A: UNITS
FUNDAMENTAL QUANTITIES
Quantity
Unit
Symbol
Capacitance (C)
Farad
Conductance (G)
Siemens
C u r r e n t (I)
Ampere
Coulomb
E n e r g y (W, E )
Joule
Force (F)
Newton
Frequency (f)
Hertz
Hz
Length (L)
Meter
Mass (m)
Kilogram
kg
Potential (V, )
Volt
Power (P)
Watt
Resistance (R)
Ohm
T e m p e r a t u r e (T)
D e g r e e s Kelvin
Time (t)
Second
DERIVED QUANTITIES
A n g u l a r Frequency ()
Radian/sec
rad/s
Conductivity ()
Siemens/meter
S/m
Dielectric constant
(Relative Permittivity)
Permittivity (, )
Dimensionless
Farad/meter
F/m
Resistivity (p)
Ohmm
m
Velocity (v)
Meters/second
m/s
Wavelength ()
Meters
465
466
Value
Symbol
Avogadro's N u m b e r
6.022 X 1 0 / m
B o l t z m a n n Constant
1.38066 X 1 0  J / K
B o l t z m a n n Constant
k/q
8.61738 X 1 0  e V / K
23
2 3
1 9
1.602 X 10
Electronvolt
eV
1.602 X 10~ J
E l e c t r o n R e s t Mass
9.109 X 1 0 " k g
Permittivity in Vacuum
M a g n i t u d e of Electronic Charge
Planck's Constant
Speed of Light in Vacuum
Thermal Voltage
19
31
8.854 X 1 0  F / m
1 2
6.626 X 1 0  J  s
2.998 X 1 0 m / s
34
V =
0.02586 V ( T = 300K,27C)
kT/q
B.2 P R E F I X E S
6
B.3 C O N V E R S I O N F A C T O R S
8
10
1A = 1 0  c m = 1 0  m
9
l n m = 10~ m = 10A = 1 0  c m
4
= 10" cm = 10" m
l e V = 1.602 X 1Q J
19
8 \
+ + j + = 0
dx
dy
dz
h
27
(C1)
'
(C.2)
467
1 dg
1 dg
g dx
8n mE
g ^f
1 dg _
g dz
Since the righthand side of E q . (C.3) is a constant, it follows that each of the
terms is a constant so that
We assume for t h e sake of simplicity that each of the sides of the cubic well
has length a. T h e b o u n d a r y conditions for are
\\> = 0atx
= y = z = 0
=y
(C.5)
= =
The solution to each of t h e equations in (C.4) contains sine and cosine terms.
Possible solutions, wherein the cosine terms d r o p out and, subject to t h e b o u n d a r y
conditions in E q . (C.5), b e c o m e
g=
x
'3
Asin^
(a)
= B s i n ^
a
(b)
g = C sin
(C.6)
(c)
,
,
,
n\ + n) + n =
8mEa
(C.7)
^h\N
2
x
N)
z
8 ma
We are actually interested in a section of the sphere of the solid that is m a d e
by joining t h e t h r e e m a x i m u m values for which the radius of t h e sphere R is
V7V2 + N + N . That section is o n e for which t h e values of n , , n are all positive.
This section forms one eighth of the sphere, shown in Fig. CF.l, having volume
vR /6 or
2
V o l u m e = ^(N
2
X
(C.9)
468
(CIO)
(Gil)
The question is, what is t h e distribution of the density of the electrons as a
function of energy? We first introduce the following terms defined as
N(E)
N(E)dE
N u m b e r of electrons
1 ./V
Volume
The term can be written as
(C.12)
where N represents the total density of states from zero energy to E.
T h e r a t e of change of t h e density of states becomes
s
(C.13)
469
(C.15)
A s an example, assume that the density of free electrons in silver is 5.9 X
1 0 m " = N/a .We
solve for E to b e 5.51eV.
28
D: EINSTEIN RELATION
By considering the expressions for the drift and diffusion current densities given by
Eqs. (4.15) and (4.16), it seems reasonable to expect that, other t h a n t h e particle
charge, there must b e some o t h e r commonality b e t w e e n these 2 equations since
both refer to t h e motion of carriers in t h e same solid. This commonality exists
b e t w e e n the mobility and t h e diffusion constant.
To determine the relation b e t w e e n t h e mobility and the diffusion constant, we
consider a slab of semiconductor that has b e e n d o p e d to a nonuniform distribution
of electrons in a region, as shown in D . F l a . For < 0 t h e material is intrinsic, for 0 <
= d t h e distribution of electrons is increasing while for > d the density of elec
trons is constant.
A t t h e r m a l equilibrium electrons diffuse from the region a r o u n d = d to that
at = 0 leaving behind positively ionized atoms. A n electric field is thus established
b e t w e e n = d and = 0 directed from = d to = 0. Electrons continue to diffuse
from = d to = 0, and since at thermal equilibrium t h e electron current is zero,
the electric field returns t h e same n u m b e r of electrons from = 0 to = d. In
accordance with E q . (4.15) t h e electron drift current is directed from right to left
and in the direction of the electric field while t h e electron diffusion current is from
left to right. For the total electron current to b e zero we have
qO dn
a
dx
(D.l)
n{x) = n exp
,(0)  ,(*)
 ,(x)
kT
, exp
kT
E (x)=
qtfx)
(b)
n(x) = n exp
i
~q<b(x)
for 0 =s d
kT
x =
=d
(a)
(b)
E; = 0
(c)
Figure D.F1 Distribution of electrons, potential and energy versus distance.
()
dx
n exp
t
kT
kT
dx
exp 
kT
so that
"
1
a
I '
471
area (cm )
B,E,C
BV
BVCEO
speed of light ( c m / s )
C.
je
CB0
C , C
d
D, D
n
D,
gs
nE
nC
B J T collectorbase capacitance
D,G,S
E ,E
E, E
=
=
E
E
Force (N)
frequency (Hz)
f (E)
f(E)
F e r m i  D i r a c distribution function
fj
smallsignal transconductance
472
tion region
h
hv
p h o t o n energy
current ( A )
/ , I~ I'
i, i, i
C B O
C E O
I, I
cs
ES
base
F E T drain current
I I
I , I
OV
D1S
D2S
p h o t o g e n e r a t e d current
SAT
J,J
Boltzman constant
wave vector
M O S F E T device p a r a m e t e r
L'
L ,L
L,
nE
pB
m*^ m*
mass
n'
quantum number
oc
QE
Qp
n(0)
0n
N,P
N ,P
473
heavily d o p e d ,
density of acceptors, d o n o r s
AC
N(E)
N, N
A
./V
N
N, N
hole density
p'
p (0)
p
m a g n i t u d e of electron charge
ph
total charge
Qn
0p
P(O\PM
dm
m a x i m u m value of Q
Q, Q
s
lattice spacing
r, R
resistance
r.,r ,r
BJT ohmicresistance;base,emitter,collector
storage time
O N
OFF
ox
thickness of insulator in M O S F E T
depress Kelvin
potential
474
B J T Early voltage
applied voltage
M O S F E T b o d y voltage
builtin voltage
V.
junction voltage = V V
b r e a k d o w n voltage
bi
br
ce> be
as
gs
bi
EB"*
CB>
V, V
V (V ),
V (V )
V,
EC
cc
BB
GS
DS
SB
cs
F E T D C pinchoff voltage
FB
M O S F E T D C flatband voltage
t h e r m a l voltage k T/q
F E T D C voltage at saturation
drift velocity
saturation velocity
t h e r m a l velocity
SAT
lh
B J T base width
W ,W
F E T depth
Greek
a
ot
=
=
relative permittivity
= permittivity of oxide
wavelength
mobility
mobility of electron, holes
frequency of light
=
=
resistivity
conductivity
B
T
V
,
(>
^,
= affinity of semiconductor, E E
475
chapter 1
ATOMIC STRUCTURE AND
QUANTUM MECHANICS
1.0 I N T R O D U C T I O N
Major advances in the fields of semiconductor materials and semiconductor devices
have t a k e n place within the last four decades. These advances have b e e n largely
responsible for the information revolution, b o t h in the processing and transmission
of intelligence.
Materials used in t h e fabrication of largescale circuits may be classified as:
good conductors, insulators, and semiconductors. T h e basic difference b e t w e e n the
three is their resistance to current flow, defined in terms of t h e resistivity of the
material. G o o d conductors, such as copper and aluminum, have a resistivity of less
t h a n 10~ ohmcm and are used for lowresistance wiring and interconnections in
electric circuits. Insulators that have resistivities greater than 10 ohmcm are used as
isolators of circuits and devices, and in the formation of capacitors.
B e t w e e n these two limits of resistivities lie the semiconductor materials: the
p u r e elemental semiconductors, silicon and germanium, and some IIVI, IVVI, and
IIIV compounds, t h e most i m p o r t a n t of which is gallium arsenide.
Their midrange resistivity, or conductivity, is not the direct reason for the
importance of semiconductors. Rather, it is the extent to which their properties are
influenced by light, temperature, and m o r e importantly, by the addition of minute
amounts of special impurities. Extensive changes take place in the resistivity of silicon
when one part of an impurity is added to a million parts of silicon. Silicon is abundant
in nature in the form of sand (silica) and clay. However, before silicon can b e used in
devices, major purification of the material is required. A n application that illustrates
the use of the three types of materials in the formation of integrated circuits follows.
Integrated circuits are fabricated on 8 to 20cm diameter circular sections of
silicon k n o w n as wafers. T h e major c o m p o n e n t s in integrated circuits are transistors,
3
Chapter 1
C R Y S T A L S A N D T H E UNIT CELL
Based on t h e internal a r r a n g e m e n t of the atoms, a solid is labeled as amorphous,
crystalline, or polycrystalline.
In crystalline solids, the atoms are arranged in an
orderly threedimensional array that is r e p e a t e d throughout t h e structure.
A m o r p h o u s solids have their atoms arranged in a very r a n d o m m a n n e r with no
r e p e a t e d pattern. The atoms in polycrystalline solids are so arranged that, within
certain sections, s o m e sort of a p a t t e r n of the atoms exists but the various sections
are randomly arranged with respect to each other.
Most semiconductors are crystalline in nature. Let us look into the internal
a r r a n g e m e n t of the atoms of semiconductors in the basic building block k n o w n as
t h e unit cell. T h e a r r a n g e m e n t inside the unit cells of silicon and g e r m a n i u m is
k n o w n as t h e diamond lattice because this a r r a n g e m e n t is a characteristic of dia
m o n d . D i a m o n d is a form of carbon, which is an element in Column I V of the peri
odic table. In the diamond lattice, shown in Fig. 1.1, t h e atoms are arranged within a
cube having dimension L, w h e r e L is k n o w n as the lattice constant.
The unit cell for t h e d i a m o n d lattice has an a t o m in each corner of the cube
(8), one at the center of each of the six faces (6), and four (4) internal to the cube
located along the diagonals. F r o m this total of 18 atoms, the a t o m at each corner is
shared by eight cells and each face a t o m is shared by two cells.
We will use the above information in the following examples to calculate the
density of atoms and the mass density of silicon.
EXAMPLE
1.1
Section 1.2
E X A M P L E 1.2
Determine the density of silicon given Avogadro's number is 6.023 10 atoms/mole.
23
SOLUTION From the periodic tabic of elements, the atomic weight ol silicon is 28.09. I he density ol
silicon atoms was found in Example I I to be 5
Mi "cm \
.
5 10 atoms/cm 28.09 g/mole
2.33 g
Densitv =
7

; f
=
6.023 X 10" atoms/mole
cm
22
Gallium Arsenide, a IIIV compound formed from Gallium (III) and Arsenic (V),
crystallizes in the form of a slightly different unit cell, known as the Zincblende Structure.
In this structure, G a and A s atoms are found at alternate locations inside the unit cell.
O n close study of the unit cell lattice of Fig. 1.1, we note that each atom is attached
to the four nearest neighboring atoms. This attachment of atoms, in fact, represents the
force that holds the lattice atoms together and is known as covalent bonding.
E a c h a t o m of C o l u m n I V elements has four electrons in its outermost shell.
T h e covalent bonding results from the sharing of electrons b e t w e e n atoms. W h e n
each atom, say A, is shown b o n d e d to four neighboring atoms, each of t h e four
neighboring atoms contributes one electron to the b o n d with A. A t o m A, therefore,
contributes one electron to each bond, so that two electrons are shared by a t o m A
with each of the four atoms. This sort of bonding accounts for some physical proper
ties of these solids.
1.2 W H A T A R E W E L O O K I N G FOR N O W ?
Eventually, we are interested in determining the current in a semiconductor device
in response to the application of a source of energy, such as an electric source or
light source. To d o that, we n e e d to k n o w t h e types and densities of t h e current carri
ers and their masses.
