Vous êtes sur la page 1sur 99

Chapter 2 Semiconductor

Device Physics Fundamentals


2.1
2.2
2.3
2.4
2.5
2.6
2.7

Semiconductor Materials
Crystal Structure
Semiconductor Models
Semiconductor Doping
Carrier Statistic
Carrier Transport
Carrier Generation/Recombination

Literature:
Pierret, Chapter 1-3, page 1-132
Acknowledgement Oliver Brand for slides
August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

2.1 Semiconductor Materials


Material Classification by Resistivity:
Material

Resistivity
[cm]

Insulators

105 <

Semiconductors
Conductors

10-3 < < 105 (108)


< 10-3
Jaeger, Blalock, Table 2.1

Literature:
August 18, 2014

Pierret, Chapter 1.1, page 3-6


ECE 3040: Chapter 2.1-2.2
Semiconductor Fundamentals

Semiconductor Materials

Semiconductors have
conductivities between those
of insulators and those of
conductors
Semiconductors are materials
whose electric properties
(e.g. conductivity & resistivity)
can be controlled over a wide
range by doping, i.e. addition
of controlled amounts of
specific impurity atoms
Semiconductor materials are
the base of all semiconductor
devices, such as diodes and
transistors (BJT and
MOSFET)

Pierret, Fig. 3.8


August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

Resistivity/Conductivity of Insulators,
Semiconductor and Conductors

Sze, Fig. 2.1


August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

What Elements are Semiconductors based on?

Elemental Semiconductors:
Silicon (Si), Germanium (Ge)
Compound Semiconductors:
III-V: e.g. Gallium Arsenide (GaAs)
II-VI: e.g. Zinc Selenide (ZnSe)
IV-IV: Silicon Carbide (SiC)
Alloy Semiconductors:
Binary:
Si1-xGex
Ternary:
AlxGa1-xAs
AlxIn1-xAs
GaAs1-xPx
Quaternary: GaxIn1-xAs1-yPy
Because of its well-developed
fabrication technology, SILICON
is by far the most important
semiconductor material

August 18, 2014

Shaded = most important semiconductor elements

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

Jaeger, Blalock, Table 2.2



5

Element and Compound


Semiconductors

August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

The Other Semiconductor Materials

Silicon is the dominating


semiconductor material
The other semiconductor
materials are utilized in
applications requiring e.g.
High speed
Optoelectronic properties
High temperature operation
Example: Ternary alloys, such
as AlxGa1-xAs, have a tunable
bandgap, enabling LEDs and
laser diodes with engineered
output spectrum
Substantial research is
performed (also at GaTech!) on
both Si and non-Si
semiconductor devices

August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

2.2 Crystal Structure


2.2.1 Unit Cell
Unit Cell & Primitive Unit Cell
Cubic Crystal Structure
2.2.2 Semiconductor Lattice
Diamond lattice
Zincblende lattice
2.2.3 Miller Indices
Crystal Directions
Crystal Planes

Literature: Pierret, Chapter 1.2, page 6-16


August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

Classification of Solids
Solids can be classified based on their degree of atomic order

Amorphous
No recognizable
long-range order

Polycrystalline
Solid is made up of
crystallites, i.e.
segments which are
completely ordered

Crystalline
Entire solid is made
up of atoms in an
orderly array
Pierret, Fig. 1.1

Most semicon. devices are based on crystalline semiconductors!


August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

2.2.1 Crystal Unit Cell

Crystal = Periodic 3-D arrangement of


atoms
Unit cell = small portion of the crystal
that can be used to reproduce the
crystal
Primitive (unit) cell = smallest unit cell
possible
Unit cells (including primitive unit cells)
are not unique
A unit cell does not need to be primitive;
often it is more convenient to have a
slightly larger unit cell with orthogonal
sides (instead of primitive cell with nonorthogonal sides)
WWW-page on crystal structures:

Which cells are unit cells?


Which cells are primitive cells?
Which cell is not a unit cell?

http://departments.kings.edu/chemlab/chemlab_v2/

August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

10

Cubic Crystal Structure

Simple cubic (sc) is the simplest


of all 3-D unit cells:
Cube with an atom
positioned at each corner
Only 1/8 of each corner
atom is inside the unit cell: 1
atom per unit cell; each
atom has 6 nearest
neighbors
The side length is called the
lattice constant a
Very few materials crystallize in
a simple-cubic lattice
How can we see the crystal
structure of a crystal?

