Basics of Semiconductors

Carrier Statistics, Energy Bands, Poisson’s Equation, Drift & Diffusion

1 Electrical Properties of Semiconductors

Before going into band theory and doping, it helps to see the “headline” electrical behavior of a
semiconductor: how well it conducts, and how that conduction depends on carrier density, mobility,
doping, and temperature. Every quantity introduced here – carrier concentrations n0 , p0 , mobility µ,
intrinsic concentration ni , doping ND , NA – is defined properly in the sections that follow; this section
previews how they combine into a single measurable number: resistivity.

1.1 Conductivity and Resistivity

Free electrons (concentration n0 ) and holes (concentration p0 ) both carry current, each moving with
its own drift mobility (µn for electrons, µp for holes) under an applied field. The total conductivity
σ sums both contributions:

σ = q (n0 µn + p0 µp ) [S/cm or Ω−1 cm−1 ]

Resistivity ρ is simply the reciprocal:

1 1
ρ= = [Ω-cm]
σ q(n0 µn + p0 µp )

Special cases
ˆ Intrinsic material: n0 = p0 = ni , so σi = qni (µn + µp ).

ˆ n-type (moderate doping, n0 ≈ ND ≫ p0 ): σ ≈ qND µn .

ˆ p-type (moderate doping, p0 ≈ NA ≫ n0 ): σ ≈ qNA µp .

In each extrinsic case, conductivity is dominated by the majority carrier, so it is set almost entirely
by the doping concentration and that carrier’s mobility. (Doping, majority/minority carriers, and
ni are covered in Secs. 1–4 below.)

1.2 Resistivity versus Doping Concentration

Since µn and µp themselves fall as doping increases (more ionized-impurity scattering), resistivity does
not scale as a perfect 1/N curve, but it still decreases monotonically as doping rises.

1
Semiconductor Device Fundamentals Consolidated Notes

Resistivity of doped silicon at 300 K

n-type
101 p-type

Resistivity ρ (Ω-cm)
100

10−1

10−2

10−3 14
10 1015 1016 1017 1018 1019 1020
Doping concentration (cm−3 )

Reading the curve

ˆ At a given doping level, p-type silicon is more resistive than n-type, since µp < µn (holes move
more sluggishly than electrons in Si).

ˆ Resistivity falls by roughly four orders of magnitude as doping rises from 1014 to 1019 cm−3 .

ˆ This ρ-vs-N curve is exactly how a measured sheet resistance (see the Sheet Resistance discussion
later in these notes) is converted into an equivalent doping concentration in practice.

1.3 Temperature Dependence

Unlike a metal, whose resistivity rises with temperature (more lattice scattering, fixed carrier count),
a semiconductor’s resistivity can fall with temperature, because ni itself grows rapidly with T :

ˆ Intrinsic / lightly-doped region: as T increases, ni increases exponentially (recall ni rises 100×

from 27°C to 100°C), so σi = qni (µn + µp ) rises and ρ falls sharply.

ˆ Extrinsic (moderate temperature) region: n0 ≈ ND (or p0 ≈ NA ) is essentially fixed by

doping, so σ changes only through the slow decrease of mobility with T – resistivity rises gently,
metal-like.

ˆ Freeze-out (low temperature) region: donors/acceptors are not fully ionized, so carrier con-
centration – and hence conductivity – drops as T decreases.

ni (T ) ∝ T 3/2 e−Eg /2kT

 
σi (T ) = qni (T ) µn (T ) + µp (T ) ,

2 What Is a Semiconductor?
Semiconductors get their name because they conduct current better than insulators but not as well
as conductors. In the discussions that follow, unless stated otherwise, we assume equilibrium con-
ditions: no illumination, no radiation, no magnetic field, uniform temperature (room temperature,
300 K), and no externally applied voltage or current (electric field = 0).

