the PN junction at equilibrium.

Figure 4: Potential in a PN junction at equilibrium

In the space charge zone, the potential varies quadratically with distance, it is
maximum at the terminals of the neutral zones of the junction.

For a semiconductor in equilibrium, the potential is related to the free carrier

density by the following relationships (Boltzmann statistic):

is the Fermi potential, it is constant and is only defined at equilibrium. The

position of the Fermi potential is determined as follows:

The value of the potential barrier is determined in the same way:

The height of the potential barrier varies little with the doping values. It is of the
order of 0.6V to 0.9V.

35 / 82
The existence of the electrostatic potential causes the energy bands in the space

charge zone to bend and, at equilibrium: the Fermi levels are aligned.

Figure 5: Alignment of Fermi levels

I.5 Width of the depletion zone

By integration of the electric field according to x, we obtain:

Hence

In the same way, you get on the P side:

Finally the width W of the depletion zone, at equilibrium, is valid:

II. POLARIZED PN JUNCTION

36 / 82
 The application of an external voltage V to the PN junction adds more or less to
the value of the electrostatic potential depending on its direction of application and
will cause a movement of charges in the junction, which is different in the two cases.

II.1 The forward-polarized PN junction

If a positive VD voltage is applied to the PN junction, the height of the potential

barrier decreases:

Figure 7: Potential in a forward-polarized PN junction

For VD>0, the electric field decreases in the depletion zone, which has the effect of
reducing the conduction current (due to the electric field) which can no longer
compensate for the diffusion current between the two sides of the junction. This
effect takes place for both types of free carriers, electrons and holes. As a result, a
current of electrons and a current of holes is created as soon as the PN junction is
forward-biased.

II.2 The reverse-polarized PN junction

A negative VD voltage applied to the PN junction increases the height of the

potential barrier:

37 / 82
The potential barrier is increased as well as the electric field in the depletion zone,
which has the effect of repelling holes more violently towards the P zone and
electrons more violently towards the N zone. As a result, there can be no current of
electrons or holes.

It is known that the density of minority carriers increases very strongly with
temperature, and therefore the reverse current, although very low, increases very
rapidly with temperature. This is not the case with direct current, because the
density of the majority carriers hardly varies with temperature (all impurities are
ionised at room temperature).

38 / 82
THE METAL-SEMICONDUCTOR CONTACT:

I. INTERFACE BETWEEN TWO DIFFERENT MATERIALS

The essential problem is the determination of the potential barrier that exists at
the various interfaces. The problem is in fact to position the interfaces in relation to
each other, the Valence and Conduction strips of the different materials. The
potential barrier is determined from the barrier that the electron has to pass
through to exit the material.

I.1 Output work

A conduction electron is subjected to a set of forces from all the ions constituting
the metal, the resultant of which is zero. As a result, this electron is free to move,
under the action of an electric field, for example. When an electron reaches the
surface of the metal, the compensation of the forces due to the ions is no longer
total, so that the electron is retained inside the metal. In order to extract it, energy
must be supplied to it. At absolute zero degrees, with all free electrons below the
Fermi level, the minimum energy required to extract an electron from the metal is
the energy required to extract an electron from the Fermi level. This quantity is

called the Metal Exit Work, it will be called , it is characteristic of the metal.

Figure 1: Metal removal work

Vacuum Level (NV) is the energy of an electron extracted from the body and without
initial velocity.

So the work involved in getting out of a metal is:

39 / 82
 Figure 2: Value of output work of different metals

If the work function is a specific parameter for a metal, it is not the same for a
semiconductor due to the fact that the Fermi level of the latter is fluctuating
according to the doping.

I.2 Electronic affinity

The semiconductor is characterized by another parameter which is the energy

that must be supplied to an electron located at the bottom of the conduction band,
to extract it from the semiconductor and bring it into the vacuum without initial
speed, it is the electronic affinity, it is called . We can define the same quantity for
an insulator.