Chapter 1
1.3
F A I L U R E O F C L A S S I C A L M E C H A N I C S AT A T O M I C L E V E L S
Two experimentally observed p h e n o m e n a evolved that could not be explained by
the theories of Classical or N e w t o n i a n mechanics. These were: blackbody radiation
and the sharp discrete spectral lines emitted by h e a t e d gases. First, we will consider
blackbody radiations.
A blackbody is a h e a t e d solid labeled as an ideal radiator of electromagnetic
waves. W h e n a solid is heated, it emits radiation over a certain b a n d of frequencies.
The h e a t e d atoms vibrate so that the amplitudes and frequencies of the vibrations
seem to resemble the radiating a n t e n n a of a broadcast station. Classical mechanics
predict that this h e a t e d solid emits radiation over a continuous b a n d of frequencies.
M e a s u r e d responses, however, indicate that this is not true. A s a m a t t e r of fact, radi
ation takes place only over a certain b a n d of frequencies. F u r t h e r m o r e , t h e fre
quency spectrum of the radiation changes as the t e m p e r a t u r e of the solid is varied,
as shown in Fig. 1.2.
A similar p h e n o m e n o n occurs w h e n certain gases are heated. It was experi
mentally observed that h e a t e d gases emit radiation in small discrete quantities at
certain discrete spectral wavelengths. Again, Classical mechanics had n o explana
tion for this.
The conclusion that was reached confirmed that Classical mechanics could not
predict p h e n o m e n a that occur at microscopic scales or at atomic levels.
1.4 P L A N C K ' S H Y P O T H E S I S
The first b r e a k t h r o u g h came about with the work of Max Planck. Planck's hypothe
sis, presented in 1901, was that light is emitted or absorbed in discrete units of
energy called photons. The vibrating atoms of a h e a t e d body emit radiation at a fre
quency and the energy is restricted to certain discrete, or quantized values, given
by
= nhv for = 0,1,2,3,.
(1.1)
Section 1.5
p
Wavelength in mm
where is the frequency of radiation and h is a factor that Planck obtained by a the
oretical fit to the experimental results. This is labeled Planck's constant and its value
is 6.62 X l f r J  s .
Planck's hypothesis confirmed that, in addition to the quantizing of electro
magnetic waves, light may also b e viewed as consisting of particles labeled wave
packets, or photons. This laid the foundation for t h e theory of t h e dual n a t u r e of
light. It also paved the way in explaining the energy levels of electrons in a solid, as
it considered electrons to be particles of m a t t e r while at the same time possessing
wavelike properties.
A t this time, Rutherford explained electron behavior as resulting from a circu
lar motion a r o u n d the nucleus.
Following u p o n t h e conclusions of R u t h e r f o r d and Planck's hypothesis, Niels
Bohr, in 1913, put forward a m o d e l confirming the planetarylike motion of elec
trons a r o u n d an a t o m and included the quantumlike theory of Planck.
We will now consider B o h r ' s theories beginning with his classical m o d e l of
t h e atom.
34
Chapter 1
control the orbit of the electron. The electrical forces are d e t e r m i n e d by Coulomb's
Law, which relates t h e force exerted by one charged object on another.
In accordance with B o h r ' s model, we will use Coulomb's L a w for t h e oneelectron hydrogen a t o m and d e t e r m i n e an expression for the total energy that the
electron possesses, assuming that it is rotating in a fixed circular orbit about the
nucleus.
E n e r g y of H y d r o g e n Electron
In a hydrogen atom, t h e electrostatic attractive force b e t w e e n the nucleus,
consisting of t h e proton, having a charge +q and the electron having a charge of q
is given by
4ire r
(12)
4 /
(1.3)
% dx
(1.4)
Assuming that the reference for zero potential is at infinity, and since t h e sys
tem is assumed to possess spherical symmetry, we determine the expression for the
potential of a point at r as
~ 1
4 " 4
0
( 1
'
5 )
= ^
(1.6)
(17)
After solving for mv , the expression for the kinetic energy becomes
K.E.
(18)
(1.9)
T h e H y d r o g e n A t o m as a R a d i a t i n g A n t e n n a
The relations we just derived indicate that the electron moves in a circular orbit
having a radius r, velocity v, and total energy E. If a hydrogen a t o m is heated, the
electron absorbs energy and moves to a higher level, corresponding to a new total
energy E. It is thus possible for the electron to occupy any orbit, depending u p o n
Chapter 1
1.6 B O H R ' S Q U A N T I Z E D M O D E L
The classical theory proposed by B o h r obviously failed w h e n it was applied to the
hydrogen atom. In trying to explain the dilemma h e faced, B o h r p r o p o s e d a model
that confirmed the concept of the quantization of energy. The m o d e l was without
proof, as e m b o d i e d in his two proposals k n o w n as postulates.
His first postulate was based on his classical theory, with the exception that an
electron could remain in a circular orbit without radiating any energy. His second
postulate stated that a q u a n t u m of radiation is emitted or absorbed when an elec
tron moves from one energy level to another. We summarize the substance of the
model as follows:
1. Electrons revolve a r o u n d t h e nucleus only in certain definite circular orbits,
every orbit corresponding to a certain level of energy.
2. These orbits are labeled stationary orbits. Electrons could stay in these station
ary orbits without radiating any energy.
3. T h e transfer of an electron from an orbit of lower energy to an orbit of higher
energy requires absorption of radiation by the atom, while a fall from a higher
energy level to a lower energy level orbit results in the emission of radiation.
This energy difference, whether absorbed or emitted, is
Section 1.6
E E
2
= hv
(1.10)
4. The only stationary orbits of the electrons are those for which the angular
m o m e n t u m is also quantized.
Applying conclusion n u m b e r 4 above to the single hydrogen atom, we have
mvr
= nh/2ir
= 1,2,3,...
(1.11)
where m is t h e electron mass, is the linear electron velocity, r is the orbital radius
for a given value of n, and is k n o w n as a quantum number. Since we are still refer
ring to circular orbits of t h e electron, E q . (1.7) applies. W h e n solved for t h e velocity
it gives
n
\i/2
\4Tte rm
0
nhs
(1.12)
2s nh
0
 =
<
( e V ) =  1 3 . 6 / n , = 1,2,3,...
(1.15)
where is t h e energy of the allowed energy level of the discrete orbit, which is
obtained by replacing with t h e relevant integer. It is obvious that t h e electron can
exist only at certain discrete energy levels. It is thus possible for an electron to move
b e t w e e n any two such levels with the consequent release of radiation, going from a
higher (higher n) to a lower (lower n) energy level. T h e corresponding frequency of
radiation is d e t e r m i n e d from E q . (1.10). The electron can also b e m o v e d to a higher
10
Chapter 1
E{eV),
0
0.56 I
0.87 
5
4
1.53
3.41
13.60
 1
Section 1.7
W a v e Mechanics
11
HIGHLIGHTS
Classical mechanics could not provide explanations of the behavior of solids, such as elec
tromagnetic radiation from certain bodies.
Max Planck hypothesized that it is possible for bodies to radiate energy at certain fre
quencies, depending upon the energy that is imparted to them. He demonstrated that light
has not only wavelike properties but particlelike properties, as well.
Neils Bohr first hypothesized that electrons can take on any value of energy in their orbit
around the nucleus. He later modified this theory.
In his attempt to provide explanations of experimental results, Bohr further postulated
that electrons can exist in certain orbits only, each orbit corresponding to a certain energy.
He further stated that the transfer of an electron from one orbit to another was accompa
nied by a release or a gain of energy. Bohr's explanation served to model the hydrogen
atom but could not be extended beyond the one electron model.
EXERCISES
El1 a) Determine the binding energy for an electron in a hydrogen atom at a quantum
number 3 level. Give the answer in eV and in joules.
b) Repeat the above determination for an electron in a silicon coulombic potential, using the
results of the hydrogen atom. The relative dielectric constant for silicon is 11.8 and the effec
tive mass of a silicon electron is assumed to be 1.18 times that of the hydrogen electron.
Ans:
a) = 1.5eV
19
b) = 0.02 X 10 J
El2 a) Use Planck's hypothesis to calculate the wavelength associated with a leV photon,
b) Repeat (a) for an electron.
Ans:
b) = 1.22 X 10" m
1.7 W A V E M E C H A N I C S
A t t e m p t s to explain electron motion in terms of wave motion were p r o m p t e d by the
experimental observations of a n u m b e r of scientists. They observed that diffraction
patterns could be observed and r e c o r d e d photographically w h e n b e a m s of electrons
were passed through crystals. These results indicated that electrons h a d characteris
tics usually associated with waves, and thus should obey wavemotion equations. It
was de Broglie, in 1924, w h o laid the foundation of Wave Mechanics. H e confirmed
B o h r ' s conclusion that t h e ratio of the energy of a wave to t h e frequency was a con
stant, h, and postulated that the product of t h e momentum
of the electron and t h e
wavelength was also equal to this constant. This constant is, again, Planck's
constant.
This relationship b e t w e e n dynamic properties of the electron, such as m o m e n t u m ,
and wave properties, such as wavelength, forms the basis of wave mechanics. Thus,
12
Chapter 1
(a)
Kmv = h
(b)
(1.16)
We want to emphasize, at this point, that we are not implying the physical exis
tence of electron waves. Rather, we are thinking of a moving particle as having wave
like properties associated with it. That a particle has wavelike properties follows
from electron diffraction experiments, which show that electron b e a m s b e h a v e like
light beams, and thus must obey t h e same wave relations as light. Before carrying
the analogy any further, let us summarize the accepted dual theories of light:
a) Light has b o t h particlelike and wavelike properties.
b) Light is c o m p o s e d of photons. Thus, a light ray consists of particles (i.e., p h o
tons) each of which possesses an energy, hv, where is the frequency of t h e
light.
c) T h e intensity of light at a location is the density of t h e p h o t o n s at that location.
The analogy b e t w e e n light and electron b e a m s is extended so that t h e travel
and diffraction of a light wave in a m e d i u m of varying refractive index is analogous
Section 1.8
13
Ap Ax > h/2
(a)
AEAt>h/2ir
(b)
(1.17)
E X A M P L E 1.3
a) Determine the uncertainty in the position of a bullet of mass 20g travelling with an uncertainty
in its velocity of lOcm/s. Planck's constant is h = 6.62 X 10~ Js.
34
3
b) Repeat part (a) for an electron that has an uncertainty in its velocity of 10 cm/s.
Solution
a) At constant mass m, AvAx
/2
34
32
31
We note, while Ax for the bullet is negligible compared to its dimensions, the Ax for the electron is
many orders of magnitude larger than the diameter of the atom.
14
Chapter 1
1.9 S C H R O D I N G E R ' S E Q U A T I O N
It was Schrodinger who, in building on d e Broglie's work, first p r o p o s e d t h e wave
equation k n o w n by his n a m e . H e incorporated the quantization theory proposed by
Planck and t h e wavelike n a t u r e of m a t t e r as p r o p o s e d by de Broglie. This equation
is as basic to wave mechanics as Newton's laws are to classical mechanics and as
Maxwell's equations are to electromagnetic
theory. Schrodinger's equation in the
steadystate expresses the probability of locating a particle at a point in space where
the wave function (,,) is a m e a s u r e of that probability and is given by
+ ^(E
~ \)
= 0
(1.18)
Section 1.9
Schrodinger's Equation
15
T h e function and its first derivative are finite, continuous, and singlevalued.
Postulate 2.
The probability per unit length (per unit volume in the threedimensional
case) of finding a particle at a particular position and at a certain instant of
time is * , where * is the complex conjugate of . reason for multi
plying the function by its complex conjugate is that the probability must b e a
positive real quantity, whereas is complex for t i m e  d e p e n d e n t cases. If we
integrate * over the entire system (entire volume), the result must be
unity. H e n c e ,
16
Chapter 1
* dxdydz=
\\ dx dy dz = 1
J space
(1.19)
* space
(1.20)
dx
where k, labeled t h e wave vector, or wave n u m b e r , is given by
k =
Vch?mE/r?
(1.21)
W~
Electron in well
has W = 0 and
energy =
x=0
Figure 1.6 Electron in a potential well.
Section 1.9
Schrodinger's Equation
17
and
at = a
and
= Asm kx
(1.22)
a 8<n mE
, ,
^2
= " ^
2
(123)
Energy
n=3
n=2
= 1
=a
=0
=a
(a)
(b)
2
(c)
Figure 1.7 Variations of (a) and (b) along the well for different values of n,
showing the standing wave patterns, (c) Allowed energy levels for = 1,2, and 3.
18
Chapter 1
= .
(1.25)
Thus, at these points, = a/n and since the distance b e t w e e n zeros is a halfwave
length (one wavelength corresponds to the spatial distance over one cycle), then
= /2
a/n
and
a = /2.