2
Semiconductor Device Fundamentals Consolidated Notes

2.1 Free Electrons and Holes

Silicon has four valence electrons, each shared in a covalent bond with a neighboring atom. Two
mechanisms move charge through the lattice:

1. Free electrons – an electron breaks loose from a bond and travels through the crystal, carrying
charge −q.

2. Holes – when an electron leaves, a vacancy remains. A neighboring valence electron can hop into
that vacancy, which looks like the vacancy itself (the “hole”) moving in the opposite direction. A
hole behaves like a fictitious particle of charge +q.

q = 1.6 × 10−19 C (elementary charge)

A free electron and a hole that meet can recombine and annihilate each other.

e− (free electron)
Si Si Si

Si Si Si
hole

3 Energy Band Diagram

An energy band diagram shows how electron energy levels are arranged in a material. The horizontal
axis is geometric distance; the vertical axis is electron energy. Two bands matter most:

ˆ Valence band (EV at its top edge) – electrons bound to atoms.

ˆ Conduction band (EC at its bottom edge) – electrons free to move and conduct.

The band gap Eg = EC − EV between them determines whether a material is a conductor (no gap /
overlapping bands), a semiconductor (small gap, Eg ≈ 1 eV), or an insulator (large gap, Eg ≈ 9 eV).

Insulator
Conductor Semiconductor EC empty
Cond. band EC empty
overlap Eg ≈ 9eV
Eg ≈ 1eV

Val. band
EV filled EV filled

4 Intrinsic Semiconductor
In a pure (intrinsic) crystal, electrons occupy the valence band up to EV . Thermal vibration can give
an electron enough energy (E ≥ EC ) to jump to the conduction band, leaving a hole behind in the
valence band. The excess energy E − EC becomes the electron’s kinetic energy.

3
Semiconductor Device Fundamentals Consolidated Notes

Thermal generation: electron–hole pairs

EC

Eg

EV

Key facts
ˆ Every thermally freed electron leaves one hole, so ni = pi (intrinsic concentration).

ˆ Silicon: Eg = EC − EV = 1.12 eV at 300 K.

ˆ At 300 K, ni ≈ 1010 cm−3 , versus ≈ 5 × 1022 silicon atoms/cm3 .

ˆ Raising temperature from 27°C to 100°C increases ni by roughly 100×.

5 Extrinsic Semiconductors (Doping)

Doping is the deliberate introduction of impurity atoms into silicon to control the number of free
electrons or holes, producing an extrinsic semiconductor.

5.1 n-type: Donor Doping

Donor atoms (phosphorus, arsenic, antimony) have one extra valence electron. This electron is weakly
bound at an energy level ED just below EC , so thermal energy easily lifts it into the conduction band.
Because all bonds remain intact, no hole is created by this departure – the donor atom is left ionized
(net +q, immobile).

n-type: donor electrons rise to EC

EC

ED

EV

5.2 p-type: Acceptor Doping

Acceptor atoms (boron, gallium, indium) have one fewer valence electron, creating a vacancy (hole).
They ”accept” an electron from a nearby silicon bond, becoming ionized (net −q, immobile) while the
resulting hole is free to move. This corresponds to an acceptor energy level EA just above EV .

p-type: valence electrons rise to EA , leaving holes

EC

EA
EV

4
Semiconductor Device Fundamentals Consolidated Notes

Carrier concentration relations

n0 p0 = n2i (mass-action law, holds regardless of doping)

n2i
n-type: n0 ≈ ND , p0 = (small)
n0
n2i
p-type: p0 ≈ NA , n0 = (small)
p0

These approximations (n0 ≈ ND , p0 ≈ NA ) hold only for moderate doping, assuming complete
ionization.

Worked Example

Silicon contains 5 × 1022 atoms/cm3 . If 2 donor atoms are added for every 106 silicon atoms,
find the free-electron concentration produced (assume complete ionization).
Solution:
2
ND = 6 × 5 × 1022 = 1017 atoms/cm3
10
Each donor contributes one free electron, so n0 ≈ ND = 1017 electrons/cm3 .