Sze, Fig. 2.3


August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

11

Cubic Crystal Structures


fcc and bcc Lattice
Body-Centered-Cubic (bcc) unit cell
Additional atom in the center of the cube
2 atoms per unit cell
8 nearest neighbors for each atom
Examples: Cr, W, Na, Fe
Face-Centered-Cubic (fcc) unit cell
Additional atoms in the center of each
face of the cube
Also known as cubic-close-packed (ccp)
4 atoms per unit cell
12 nearest neighbors for each atom
Examples: Ni, Ag, Au, Cu, Al
August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

Sze, Fig. 2.3


12

2.2.2 Diamond Lattice


Silicon Crystal Structure

(1 = 10-10 m = 10-8 cm = 0.1 nm)

Sze, Fig. 2.4


Calculate: How many Si atoms do
we have per cm-3?
August 18, 2014

Silicon and germanium have a


diamond lattice, a special cubic
crystal structure
The diamond lattice consists of
two interpenetrating facecentered-cubic (fcc) lattices
shifted by one-quarter of the
cube body diagonal
Each silicon atom has 4 nearest
neighbors forming a tetraeder
structure
The silicon unit cell has a lattice
constant a = 5.43 (T = 300 K)

Each silicon atom forms four


covalent bonds with its nearest
neighbors (see Chapter 2.3)

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

13

Diamond Lattice along <110> Axis

August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

14

Zincblende Lattice
GaAs Crystal Structure

Most III-V semiconductors (e.g.


GaAs) have a zincblende lattice,
again a cubic crystal structure
Similar to the diamond lattice,
the zincblende lattice consists of
two interpenetrating facecentered-cubic (fcc) lattices
shifted by one-quarter of the
cube body diagonal, but with
different atoms occupying each
fcc sublattice
The GaAs unit cell has a lattice
constant a = 5.65 (T = 300 K)
(1 = 10-10 m = 10-8 cm = 0.1 nm)

Sze, Fig. 2.4


August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

15

2.2.3 Miller Indices

Miller indices are used to specify


crystallographic planes and
directions
How do I find the Miller index for a
plane?
Record where the plane
intersects coordinate axes in
units of a: e.g. 1,2,3
Invert intercept values:
e.g. 1, 1/2, 1/3
Multiply with constant to get
smallest possible set of whole
numbers: e.g. 6,3,2
Enclose set with curvilinear
bracket: e.g. (632)
Note: (6 32) (63 2)
What are the other planes indices?

(632)

(001)

(22 1)
Pierret, Fig. 1.6

August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

16

More on Miller Indices


Miller indices for directions are found analogous to the
components of a vector
In a cubic crystal, the planes

(100),(010),(001),(1 00),(0 1 0),(00 1) {100}


are equivalent because of symmetry and summarized as {100}

(hkl)

Particular crystal plane

{hkl}

Equivalent planes

[hkl]

Particular crystal direction

<hkl>

Equivalent directions

For cubic crystals, the direction [hkl] is normal to the plane (hkl)
August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

17

Crystal Planes and Directions


Cubic Crystal Structure

Pierret, Fig. 1.7



August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

18

Wafer Flats
Silicon wafers used for device processing have flats (or a notch)
indicating crystal direction and doping type
A {100} wafer has
{100} surface
For {100} wafer,
the surface direction
perpendicular
to primary flat
is [011]
How is a wafer
fabricated?

[011]

August 18, 2014

ECE 3040: Chapter 2.1-2.2


Semiconductor Fundamentals

19

2.3 Semiconductor Models


2.3.1 Electron States in Atoms
2.3.2 Semiconductor Bond Model
Covalent Bonds
2.3.3 Semiconductor Band Model
Energy Bands
Band Gap
Electron and Holes
Band Structure & Effective Mass
Simplified Semiconductor Band Model
Literature: Pierret, Chapter 2.1-2.3, page 23-40
Pierret, Appendix A, page 733-748

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

20

Silicon Atomic Structure


Silicon with its 14 electrons
has (in its ground state) filled
n = 1 (2 electrons) and n = 2
(8 electrons) levels; these
are deep-lying energy levels,
tightly bound to the nucleus
In case of the n = 3, the
possible states (2 s- and 6 pstates) are filled with the 4
remaining electrons; these 4
rather weakly bound
electrons are called valence
electrons and participate in
chemical reactions and
atom-atom interaction

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

Pierret, Fig. 2.2


21

2.3.2 Semiconductor Bond Model


Bonding to next neighbors is
achieved by sharing valence
electrons (electrons in
outermost shell), forming
covalent bonds
In case of silicon:
Atomic number
= number of electrons
= 14
Number of valence
electrons = 4
Covalent tetraeder
bonding (diamond lattice)
Sze, Fig. 2.11
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

22

Semiconductor Bond Model


Valence electrons are
weakly bound (compared to
electrons in inner shells)
Thermal energy at room
temperature can break
covalent bonds and free
electrons which now
contribute to the materials
conduction
Remaining are missing
bonds/electrons, which are
called holes (and also
contribute to the conduction,
but why?)
Sze, Fig. 2.12
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