Figure 3: Electronic Affinity of a Semiconductor

Semiconduct If Ge GaP GaA GaS SiO2

or s b
(eV) 4.01 4.13 4.3 4.07 4.06 1.1

40 / 82
 Figure 4: Electronic affinity value of different semiconductors

II. THE METAL-SECONDUCTOR CONTACT

Thus to extract a conduction electron from the metal and bring it into the

vacuum, it is necessary to supply it with energy ; to extract a conduction electron

from the semiconductor and bring it into the vacuum, it is necessary to supply it
with energy . Let us imagine that the metal and the semiconductor are separated
by a very small interval which is made to tend towards an interatomic distance. It is

necessary to provide energy to extract the electron from the metal, this electron
restores energy by entering the semiconductor. As a result, at the interface, the
potential barrier that the electron must cross to pass from the metal into the
semiconductor is given by :

The strip structure in the vicinity of the interface is therefore conditioned by the

possible difference in the metal ( ) and semiconductor ( ) output work.

II.1

41 / 82
 Figure 5: Metal-semiconductor contact with

Since the Fermi levels are aligned in the absence of contact, so that
thermodynamic equilibrium is achieved without any exchange of electrons.

This energy diagram is valid regardless of the type (n or p) of semiconductor, as long

as the condition is respected. The bands are horizontal, the system is said to be
in a flat band regime.

II.2

Figure 6: Metal semiconductor contact with >

When the two materials are brought into contact, as the work function of the
semiconductor is less than that of the metal, the electrons pass from the
semiconductor into the metal. The system stabilizes at an equilibrium regime
defined by the alignment of Fermi levels. The resulting energy pattern is different
depending on the type of semiconductor.

a. N-type semiconductor

- At thermodynamic equilibrium

The electrons that pass from the semiconductor to the metal cause energetic
changes in each of the materials.

42 / 82
 Figure 7: Metal semiconductor contact with > at thermodynamic equilibrium

In the semiconductor, a depletion zone is created, the ND+ ionized donor ions are
no longer compensated by electrons, a positive space charge zone appears.

On the other hand, the Fermi-level conduction band distance, which reflects the
electronic population, is greater in the vicinity of the interface than in the neutral

region of the semiconductor > . Since the Fermi level is horizontal, this results in
an upward curvature of the bands. In the metal, an accumulation of electrons
appears at the surface.

This double space charge is associated with an electric field and a diffusion potential

which, as in the case of the PN junction, balances the diffusion forces and
determines the state of equilibrium.

The number of positive charges developed in the semiconductor is equal to the

number of negative charges developed in the metal. The latter are accumulation
charges, the density of states in the metal being of the order of 1022 cm-3, these
charges develop on the surface of the metal. In the semiconductor, these charges
are depletion charges due to donor ions, the density of these donors being typically
1016 to 1018
of the order of cm-3, this space charge is relatively spread within the
semiconductor, resulting in band curvature.

- Polarization: V< 0

43 / 82
 Figure 8: Metal semiconductor contact with > under bias V<0

The conduction band of the semiconductor rises by , the curvature decreases.

The semiconductor barrier → metal rises and decreases while the semiconductor
barrier métal→ remains unchanged. The equilibrium is broken, the electrons diffuse
from the semiconductor to the metal and create a current I from the metal to the
semiconductor. If the bias voltage is increased further, the flat-band regime is

reached when .

- Polarization: V> 0

Figure 9: Metal semiconductor contact with > under V>0 bias

The conduction band of the semiconductor is lowered, increasing the height

of the barrier against electron scattering. So there's no more current. The structure is

reverse-biased. The metal-semiconductor(n) structure with constitutes a

rectifier contact. It is a Schottky diode.

44 / 82
 b. P-type semiconductor

Figure 10: p-type metal-semiconductor contact with > at thermodynamic

equilibrium

When the two materials are brought into contact, the electrons diffuse from the
semiconductor to the metal until the Fermi levels are aligned. A negative space
charge zone appears in the metal and a positive space charge zone appears in the
semiconductor. It is also accompanied by a curvature of the strips.

The space charge in the metal is due to an accumulation of electrons on the surface.
In the case of the space charge zone in the semiconductor, the upward curvature of
the strips shows that the Fermi-valence band level distance decreases near the
surface, indicating an accumulation of holes. Like electrons in metal, holes in the
semiconductor are free charges that will accumulate on the surface of the
semiconductor. The spread of positive charges is less than in an N-type
semiconductor.