In n-type material, electrons are the majority carriers and holes the minority carriers (and vice-
versa for p-type).

6 The Fermi Level

The Fermi level Ef is the energy at which the probability of an electron occupying a state is exactly
50%, given by the Fermi–Dirac distribution:

1
f (E) =
1+ e(E−Ef )/kT

Its position tracks the doping type:

Degenerate (n+ )
Intrinsic n-type p-type
EF
EC EC EC
EF
EF = Ei
Ei Ei Ei

EF
EV EV EV

ˆ Intrinsic: EF lies midway between EC and EV (EF = Ei ).

ˆ n-type: EF shifts toward EC (more electrons).

ˆ p-type: EF shifts toward EV (more holes).

ˆ Degenerate (very heavy doping): EF moves inside EC (n-type) or EV (p-type); the material starts
behaving like a metal.

5
Semiconductor Device Fundamentals Consolidated Notes

6.1 Carrier Concentrations from the Fermi Level

Using Maxwell–Boltzmann statistics, with the thermal potential ϕT = kT /q and Fermi potential ϕF =
(Ei − Ef )/q:

p0 = ni e(Ei −Ef )/kT = ni eϕF /ϕT n0 = ni e(Ef −Ei )/kT = ni e−ϕF /ϕT
p0 n0 NA ND
ϕF = ϕT ln = −ϕT ln =⇒ ϕF ≈ ϕT ln (p-type), ϕF ≈ −ϕT ln (n-type)
ni ni ni ni

where k = 1.3807 × 10−23 J/K = 8.62 × 10−5 eV/K is Boltzmann’s constant and T is absolute
temperature.

7 Equilibrium in the Presence of an Electric Field

So far we assumed the bands are flat (no field). Inside real devices (MOSFETs, pn junctions) an electric
field bends the bands – yet the semiconductor can still be in equilibrium: no net current, no energy
exchange, and a constant Ef throughout. To keep n and p consistent, Ei (and hence EC , EV ) must
bend along with the field.
∆EC
EC

EF

Ei

Ψ(x) EV
∆Ψ
x

p = ni e(Ei −Ef )/kT n = ni e(Ef −Ei )/kT n p = n2i (still holds, even with a field)

The field is the negative gradient of electrostatic potential: E = −dΨ/dx. If ∆EC = ∆Ei is the
potential-energy change for an electron (charge −q) between two points, the corresponding electrostatic
potential difference is:
∆EC
∆Ψ =
−q

7.1 Carrier Ratio Across a Potential Step

If two points have a potential difference ψ12 between them, carriers redistribute by Boltzmann statistics:

n1 p1
= eψ12 /ϕT = e−ψ12 /ϕT
n2 p2

ψ12
n1 n2

6
Semiconductor Device Fundamentals Consolidated Notes

7.2 Nonequilibrium: Quasi-Fermi Levels

When there is energy exchange with the outside world (e.g., a bias voltage across a pn junction), the
system is out of equilibrium. Two separate quantities, the quasi-Fermi levels Ef n (electrons) and Ef p
(holes), replace the single Ef :

p = ni e(Ei −Ef p )/kT n = ni e(Ef n −Ei )/kT

In equilibrium, Ef n = Ef p = Ef and np = n2i . Out of equilibrium, Ef n ̸= Ef p and np ̸= n2i .

8 Charge Density and Poisson’s Equation

8.1 Charge Density
The net charge density ρ in a doped semiconductor sums contributions from mobile carriers and ionized
(fixed) dopants:

ρ = q (p − n + ND − NA )

ˆ n-type (ND > NA ): effective doping Nef f = ND − NA .

ˆ p-type (NA > ND ): effective doping Nef f = NA − ND .