23

2.3.3 Semiconductor Band Model


Sze, Fig. 2.15

Single atom:
according to Bohrs
atom model, the
electrons occupy
well defined,
discrete energy
levels
Crystal: due to the interaction of neighboring atoms, the
discrete energy levels split up and a bandstructure is
developing featuring allowed and forbidden energy bands
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

24

From
Isolated
Atoms
to a Crystal

Semiconductor crystal has valence


band and conduction band
separated by bandgap Eg
At T = 0 K:
Completely filled valence band
Completely empty conduction
band
At elevated T, some electrons can
be excited from the valence to the
conduction band
Note: Electrons in bands are not
associated with particular atom

Pierret, Fig. 2.5


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

25

Semiconductor Band Model


Reason for energy level splitting is the Pauli principle stating
that no 2 electrons can occupy the same state (with respect to
energy, momentum, spin), i.e. have the same quantum numbers
n, l, m, and s
Simplified Model: valence and conduction band are separated
by a band gap; EC is the lowest energy of the conduction band,
EV the highest energy of the valence band
EC
Eg

Band Gap
EV

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

26

Carriers: Electrons and Holes

Completely filled band and empty band do not contribute to current


conduction
If an electron is excited from the valence into the conduction band,
the additional electron in the conduction band contributes to the
current conduction; similarly, the missing electron in the valence band
(like a bubble in a liquid), a so-called hole, contributes to the current
conduction; the (thermal) excitation has created an electron/hole pair

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

27

Band Gap & Material Classification

Insulator

Semiconductor

Metal
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

Pierret, Fig. 2.8


28

Indirect vs. Direct Bandgap Semiconductor


Simplified Si & GaAs Band Structure
Using quantum
mechanics by
solving the 3-D
Schrdinger
equation in the
periodic potential
of the nuclii
Numerical
calculations deliver
the band diagrams,
i.e. the E(k), E(p)
relationships with
wavenumber k and
crystal momentum
p

Indirect Bandgap

p 0

Direct Bandgap

p = 0

Sze, Fig. 2.18


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

29

Characteristics of Si and GaAs


Silicon Si
Element semiconductor
Diamond lattice structure
Indirect Semiconductor
Generation of electron-hole
pairs requires E and p
Bandgap: Eg = 1.12 eV at
room temperature
Lattice constant: a = 5.43

August 18, 2014

Gallium Arsenide GaAs


III-V semiconductor
Zincblende lattice structure
Direct Semiconductor
Generation of electron-hole
pairs requires E only,
making GaAs suitable for
photonic applications
Bandgap: Eg = 1.42 eV at
room temperature
Lattice constant: a = 5.65

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

30

Simplified Semiconductor
Bandstructure
mn*

Eg

Band Gap
mp*

August 18, 2014

EC

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

EV

31

Effective Mass meff (mn*, mp*)


Definition 1-D:

meff

1 2E(k)
!2 k 2

Definition 3-D:

" 1 %
1 2E(k)
$
' = 2
# meff & ij ! k ik j

Electron kinetics in single-crystal material is similar to


free electron behavior but with effective mass meff

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

32

Effective Mass in Si, Ge & GaAs


Material

mn*/m0

mp*/m0

Si

1.18

0.81

Ge

0.55

0.36

GaAs

0.066

0.52

Free electron mass: m0 = 9.1 10-31 kg


Shown are average electron and hole effective masses as
needed for the density of states at 300 K (see Chapter 2.5),
but keep in mind that mn* and mp* are tensors!
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

33

Effective Mass Concept


The movement of carriers (electrons and holes)
inside the semiconductor crystal can be described by
classical (Newtonian) mechanics relations just by
replacing the free-electron-mass m0 with the effective
mass mn* (or mp* for holes)
F

= mn*

dvn
dt

and F

= mp*

dvp
dt

The effective masses are however tensors, i.e. the

carrier
acceleration generally varies with direction of
travel in the crystal
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

34

2.4 Semiconductor Doping


Example: Silicon Doping:
Doping with donor atoms,
phosphorous P or arsenic
As, having both 5 valence
electrons, yielding N-type
silicon
Doping with acceptor
atoms, boron B, having 3
valence electrons, yielding
P-type silicon
With doping, the electrical
conductivity of the
semiconductor material is
adjusted over a wide range
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

35

Semiconductor Doping
Doping in Bond Model:
N-Type Doping:
fifths valence electron does not
participate in covalent bonds to
next neighbors; only small energy
is required to free this electron,
thus contributing to the materials
conductivity
P-Type Doping:
due to doping atom with three
valence electrons, one electron is
missing for the covalent bonds to
the four next neighbors; this
missing electron (called a hole)
can hop from covalent bond to
bond, thus also contributing to the
materials conductivity