- Polarization: V< 0

Figure 11: P-type metal-semiconductor contact with > under bias V<0

The essential difference between this metal-semiconductor (p) structure, and the
previous metal-semiconductor (n) structure, lies in the fact that the space charge in

45 / 82
the semiconductor corresponds to an accumulation regime and not to a depletion
regime. As a result, there is no carrier void, and therefore no insulating zone at the
interface. Thus when this structure is biased, the applied voltage is no longer
localized in the space charge region of the semiconductor, as in the previous case,
but distributed throughout the semiconductor, which is more resistive than the rest
of the structure. At the interface, the arrival or departure of a hole in the
semiconductor is immediately compensated for by the arrival, or departure, of an
electron in the metal. As a result, current flows freely in both directions at the
contact. The contact is ohmic.

- Polarization: V> 0

Figure 12: P-type metal-semiconductor contact with > under V>0 bias

II.3

Figure 13: Semiconductor metal contact with <

When the two materials are brought into contact, the work function of the metal
being less than that of the semiconductor, the electrons leave the metal and enter
the semiconductor. The system stabilizes at an equilibrium regime defined by the

46 / 82
alignment of Fermi levels. The resulting energy pattern is different depending on the
type of semiconductor.

a. N-type semiconductor

Electrons passing from the metal into the semiconductor cause a localized
electron deficit on the surface of the metal, and in the semiconductor a sparsely
spread accumulation zone. This results in a downward curvature of the valence and
conduction bands.

Figure 14: Semiconductor metal contact (n) with < at thermodynamic

equilibrium

If the structure is biased, the bias voltage is distributed throughout the

semiconductor. Any electron that arrives at the interface in the semiconductor
passes freely through the metal and vice versa. The contact is ohmic.

Figure 15: Metal semiconductor contact with < under bias

47 / 82
 b. P-type semiconductor

The electrons pass from the metal into the semiconductor. A deficit of electrons
appears on the surface of the metal. In the semiconductor, the electrons coming
from the metal recombine with the holes creating a depletion zone due to the
presence of NA- ions which are no longer compensated for by the holes. This creates a
space charge zone in the semiconductor. The system evolves until the resulting
electric field and scattering voltage stop the diffusion of electrons.

Figure 16: Semiconductor metal contact (p) with < at thermodynamic

equilibrium

The polarisation voltage is located at the depletion zone, which is insulating. If the
metal semiconductor voltage is negative, the conduction and valence bands rise, the
curvature of the bands increases. The potential barrier is increased, the current does
not flow. The structure is reverse biased.

Figure 17: Metal semiconductor contact (p) with < under bias

If the bias voltage of the semiconductor is positive, the stripes are lowered and
the potential barrier that the holes must pass through to get through the metal is

48 / 82
reduced. The current flows freely and the structure is polarized in the forward
direction.

The p-type metal semiconductor structure with is therefore a rectifier

contact, it is a Schottky diode.

In summary, the metal-semiconductor contact is ohmic or rectifier depending on

the difference in output work and the type of semiconductor.

With > , the metal-semiconductor contact (N) is rectified,

the metal-semiconductor (P) contact is ohmic.

With < , the metal-semiconductor (N) contact is ohmic,

the metal-semiconductor contact (P) is rectified.

49 / 82
III. BALANCED SPACE LOAD ZONE

III.1 Electrical field and potential

For the following calculations, only the case of an N-type semiconductor with >
. Moreover, the semiconductor is considered to be homogeneous.

The excess carrier density is equal to . Thus the charge density in the
semiconductor is written as :

In the space load zone (W): 0 < x < W

Outside the space load zone: x > W

The electric field E is obtained by integration of the charge density according to the
relationship:

being the charge density

The electric field is maximum at and is worth:

The potential V is obtained by integration of the electric field because:

We get it,

50 / 82
 The diffusion voltage results from the difference in the output workings of the
metal and the semiconductor :

. This voltage corresponds to the potential difference between the surface

of the semiconductor and its volume, i.e. at the terminals of the space charge zone.

III.2 Width of space load zone

From the diffusion voltage , the width of the space charge zone W at equilibrium is
easily determined.

51 / 82
 THE METAL-INSULATOR-SEMICONDUCTOR CONTACT:

THE MOS CAPACITY

I. METAL-VIDE-SEMICONDUCTOR STRUCTURE

Either an output working metal and an output working semiconductor

separated by an insulator consisting of a relatively small void thickness.

If the metal and the semiconductor are electrically connected, they form a single
thermodynamic system, their Fermi levels align and a potential difference is created
by the differences in output work.