(1.26)
hk
E
  s
(1.28)
= /2m
(1.29)
.*'
...**
k
>p
Section 1.9
Schrodinger's Equation
19
In accordance with Eq. (1.22) and as indicated earlier, can take on b o t h pos
itive and negative integer values. T h e wavefunction takes on positive and nega
tive values so that
= ,
fl
rtl
X
\
Application to a Potential W a l l
This application, which is illustrated by the solution of the following example, is
used to d e t e r m i n e the probability of locating an electron having energy in a
region of potential energy, W, w h e r e W is higher t h a n E.
E X A M P L E 1.4
An electron with a total energy moves, as shown in Fig. 1.9, in the onedimensional Region 1 in
which the potential energy may be taken as zero, so that W  0 for < 0. At = 0, there is a poten
tial energy barrier of height W > /.', as shown in the accompanying figure.
a) Verify that the solution of the Schrodinger equation in Region l is
j = C sin ax i D cos ax and in Region 2 is P, Ac '<
' > I lic^'K where a and d are real
numbers.
b) Determine the constants B. C, and I.) in terms of .
V
{)
Solution
a) Schrodinger's equation in Region i becomes
Incident wave
of electron having
energy
1 ^*
Potential wall
Region 1
x=0
Region 2
20
Chapter 1
dx
Ir
,,
., 8TTW
7/:
,,
 + k,, = 0, where kf =
dx
The expression, , = C sin ax + D cos ax, is a solution to the differential equation where a
In Region Schrodmgcr">; equation becomes
2
W) , = 0 for < W
or
 k:> = 0 where k = * (  /.)
:
xkl
d,
In order for to remain finite as > , must equal zero. The solutions at 0 must match
that
C sin + D cos x
xl!i
T= 0
= Ae~ o\
therefore,
D= A
By matching the derivatives at = 0, we have
it
aC  d
,
and
/I
C
ad
0
Hence,
A
0,> = /I, and C
A
ad
sin ax + A cos ex
0
= Ae""*^
2
Section 1.9
Schrodinger's Equation
21
Reflected wave
Incident wave
Incident wave of
electron having
energy
Potential wall
W>E
x= 0
Region 2
The fact that has a n o n z e r o value implies that it is possible for an electron having
energy to exist in a region w h e r e t h e barrier energy, W, is greater t h a n t h e energy of the
electron.
This example can b e modified to include a third region, labeled Region 3, identical to
Region 1 but placed to the right of R e g i o n 2. It can b e shown that it is possible for a parti
cle having energy in Region 1 to cross into R e g i o n 3 while going through a potential
barrier of energy W, where W > E. The probability of this occurring is higher if the thick
ness of Region 2 is reduced. H e n c e , the particle is said to tunnel through t h e barrier of
Region 2. This concept is applied in t h e study of the E s a k i diode, also k n o w n as t h e tunnel
diode, and in the formation of ohmic contacts.
Having established the significance and plausibility of Schrodinger's equation, we
will use it in the next chapter to establish the q u a n t u m n u m b e r s and, hence, t h e energy
levels that a hydrogen electron can have. We will t h e n extend t h e results to t h e manyelectron sample, such as silicon.
2
REVIEW QUESTIONS
Ql6
Ql7
Ql8
Ql9
HIGHLIGHTS
The properties of semiconductors are determined by phenomena on an atomic scale and,
hence, require the explanations provided by quantum mechanics' concepts.
Wave Mechanics established the intimate link between the dynamic and the wave proper
ties of the atom. Herein lies the analogy between a beam of electrons and a light beam.
De Broglie suggested that quantummechanical concepts in their duality applied not only
to electromagnetic waves but to electrons as well.
22
Chapter 1
Another link between particlelike properties and wavelike properties of matter was pro
vided by the de Broglie relationship, which expresses wavelength in terms of momentum.
The most farreaching step was taken when Schrodinger developed the extremely impor
tant and complex equations that describe the properties of electrons in a physical system.
It describes the large number of energies that an electron can have in a region where the
energy is confined.
EXERCISES
El3 A tennis player serves a 75g ball at a speed of 200 km/h. If the uncertainty in the veloc
ity of the ball is lOcm/s, determine the uncertainty in its position.
Ans:
32
= 1.4 X 10" m
El4 Use Schrodinger's second postulate to determine the expression for A in the solution
of Schrodinger's equation, in one dimension, for the infinitely deep potential well.
Ans: A =
Vl/a
El5 Determine the energy in eV of an electron for = 1 in the infinitely deep well when
a = 50A.
3
PROBLEMS
1.1 Determine the density of GaAs given the lattice constant L = 5.65A, and the molecu
lar weight is 144.63/mole. Avogadro's number is 6.02 10 atoms/mole.
1.2 Use the Bohr model for the hydrogen atom to plot potential energy and total energy
as the radius r from the proton increases. Clearly identify the magnitudes of the
kinetic energy at two separate radii.
23
1.3 Determine the velocity of an electron in the ground state of the hydrogen atom.
1.4 The laws of classical physics apply to the motion of a particle provided the dimensions
of the system are much larger than the deBroglie wavelength. For the following elec
trons, determine whether the laws of classical physics apply:
a) An electron is accelerated in the beam of a cathoderay tube that has an acceler
ating voltage of 30KV.
b) An electron that is accelerated by a potential of 100V in a device whose dimen
sions are of the order of 2cm/s.
c) An electron in a hydrogen atom.
1.5 Determine the energy of a photon having wavelengths = , and = 10A.
Express the energy in eV and J.
1.6 The antenna of an AM radio station transmitter radiates 100KW of power at
lOOOKHz.
a) Calculate the energy of each radiated photon.
b) Calculate the number of photons radiated per second.
1.7 An oscillator is operating at a frequency of 10MHz.
a) Calculate the energy of the quantum of radiation of the oscillator.
Section 1.9
Schrodinger's Equation
1.8
1.9
30
1.10
1.11
1.12
1.13
1.14
1.15
1.16
1.17
1.18
23
30
chapter 2
ENERGY BANDS AND
CURRENT CARRIERS IN
SEMICONDUCTORS
2.0
INTRODUCTION
We concluded in C h a p t e r 1 that in accordance with the Uncertainty Principle it is
not possible to specify, at the same time, the location or the m o m e n t u m of an elec
tron in a solid. This Principle points to one of the weaknesses of the B o h r hypothe
sis, which assumed that electrons could b e assigned to certain orbits, which in
essence implied that their position was known. We then decided that Wave
Mechanics' concepts are n e e d e d to explain t h e behavior of electrons in solids.
In this chapter, we will apply Schrodinger's equation to obtain information on
the hydrogen atom. We then project the results, with the necessary modifications, to
the manyelectron solid. We establish the existence of discrete energy levels from
which, because of their large n u m b e r and t h e closeness of these levels, energy bands
result. T h e highest valence b a n d and t h e lowest conduction b a n d are of p a r a m o u n t
interest. These b a n d s are separated by a region in which n o electrons of t h e semi
conductor can exist: This is the forbidden
band.
A t each of these two t o p bands, a different carrier is said to exist. These carri
ers are the electron and a vacant space k n o w n as t h e hole.
We will use Schrodinger's equation to formulate the conditions existing in the
hydrogen atom.
24
Section 2.1
25
This is not acceptable as it implies that t h e position of the electron at a given time is
k n o w n and this violates Heisenberg's Uncertainty Principle. Second, it is not possi
ble to apply B o h r ' s theory to an a t o m that has m o r e t h a n one o u t e r electron. We will
therefore consider a quantummechanical solution since it is not constrained by
these limitations. We will apply the results of the solution of the Schrodinger equa
tion to d e t e r m i n e the possible energy states of the hydrogen electron and to study
the manyelectron problem.
Quantum Numbers
T h e hydrogen atom is readily approximated by a nucleus of charge +q, fixed in
position, in whose field the electron orbits. The electron, w h e n located at distance r
from the nucleus has potential energy, given by E q . (1.6), as q /Airs r.
The poten
tial energy distribution of the hydrogen a t o m is shown in the sketch of Fig. 2.1(a).
The potential distribution is, in a way, similar to t h e potential well we consid
ered in C h a p t e r l . T h e potential energy (negative) is highest at infinite r and lowest
at t h e nucleus. A major difference b e t w e e n t h e two is the m o r e complex b o u n d a r y
2
Radial
distance r
Atom core
(a)
Chapter 2
(r, , ) = /?(/)/()/()
Section 2.1
27
In the solution of the equation for , the magnetic orbital quantum number, m
appears, determines t h e direction of t h e orbital angular m o m e n t u m , and has inte
gral values that vary from i to + , including m = 0. The word magnetic results
from the fact that an electron in an orbit represents an accelerating charge and,
h e n c e an electric current that has a magnetic field.
T h e existence of these three q u a n t u m n u m b e r s was predicted earlier by
Sommerfeld. Later, G o u d s m i t and Uhlenbeck, guided by their knowledge of spectra
m o r e complicated than that of hydrogen, predicted that the electron itself possesses
a magnetic m o m e n t and an angular m o m e n t u m quite i n d e p e n d e n t of its rotation in
an orbit a r o u n d the nucleus.
F u r t h e r m o r e , it was observed that in a manyelectron atom, the actual n u m b e r
of q u a n t u m states is twice that predicted by the combination of n, i and m . It was
then postulated that t h e electron, in its orbit a r o u n d the nucleus, spins on its own
axis in two unique a n d opposite directions with respect to its orbital m o m e n t u m .
Thus, each electron can exist in two spin states. A s a result, a fourth q u a n t u m n u m
ber, labeled the spin quantum number, m was included, which has a value of 1 / 2 .
This intrinsic spin with respect to the orbital angular m o m e n t u m results in t h e dou
bling of the n u m b e r of q u a n t u m states. Therefore, every q u a n t u m state is identified
by four q u a n t u m numbers, designated by n, , m , and m .
(
Chapter 2
space. Thus, any q u a n t u m state, identified by the four q u a n t u m numbers, can accom
m o d a t e n o m o r e t h a n o n e electron. Actually, the q u a n t u m n u m b e r s represent
a n o t h e r way of defining the wavefunction of a given electron.
To comply with t h e Pauli Exclusion Principle, it follows that each electron in
the solid, regardless of the n u m b e r of electrons, is identified by four q u a n t u m num
bers.
In Table 2.1 are listed t h e various q u a n t u m n u m b e r s together with the maxi
m u m n u m b e r of electrons that can be a c c o m m o d a t e d in an isolated atom.
The r e a d e r will question our reference to m a n y electrons, whereas the origin
of t h r e e of the four q u a n t u m n u m b e r s resulted from a solution of Schrodinger's
equation for the oneelectron m o d e l . In fact, the levels shown in Table 2.1 can be
interpreted as the possible states that the hydrogen electron can have. F u r t h e r m o r e ,
if t h e results for t h e manyelectron p r o b l e m are to b e derived from the single elec
tron model, then the various levels defined by the q u a n t u m n u m b e r s and can
a c c o m m o d a t e the n u m b e r of electrons shown in the table.
TABLE 2.1
Shell
n>0
0< s  1
<m <
m =l/2
1/2
:l/2
:l/2
All
1/2
0
0
1
+1
Levels
Maximum
Number of
Electrons
Is
2s
2p
3s
3p
3d
2
6
10
Section 2.1
Figure 2.2
29
T h e M a n y  E l e c t r o n Solid
The transition from the oneelectron p r o b l e m to the manyelectron a t o m does not
require new solutions. All that must be d o n e is to assign to the electrons t h e energy
levels d e t e r m i n e d by the oneelectron model. In addition to this, o t h e r effects
a p p e a r as a consequence of having m a n y electrons a r o u n d the nucleus.
Naturally, we expect that these electrons will exert forces on one another. If
the attractive force of the nucleus is assumed to be zero at infinity and negative else
where, then the repulsive forces exerted by the o t h e r electrons are positive in
nature. This, therefore, raises the values of t h e energy levels that are calculated by
using t h e oneelectron model. The forces referred to above are due to t h e electrons
of the same atom. The forces m e n t i o n e d above, in conjunction with the Exclusion
Principle, have the following effects:
1. E n e r g y levels calculated using t h e oneelectron m o d e l and a certain q u a n t u m
n u m b e r are lower than the actual levels. The n u m b e r also denotes the shell
identification as shown in Table 2.1.
2. Electrons having the same q u a n t u m n u m b e r but different / n u m b e r s do not
all have t h e same energy levels. Within a given shell, with a given value of n, the
5 electron (/ = 0), has t h e lowest energy, the electron (/ = /) the next higher,
the d higher, and so on. The smaller the value of I, t h e closer t h e wave function
p e n e t r a t e s to the nucleus. The energy levels are actually depressed. Those with
high I do not p e n e t r a t e into the interior of the a t o m at all. They find themselves
in a field of charge, in which the nucleus is shielded by all o t h e r electrons and
they have energies that are almost hydrogenlike. Figure 2.3 shows the distribu
tion of these levels schematically for a single atom.
3. Electrons having t h e same and I q u a n t u m n u m b e r s but different m and m
q u a n t u m n u m b e r s have slightly different energies corresponding to different
energy levels.