In a neutral region at equilibrium, ρ = 0, giving the charge-neutrality condition:

n + NA = p + ND

Combining charge neutrality with the mass-action law n0 p0 = n2i gives the exact equilibrium con-
centrations:

   
1 1
q q
n0 = 2 2
(ND − NA ) + (ND − NA ) + 4ni p0 = 2 2
(NA − ND ) + (NA − ND ) + 4ni
2 2

When doping ≫ ni , these reduce to the familiar n0 ≈ ND − NA (n-type) or p0 ≈ NA − ND (p-type).

8.2 Poisson’s Equation – Derivation in Three Steps

Step 1 (Field from potential).
Z y
dΨ(y)
E(y) = − ⇐⇒ Ψ(y) = Ψ(y0 ) − E(y) dy
dy y0

Step 2 (Gauss’s Law: field from charge density).

Z y
dE(y) ρ(y) 1
= =⇒ E(y) = E(y0 ) + ρ(y) dy
dy εs εs y0

where εs = kε0 is the material permittivity (ε0 = 8.854 × 10−14 F/cm; for silicon k = 11.9 ⇒ εs =
1.05 × 10−12 F/cm).
Step 3 (Combine). Substituting Step 1 into Step 2:

7
Semiconductor Device Fundamentals Consolidated Notes

d2 Ψ(y) ρ(y)
2
=− Poisson’s Equation
dy εs

This single equation is the starting point for essentially every electrostatic device model — depletion
regions, MOS capacitors, pn junctions — since once ρ(y) is known, Ψ(y) and E(y) follow by integration.

9 Transit Time and Drift

9.1 Transit Time
Transit time τ is the average time a carrier takes to cross a region under steady current flow. If total
mobile charge inside the region is |Q| and the current is I:

|Q| |Q|
I= τ=
τ I

9.2 Drift Velocity

With no applied field, carriers move randomly with zero net displacement. An applied field superim-
poses a net drift on this random motion, with average drift velocity vd :
a a
τ= vd =
vd τ
where a is the length of the region traversed.

Bar of semiconductor (a × b × c)
vd

|Q| I |Q|
|Q| = nq(abc) I= = nq bc vd J= = nq vd |Q′ | = = nqc (sheet charge)
τ A ab

9.3 Low-Field Mobility

At low fields (< 3 × 103 V/cm for electrons), drift velocity is proportional to field, with proportionality
constant µB (bulk mobility):
V
vd = µB E, E=
a

8
Semiconductor Device Fundamentals Consolidated Notes

Bulk mobility in Si at 300 K

Velocity saturation at high field 1,500
Electrons

µB (cm2 /V·s)
Holes

1,000
|vd |

500

0
1014 1015 1016 1017 1018 1019
|E| (field) Doping N (cm−3 )

b a a a2
I = b|Q′ |vd = b|Q′ |(µB E) = µB |Q′ | V τ= = =
a vd µB E µB V

Physically, µB = qτc /mef f , where τc is the mean time between scattering events and mef f is the
carrier’s effective mass. Mobility decreases with rising temperature and doping concentration.

10 Conductance, Resistance, Sheet Resistance

Since I = b|Q′ |(µB E) = µB |Q′ | ab V has the form I = GV (Ohm’s law):

b b
G = µB |Q′ | = µB (nqc) J = nqµB E = σE
a a

Resistance is R = 1/G:

1 1 a a 1
R= = ′
· = Rs Rs = (Ω/□, “ohms per square”)
G µB |Q | b b µB |Q′ |

a/b is simply the number of unit squares in the current path — a standard layout parameter in IC
resistor design. Two bars of the same material and same Q′ but different absolute sizes have identical
resistance if they have the same a/b ratio.