This potential difference is associated with an electric field and a charging zone
(ZCE).

Fig. 1: Balancing a Metal-Insulator-Semiconductor System

The structure behaves like a planar capacitor with a constant voltage between the
armatures. The charge depends on the capacitance between the armatures. If the
distance decreases, the capacitance increases and so does the charge.

52 / 82
The total charge developed in the metal is equal and opposite to the total charge
developed in the semiconductor. These charges are superficial in the metal and
extend further into the semiconductor.
- The space charge in the metal results from the variation of the electron
density in the vicinity of the surface.
- The space charge in the semiconductor results from the variation in the
density of free carriers, electrons and holes in the vicinity of the surface.
The nature of the space charge and the curvature of the tapes are a function of the
type of semiconductor and the difference in the output work.

Consider a P-type semiconductor.

If the work function of the metal m is greater than the work function of the
semiconductor sc, positive charges develop in the semiconductor and negative
charges in the metal. The positive charges are due to the accumulation of holes (free
charges) that will accumulate on the surface of the semiconductor. Negative
charges are due to the accumulation of electrons on the surface of the metal.

⇒ The structure is in an accumulation regime

Fig.2: Metal-Vacuum-Semiconductor System with m > σχ

When the work function of the metal m is equal to the work function of the
semiconductor sc, the diffusion voltage is zero, no charge appears.
⇒ The structure is in flat strip mode

53 / 82
 Fig. 3: Metal-Vacuum-Semiconductor System with m = sc

If the work function of the metal m is smaller than the work function of the
semiconductor sc and the potential gradient is relatively small, negative charges
develop in the semiconductor and positive charges in the metal. The electrons
coming from the metal recombine with the holes creating a depletion zone free of
mobile charges and negatively charged due to the presence of the fixed ions NA- .
The electron density (minority carriers in type P) is less than . Acceptor ions make
up the bulk of the space charge zone.

⇒ The structure is in a depletion regime.

Fig.4: Metal-Vacuum-Semiconductor System with m < σχ

If the potential gradient becomes larger, the hole deficit increases, the electron
density increases as well. When the electron density becomes greater than , the

54 / 82
semiconductor becomes N-type near the surface. Most of the charge is still due to
the acceptor ions.

⇒ The structure is in a reversal mode.

Fig.5: Metal-Vacuum-Semiconductor System with m << sc

If the potential gradient is such that the electron density exceeds the hole

density ( ), the charge becomes conditioned by the presence of free

electrons near the surface.

⇒ The structure is in a strong inversion regime

Fig.6: Metal-Vide-Semiconductor System with m << sc

As far as the space load zone is concerned, its extent depends on the regime in
which the structure is located. In metal, the space load zone is very small because

55 / 82
the density of states in the material is very high (1022 cm-3). In semiconductors,
when the structure is in the accumulation and inversion regime, the space charge
zone is small (a few nm) because it is due to moving carriers. In depletion regime,
the space charge zone is due to fixed charges (ions), so it is larger.

II. METAL-INSULATOR-SEMICONDUCTOR STRUCTURE

In the MIS structure, the gap between the metal and the semiconductor is filled
with an insulator (SiO2).

As in the previous study, only type P is considered. The N type study being exactly
the same (electrons are replaced by holes).

It is also assumed that the insulation is perfect, i.e. there are no fillers present in the
material.

Fig.7: Metal-Insulator-Semiconductor System with m < σχ

The potential barrier extends partly into the semiconductor, partly into the
insulator. The shape of the potential barrier in the semiconductor depends on the
distribution of the space charge zone (strip curvature). In the insulator, which is
considered perfect, there is no charge, so the potential varies linearly.

The different operating regimes presented above apply in exactly the same way.
These diagrams are a function of the potential gradient that exists between the
metal and the semiconductor. This gradient has two origins:

- The difference in exit work,

- The difference in polarization potential.

56 / 82
III. METAL-INSULATOR-SEMICONDUCTOR STRUCTURE UNDER POLARIZATION

In order to highlight only the influence of polarization, the structure is

considered to be in flat band (m = sc) regimes with zero polarization.

Under the action of polarization, the structure evolves from an accumulation

regime to an inversion regime, through flat band and depletion regimes.