(
30
Chapter 2
Figure 2.3
2.2 E N E R G Y L E V E L SPLITTING
We saw in t h e last section that while the major subdivision of t h e energies of elec
trons are d e t e r m i n e d by t h e q u a n t u m n u m b e r n, every electron in an a t o m or in a
complete crystal occupies a q u a n t u m state defined by the four q u a n t u m n u m b e r s
and has its own energy level.
A s a m e n t a l exercise, let us isolate a single a t o m of a solid and consider the
effects on the energy levels as a n o t h e r a t o m of t h e same solid is m o v e d from a fur
ther distance away and closer to the first atom. Originally, the electrons of these two
atoms h a d energy levels that were identical. A s t h e second a t o m is m o v e d closer to
the first atom, the outermost electronic orbits tend to overlap. The energy levels of
t h e two atoms are slightly modified so as to a c c o m m o d a t e t h e Exclusion Principle.
W h e n a solid is formed from atoms, theoretically, energy levels are
formed for each energy level that exists in one atom. This p h e n o m e n o n is k n o w n as
energy level splitting, and at t h e n o r m a l spacing of the atoms in a solid, takes place
mainly for t h e higher energy levels. Obviously, t h e degree to which splitting takes
place increases with the extent of interactions of t h e atoms.
In Fig. 2.4, we illustrate energy splitting by showing the effects of bringing five
atoms of the s a m e solid close to one another. We note that t h e higher energy levels
associated with t h e higher q u a n t u m n u m b e r s and farthest away from the nucleus
are affected first as t h e orbits begin to overlap. The electrons in the lower energy
levels are initially shielded, but as t h e atomic spacing is decreased, lower levels
begin to split.
Section 2.3
31
n =3
=2
= 1
Atomic separation
Figure 2.4 The splitting of an energy level as five atoms of the same solid are
brought together.
2.3 E N E R G Y B A N D F O R M A T I O N
Realizing that each a t o m of silicon has 14 energy levels, corresponding to 14 elec
trons, and that t h e r e are approximately 1 0 atoms in a cubic centimeter, it is easy to
imagine the consequences of the splitting of energy levels when t h e atoms come
together to form a cubic centimeter of silicon. G r a n t e d , the splitting (of major con
sequence) in silicon takes place mainly at the 3p and 3s levels. The result of the split
ting is the formation of bands of energy at t h e n o r m a l atomic spacing R . This effect
is illustrated in Fig. 2.5 for a metal, an insulator, and for silicon. The b a n d gap energy
refers to the width of the forbidden b a n d separating the conduction and valence
bands.
In Fig. 2.5, we have shown the t o p two or three bands. It is worthwhile noting
that, at the n o r m a l atomic spacing in b o t h the insulator and silicon, the t o p b a n d is
separated from the next lower b a n d by a region where n o energy levels of the solid
exist. This region is k n o w n as the forbidden band or band gap. N o such b a n d exists
in copper.
The top b a n d is k n o w n as the conduction band, followed at a lower level by
the forbidden band and the valence band. A t the t e m p e r a t u r e of absolute zero, elec
trons occupy the lowest energy levels so that there are n o electrons in the conduc
tion band, although t h e levels are there. It is a p p a r e n t from the electronvolt width
of t h e gap that it requires less energy, t h e r m a l or otherwise, to move an electron
from the valence b a n d to the conduction b a n d of silicon than is n e e d e d to accom
plish the same in the insulator.
The specific characteristics of t h e bands, including the width of t h e forbidden
band, t h e b a n d gap energy, fundamentally determine the properties of a material.
We observe that for copper, as shown in Fig. 2.5 (a), the energy b a n d s overlap. Thus,
it is fairly easy for an electron to m o v e a r o u n d in copper; this electron can m o v e
from filled levels to higher empty levels. O n e can supply the energy to m o v e t h e
electron by raising t h e t e m p e r a t u r e or by subjecting t h e metal to a source of light.
(We are assuming that the lower energy levels are occupied first and that conduc
tion can occur only in t h e partially filled bands.)
22
Chapter 2
Silicon
gap = 1.12 eV
at T=300K
Lattice spacing
(a)
Figure 2.5 The splitting of energy levels as atomic spacing decreases for (a) a
metal such as copper, (b) an insulator such as Si0 , and (c) silicon. R is the normal
atomic spacing.
2
We also note, for an insulator in Fig. 2 . 5 (b), t h e existence of a wide gap sepa
rating the filled b a n d s from t h e empty bands. This forbidden gap, in which n o elec
trons can exist, is t h e reason for the p o o r conductivity of insulators. A great a m o u n t
of energy is required to excite the electrons from the filled levels to t h e higher
empty levels.
T h e b a n d structure of silicon indicates the existence of a forbidden gap, which
is relatively n a r r o w c o m p a r e d to that of an insulator, making silicon a b e t t e r con
ductor t h a n the insulators because the electrons require a m u c h smaller a m o u n t of
energy to move from the filled lower bands to the unfilled u p p e r bands. Electrons at
absolute zero t e m p e r a t u r e occupy the lowest levels of energy in accordance with
the Pauli Exclusion Principle so that each is theoretically stacked, in energy, over
a n o t h e r until all electrons are accommodated.
Let us consider silicon in m o r e detail. The a t o m has Is , 2s , 2p , 3s , and 3p ,
where all states u p to those in the 3p level ( = 1) are filled to the m a x i m u m num
ber of electrons that they can accommodate. The 3p level has six states, which
require six electrons to fill all the states. Only two of t h e six states are filled with
electrons. A t higher temperatures, some electrons will be raised from the 3s level to
empty states at 3p.
2
Section 2.3
33
3p 2N electrons
i= 1
n=3
3 s
6N states
2N electrons
i =
2N states, 2N electrons
6N states, 2N electrons
>
2/7
2s
Rr
normal spacing
Figure 2.6
34
Chapter 2
Low energy levels p r o d u c e n a r r o w b a n d s with a wide gap and high levels pro
duce wide b a n d s with a narrow gap.
Material
GaAs
Si
Ge
(0)^^(eV),rinK
g
,(0)
(10 )
1.519
1.17
0.7437
5.405
4.73
4.774
204
636
235
1.422
1.125
0.663
REVIEW QUESTIONS
Q21 Where do the quantum numbers come from?
Q22 If each electron in a solid is permitted to occupy only one energy level, estimate the
separation in eV between two adjacent levels in silicon, assuming that an energy band
is about leV.
Q23 What phenomena cause energy levels to split?
Q24 When atoms are brought together to form a solid, which energy levels split first: the
high levels farther from, or the one closer to, the nucleus and why?
Q25 What particular aspects completely specify the electrical conducting properties of a
solid?
Q26 If is the atomic number of a solid, what is the charge of the nucleus?
Section 2.3
After splitting
iVisolated
atoms
35
Conduction
band
4JVstates
no electrons
atT=0K
6N states in 3p
2N states in 3s
4N electrons
4N states
4N ectrons
Valence band
()
Figure 2.7 Development of the Band Model from Fig. 2.6: (a) available states and
electrons in atoms, (b) splitting causes redistribution of states and electrons, (c)
final band formation for > 0. The number of electrons available in the
conduction band depends on temperature. At = OK, there are none.
HIGHLIGHTS
EXERCISES
E21 A certain atom has 12 electrons. Identify, in increasing order of energy, the set of quan
tum numbers that correspond to each of the 12 electrons.
E22 How many of the electrons in an atom of silicon, at thermal equilibrium, form each of
the conduction band and the valence band?
E23 Determine the electron configuration, in ascending order of energy, for the following
elements: Ge, Ga, As, and P.
36
Chapter 2
2.4 M A T H E M A T I C A L M O D E L O F B A N D F O R M A T I O N
In the last section, we d e t e r m i n e d that an electron in a hydrogen a t o m is allowed to
possess discrete values of energy, E, with each energy level identified by a set of four
q u a n t u m numbers. N o information was provided as to the preclusion of t h e exis
tence of electrons in certain energy bands, such as the forbidden b a n d separating
the conduction and valence bands. N o r was information available as to the shape of
the energy distribution as a function of t h e wave n u m b e r
k(mr/a).
In this section, we will consider a mathematical m o d e l for the potential energy
distribution in an array of atoms and present the results of this approximate model.
The electronic potential energy of t h e hydrogen atom is given by E q . (1.6) and
plotted in Fig. 2.8(a) for one a t o m and plotted in Fig. 2.8(b) for two atoms. Assuming
a onedimensional hydrogen crystal, t h e potential energy sketched as a function of
distance, would a p p e a r as shown in Fig. 2.8(c) with a lattice constant, L. The precise
w
(c)
Figure 2.8 (a) Potential energy of an electron in the vicinity of one hydrogen
atom; (b) identical to instance (a) except two atomic cores are included; and (c)
identical to instance (a) except an array of atoms are shown.
Section 2.4
37
w
, r
b
X
L
Figure 2.9 One dimensional KronigPenney model of the potential energy
distribution.
Chapter 2
and from the equations resulting from substitution of the expressions for the wave
functions in Schrodinger's two equations. The two i m p o r t a n t solutions consist of an
expression on t h e lefthand side of each of the equations, which include and W ,
and one term on t h e righthand side of each equation, given by cos k {a + b), whose
value must lie b e t w e e n 1 and + 1 . Certain values of energy, E, m a k e the lefthand
side of these equations greater t h a n 1 or less than  1 . Those ranges of energies
result in t h e forbidden bands. The o t h e r energy ranges are those that are either
occupied by electrons or that could be occupied by electrons. A plot of these two
equations is exhibited in Fig. 2.10(a). For the sake of comparison, we show in Fig.
2.10(b) the distribution of energy, E, for t h e potential well discussed in C h a p t e r 1.
T h e following observations can b e m a d e from the plots:
0
T h e lefthand sides of the equations that are used to sketch Fig. 2.10 are not
changed if kL changes by 2. For reasons related to the masses of t h e parti
cles and that will b e c o m e m o r e clear later, it is convenient to shift the curves
of t h e second and third allowed b a n d s by 2 T T / L along the jcaxis.The values
Figure 2.10 (a) Energy versus wave number k using the KronigPenney model. The dashed curve that
is superimposed is the free particle solution, (b) Energy versus wave number for the potential well.
Section 2.4
39
of the second and third allowed b a n d for positive k are shifted by 2TT/L and
those for negative k are shifted by 2TT/L. In this manner, the sketch of Fig.
2.10 is limited to the region /L to /L, as shown by the dotted curves
m a r k e d c',d',e',
a n d / ' , replacing the curves c, d, e, a n d / . T h e new curves are
shown sketched in Fig. 2.11 and they represent the energy profile for the
valence and conduction bands corresponding to = 1,2, and 3.
We have shown in Fig. 2.11 energy curves for three values of n. Obviously, by
use of t h e two equations referred to earlier, it is quite possible to extend the
sketches to higher values of n. H o w e v e r , the ones shown in the figures are a d e q u a t e
to illustrate t h e relevance and benefits of the K r o n i g  P e n n e y model.
It is important to realize that the onedimensional K r o n i g  P e n n e y model
bears a very general resemblance to the actual conditions existing in a crystal. In a
real threedimensional crystal, the Ek relationships are m u c h m o r e complicated
than those obtained in Fig. 2.10. H o w e v e r , the K r o n i g  P e n n e y m o d e l results have
exhibited two properties that are extremely important. First, bands exist in which
electrons cannot exist and, second, the shape of the Ek curves at the locations of
the forbidden b a n d s indicate that an opposite concavity exists b e t w e e n the shapes
of t h e allowed b a n d s above and below a certain forbidden band. We will refer to this
p r o p e r t y in relation to the effective mass later in this chapter.
Direct a n d Indirect S e m i c o n d u c t o r
The actual b a n d structures of semiconductors are m u c h m o r e complex than those
shown in Fig. 2.11. O n e distinguishing feature of semiconductors is the location of
the conduction b a n d energy minimum with respect to the valence b a n d m a x i m u m
on t h e Ek diagrams. In silicon and germanium, and as shown in Fig. 2.12(b), the
valence b a n d m a x i m u m does not occur at the conduction b a n d minimum. The
valence b a n d m a x i m u m in all semiconductors occurs at k = 0, whereas the conduc
tion b a n d m i n i m u m for Si and G e occurs at a different k, indicating a difference in
m o m e n t u m b e t w e e n these two points. In gallium arsenide, the conduction b a n d
minimum and the valence b a n d m a x i m u m occur at k = 0, as shown in Fig. 2.12(a).
 3 allowed band
Forbidden band
= 2 allowed band
Forbidden band
= 1 allowed band
40
Chapter 2
Gallium arsennide
(a)
Silicon
2.5 C O V A L E N T B O N D M O D E L
A representation that complements the energy b a n d diagram known as the covalent
bond model is shown in Fig. 2.13. This diagram is a twodimensional form of the dia
m o n d lattice structure shown in Fig. 1.1 in which each a t o m is b o n d e d to its four
nearest neighbors.