3 squares ⇒ R = 3Rs

11 Diffusion Current
Diffusion is carrier motion driven purely by a concentration gradient (no field needed) — from high
concentration to low. If electron concentration n(x) decreases with x, electrons flow left-to-right, so
conventional current flows right-to-left:

9
Semiconductor Device Fundamentals Consolidated Notes

n(x)
x
 
dn kT
I = Dq(bc) − D = µB ϕ T = µB
dx q

The analogous hole-diffusion current uses a (generally different) diffusion coefficient Dp :

dp
I = Dp q(bc)
dx
If n(x) is not linear, the current itself becomes position-dependent:
 
dn(x)
I(x) = Dq(bc) −
dx

11.1 Einstein Relation

Drift (mobility µB ) and diffusion (D) both arise from the same random thermal motion, so they are
locked together:

kT
D = µB V T VT = ≈ 25.9 mV at 300 K
q

This is one of the most important results in device modelling: a single measured quantity (mobility)
fixes both drift and diffusion behavior.

12 The pn Junction
The pn junction is the basic building block on which the operation of nearly all semiconductor devices
depends. It is formed when a single crystal of semiconductor is doped with acceptors on one side and
donors on the other, so that a junction separates a p-type region from an n-type region.

12.1 The Open-Circuited Junction

Fig. 1 shows the equilibrium condition once the junction has been formed. Donor ions (small circles,
“+”) and acceptor ions (small circles, “−”) are fixed in the lattice; holes (large “+” circles) and electrons
(large “−” circles) are the mobile carriers. It is assumed the junction has uniform cross section.

p type n type
Junction

+ + − − − acceptor ion (fixed)

− − − − + + + + + hole (mobile)
+ − + donor ion (fixed)
− − − − + + + +
+ + − − − electron (mobile)
− − − − + + + +

−Wp Wn
x

10
Semiconductor Device Fundamentals Consolidated Notes

Figure 1: Schematic representation of a pn junction. Circled signs = fixed dopant ions; uncircled signs
= mobile carriers (holes and electrons).

Initially, a concentration gradient exists across the junction, causing holes to diffuse to the right and
electrons to diffuse to the left. As they cross, holes neutralize acceptor ions near the junction on the
p side, and electrons neutralize donor ions near the junction on the n side. The neutralized (ionized)
dopants left behind cannot move, so an uncovered charge builds up: negative on the p side, positive
on the n side.

The depletion region

The region straddling the junction that has been stripped of mobile carriers is called the depletion
region (or transition region). Its width is typically only a few tenths of a micrometer. Outside
the depletion region, carriers exist predominantly as: holes on the p side (p ≈ NA ) and electrons
on the n side (n ≈ ND ).

12.2 Charge Density, Field, and Potential in the Depletion Region

Figure 2 shows, from top to bottom: the space-charge density ρv (x), the resulting electric field intensity
E (x), the electrostatic potential V (x), and the potential-energy barrier seen by an electron.

Charge density ρv
−WDepletion
p regionWn
(a)
Distance from junction, x

p side n side

Field intensity E
−Wp Wn
Distance from junction, x
(b)

Electrostatic potential V
−Wp Wn
(c)
V0
V =0 Distance from junction, x

Potential barrier of electrons

−Wp
V =0
(d)
V0 Wn
Distance from junction, x

Figure 2: (a) Charge density, (b) electric field intensity, (c) electrostatic potential, (d) potential barrier
for electrons in the depletion region of a pn junction.

The charge distribution is zero at the junction itself and forms an electric dipole: negative on the
p side, positive on the n side (the exact shape of ρv (x) depends on how the junction is graded).

11
Semiconductor Device Fundamentals Consolidated Notes

12.3 Poisson’s Equation Across the Junction

The field and potential follow directly from the charge distribution via Poisson’s equation:

d2 V ρv
2
=− (2-1)
dx ε

where ε = εr ε0 is the permittivity of the medium (εr = relative permittivity, ε0 = permittivity of

free space). Using E = −dV /dx and integrating Eq. (2-1) gives the field profile:

x
ρv (x′ ) ′
Z
E (x) = dx (2-2)
−Wp ε

As Fig. 2(b) shows, E is negative (directed from right/plus to left/minus), and E (−Wp ) = E (Wn ) =
0 – i.e., no field exists outside the depletion region.
The electrostatic potential of Fig. 2(c) is the negative integral of E (x); it rises across the depletion
region by the contact potential V0 . Panel (d), the potential-energy barrier seen by electrons, is similar
but inverted, since an electron’s charge is negative.