In the general case, i.e., m different from sc, the flat strip speed is not obtained for
Vg=0 but for :

In reality, the presence of localized charges on the surface also changes the potential
(positive charges).

with Qss : Interface load density (C.cm-2),

Ci Insulation capacity per unit area (F.cm-2).

Fig.8: Metal-Insulator-Semiconductor system with negative bias (Vg<0)

Under negative polarization, the structure is in an accumulation regime.

57 / 82
 Fig.9: Metal-Insulator-Semiconductor system with zero polarization (Vg=0)

Under zero polarization, we find the flat band regime.

When the polarization becomes slightly positive, the depletion regime sets in.

Fig.10 : Metal-Insulator-Semiconductor system under polarization (Vg>0)

If the voltage increases further, the structure will invert.

Fig.11 : Metal-Insulator-Semiconductor system under polarization (Vg>>0)

58 / 82
Even under polarization, the Fermi level remains constant throughout the
semiconductor. This reflects the absence of current resulting from the presence of
the insulator.

If the intrinsic Fermi level (undoped silicon) is positioned on the previous

diagrams, it is easier to visualize the inversion phenomenon, i.e. the change from P
to N type.

Fig.12: Highlighting the inversion phenomenon from the intrinsic Fermi level

Indeed, if represents the position of the Fermi level of the intrinsic semiconductor
(i.e. about the middle of the band gap) when :

- we got a type N,

- we got a type P.

IV. EQUILIBRIUM SPACE LOAD ZONE

IV.1 Space load

The potential at the insulation-semiconductor junction, i.e. the potential at the

abscissa 0, is called the surface potential VS. It is this potential that conditions the
different operating regimes of the structure.

59 / 82
 Fig.13: Metal structure - Insulator-Semiconductor

- For VS < 0, the hole density on the surface is greater than in volume, the structure
is in an accumulation regime.

nS (area) < n0 (volume)

pS (area) > p0 (volume)

- For VS=0, the structure is in flat strip mode.

nS (area) = n0 (volume)

pS (area) = p0 (volume)

- For , the surface electron density increases and the hole density
decreases, but the semiconductor remains P-type, the structure is in depletion
regime.

nS (area) > n0 (volume)

pS (area) < p0 (volume)

pS > nS

- For , the electron density on the surface is greater than the hole density,
the structure is in an inversion regime.

nS (area) > n0 (volume)

pS (area) < p0 (volume)

60 / 82
 pS < nS

If then pS = nS = ni (surface intrinsic semiconductor).

- For , the electron density at the surface is even greater, the structure is in
a strong inversion regime.

pS << nS

One can also study the evolution of the space load in the different operating
regimes.

- For , the structure is in an accumulation regime, therefore is positive.

The space charge is positive and increases exponentially with the bias voltage.
Since this charge is a storage charge, its spatial extension is very small.

- For , the structure is in depletion mode and then in low inversion

mode, so is negative.

The charge varies in . This charge being due to the acceptor ions its spatial
extension is important.

- For , the structure is in a strong inversion regime.

The space charge increases again exponentially with the applied voltage. This
charge constitutes the inversion charge, which is added to the depletion charge,
but with a much smaller spatial extension.

61 / 82
 Fig.14: Space charge in the semiconductor as a function of surface potential

(Silicon type P with NA=4.1015 cm-3)

In low inversion regime, the space charge is conditioned by the depletion

charges because the inversion charges are negligible compared to the acceptor
density (NA). In high inversion regime, on the contrary, the charge is conditioned
by the electrons whose surface density is much greater than that of the
acceptors. The threshold VT of the strong inversion regime corresponds to the
surface potential VS for which n=NA. This condition leads to :

and so as we know ,

you get .

When the gate voltage is increased beyond the value causing VS=VT the electron
concentration at the interface increases exponentially while the depletion charge

density increases in . As a result, the variation of the latter becomes negligible.

When the strong inversion regime is reached the width of the space charge zone
remains practically constant.

62 / 82
 IV.2 Potential and electric field

The analytical calculation can be further developed in the case of the depletion
regime, on the assumption, already used for the PN junction and the Schottky diode,
of total depletion. That is to say, assuming the carrier empty space load zone. Indeed
the calculation is relatively complicated in the case where the space charge results
from an accumulation regime because the potential is a function of the charge
distribution and the charges being free carriers, their distribution is a function of the
potential.