The bonding is a result of the fact that each a t o m shares four outermostorbit
electrons with four adjacent atoms. These four electrons occupy the 3s and 3p levels
in the energy b a n d representation, shown in Fig. 2.6. A t = OK, t h e r e will be AN
Section 2.5
\/ \/ \/
'
"
'
\/
%
"
41
XXXX
//
Free electrons
(ft)
Figure 2.13 Covalent bond model of a semiconductor (a) at low temperature and
(b) at room temperature.
lines in Fig. 2.13(a), being the n u m b e r of atoms in the crystal, which normally fill
the lower b a n d (the valence b a n d ) shown in Fig. 2.6 and Fig. 2.7(b).
T h e lines shown in the figure represent one electron each and the circles rep
resent the a t o m core, which includes t h e nucleus and all other electrons of the atom
except the ones in the outermost orbits, such as t h e 3s and the 3p levels in silicon. It
is of interest to realize that it is the covalent bonding that imparts hardness to
G r o u p I V semiconductors, such as G e and Si. Of course, carbon is the hardest m a t e
rial. In IIIV semiconductors, such as G a A s , t h e binding forces represent the cova
lent bonding in addition to t h e ionic forces so that the strength of the bonding is
high. This translates into a higher b a n d gap energy and a higher melting point w h e n
c o m p a r e d to G r o u p I V semiconductors having the same atomic n u m b e r and lattice
constant.
A t the n o r m a l atomic spacing, R , and for a semiconductor at OK, all electrons
are attached to their cores and are said to occupy their lowest possible levels of
energy in the b a n d diagram. A n increase of t e m p e r a t u r e imparts energy to the elec
trons, and at 300K ( r o o m t e m p e r a t u r e ) , m a n y electrons are m o v e d from the top of
the valence b a n d to the b o t t o m of the conduction band. If the imparted energy is
exactly equal to t h e b a n d gap energy, the electrons at t h e top of t h e valence b a n d
are transferred to the b o t t o m of the conduction b a n d and possess potential energy
E For energy greater than t h e b a n d gap , t h e electrons at the t o p of the valence
b a n d m o v e u p to t h e conduction b a n d and acquire kinetic energy; their total energy
places t h e m above E
A t = 300K, approximately one in 1 0 silicon electrons escapes a covalent
b o n d and becomes what is k n o w n as a free electron able to travel t h r o u g h o u t the
structure. In the energy b a n d model, an electron becomes free w h e n it has b e e n
m o v e d from the t o p of t h e valence b a n d to the b o t t o m of the conduction band.
With this in mind, we will n o t e one distinction b e t w e e n metals and semicon
ductors. T h e conductivity of metals decreases with increasing t e m p e r a t u r e , whereas,
0
12
42
Chapter 2
20
2.6 C U R R E N T C A R R I E R S E L E C T R O N S A N D H O L E S
A unique feature of semiconductors that results from their particular energy b a n d
diagram is that two types of carriers exist: electrons in the conduction b a n d and
holes in the valence band. In metals, conduction consists of the controlled motion of
electrons only. So, where d o t h e holes c o m e from a n d h o w d o they b e h a v e as current
carriers?
T h e Hole
W h e n a p u r e semiconductor, such as silicon, which is initially at = OK, acquires
t h e r m a l energy equal to or greater than the b a n d gap energy, , electrons are
excited from the top of the valence b a n d and into the conduction band. They
b e c o m e free electrons. For every electron that leaves the valence band, a vacancy in
the covalent bonding is left behind into which another electron in that valence b a n d
may move. W h e n an electric field is applied to the silicon, the electrons in the con
duction b a n d acquire velocity in a direction opposite to that of the field. Similarly,
the electrons in the valence band, which m o v e d to fill the vacancies, gain velocity. By
having the vacancy occupied by a n o t h e r electron, the vacancy moves in t h e direc
tion of the field. This vacancy is the hole. Thus, b o t h the electron and the hole cause
electric current in the same direction with the hole moving in the direction of the
field and the negatively charged electron moving opposite to the direction of the
electric field.
O n e crude analogy is that of a twolevel parking garage where, initially, all the
cars are arranged in a single row in the ground floor and the u p p e r floor is com
pletely empty. Until a car is m o v e d to the first floor from the ground floor, there can
be n o motion of cars in either floor since the ground floor is full and the first floor is
empty. W h e n the front car is m o v e d upstairs, then all the cars behind the space that
the front car occupied can now move one carlength forward, thus causing the
empty space to move towards the back. The car that was moved upstairs is available
for motion and it is analogous to our electron in t h e conduction band. The vacancy
that was created on the ground floor is analogous to the hole and it moves in a
direction opposite to the motion of the cars.
A n o t h e r analogy is represented by the cylinders shown in Fig. 2.14.
The b o t t o m cylinder in Fig. 2.14(a) is completely filled with a liquid and the
top cylinder is empty. These cylinders represent the valence and conduction bands
Section 2.6
43
Liquid motion
(a)
(b)
Figure 2.14
respectively of silicon at absolute zero. In this state, there can b e n o motion of liquid
in t h e b o t t o m cylinder and, obviously, n o n e in the top cylinder. W h e n a small vol
u m e of liquid is transferred from t h e b o t t o m cylinder to the t o p cylinder, a possibil
ity exists for motion of liquid in b o t h cylinders: the small volume of liquid in the t o p
and the b u b b l e left behind in t h e b o t t o m cylinder.
By tipping b o t h cylinders to the right side in Fig. 2.14(b), gravity forces the liq
uid in the t o p cylinder to move right and the bubble in the b o t t o m cylinder to m o v e
left. T h e small volume of liquid and t h e bubble move in opposite directions. The
force of gravity is analogous to an applied electric field, t h e small volume is analo
gous to the free electron and t h e b u b b l e is analogous to t h e hole.
A n a l y t i c a l Description of t h e Hole
A m o r e analytical description of the hole is d e t e r m i n e d as follows:
Electrons are thermally excited from the valence b a n d to the conduction
band, leaving empty states in the valence band. W h e n an electric field is applied,
electrons in t h e conduction b a n d are accelerated, and so are the electrons in the
valence b a n d as they m o v e into the empty states. T h e current density ( a m p s / m ) of
electrons in the valence band, J , can b e d e t e r m i n e d by a s u m m a t i o n of t h e motion
of all the electrons in the valence b a n d as:
2
VB
Kb =
" 1d
(23)
^~qv
filled band
d
0 0
_^~qv
empty states
d
Chapter 2
J =
VB
(25)
\/ \/ V
(a)
P.E. of electron
1
K.E. of electron
on on
.a
a
K.E. of hole
"
xt
Increasing
electron
energy
/y
vYyv
vVV v
Hole
Electron
(b)
P.E. of hole
(c)
Figure 2.15 Generation of electron/hole pair; (a) energy band model, (b) covalent bond model, and (c)
directions of energy increases of electrons and holes.
Section 2.7
Effective Mass
45
Electrons located above E and holes below E , as shown in Fig. 2.15(c), are
said to have kinetic energy given by the difference b e t w e e n their energy location
and the respective b a n d edge energy.
c
2.7 EFFECTIVE M A S S
We have just concluded that a hole possesses a positive charge, q, and t h e electron's
charge is q where q = 1.6 X ICC coulombs. H o w e v e r , the mass of t h e electron in
t h e conduction b a n d and t h e mass of the hole in the valence b a n d are quite different
and b o t h differ from the mass of an electron in a vacuum.
The mass of a particle will b e defined as the ratio of t h e net force on t h e parti
cle to t h e acceleration that t h e particle experiences. In using t h e mass of an electron
in v a c u u m or free space, one completely ignores the effect of the crystal in which the
electron is immersed. In fact, the forces within the crystal are much greater than t h e
force exerted by an electric field of the strength normally applied to an electron.
The forces in t h e crystal are k n o w n as lattice forces.
In o r d e r to account for the lattice forces in the equation of motion, we intro
duce t h e following two equations:
19
dv
F + lattice forces = m
"dt
dv
n
F=
(2.6)
E = f~
2m
(2.7)
0
Chapter 2
hk/2ir
(2.9)
h k /8TT m*
(2.10)
2
q%/m
(2.12)
Section 2.8
47
m*/m
m*/m
Si
Ge
GaAs
1.18
0.81
0.55
0.37
0.065
0.52
2.8 C O N D U C T O R S , S E M I C O N D U C T O R S , A N D I N S U L A T O R S
The feature that distinguishes these three types of materials is the extent of their
ability to conduct electrical current. For conduction of current to take place, t h e fol
lowing r e q u i r e m e n t s must b e met:
1. T h e r e must b e energy b a n d s that are partially filled with electrons. Since elec
trons occupy the lowest b a n d s first, these partially filled b a n d s are located at or
near the top of an energy band.
2. A n electric field must b e applied to accelerate the electrons in the partially
filled bands. In being accelerated, electrons gain energy, but c o m p a r e d to the
energies separating the bands, if a forbidden b a n d exists, this energy is very
small.
For energy bands to b e c o m e partially filled with electrons, either electrons are
lifted, by acquiring energy, from a completely filled b a n d to a completely empty
band, or there are b a n d s of energy that are empty and overlap filled b a n d s without
the existence of forbidden bands.
The structural difference b e t w e e n insulators and semiconductors is that insu
lators have a very wide b a n d separating the valence b a n d from the conduction
band, whereas semiconductors have a m u c h smaller b a n d gap energy. In conductors,
the t o p occupied b a n d is only partially filled with electrons. We show in Fig. 2.16 the
relationship b e t w e e n conduction and valence b a n d s in t h e three materials.
Chapter 2
Overlap
Overlap
Filled
Filled
Partially filled
I'anialK filled
Filled
Filled
Empty
Partially filled
Metals
Semiconductors
E <2eV
g
tilled
Mostly filled
Empty
Empty
Insulators
E >3eV
g
Filled
( a ) T = 0K
Filled
(6)=300
Section 2.8
49
REVIEW QUESTIONS
Q27
Q28
Q29
Q210
Q211
Q212
Q213
HIGHLIGHTS
No forbidden band exists in a metal; a wide gap (several eV) exists in insulators. The
gap is 1.12eV in silicon and 1.41eV in gallium arsenide.
The effective mass of an electron is inversely proportional to the second derivative of
the expression for energy as a function of the wave vector. Thus, at the top of the
valence band, the mass is negative; it is positive at the bottom of the conduction band.
A negativemass, negativecharge electron that is subjected to an electric field is accel
erated in the direction of the electric field, opposite to that of a negativecharge, posi
tivemass electron. The negativemass, negativecharge electron is interpreted to be a
hole, which has positive charge and positive mass.
EXERCISES
E24
50
Chapter 2
Ans: a) / = 3.43 X 10 Hz
b) = 8.74 X 10" m
10
E2S At room temperature, it is calculated that there are 10 electrons per cm that have
moved from the valence band to the conduction band of silicon.
a)
What is the density of holes?
b)
What fraction of the electrons have moved into the conduction band?
Ans: a)
10
10 cm"
b) 1 in 10
13
PROBLEMS
2.1 A certain atom has the following subshells:
2/,3d,3/,3g,5g
Determine:
a)
The values of the quantum numbers and I that correspond to each of the
subshells.
b)
Which of the subshells are allowable?
2.2 For the subshells of Problem 2.1, determine the number of electrons in each.
2.3 The expression for the potential energy of an electron in a onedimensional crystal
lattice is given by Eq. (1.6). Plot the potential energy of the electron inside the lattice
considering three atomic cores (Zq) located at = 0, = a and = 2a. Choose a suit
able scale for potential energy and cover the region from = 0 to = 3a.
2.4 The Ek diagram for a particular energy band of a certain material is shown in the fig
ure below. An electric field is applied to the material in such a direction as to cause a
force in the negative kdirection. Determine:
a)
b)
c)
Chapter 2
Problems
51
2.5 The Ek diagram for a certain material is shown in the figure below. The diagram for a
free electron is shown dotted. Sketch as accurately as possible:
a)
dE/dk versus k.
b)
d E/dk versus k.
2
2.6 The wavefunction represented by the timeindependent equation for the hydrogen
atom, in spherical coordinates, is given by:
1 d I^ 3\
fdr { drj
a)
b)
1 J)_ I
V
3\
3 J
r sin 6
2
c)
d)
ax
x e
Jo
dx = n\/a
n+1
chapter 3
INTRINSIC AND
EXTRINSIC
SEMICONDUCTORS
3.0 I N T R O D U C T I O N
We learned in the last chapter that each electron in a solid is assigned a specific level
of energy and that these levels are so close to each other that they m e r g e into bands
of energy. We also found that these b a n d s of energy that electrons may occupy are
separated by forbidden energy gaps w h e r e n o electron of t h e solid can exist. In
semiconductors, the t h r e e t o p b a n d s are the valence band, t h e forbidden gap, and
the conduction band.