Sze, Fig. 2.23

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

36

Semiconductor Doping
Doping in Band Model:
Doping atoms distort the periodicity of
the single-crystal crystal structure, thus
creating allowed (localized) energy
states in the band gap
N-Type Doping: additional electron
states ED close to the bottom of the
conduction band EC; at room
temperature, the electrons from these
states are excited into the conduction
band, increasing the conductivity
P-Type Doping: additional electron
states EA close to the top of the valence
band EV; at room temperature, these
states are occupied by electrons from
the valence band, increasing the
number of holes in the valence band
and, thus, the conductivity
August 18, 2014

EC - ED resp. EA - EV
kT = 0.026 eV
at room temperature

ED

EC
Ei
EV

P in Si: EC - ED = 0.045 eV

EC
Ei
EA

EV

B in Si: EA - EV = 0.045 eV

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

37

Carrier Concentration
Intrinsic Semiconductor (no doping: NA, ND = 0)
Carrier Concentration: n = p = ni
Extrinsic Semiconductor (doping concentration NA, ND 0)
Assumption: Complete Ionization (see Chapter 2.5)
Majority Carrier Concentrations:
N-Type Semiconductor:
Density of donor atoms: ND [cm-3]
Density of electrons (ND ni) : n ND [cm-3]
P-Type Semiconductor:
Density of acceptor atoms: NA [cm-3]
Density of holes (NA ni) : p NA [cm-3]

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

38

2.5 Carrier Statistic


2.5.1 Density of States
2.5.2 Fermi Function & Fermi Energy
Physical Interpretation
Characteristics
2.5.3 Carrier Densities
Intrinsic/Extrinsic Semiconductor
Intrinsic Fermi Energy
Mass Action Law
Temperature Dependence
2.5.4 Charge Neutrality Relationship
2.5.5 Non-Complete Ionization
Pierret, Chapter 2.4-2.6, page 40-68
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

39

2.5.1 Density of States


From quantum mechanics, we not only obtain the band
structure, i.e., the E(k) relations, but also the density of
states g(E)dE, i.e., how many allowed states are in the
range EE+dE:

gC (E)dE =
gV (E)dE =

August 18, 2014

mn* 2mn* E EC

mp*

) dE E E 1/ 2
( C)

2! 3
2mp* EV E

2! 3

) dE E E 1/ 2
(V )

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

40

Density of States
E

gC(E)

EC
Band Gap

g(E)
EV
Units of g(E)dE are [cm-3]
gV(E)

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

41

2.5.2 Fermi Function & Fermi Energy


What determines whether an
allowed state is occupied by
an electron or not?
Fermi Function f(E):
f(E) is a probability function
which gives the probability
whether a state is occupied
or not

f(E) =

1
1+ e(EEF )/ kT

EF is the Fermi Energy

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

42

Characteristics of Fermi Function


1

August 18, 2014

EF = 1 eV
kT = 0.0259 eV
Fermi Function f(E)

Because of Pauli principle:


0 f(E) 1
f(E = EF) = 0.5
The probability that a state is
occupied at the Fermi
energy is 50%
f(E) is symmetric around EF:
f(EF + E) = 1 f(EF E)
For T = 0 the Fermi function
becomes a step function, i.e.
all states below EF are
occupied, all states above EF
empty

0.5

0
0

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

Energy E [eV]

43

Physical Interpretation of
Fermi Energy EF
The Fermi energy EF has the function of a
thermodynamic potential
In thermodynamic equilibrium, EF is constant
across the device!!!
A gradient in EF indicates non-equilibrium, resulting in
a net flow of carriers

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

44

Boltzmann/Classical
Approximation
Away from E = EF, i.e. for |E EF| > 3kT, the Fermi function can
be approximated by exponential functions:

(
)
(EF E) > 3kT

E EF > 3kT

f(E) e(EEF )/ kT

f(E) 1 e(EEF )/ kT

This approximation is called the Boltzmann or Classical


approximation

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

45

2.5.3 Carrier Density n and p


The carrier densities of electrons n [cm-3] and holes p [cm-3] can
be calculated from the density of states g(E) and the Fermi
function f(E):

Top

n=

gC (E) f(E) dE

EC

p=

EV

gV (E) 1 f(E) dE
!
#"#
$
Bottom

a hole is a non-occupied electron state

The solution of the integral generally requires numerical


methods; in case of the Boltzmann approximation however, the
integrals
can be solved analytically!
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