The VG potential applied to the metal is distributed between the insulation ( Vi) and
the space charge zone (Vs).

In the insulation, there are no fillers, so;

In x=0, we have

In x=-d, we have

You get:

gold and

so

Since the electric field is derived from the potential, we have , in x=0 :

In the semiconductor, assuming total depletion in the space charge zone (W) and
homogeneous doping, the depletion charge density is written as ;

63 / 82
The fish equation is of the form:

In the same way, the following potential and electric field are obtained.

in x=0, V=VS,

The space load in the depletion zone is given by

Fig. 14: Potential and electric field variations

The electric field at the interface has a discontinuity resulting from the
discontinuity of the dielectric constant.

64 / 82
 IV.3 Threshold voltage

The threshold voltage of the structure is defined as the bias voltage of the metal
electrode required to establish the strong inversion regime. It is therefore the value

of the voltage VG causing .

We know that:

and thus

Insofar as the threshold of strong inversion corresponds to the regime for which
inversion loads become preponderant, it can be said that below the threshold the

loads are essentially depletion loads. So that for , we write

Thus

The threshold voltage VT of the structure is given by the value of VG for , i.e. ;

The threshold voltage corresponds in fact to the value of the voltage VG from which
it is no longer possible to neglect in the load of esp

65 / 82
66 / 82
 HETEROJUNCTIONS

INTRODUCTION

As early as 1951, William Shockley proposed the use of an abrupt

heterojunction in a bipolar transistor. Since then, heterojunctions have been used in
many fields: HBTs, light-emitting diodes, laser diodes, photodetectors, solar cells,
etc...

I. ENERGY BAND DIAGRAM

A heterojunction is obtained when two different semiconductor materials are

combined. Depending on the type of doping used for semiconductors, two
heterojunctions differ:

- Isotype heterojunctions, where both semiconductors have the same type

of doping.

- Anisotype heterojunctions in which the two semiconductors are doped

differently.

When two different semiconductors are in contact, a potential barrier appears at the
interface,

where and represent the electronic affinities of semiconductors. The shape of

this barrier and the corresponding energy diagram depend on the gaps in the
semiconductors and their doping.

67 / 82
 Fig. 1- Energy diagram of separate semiconductors

I.1 Energy diagram away from the junction

When the two semiconductors are brought into contact, they exchange
electrons to align the Fermi levels. This exchange takes place in the vicinity of the
junction and reveals, as in the PN junction, a space charge zone to which is
associated a potential barrier (the diffusion voltage Vd) which stops the diffusion of
the carriers and defines the state of equilibrium.
But first let's consider the energy diagram far from the junction, in the neutral
region of each of the semiconductors.
The difference in potential energy between the electron in vacuum in the vicinity
of semiconductor 1, and the electron in vacuum in the vicinity of semiconductor 2 is :

As a result, the potential difference between the two semiconductors, i.e. the
diffusion voltage, is given by :

If > leads to a negative potential, the potential of semiconductor 2 is lower than

that of semiconductor 1. A positive potential difference is established between the
semiconductor with low work function and the semiconductor with high work
function.

-Vd

68 / 82
 Fig. 2- Energy diagram away from the interface

The difference in energy between the conduction bands of the neutral regions:

If the difference of the gaps is different from , the complement results in a

difference in valence band energy. The result is :

The difference in energy of the valence bands is therefore given by :

The differences in the energy of the conduction and valence bands of the neutral
regions are composed of two terms, one of which is specific to the intrinsic
properties of the materials and the other is a function of their respective doping.

If we pose,

we get the following relationships:

I.2 Energy diagram in the vicinity of the junction

69 / 82
 Due to the difference in work function, the electrons diffuse from the
semiconductor with lower work function to the other. This scattering causes a
positive space charge region to appear in the lower work function semiconductor
and a negative space charge region to appear in the other. As in the homojunction
(PN junction), the diffusion voltage increases to the value that stops diffusion and
defines the steady state.

Due to the continuity of potential at the interface , we have..,

In the vicinity of the interface, the energy diagram varies according to the nature of

the semiconductors, i.e. from and to and according to their doping, i.e. the
difference in their output work.

a.

In the absence of polarization, we have,

Since the output works of the semiconductors are equal, the potential difference

is zero, so that the differences in energy of the conduction and valence bands
are the same at the interface the structure is in a flat band regime.