In this chapter, we will d e t e r m i n e expressions for t h e densities of electrons
and holes in the conduction b a n d and valence b a n d respectively, for a semiconduc
tor into which n o impurities have b e e n deliberately introduced. We will t h e n deter
mine t h e densities w h e r e small traces of impurities are added to the semiconductor.
So that we can d e t e r m i n e currentvoltage relationships, we n e e d to explain the
processes that cause electrons and holes to move. These we will study in C h a p t e r 4.
3.1 D E N S I T Y O F STATES
The densities of electrons and holes in the conduction and valence b a n d s respec
tively are d e p e n d e n t on two factors. First, we n e e d to k n o w the density of states
available for occupancy. Then, we d e t e r m i n e the probability of occupancy of the
various states at their respective energy levels.
The density of states refers to the n u m b e r of available electron states for unit
volume per unit energy at a certain energy level. The expression for t h e density of
states in a metal, derived in A p p e n d i x C, is given by
52
Section 3.1
M*) = i(ffV
Density of States
53
(3.1)
N(E)
0
54
Chapter 3
3.2 F E R M I  D I R A C D I S T R I B U T I O N F U N C T I O N
The relations derived in A p p e n d i x C describe conditions occurring at absolute zero.
A s we raise t h e t e m p e r a t u r e , we expect some electrons to move to higher energy
levels and thus to energy levels previously unoccupied. The question then is: W h a t is
the distribution of t h e electrons with respect to energy as t h e t e m p e r a t u r e is raised
above zero, subject to t h e assumption that the density of states is i n d e p e n d e n t of
temperature?
We also n o t e d that t h e Pauli Exclusion Principle states that a q u a n t u m state
can b e occupied by only two electrons that have opposite spin. A t absolute zero, the
electrons therefore occupy t h e lowest energy levels possible. Because of the Pauli
Exclusion Principle, the distribution of electrons is governed by Fermi statistics. The
Fermi statistics describe what occurs as t h e t e m p e r a t u r e is raised above zero. It
describes t h e probability that a state of energy is filled by two electrons of o p p o
site spin.
T h e FermiDirac distribution predicts that as t h e t e m p e r a t u r e increases, an
energy state corresponding to energy will have a higher probability of being occu
pied t h a n at a lower t e m p e r a t u r e . The Fermi function is given by
1
A
e x p [ (  E )/kT]
F
( 3
+ 1
2 )
where f(E) is t h e probability that a state with energy is occupied, E is the Fermi
energy, k is B o l t z m a n n ' s constant, and is the absolute t e m p e r a t u r e . We will discuss
shortly t h e physical significance of E .
The above expression can also b e interpreted to m e a n that at a certain tem
p e r a t u r e T, t h e probability of occupancy of a state is lower, the higher the energy
level of that state. Thus, the states at higher energy levels are less likely to b e occu
pied than states at lower energy levels. Since we observe from E q . (3.1) that t h e den
sity of states is higher at the higher energy level, we can conclude that at the higher
energy levels, w h e r e the states are m o r e numerous, the probability of occupancy of a
state is m u c h lower than at t h e lowerlevel, lessnumerous states. T h e r e a s o n for this
seemingly a b n o r m a l behavior results from the fact that electrons initially occupy
the lowestlevel states w h e r e the Pauli Exclusion Principle allocates only two elec
trons of opposite spin to each state.
A l t h o u g h E is defined in A p p e n d i x C, we shall relate its significance in this
context as well. A t t e m p e r a t u r e s approaching absolute zero and for < E , E q .
(3.2) indicates t h a t / ( ) will b e equal to 1 and, for > E ,f(E)
will b e zero. Thus, at
absolute zero, E is t h e dividing energy below which all states are occupied with
p
Section 3.2
55
f(E)
electrons and above which all states are vacant. This is the same definition for E we
arrived at in A p p e n d i x C.
A s t h e t e m p e r a t u r e increases, some electrons acquire enough energy to move
into states above E . Figure 3.2 shows t h e FermiDirac distribution function for
three temperatures. N o t e t h e roundingoff of the curve at the higher energy levels.
Also n o t e that at = E , f(E) is always 1/2. We can therefore define E as being
that energy level for which the probability of occupancy is 1/2; or we can say that
over a long period of time, the probability is that half t h e states are filled at the level
E.
F
We have seen from E q . (3.1) that N(E) gives the density of the states at energy
E. A t t h e r m a l equilibrium, the density of electrons, dn, having energy b e t w e e n
and + dE is t h e n
dn = f(E)
N(E) dE
(3.3)
By substituting f o r / ( ) from E q . (3.2) and for N(E) from E q . (3.1) and using
the effective mass for a semiconductor, we can then write E q . (3.3) as
2
_^/2)(8m*/h r E^dE
a
1 + e x p [ (  E )/kT]
56
Chapter 3
though it is in t h e forbidden gap. The answer is that energy states obviously exist in
the forbidden b a n d but these states do not accommodate electrons of t h e semicon
ductor.
Since t h e probability of occupancy of a state i s/ ( / ? ) , the probability of vacancy
of a state b e c o m e s [1/()].
In the following example, we illustrate the application of the Fermi distribu
tion function to a metal.
E X A M P L E 3.1
(a) Determine the probability of occupancy of a state that is located at 0.259eV above E at:
F
i) = OK
ii) = 300K
iii) = 600K
(b) Determine the probability of vacancy of a state that is located at 0.4eV below E atT=
p
300K.
(c) Repeat part (b) if the state is at O.OleV above /.', at I 300K.
Solution
(a) i) At T= OK
M
Sx259 =
1 + exp 0
ii) A t T = 300K
eXP
4.54
1 0
5
075z5v
iii) At = 600K
jtE)
J X
'
 = 6.69 X 10~
0.259
1 + exp . .
0.0518
n r
1/( ) = 1
' o.o,  1
5 9 5
025^
3.3 C A R R I E R DENSITIES
Semiconductors b e c o m e useful only after special impurities are a d d e d to t h e m . In
its almost p u r e form, and w h e n n o impurities are added, the semiconductor is
labeled as intrinsic. It is extrinsic w h e n selected impurities are a d d e d and t h e semi
conductor is said to be doped with impurities.
Section 3.3
Carrier Densities
57
Densities of S t a t e s in t h e Conduction a n d V a l e n c e B a n d s
Initially, we will d e t e r m i n e expressions for t h e electron and hole densities in an
intrinsic semiconductor. We will assume that the expression for the density of states
given by E q . (3.1) applies for all conduction b a n d and valence b a n d energies with
the a p p r o p r i a t e effective masses substituted.
T h e r e are about 5 1 0 atoms/cm in silicon. A t = OK, the AN q u a n t u m
states in t h e valence b a n d are filled with electrons and the AN states in t h e conduc
tion b a n d are empty. These states are distributed t h r o u g h o u t t h e bands. The states of
i m p o r t a n c e in semiconductor devices, as we shall see later, are the ones near the top
of t h e valence b a n d and t h e states near t h e b o t t o m of the conduction band.
To obtain the densities of states n e a r the b a n d edges, we let E represent t h e
m i n i m u m electron energy in the conduction b a n d and E represent the m a x i m u m
hole energy in the valence band. We recall that an increase of t h e energy of an elec
tron in the conduction b a n d corresponds to t h e electron moving u p on t h e energy
scale in the conduction band. A n increase of the energy of a hole in the valence
band, then, corresponds to the hole moving down into the valence band. Thus, we
change t h e variable for integration so that E q . (3.1) b e c o m e s (E E ) for t h e con
duction b a n d and it b e c o m e s (E E) for t h e valence b a n d states. Therefore, the
functions for t h e densities of states in t h e conduction and valence bands, as shown in
Fig. 3.3, b e c o m e
22
(3.5)
where (E) is t h e n u m b e r of states per unit volume per unit energy at in t h e con
duction band. The corresponding density of hole states in t h e valence b a n d at is
(3.6)
where (E) is t h e density of states in t h e valence b a n d and is assumed to be
located at E. Using the above expressions, we d e t e r m i n e t h e electron density in the
conduction b a n d from
(3.7)
N(E)
Chapter 3
where f (E) is given by E q . (3.2). The hole density, which represents the density of
vacant states, is given by
c
i
where f (E)
v
Ey
N (E)f (E)dE
p
(3.8)
oo(BOTTOM)
= exp [  (  E )/kT]
for > E
(3.9)
T h e reason for restricting the energy level to values greater than E is that an
energy level < E in E q . (3.9) makes the probability greater than unity, which is
meaningless.
Let us briefly digress to investigate the condition for which t h e approximation
of E q . (3.9) is fairly valid. U p o n comparing f(E) from Eqs. (3.2) and (3.9) at various
values of (E  E ), we n o t e that for (E  E ) = 3kT,f(E) calculated from E q . (3.9)
differs from that of E q . (3.2) by about 5 percent and at (E  E ) = 4kT t h e differ
ence is a b o u t 1.8 percent. We will arbitrarily state that E q . (3.9) is valid provided the
Fermi level is at least 3kT below the b o t t o m of the conduction b a n d where all free
electrons reside.
For energy levels below E , the probability of a vacancy (or hole occupancy)
can b e written from E q . (3.2) as
F
_ f)
(F
1
exp [(E  E )/kT]
F
+ 1
ex [(E
E )/kT]
1 + exp [(E  E )/kT]
V
'
1  f (E)
c
= exp  [(E
 E)/kT]
(3.11)
The restriction to E q . (3.11) implies that t h e Fermi level is at least 3kT above
the t o p of the valence band.
Since the electrons occupy states in the conduction b a n d and holes represent
unoccupied states in t h e valence band, we conclude that t h e validity of the
B o l t z m a n n approximation is restricted to t h e range of Fermi energies extending
from 3kT above the top of t h e valence b a n d to 3kT below the b o t t o m of the con
duction band.
Section 3.3
Carrier Densities
59
= exp  ( E 
f (E)
= exp(E E)/kT)
Ep/kT)
(a)
(b)
(3.12)
Sketches graphically describing the operations in Eqs. (3.7) and (3.8) are
shown in Figs. 3.4 and 3.5 for t h e electron and hole distributions in the conduction
and valence b a n d s respectively, for different locations of E .
Before we p r o c e e d with evaluating the integral, we will clarify two questions
that an interested r e a d e r may raise. T h e first concerns changing the limits of inte
gration. We have replaced the t o p of t h e conduction b a n d by plus infinity and the
b o t t o m of the valence band by minus infinity. The reason is to simplify the integration.
However, t h e change is quite justifiable since t h e exponential functions f {E) and
f(E) decrease so rapidly as we m o v e away from the b a n d edges that t h e densities of
the carriers b e c o m e negligible as we move a few kTs away from the edges of the
conduction and valence band edges. The second question concerns the effective masses
We confirmed at t h e end of C h a p t e r 2 that the effective mass, m*, is a function
F
)xfc(E)
)[i/c(
Figure 3.4 The product of the distribution of energy states and the Fermi function gives the
distribution of electrons and holes for E midway between E and E .
F
60
Chapter 3
N (E)xf (E)
N(E)
N {E)
N (E)x[lf (E
Ey
N (E)xf (E)
n
Ei
E
F
N(E)x[lf (E
c
N (E)
~lfc(E)
Figure 3.5 This is Fig. 3.4 repeated except that (a) the Fermi level is assumed to be above the
middle of the gap and (b) the Fermi level is assumed to be below the middle of the gap. Note
that at = E f(E) = 0.5.
p
of the energy level. In Eqs. (3.13) and (3.14) we have, however, assumed that this
mass is constant and we placed it outside the integrand. The justification is that since
the exponential in t h e expressions decays very rapidly as we move away from the
conduction and valence b a n d edges, E and E , our interest is in the carrier densi
ties over t h e n a r r o w portions above E a n d below E . In these small regions, t h e
effective mass is relatively constant as we indicated earlier.
To integrate E q . (3.13), we l e t x = ( E E )/kTso
that t h e integral becomes
c
v _
{xkf)
kT e x p ^
(xkT
+ E
kT
dx
(3.15)
The integral has a value of
/ 2
and becomes
2
1/2
n=im2\ \^4^^){kT)^
h
2
kT
2~
(3.16)
Section 3.3
Carrier Densities
61
(E
Ey\
(317)
(318)
(3.19)
where TV and TV are given by
(3.20)
(3.21)
3
the effective masses of electrons and holes. Relative masses of electrons and holes
are shown in Table 2.3.
We r e m i n d t h e r e a d e r that t h e above relations were derived subject to the
approximation that E was n o closer than 3kT units of energy to either the conduc
tion b a n d or valence b a n d edges.
F
TABLE 3.1 Effective density of states and band gap energy at = 3Q0K.