Equilibrium condition
Under open-circuit conditions the net hole (and net electron) current across the junction must be
zero – otherwise carriers would pile up indefinitely, which is physically impossible. A large hole
diffusion current (p to n) is exactly balanced by a hole drift current (n to p) driven by the built-in
field, and likewise for electrons. This equilibrium condition fixes the contact potential V0 in terms
of the doping concentrations ND and NA ; for typical doping levels, V0 is of the order of several
tenths of a volt.

13 The Biased pn Junction

Applying an external voltage VD across the junction disturbs this equilibrium and gives the pn junction
its essential property: current flows easily in one direction and is almost completely blocked in the
other (rectifier, or unilateral, action).

p n p n

ID depletion region depletion region IS

−V + +V −
D D
(a) Forward-biased (b) Reverse-biased

Figure 3: (a) Forward-biased pn junction: VD lowers the barrier, giving a large current ID . (b) Reverse-
biased pn junction: VD raises the barrier, leaving only the small reverse saturation current IS .

13.1 Forward Bias

In Fig. (a), VD is applied + to the p side and − to the n side. Because no field is assumed to exist
outside the depletion region, the applied voltage appears almost entirely across the junction, reducing
the potential barrier by qVD . This disturbs the balance between diffusion and drift: holes now diffuse
more easily from p to n, and electrons from n to p. The resulting current is the sum of the hole and

12
Semiconductor Device Fundamentals Consolidated Notes

electron currents; once carriers cross the junction they become injected minority carriers in the
other region.

Forward bias
Because the available majority-carrier population is substantial, this diffusion current can be large.
The polarity that reduces the barrier and produces this large current is called forward bias.

13.2 Reverse Bias

In Fig. (b), the polarity of VD is reversed relative to (a). This increases the potential barrier by qVD ,
which:

ˆ Decreases the flow of majority carriers (holes in p, electrons in n), since they must now climb a
higher potential hill.

ˆ Leaves the flow of minority carriers (electrons in p, holes in n) essentially unaffected, since they
“fall down” the barrier regardless of its height.

Reverse saturation current

The small resulting current, denoted IS and called the reverse saturation current, flows from
n to p (opposite to forward-bias current) and is set only by the rate at which minority carriers are
generated – so it is essentially independent of the magnitude of the reverse bias, as long as minority
carriers are available.

14 Summary of Key Relations

13
Semiconductor Device Fundamentals Consolidated Notes

Quantity Relation
1
Fermi–Dirac probability f (E) =
1+ e(E−Ef )/kT
Carrier concentrations p0 = ni e(Ei −Ef )/kT , n0 = ni e(Ef −Ei )/kT
Mass-action law n0 p0 = n2i
Conductivity σ = q(n0 µn + p0 µp )
Resistivity ρres = 1/σ
Charge density ρ = q(p − n + ND − NA )
Charge neutrality n + NA = p + ND
Potential ↔ field E(y) = −dΨ(y)/dy
Poisson’s equation d2 Ψ/dy 2 = −ρ(y)/εs
Transit time τ = |Q|/I
Drift velocity v d = µB E
Current density (drift) J = nqvd = σE
Sheet resistance R = Rs (a/b), Rs = 1/(µB |Q′ |)
Diffusion current I = Dq(bc)(−dn/dx)
Einstein relation D = µB VT , VT = kT /q
Poisson’s eq. (pn junction) d2 V /dx2 = −ρv /ε
Rx
Field from charge (pn junction) E (x) = −Wp ρv (x′ )/ε dx′
Forward bias barrier → V0 − VD , large diffusion current
Reverse bias barrier → V0 + VD , small IS (drift-limited)

— End of Notes —

14