46

Carrier Concentration
Intrinsic

Extrinsic
N-Type

Extrinsic
P-Type

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

47

Carrier Density Boltzmann


Approximation
For electrons: (EC EF) 3kT
n = NC e

(EF EC )/kT

EC

3/2

# m* kT &
NC = 2 % n 2 (
%$ 2! ('

19

2.86 10 cm
"$$#$$%

EV

Si @ 300K

For holes: (EF EV) 3kT


p = NV e

Ei

EC3kTEFEV+3kT

(EV EF )/kT

From Pierret:
3/2

3/2

# m* kT &
NV = 2 % p 2 (
%$ 2! ('
August 18, 2014

2.66 1019 cm3


"$$#$$%
Si @ 300K

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

NC,V

" m* %
19
= 2.5110 $ n,p '
$# m0 '&

NC 3.2 1019 cm3


NV 1.8 1019 cm3

cm3

for Si @ 300K

48

Carrier Density General Solution


Range
for EF

EC
Ei

For electrons:
EV
2
n = NC
F1/ 2 ( C ) with C = (EF EC ) / kT

For holes:
2
p = NV
F1/ 2 ( V ) with V = (EV EF ) / kT

F1/2(c) is the tabulated Fermi-Dirac integral

of order 1/2:
1/ 2
F1/ 2 c =
d
c
0 1+ e

( )

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

49

Intrinsic Semiconductor
No doping
Carrier concentration n = p = ni (for every electron in the
conduction band, we have a hole, i.e. missing electron,
in the valence band)
Fermi energy EF = Ei close to center of band gap

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

50

Intrinsic Semiconductor
No doping, i.e. intrinsic semiconductor with
n = p = ni and EF = Ei
From Boltzmann approximation with EF = Ei, we
obtain
n = NC e(Ei EC )/ kT = ni
p = NV e(EV Ei )/ kT = ni

Thus, solving for ni yields

n p = ni2 = NCNV e(EV EC )/ kT


= NCNV e

Eg / kT

ni = NCNV e

Eg / 2kT

1010 cm3
!#"#$
Si @ 300K

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

51

Intrinsic Semiconductor
Carrier Concentration ni(T)

GaAs: ni,(300K) = 2.25 106 cm-3


Ge: ni,(300K) = 2.5 1013 cm-3
Si: ni,(300K) = 1.0 1010 cm-3
Pierret, Fig. 2.20
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

52

Location of Ei
From the intrinsic carrier concentrations

p = ni = NV e(EV Ei )/ kT
n = ni = NC e(Ei EC )/ kT
With n = p, the intrinsic Fermi level Ei can be extracted

NV e(EV Ei )/ kT = NC e(Ei EC )/ kT

*&
#
m
E + EV 3
p
Ei = C
+ kT ln% (
2
4
%$mn* ('

The intrinsic Fermi level is only in the center of the band


gap if mn* = mp* !!

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

53

Extrinsic Semiconductor
With doping
Carrier concentration n p ni
Example: n-type semiconductor

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

54

Extrinsic Semiconductor
From Boltzmann approximation

n = NC e(EF EC )/ kT = NC e(Ei EC )/ kT e(EF Ei )/ kT


!##"##$
=ni

n = ni e(EF Ei )/ kT
Analogous for the holes

p = ni
August 18, 2014

np

2
= ni

Mass-Action-Law

e(Ei EF )/ kT

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

55

Fermi Energy EF Ei [eV] as a


Function of Doping and Temp.

Sze, Fig. 2.28


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

56

2.5.4 Charge Neutrality Relation


In equilibrium, charge neutrality is fulfilled:

p
!

holes

n
!

electrons

+
+ N
D
!
ionized
donors

N
A
!

=0

ionized
acceptors

Assuming complete ionization NA = NA and ND+ = ND:

p n + ND NA = 0

np
August 18, 2014

2
= ni

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

57

Electron and Hole Concentration


From mass-action-law and charge neutrality:
2

#ND NA &
ND NA
2
n=
+ %
+
n
i
(
2
2
$
'
2

#NA ND &
NA ND
2
p=
+ %
+
n
i
(
2
2
$
'

Special Cases:

ND >> NA

and ND >> ni :

p = ni2 / ND
NA >> ND and NA >> ni :

August 18, 2014

n = ND
p = NA

n = ni2 / NA

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

58

Measurement of Carrier Concentration


Hall Effect

P-Type Semiconductor
The carrier concentration n or p
can be measured using the Hall
effect
Assume: p-type semiconductor
+
plate with L > W t (thin plate)
x
with current I applied in x-direction VH
W

and a magnetic induction Bz


L
applied along z-direction
Holes experience a Lorentz force
in y-direction:
! !

+
F = q (v B) = (0,qv xBz ,0)

Bz

z
x

Hole accumulation creates electric field in y-direction (electrostatic force =


Lorentz force, because no current flow in y-direction), resulting in a Hall
voltage VH:
J
1 IBz
VH = y W = v xBz W = x Bz W =
qp
qp
! t
RH
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

Hall Coefficient
59

2.5.5 Temperature Dependence of


Carrier Density

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

60

Temperature
Dependence
of Carrier
Density

Pierret, Fig. 2.22


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

61

Carrier Statistics Formula Summary


Charge Neutrality & Mass-Action Law:
Intrinsic Semiconductor

Extrinsic Semiconductor
Linking Doping and Carrier Concentration

Linking Band Structure and Carrier Concentration

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

62

2.6 Carrier Transport


2.6.1 Carrier Drift
Drift Current
Mobility
Resistivity
Band Structure under Applied Field
2.6.2 Carrier Diffusion
Diffusion Current
2.6.3 Total Current Equations
2.6.4 Einstein Relations
Pierret, Chapter 3.1-3.2, page 75-104
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