Semiconductor
Si
Ge
GaAs
iV (cm~ )
N (cm~ )
E (eY)
19
3.22 x 10
1.03 x 10
4.21 1 0 "
19
1.83 x 10"
5.35 x 10
9.52 10
18
18
1.12
0.66
1.42
By taking the product of and from Eqs. (3.18) and (3.19), we obtain the
interesting result
np = N N e x p  (
c
y
r
(322)
62
Chapter 3
band, a hole is created in t h e valence b a n d . We define this density as n and nf, from
E q . (3.22) b e c o m e s
i
np = nf = N N e x p  ( ^ ^ )
c
(323)
The intrinsic carrier density, , assuming that t h e effective masses d o not change
with t e m p e r a t u r e and using Eqs. (3.20) and (3.21) in E q . (3.23), becomes
;
n = K 7/
t
3 / 2
exp^
(3.24)
1 0
 3
 3
22
23
10
13
. = f f f OWT expg)
(3.25)
Section 3.3
Carrier Densities
Solution
a) n?  N N
e~ g
Using the values of N and /V. from Table I and using an of l.lleV, we determine nf as
E
/kT
nf = 0.9772 X 1 0 W
,J
n. = 9.885 X 10 cm
b) At = 600K, we determine E , from the relation in Table 22, to be 1.032eV. Using Eq. (3.24)
and (300K) = 10'cm " , we have
g
(300K)
UoO/
n,.(600K) = 3.337 x 10 cm
15
= 1.137 10
63
64
Chapter 3
We highlight the influence of in the exponential by (a) assuming that /. does not change
with temperature and (b) comparing the factors by which each of the temperature terms, in the
expression for n , affects n For an increase of from 300K to 600K, the T term increases n by a
factor of 2.8, whereas the same increase in causes the exponential tei m to increase //. by about a
factor of 50,000.
It is important to indicate that there are discrepancies among various sources in the values of
the intrinsic carrier density. The reason for this uncertainty in the calculated value is a result of the
very complex relations from which the values of the effective masses are determined.
In this book and for silicon at / 300K. we will use a value of 1 X 10' carriers per cubic cen
timeter. This value agrees fairly well with the calculation using the values of and \ listed earlier.
There is no sacrifice of accuracy in this approximation. In fact, the advantage is that it is a number
that can easily be remembered.
2
TV,
The expression for E
E
\m*l
(3.27)
\m*
becomes
, _ ^2
WS_SLSW!S)
4
m*
< 3
2 8 )
\m*j
E = E =E
i
f kT<n(^j
(3.29)
Section 3.3
Carrier Densities
65
E X A M P L E 3.3
Determine the location of the Fermi level with respect to the middle of the band gap in intrin
sic silicon and intrinsic gallium arsenide at = 300K.The value of k is given as 8.61 x 10 eV/K.
5
Solution
+ E)
v
19
\ ._
8.61 x " X 300
1.83 10
2 +
n;
2
3.22 l o "
E, = (E + E )I2  0.0073eV
c
+ E )/2
E = (
i
9 52 10
+ 0.012915 in ^
18
1 Q l 7
+ E )/2 + 0.0403eV
v
We note that the Fermi level at 300K in intrinsic silicon is 0.0073eV below the midgap, while in
gallium arsenide, is 0.0403eV above the midgap.
REVIEW QUESTIONS
Q31
Q32
Q33
Q34
Q35 In Eq. (3.15), the upper limit of integration was set at infinity. Why?
Q36 Briefly explain why the narrower the band gap, the higher is the intrinsic carrier den
sity in a semiconductor.
Q37 Briefly explain why the intrinsic carrier density increases with an increase of tempera
ture.
HIGHLIGHTS
The density of states function N(E) represents the number of free electrons in a metal
per unit volume per unit energy at energy E.
The FermiDirac function f(E) represents the probability that a state located at energy
is occupied by an electron. Therefore, (1 / ( ) ) represents the probability that a
vacancy exists at energy E.
The energy level E is that energy at which the probability of occupancy is onehalf.
The product oif (E) and (E) given by Eqs. (3.6) and (3.5) represents the number of
electrons per unit volume per unit energy at E.
F
66
Chapter 3
The intrinsic carrier density represents the density of each of the electrons in the con
duction band and holes in the valence band.
EXERCISES
E31 Determine the number of electronic states in the conduction band of silicon located
16
13
n = 2.17 X 10 cm~
E33 a) Determine the location of the Fermi level in intrinsic Ge, with respect to the bottom
of the conduction band,
b) Determine the location of E when measured from the top of the valence band.
F
Ans:
a) E Q
E=
p
0.338eV
3.4 E X T R I N S I C S E M I C O N D U C T O R S
Extrinsic semiconductors are formed by t h e addition of small a m o u n t s of selected
impurities to p u r e semiconductors. These impurities are elements in C o l u m n III or
Column V of the periodic table. The addition of as little as one impurity a t o m to a
million semiconductor atoms has considerable effect on the conducting properties
of the semiconductor. The atomic dimensions and the electronic structures of these
impurities are similar to those of the semiconductor. The effect of adding the impu
rity is to increase the density of one of the two carriers and, hence considerably alter
t h e electrical conductivity. This process cannot b e effective in good conductors, such
as copper, since t h e density of electrons is so large that it can hardly be changed by
the addition of impurities.
These impurities are also k n o w n as dopants and the process of adding t h e m is
k n o w n as doping the semiconductor. The result of the doping generates either of two
types of extrinsic semiconductors, identified by the type of the carrier whose density
is increased. They m a y be Ntype or Ptype, where the electron density is increased
in the Ntype and the hole density is increased in the Ptype.
It is important to indicate here, a fact which will be explained later, that the semi
conductor is charge neutral whether in the intrinsic condition or after impurities are
added and the semiconductor
becomes
extrinsic.
Ntype S e m i c o n d u c t o r
W h e n elements from C o l u m n V of the periodic table are added to silicon, an jVtype
semiconductor results. Typical dopants from C o l u m n V elements are phosphorous,
arsenic, and antimony. These have five valence electrons in their outermost orbit
and are k n o w n as donor impurities because they have one electron in excess of what
is n e e d e d for the covalent bonding.
Section 3.4
%
%
//
<
Extrinsic Semiconductors
67
An escaping electron
from a phosporus
atom in a forbidden
band to the conduction
band in order to become free
Phosporus atom
 becomes positive
ion replacing Si atom
Figure 3.7 The addition of a 5electron donor phosphorus atom to the covalent
bond structure generates a free electron since only four are required to complete
the bond.
Chapter 3
~m*
m (11.8)
where m*/m is the relative effective mass in silicon and is equal to 1.18.
A n order to magnitude value for E is obtained by assuming the masses to be
equal so that E becomes,
Q
13.6
E
=(iW
"
l e V
The above indicates that an energy of O.leV is sufficient to excite the fifth
d o n o r electron into the conduction b a n d so that it becomes a free electron leaving
behind an ionized d o n o r atom. Thus, we can state that the d o n o r energy level in sili
con is very close to the conduction b a n d edge.
M o r e accurate values for ionization energies are listed in Table 3.2.
A n energy b a n d representation of the condition of d o n o r atoms in silicon is
shown in Fig. 3.8, clearly illustrating t h e proximity of t h e d o n o r energy level to the
conduction band. Since electrons in the d o n o r impurity occupy different energy lev
els, t h e higher the t e m p e r a t u r e of the sample (semiconductor + impurity), the
greater is the t h e r m a l energy and the larger is the n u m b e r of electrons that can be
elevated to the conduction band.
The i m p o r t a n t difference b e t w e e n this ionization mechanism to p r o d u c e
d o n o r electrons and the intrinsic process to p r o d u c e electrons is that the ionized
impurities are fixed charges in t h e lattice and n o holes are produced.
Thus, o n e can state that at = 300K, each d o n o r a t o m contributes a free elec
tron to the conduction b a n d leaving behind the ionized d o n o r atom. The question
0.044eV phosphorus
C
D
.....
..._.
_<3
E.
1.12eV
v
=0
=100
T=300K
Figure 3.8 Energy band representation of donor impurity in silicon and the effect
of temperature. The phosphorus donor atoms have energy about 0.044eV below E
and are completely ionized at = 300K.
Section 3.4
Extrinsic Semiconductors
69
becomes: Is the total resulting density of electrons, intrinsic and donor, equal to t h e
sum of the d o n o r density and the intrinsic carrier density? The answer is n o and we
shall investigate this in a later section. Suffice it to recall h e r e that the semiconduc
tor is chargeneutral.
PType S e m i c o n d u c t o r
A semiconductor is said to be Ptype when an element from Column III of the peri
odic table is added to it. The c o m m o n dopants are: boron, aluminum, gallium, and
indium. T h e atoms of these dopants have t h r e e electrons in their outermost orbit
and w h e n a d o p a n t a t o m replaces a semiconductor atom, in the lattice structure, a
space in t h e covalent b o n d is available into which an electron of the semiconductor
can move. H e n c e , the d o p a n t a t o m accepts an electron. This type of d o p a n t is k n o w n
as an acceptor
impurity.
The electron that occupies t h e available space in the acceptor bonding comes
from the electrons that are in the valence b a n d of the semiconductor and, once cap
tured, this electron causes t h e acceptor a t o m to b e c o m e negatively ionized. A
sketch of the resulting covalent b o n d a r r a n g e m e n t is shown in Fig. 3.9.
T h e condition for ionization of acceptor atoms is analogous to that of donor
atoms. T h e acceptors' atoms are located at energy levels slightly above the valence
band. A t r o o m t e m p e r a t u r e , there is sufficient thermal energy to excite electrons
from the valence b a n d into t h e acceptor level. The absence of an electron from t h e
top of the valence b a n d generates a hole. Thus, for every acceptor atom, a hole is
g e n e r a t e d at r o o m t e m p e r a t u r e and the acceptor a t o m becomes negatively ionized.
This condition is illustrated in Fig. 3.10.
Thus, at r o o m t e m p e r a t u r e , each d o n o r a t o m provides a free electron to the
conduction b a n d and each acceptor a t o m generates a hole in t h e valence band. A t
Hole made
available
Electron escapes
from valence
band to join ~
boron atom
Boron atom
substituted
in Si lattice
>
\
.
/
>< <
Figure 3.9 A boron atom accepts an electron to complete the covalent bond and
hence creates a vacancy, a hole, into which another electron can move.
Chapter 3
E
0.045eK boron
f
=0
tttt
=100
=300
Figure 3.10 The acceptor energy level in the energy band diagram of silicon and
the effect of temperature on the ionization of the acceptor atoms. At = 300K all
acceptor atoms are ionized by the addition of electrons from the valence band.
Element
Function
Ionization
energy (eV)
Boron
Aluminum
Gallium
Phosphorus
Arsenic
Antimony
Acceptor
Acceptor
Acceptor
Donor
Donor
Donor
0.045
0.057
0.065
0.044
0.049
0.039
Section 3.6
3.5
71
THERMAL EQUILIBRIUM
In an earlier section, we derived an expression for the intrinsic carrier density, the
electron density, and t h e hole density, as a function of the t e m p e r a t u r e and the band
energy gap. We concluded that at r o o m t e m p e r a t u r e t h e intrinsic carrier density in
silicon is approximately 1.0 X 1 0 c m . T h i s carrier density results from the transfer
of electrons to the conduction b a n d as they receive thermal energy. T h e question
that comes to mind is: A s long as t h e thermal energy is available, why is it that the
density of the carrier does not continuously increase, but rather reaches a fixed
value?
The answer is that as carriers are continuously generated, and w h e n t h e r e is n o
other source than t h e r m a l energy, electrons and holes continuously recombine and
pairs of carriers disappear. In moving through the lattice, they encounter obstacles
causing t h e m to lose energy and disappear. Thus, the rate of generation is accompa
nied by the equal and opposite r a t e of recombination. This condition exists at ther
mal equilibrium. Thermal equilibrium is therefore defined as the state in which a
process is accompanied by an equal and opposite process, while the system is held at
constant temperature, and no external source of energy acts on it.
In t h e next chapter, we will present analytical relationships for the processes
of carrier generation and recombinations.
10
3.6
_3
D E N S I T I E S O F C A R R I E R S IN E X T R I N S I C S E M I C O N D U C T O R S
W h e n dealing with d o p e d semiconductors, we refer to electrons in Ntype semicon
ductors, t h e m o r e n u m e r o u s carriers, as the majority carriers. Holes in Ntype semi
conductors are minority carriers. In Ptype semiconductors, holes are the majority
carriers and electrons are t h e minority carriers.
We stated earlier that the density of electrons, w h e n donors are added to sili
con, is not equal to the sum of the intrinsic electron density and the density of the
ionized d o n o r atoms. What, then, is t h e resulting density?
In an earlier section, and prior to determining the intrinsic carrier density, we
found that the product of and p, n\, is a constant for a certain semiconductor at a
certain t e m p e r a t u r e and is given by np = nj = N N exp (E /kT).