The condition leads to the existence of four cases depending on whether

and are negative or positive, i.e. depending on the relative values of the electronic
affinities and gaps of the semiconductors.

The condition results in

The condition results in

70 / 82
In each case, different types of semiconductor doping can be represented by
moving the Fermi level vertically. In this way, heterostructures of NN, NP, PP or PN
types can be realized.

Fig. 3- Energy diagram of a heterojunction between two different semiconductors

having the same output work function

b.

If the output works are different, the electrons scatter from the semiconductor
with low output work to the other. A space charge zone is established in the vicinity
of the interface.

We will assume that the work function of semiconductor 2 is smaller than the work

function of semiconductor 1 ( ). The electrons diffuse from semiconductor 2 to

semiconductor 1, and vice versa for the holes. A positive space charge appears in
semiconductor 2 and a negative space charge appears in semiconductor 1. This
space charge has a completely different nature and spatial extension depending on
the type of each semiconductor.

71 / 82
In semiconductor 1 the space charge is negative. If this semiconductor is N-type, this
space charge is due to an increase in electron density in the vicinity of the interface.
The semiconductor is in an accumulation state. The space charge is therefore small.

If the semiconductor is P-type, the electrons diffusing from semiconductor 2

recombine with the holes in semiconductor 1, and a space charge zone resulting
from the acceptor ions not compensated by the holes appears. The space charge
results from the disappearance of the holes, semiconductor 1 is in depletion regime.
The space charge is therefore more extensive.

In semiconductor 2, the space charge is positive. If the semiconductor is N-type, a

depletion regime with a certain spatial extension of the charge density is established
in the vicinity of the interface. If the semiconductor is P-type, an accumulation
regime is established.

The different cases, shown in the following figure, have particular characteristics.

In figure a. the electrons diffuse from semiconductor 2 to semiconductor 1 and vice

versa for the holes.

In figure b. the electrons diffuse from semiconductor 2 to semiconductor 1. Due to

the holes cannot diffuse from semiconductor 1 to semiconductor 2.

In figure d. the holes diffuse from semiconductor 1 to semiconductor 2. Because of

the electrons cannot diffuse from semiconductor 2 to semiconductor 1.

In figure c. given the barriers and neither electrons nor holes can diffuse.
However, some carriers can diffuse either by crossing the potential barrier by
thermal agitation or by tunneling. Structure equilibrium is achieved in each of the
semiconductors by transferring carriers from the interface to the volume and vice
versa.

In each case, the diffusion voltage is partially established in each of the

semiconductors.

72 / 82
Fig. 4- Energy diagram of a heterojunction between two different semiconductors

73 / 82
II. HETEROJUNCTION AT THERMODYNAMIC EQUILIBRIUM

The calculation of the potential distribution in the vicinity of the interface is done
by integration of the fish equation. This calculation is relatively complicated in the
case where the space charge results from an accumulation regime. On the other
hand, the calculation can be developed in depletion regime in the case where the
space charge zone is devoid of free carriers.

In order for the space charge depletion zone to be depleted in each of the
semiconductors, the semiconductors must be of different types.

Fig.5: Heterojunction Ge(n)-GaAs(p)

Poisson's equation is written as follows

In semiconductor 1 :

74 / 82
 Fig.6: Electric field and potential at the interface of a heterojunction Ge(n)-GaAs(p)

In semiconductor 2 :

Since we have the same amount of load on each side of the junction, we have

Moreover, by considering the continuity of the potential in x=0, we can determine,

The total space load zone is therefore :

75 / 82
III. POLARIZED HETEROJUNCTION

The various energy diagrams show that there are two types of discontinuity for
the conduction and valence bands.

The first corresponds to the case where the diffusion voltage is added to the energy
difference of the bands considered. This discontinuity is called pseudo-continuity
(valence band in Fig. 4.b). The second type corresponds to the case where the
diffusion voltage and the gap discontinuity play opposite roles. This discontinuity is
called strong discontinuity (conduction band in Fig. 4.b).

The electron and hole currents can only be significant when there is a strong
discontinuity in the strip in play. In this case, the diffusion voltage and the barrier
height resulting from the difference in the gaps partially compensate each other.
Therefore, the carriers responsible for the current at a heterojunction are those that
see a barrier with a high discontinuity.