63

2.6.1 Carrier Drift


Drift = Charged particle motion in response to an
applied electric field
Macroscopic Definition: Carriers of a given type
(electrons or holes) move along at a constant
velocity, the drift velocity, parallel or antiparallel to the
applied electric field
Drift Current Densities [A/cm2]:

!
!
Jp,drift = q p vd,p
!
!
Jn,drift = q n vd,n
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

64

Carrier Drift
For small electric fields, the drift velocity is
proportional to the applied electric field with the
mobility as proportionality factor:
vd,p = p
vd,n = n

The mobility is the central parameter characterizing


the carrier drift, resulting in the following current
densities: !
!

Jp,drift = q p p
!
!
Jn,drift = q n n
!
!
!
Jdrift = Jn + Jp = q nn + pp

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

!
]
65

Drift Velocity in Undoped Si at


Room Temperature

Pierret, Fig. 3.4


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

66

Carrier Mobility
Definition of the carrier mobility [cm2/Vs]:

vd,p

vd,n

Room-temperature mobility of Si and GaAs:

Silicon (low-doped)

n 1360 cm2/Vs
p 460 cm2/Vs

Gallium Arsenide (low-doped)

n 8500 cm2/Vs
p 400 cm2/Vs

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

67

Drift Microscopic Picture


Charged particles are accelerated by the electric field
(F = q ) and loose energy by collisions with
impurities and the crystal lattice (impurity scattering
and lattice scattering)

The drift of a particle is superimposed on the random


thermal motion

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

68

Scattering
Mechanisms I
(a) Impurity and (b) lattice
scattering limit the carrier
mobility
mean free time
q
between collisions
=
m*
effective mass

The mobility decreases with


increasing total doping
concentration (NA + ND)
Pierret, Fig. 3.5
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

69

Room Temperature Mobility in Si as a


Function of Doping Concentration

Pierret, Fig. 3.5a


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

70

Scattering
Mechanisms II
T-dependence of the mobility:
Lattice Scattering:
in Si: ~ T(2.22.3) (experiment)
Reducing the temperature means
less thermal lattice vibration, i.e.
less interaction with carriers, i.e.
higher mobility
Impurity Scattering:
in Si: ~ T+1.5 (theory)
Dominant scattering mechanism at
low temperatures; interaction is
reduced at higher temperatures,
i.e. higher thermal velocities,
because carrier is less time in
close proximity to impurity
= (T, NA+ND)
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

Pierret, Fig. 3.7


71

Resistivity
Definition of conductivity
and resistivity [cm]
tensor:

!
! 1!
J==

From the drift current, the


resistivity is given by:

1
=
q nn + pp

)
Pierret, Fig. 3.8

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

72

Four-Point Probe
The resistivity of a
semiconductor can be
determined by a currentvoltage measurement using
a four-point probe

V
= W CF
I

W is the substrate thickness,


CF a correction factor (see
graph)

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

73

Band Bending by Electric Field


The presence of an electric
field results in a band
bending, i.e. EC and EV are
no longer constant
Electrostatic Potential V:
V=

1
EC Eref
q

hole
movement

Note: V and E are arbitrary


to within a constant!

electron
movement acceleration
collision
loss of energy

EC
Ei
EV
x

Electrostatic Field :

= V
= 1 dEC
q dx

August 18, 2014

1 dEV 1 dEi
=
q dx
q dx
ECE 3040: Chapter 2.1-2.2 Semiconductor
Fundamentals

x
74

Band Bending and Electric Field

Pierret, Fig. 3.10


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

75

2.6.2 Carrier Diffusion


Due to their random thermal motion, particles (in our case
electrons and holes) tend to spread out or redistribute, i.e. on a
macroscopic scale, they diffuse from areas of higher
concentration to areas of lower concentration
Diffusion = particle motion due to a gradient in the particle
concentration
Diffusion is described by Ficks law, stating that the particle flux
density F [particles/(cm2s)] is proportional to the negative
gradient of the particle concentration grad(),

F = D
with the diffusion coefficient D [cm2/s] as proportionality constant

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

76

Diffusion Current
In case of diffusion of charged particles, i.e. electron
and holes in our case, a diffusion current is resulting
(equaling the product of flux density and carrier
charge)

Jp,diff = q Dp p
Jn,diff = q Dn n

with diffusion coefficients Dn and Dp [cm2/s]


Why has the electron diffusion current no sign?

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

77

2.6.3 Total Current Equations


Summing up drift and diffusion current densities, we
obtain the total current density equations:

!
!
!
Jp = q p p q Dp p
!
!
!
Jn = q n n + q Dn n
! !
!
J = Jn + Jp
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

78

2.6.4 Einstein Relations


Einstein Relations relate drift
() and diffusion (D)
Simplified Derivation:
Assume non-uniformly doped
n-type semiconductor in
equilibrium (i.e. EF = const.
across the material)
In equilibrium, the total electron
and hole current densities are
zero:
!
!
!
Jp = q p p q Dp p = 0
!
!
!
Jn = q n n + q Dn n = 0
Pierret, Fig. 3.14
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

79

Einstein Relations
From
we obtain

n = ni e(EF Ei )/ kT
dn
ni (EF Ei )/ kT dEi
q (EF Ei )/ kT
= e
= nie
#"##
$
dx
kT
dx
kT !#
=n

knowing that =

1 dEi
q dx

and

dEF
=0
dx

Inserting into the current density equation for the


electrons
yields the Einstein Relations relating drift
(n) and diffusion (Dn)
(similar derivation
for Dp and p)
August 18, 2014

kT
kT
Dn =
n and Dp =
p
q
q

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

80

Drift & Diffusion in Si (T = 300 K)

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

81

2.7 Carrier Generation & Recombination


Perturbation of semiconductor, i.e. an excess or deficit in
carrier concentration with respect to the equilibrium values is
created, resulting in non-equilibrium conditions
If the perturbation is removed, recombination/generation (R-G)
processes will restore equilibrium conditions; if perturbation is
maintained, R-G processes will stabilize the (non-equilibrium)
carrier concentrations
2.7.1 Generation and Recombination Processes
2.7.2 R-G Statistics
2.7.3 Continuity Equations
Literature: Pierret, Chapter 3.3-3.5.1, page 105-132
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

82

Generation

Band-to-band R-G
processes only involve
electron in conduction band
and hole in valence band
R-G center generation/
recombination requires
R-G center (lattice defect or
impurity which generates
states in the bandgap)
Auger recombination
requires 3 carriers (either 2
holes and 1 electron or 2
electrons and 1 hole)
All processes occur at all
times (even in equilibrium),
with the process having the
highest rate dominating

Recombination

2.7.1 Generation/
Recombination
Processes

Pierret, Fig. 3.15


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

83

Momentum Considerations
Band-to-Band vs. R-G Center R-G Processes

What is the dominant R-G process, i.e. the


process proceeding at the fastest rate, in a
particular system?
When do band-to-band R-G processes dominate
over R-G center R-G processes and vice versa?
Photons carry energy but very little momentum
Phonons carry momentum but very little energy
In indirect semiconductors, band-to-band
recombination requires the emission/absorption
of a phonon (for momentum conservation) in
addition to photon emission, making this process
negligible compared to R-G center recombination
In direct semiconductors, band-to-band
recombination proceeds at a much faster rate,
making these materials suitable for light emitting
devices, e.g. LEDs and lasers

Direct
Semiconductor
(e.g. GaAs)

Indirect
Semiconductor
(e.g. Si or Ge)
Pierret, Fig. 3.18

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

84

Perturbation:
Carrier Injection/Generation

2
p n > ni
Processes:
Photogeneration
Operation of diode in forward direction
Impact ionization

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

85

Perturbation:
Carrier Extraction/Recombination

2
p n < ni
Process:
Operation of diode in reverse direction

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

86

Carrier Generation/Recombination
After switching off the distortion, the system returns to
the equilibrium state (p n = ni2) with a characteristic
time constant, i.e. excess carriers will recombine (in case
p n > ni2) or additional carriers will be generated (in case
p n < n i2)
Notation:

August 18, 2014

nn
pn
nn0, pn0

electron density in n-type semiconductor


hole density in n-type semiconductor
equilibrium carrier densities

np
pp
np0, pp0

electron density in p-type semiconductor


hole density in p-type semiconductor
equilibrium carrier densities

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

87

Low-Level/High-Level Injection
Example: Injection by photogeneration (n = p) in n-type material
Low-Level Injection:

High-Level Injection:

n = p << ND

n = p ND

nn = nn0 + n nn0

nn = nn0 + n > nn0

pn = pn0 + p p < nn0

pn = pn0 + p p > nn0

Example: n-type Si, ND = nn0 = 1015 cm-3, p = ni2/n = 105 cm-3


n = p = 1012 cm3

nn 1015 cm3

nn 1017 cm3

pn 1012 cm3

pn 1017 cm3

August 18, 2014

n = p = 1017 cm3

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

88

Low-Level/High-Level Injection
Example:
N-type semiconductor
ND = 1015 cm-3
Equilibrium:
nn0 = ND = 1015 cm-3
pn0 = ni2/n = 105 cm-3
Low-Level Injection:
n = p = 1012 cm-3
nn nn0 = 1015 cm-3
pn p = 1012 cm-3 < nn0
High-Level Injection:
n = p = 1017 cm-3
nn 1017 cm-3 > nn0
pn p = 1017 cm-3 > nn0
Equilibrium
August 18, 2014