As in the case of metal-semiconductor contact, the current is conditioned at the

interface by the thermal emission mechanism of the carriers and in each of the
semiconductors by the diffusion of these carriers. These two processes constitute
two different forms of the same current and are therefore conditioned by each other.

III.1 Thermoelectronic emission model

a. Thermal electron emission current

In the absence of polarisation, the thermoelectronic emission current is the same

in both directions, the resulting current is zero.

If semiconductor 1 is biased with respect to semiconductor 2 by a voltage

. This bias voltage is distributed between the two space charge regions.

76 / 82
The semiconductor barrier 1 → σεμιχονδυχτορ 2 increases and the semiconductor
barrier 2 → σεμιχονδυχτορ 1 decreases, the equilibrium is broken, a flow of electrons
passes from semiconductor 2 to semiconductor 1, resulting in an electric current in
the other direction.

The semiconductor barrier 1 → σεμιχονδυχτορ 2 decreases, and the semiconductor

barrier 2 → σεμιχονδυχτορ 1 increases, the current flows in the other direction.

The resulting energy diagrams are shown in the following figure for positive and
negative polarizations.

Fig.7: Polarised heterojunction

The variation of the current with the voltage is a function of the nature of each space
charge zone, i.e. the type of each semiconductor.

77 / 82
Semiconductors: P-P :

The electron current is negligible.

Semiconductors: N-N :

There is a storage charge in semiconductor 1 (narrow gap charge zone and

conductive). There is a depletion charge in Semiconductor 2 (wide space charge
zone and insulating). The bias voltage is established in the depletion space charge
zone.

with

If , the structure is reverse biased .

Yes, the structure is polarized in the passing direction.

Semiconductors: P-N :

We've got a depletion charge in each of the two semiconductors. The voltage V
builds up in each of the space charge regions.

with

Semiconductors: N-P :

We have a storage charge in each semiconductor. The contact between the two
semiconductors is ohmic.

b. Hole thermal emission current

78 / 82
 There is a strong discontinuity in the valence band only in the case of . In
the absence of polarization, the hole thermal emission current is the same in both
directions, the resulting current is zero.

If semiconductor 1 is biased with respect to semiconductor 2 by a voltage

. This bias voltage is distributed between the two space charge regions.

The semiconductor barrier 1 → σεμιχονδυχτορ 2 decreases and the semiconductor

barrier 2 → σεμιχονδυχτορ 1 increases, the equilibrium is broken, a flow of electrons
passes from semiconductor 1 to semiconductor 2, resulting in an electric current in
the other direction.

Semiconductor barrier 1 → σεμιχονδυχτορ 2 increases, and semiconductor barrier 2 →

σεμιχονδυχτορ 1 decreases, the current flows in the other direction.

The resulting energy diagrams are shown in the following figure for positive and
negative polarizations.

79 / 82
 Fig. 8: Polarised Ge-GaAs heterojunction

In contrast to the electron current, the hole current can only be large if both
semiconductors are P-type.

Semiconductors: P-P :

The space charge zone is then an accumulation zone in Semiconductor 1 and a

depletion zone in Semiconductor 2. The bias voltage is established only in the space
charge zone of semiconductor 2.

Yes, the structure is polarized in the passing direction.

If , the structure is reverse biased .

Semiconductors: N-P :

We've got a depletion charge in each of the two semiconductors. The voltage V
builds up in each of the space charge regions.

Semiconductors: P-N :

We have a storage charge in each semiconductor. The contact between the two
semiconductors is ohmic.

Semiconductors: N-N :

The hole current is negligible.

III.2 Dissemination model

80 / 82
 The calculation of the current in the diffusion model is a function of the
respective types of each of the semiconductors. If the heterojunction is carried out
with two semiconductors of the same type, one of them is necessarily in
accumulation mode and the other in depletion mode. The bias voltage is then
essentially established in the depletion zone.

If both semiconductors are of the same type, the carriers that diffuse into the
space charge region are majority carriers. The minority carrier current can therefore
be neglected, so that the majority carrier current constitutes the total current.

If, on the contrary, the heterojunction is of the PN type, the carriers that are
thermally emitted at the interface from one semiconductor become minority
carriers in the other. As a result, they no longer make up the entire current.

The same current expressions are obtained in the carrier thermal emission model
and in the diffusion model, with different values of the carrier velocities.

81 / 82
82 / 82