LL Injection

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

HL Injection
89

2.7.2 R-G Statistics


Band-to-Band Recombination

Likely in direct (band gap)


semiconductors, such as GaAs
Unlikely in indirect (band gap)
semiconductors, such as Si,
because of required momentum
conservation
In thermal equilibrium, the
thermal generation rate Gth
equals the recombination rate Rth:

Equilibrium:

Sze, Fig. 3.10

Non-Equilibrium:

Gth = Rth = nn0 pn0

The thermal recombination rate is


proportional to the electron and
hole carrier densities with
being the proportionality factor
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

90

Band-to-Band Recombination
In non-equilibrium, the change in minority carrier concentration
is:
dpn (t)
= GL + Gth R
dt
Assuming steady state, dpn/dt = 0, we obtain
GL = R Gth U

with the net-recombination


rate U
The recombination rate and the thermal generation rate are
proportional tothe available carrier densities (see equilibrium):
U = nn pn nn0 pn0
!"# !$"$
#
=R

[(

=G th

)(

= nn0 + n pn0 + p nn0 pn0

= nn0 + pn0 + p p
Assuming low-level injection, i.e. p and pn0 << nn0, we obtain
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

91

Band-to-Band Recombination
U = nn0 p =

The minority carrier lifetime p controls the recombination velocity


after the distortion is switched off

nn0

pn pn0 pn pn0
=
1
p
nn0

dpn (t)
pn pn0
pn
= GL U = GL
= GL
dt
p
p

The minority carrier lifetime in case of the direct recombination is


inversely proportional to the majority carrier equilibrium density nn0
Excess carriers recombine after switching off GL (GL = 0 in DE) with the
time constant p:

pn (t) = pn0 + p GL e

t / p

How can we measure the minority carrier lifetime?

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

92

Measurement of p
Illuminate semiconductor
sample with light and
measure its electrical
resistance (or the voltage
drop across the sample):
a. n-doped sample under
constant illumination
b. Minority carrier
concentration in steady
state (t < 0) and after
switching off the illumination
c. Experimental setup to
determine the minority
carrier lifetime p
August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

Sze, Fig. 3.11

Initial slope:
-p/p

93

R-G Center Recombination


Sze, Fig. 3.12

Dominant recombination
process in indirect
semiconductors, such as Si
Indirect recombination &
generation of electron-hole
pairs via localized energy
states in the band gap, so
called recombination
centers or R-G centers
The R-G centers are
characterized by their
energy Et and their density
Nt

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

94

Location of Impurity Atom Energy States

Sze, Fig. 2.24


August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

95

R-G Center Recombination

In a more involved derivation, an expression similar to that of the bandto-band R/G case can be obtained for R-G center recombination/
generation (again assuming minority carriers and low-level injection) in
response to a perturbation:

dpn (t)
p
= GL U = GL n
dt
p

and

dt

= GL

np
n

The minority carrier lifetimes are now inversely proportional the R-G
center concentration NT:

1
n
NT

dnp (t)

1
and p
NT

In case of arbitrary injection levels and for both carrier types in a nondegenerate semiconductor, we find:

U=

August 18, 2014

pnnn ni2

(E E )/kT
(E E )/kT
p #$pn + ni e i t %& + n #$nn + ni e t i %&
ECE 3040: Chapter 2.1-2.2 Semiconductor
Fundamentals

96

Example: Gold R-G Centers

August 18, 2014

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

97

2.7.3 Continuity Equations


putting it all together
Net change of carrier concentration due to currents and R/G:

n n
n
n
n
=
+
+
+
t t drift t diffusion t thermal RG t other processes
"$$#$$% "$#$% "$
$#$$
%
1 !
= Jn
q

n
n

e.g. GL

p p
p
p
p
=
+
+
+
t t drift t diffusion t thermal RG t other processes
"$$#$$% "$#$% "$
$#$$
%
1 !
= Jp
q

August 18, 2014

p
p

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

e.g. GL

98

1-D Continuity Equations


For minority carriers in 1-D case: Jp = q p p q Dp p

Jn = q n n + q Dn n

np (x,t)
t

2np

np np0

= +np n
+
+ Dn 2 + GL
n
x
x
n
x
np

pn (x,t)
pn
2pn
pn pn0

= pn p
p
+Dp 2 + GL
%
t
x"###
x
p
!###
$ !#"x
#
$ Perturb. !#"
#
$
Drift

August 18, 2014

Diffusion

ECE 3040: Chapter 2.1-2.2 Semiconductor


Fundamentals

thermal RG

99

Vous aimerez peut-être aussi