T h e r e is n o con
dition in the expression that restricts it to intrinsic semiconductors because E does
not change with impurity concentration and N and N are constants. This product is
therefore a constant equally valid for intrinsic as well as for d o p e d semiconductors,
provided that we refer to the carrier densities at t h e r m a l equilibrium (a condition
w h e n there are n o sources of energy other than thermal energy). We label the
extrinsic values of the electron and hole density at t h e r m a l equilibrium as n and p
respectively so that we can write,
c
oPo
(3.30)
Chapter 3
C h a r g e Neutrality
We consider the general case w h e n both donors and acceptors are a d d e d to a semi
conductor. Assuming t h e r m a l equilibrium, t h e material includes particles that have
positive charges and others that have negative charges. The negative charges are
m a d e u p of t h e electrons in the conduction b a n d and the acceptor atoms which are
ionized. T h e holes in the valence b a n d and the ionized donor atoms constitute the
positive charges. Since the intrinsic silicon is charge neutral and the a d d e d impuri
ties are neutral, t h e resulting mix is also neutral so that, charge neutrality requires,
q(n
+ N~) = (p
+ N )q
(3.31)
where N denotes the density of acceptor atoms and Np refers to the density of
d o n o r atoms, assuming they are all ionized and n and p are t h e t h e r m a l equilib
rium values of electron and hole densities respectively. By replacing p in E q . (3.30)
by its equivalence from E q . (3.31), a quadratic equation in n is obtained. The solu
tion to the quadratic equation is given by E q . (3.32) where the positive sign has been
selected for the t e r m u n d e r t h e square root because that t e r m has a larger magni
t u d e than the first term and n cannot be negative.
A
1/2
NDNA
(3.32)
We have d r o p p e d the positive and negative superscripts from the symbols for impu
rity densities since we have already assumed that all impurity atoms are ionized at
room temperature.
For an Ntype semiconductor, either N
N or N is zero. F r o m E q . (3.32),
we find n for an TVtype semiconductor to be,
D
1/2
=
2
N\
D
(3.33)
(334)
n = N
0
and p
n
= j
(3.35)
D
16
For a d o n o r density of 1 0 c m ~
1 0 c m ~ , we find,
10
1 6
n = 10 cm
Q
 3
and p = 1 X 1 0 c m " .
Q
Section 3.6
73
in
Extrinsic region
full ionization
Intrinsic
Extrinsic region
all donors ionized
Extrinsic regie
partial ionizatii
in r
Intrinsic region
100
_L
200
300
400
_L
600
500
T(K)
(l/:
Figure 3.11 Variation of majority carrier density with temperature. In the logarithmic
sketch, the slope of the line in the intrinsic region is a measure of the band gap energy .
74
Chapter 3
E X A M P L E 3.4
17
16
To a sample of intrinsic silicon we add 10 atoms per cc of phosphorus and 9 10 atoms per
cc of boron.
Calculate the density of the majority and of the minority carriers at:
a) 1 = 300K,
b) T= 600K
Solution
a) at = 300K, we assume that all impurities are ionized so that,
np = n] = 1 10
20
(1)
= + N
x 10 = + 10"
16
(2)
= 10 cm~
10 c ,~3
We observe that the electrons supplied by the donor atoms dominate the resulting electron
density.
4
b) at = . , = 11.137 X 10 cm
2
30
so that
np = 11.137 X 103
+9
x 10 = + 10
16
(3)
17
(4)
= 1.1 x 1 0 c n r
M
Comparing the results to those of part (a), we note the effect of the thermally generated elec
trons and holes.
E X A M P L E 3.5
17
3
A sample of intrinsic silicon is doped with 10 atoms/cm of phosphorus. Calculate the den
sity of the electrons and the holes at:
Section 3.6
a) = 300K
75
b) = 600K
Solution
a) At = 300K, we find that
10
n = 1 X 10 crrr
np
1 x HPcm *
(1)
+10"
(2)
n * 10 cm \p =  = l x
10 cm"
b) At = 600K,
2
30
= 11.137 X 10 cnT
m:4
= /> + 1 0
1 7
(4)
n = 1.0011 X 10 cm
1.11 X 10"cm
Note the large increase in the hole density while the electron density remained practically
constant.
A d d i t i o n a l expressions f o r n a n d p
0
n, =
^^^
(3.36)
76
Chapter 3
N zx (^B
(3.37)
/V exp^ = n exp^
(3.38)
7V exp^ = n exp^;
(3.39)
n =
and
t )
n = n e x p p T ^ ) =
(^ =
40
( )
(3.41)
3.7 F E R M I L E V E L IN E X T R I N S I C S E M I C O N D U C T O R S
F r o m E q . (3.40), we find t h e location of t h e Fermi level in an Ntype semiconductor,
where = n = N , as
Q
 Ei = kTn^
n
= kT in
i
(3.42)
Ei  E =
F
kTln^
= kT In ^
(3.43)
We observe from Eqs. (3.42) and (3.43) that for an TVtype semiconductor the Fermi
level is higher t h a n E and moves u p closer to the conduction band. For a Ptype
semiconductor, t h e Fermi level moves d o w n towards the valence b a n d since E is
less t h a n E.. We recall that E the Fermi level for an intrinsic semiconductor, is
approximately midway b e t w e e n the conduction and valence bands. Calculations for
the locations of the Fermi level in extrinsic silicon are carried out in t h e following
example.
i
Section 3.7
77
E X A M P L E 3.6
16
(a) Determine the location of the Fermi level, with respect to E and , when 10 phosphorus
atoms per cc are added to a sample of intrinsic silicon at  300K.
(b) Repeat part (a) if 10 boron atoms per cc replace the phosphorus atoms.
c
15
AT
t,
I, I ni    0.025N5 i
n
I X 10
.358et/ =
0.35S.1
10
c + E
+ 0.3507eV
From Table 2.2, (at 300K) = l.lleV  E .  F. . by using this in the expression for above,
we obtain E = E ~~ 0.209eF = E + 0.9107.
(
 E = 0.02583 in 7  ^ 0 =  98 V
F
v1
E **E
F
,
 0.0073, also E = E + 1.12, we find
15
3 3 7
18
1 Q
10
 0.085e V
This example shows that the Fermi level moves toward the conduction band level, E , when
donor atoms are added and moves towards the valence band when acceptors are added. As the
temperature is increased, for both donors and acceptors, the Fermi levels move back towards the
middle of the gap as the semiconductor approaches the intrinsic state.
Keeping in mind that the Fermi level for an intrinsic semiconductor is located approximately
midway between the bottom of the conduction band and the top of the valence band, the following
conclusions can be drawn from the example:
c
The Fermi level moves toward E when donor impurities are added. It moves toward E
when acceptors are added.
(
Increasing the temperature beyond a certain value causes the semiconductor to become
intrinsic and the Fermi level to move toward the middle of the band gap.
78
Chapter 3
Increasing the doping density in part (d) caused E. to move yet closer to E Further
increase of the doping to 1 0 cm~ causes the conductor to become degenerate, since E is
less than 3ATfrom E.
In conclusion, we can stale that, at a fixed temperature, the location of /.', is a measure of the
doping of a semiconductor.
r
19
REVIEW QUESTIONS
Q38
Q39
Q310
Q311
Q312
Q313
Q314
22
23
HIGHLIGHTS
Which Semiconductor?
Section 3.8
79
EXERCISES
E34 A sample of semiconductor is doped with N
is 10 cm~ . Determine and p.
13
13
13
E35 If (N N )  is theoretically the largest possible value of intrinsic carrier density that can
be generated, determine as a fraction of this value, the number of electrons that can be
thermally excited into the conduction band at 300K in germanium.
Neglect changes of with temperatures given = 0.66eV.
c
Ans:
3 1
10
E37 Determine the density in cm and m of free electrons in silicon if the Fermi level is
0.2eV below E at 300K.
c
16
Ans: E c
E = 0.16eV
F
3.8 W H I C H S E M I C O N D U C T O R ?
T h r e e s e m i c o n d u c t o r s h a v e b e e n used in t h e fabrication of devices since t h e early
days of t h e s e m i c o n d u c t o r industry. T h e y are g e r m a n i u m , silicon, a n d gallium
arsenide. G e r m a n i u m was used early on but since t h e 1960s silicon has b e e n t h e
d o m i n a n t semiconductor. G a l l i u m arsenide has just recently acquired great
i m p o r t a n c e for certain devices. T h e c o m p a r i s o n a m o n g t h e t h r e e throws a light
on t h e particular p r o p e r t i e s that m a k e a s e m i c o n d u c t o r attractive for device
fabrication.
Semiconductors can b e classified into two major categories: elemental and
c o m p o u n d semiconductors. E l e m e n t a l semiconductors, such as g e r m a n i u m and sili
con, are classified in G r o u p I V of the Periodic Table of elements and have four
valence electrons. C o m p o u n d semiconductors are composed of a combination of
G r o u p III and G r o u p V elements or G r o u p II and G r o u p V I elements. E x a m p l e s of
important c o m p o u n d semiconductors are gallium arsenide, gallium phosphide,
indium phosphide, and indium arsenide.
G e r m a n i u m has two distinct attractive qualities: In the first place, it can b e
refined and processed m o r e easily than the others. In t h e second place, b o t h elec
trons and holes have higher mobilities than the corresponding carriers in silicon, as
can be seen in Table 3.3. The higher mobility translates into faster switching and
higher operating frequency limits.
A disadvantage of g e r m a n i u m is its high sensitivity to t e m p e r a t u r e because of
its relatively n a r r o w b a n d gap, which may cause instability in a device. Instability
Chapter 3
Unit
Density of atoms
Energy gap
Effective mass m*/m
electron
hole
Effective density of states
conduction band
valence band
Intrinsic carrier density
Mobility at low doping
electron
hole
Breakdown field
Relative permittivity
Melting point
cirr
eV
Si
5 x 10
1.12
Ge
22
4.4 x 10
0.66
GaAs
22
2.2 10
1.42
22
1.182
0.81
cm
0.0655
0.524
 3
3.22 10
1.83 10
1 10
cnr
cm (V  s )
2
0.553
0.357
10
19
19
1.03 10
5.35 10
2.17 x 10
19
18
13
4.21 10
9.52 x 10
2.49 x 10
 1
V/cm
dimensionless
C
1350
480
3
11.8
1410
3900
1900
10
15.8
940
5
8800
400
4 10
13.1
1240
17
18
Chapter 3
Problems
81
PROBLEMS
3.1 Draw the energy band diagrams, showing E , E , E , and E , for the following; assum
ing all impurities are ionized,
a)
Intrinsic silicon at 300K.
b)
Silicon doped with 10 boron atoms c m at 300K.
c
17
 3
16
 3
3.2 A silicon sample is doped with 10 c m of phosphorus atoms. Assume that all phos
phorus atoms are ionized at 300K and determine:
i) The electron density n .
ii) The hole density p .
iii) The location of the Fermi level with respect to E
Q
18
3.6 The effective conduction band density of states is N = 3.22 X 10 cm~ and the effec
tive valence band density of states is N = 1.83 X 10 cm~ . Assume that N and
N are located at the conduction band and valence band edges respectively. Let
= 300K.
c
19
a)
b)
c)
3.7 In a silicon sample at equilibrium the Fermi level is located above the middle of the
band gap by 0.38eV. The phosphorus donor states are located 0.04eV below the
conduction band. Determine the percentage ionization of the phosphorus atoms at
= 300K.
15
10
16
3
16
= 0.8 X 10 cm .
Chapter 3
3.11 Determine the electron and hole thermal equilibrium densities and the location of the
Fermi level, with respect to E , for a silicon sample at 300K that is doped with
a)
1 X 10 cm" of boron.
b)
3 X 10 cm~ of boron and 2.9 X 10 cm" of phosphorus.
3.12 Determine the approximate donor binding energy for silicon given e = 11.8 and
m*/m = 0.26.
3.13 An TVtype silicon sample has an arsenic dopant density of 10 cm~ . Determine:
a)
The temperature at which half the impurity atoms are ionized.
b)
The temperature at which the intrinsic density exceeds the dopant density by
a factor of 10. Assume E does not change with T.
c)
Assuming complete ionization, calculate the minority carrier density at 300K
and the location of the Fermi level referred to E
3.14 A silicon sample has the energy band diagram shown in the figure below. Given =
1.12eV,and. = 10 cm .
a)
Sketch the potential V as a function of x.
b)
Sketch the electric field % as a function of x.
c)
Determine the values of and p at = Xj and =
3.15 A silicon sample is doped with 10 donor atoms/cm . Draw an energy level diagram
showing the location of the Fermi level with respect to the middle of the band gap for:
a)
77K (liquid nitrogen).
c
15
16
16
17
10
16
b)
300K.
c)
600K.
Neglect the change of with temperature.
3.16 a)
At = 300K, what percentage of the electrons in a cm of silicon are located
in the conduction band?
b)
Repeat for = 500K.
3.17 On the energy band diagram, electron energy is measured upwards. Explain why hole
energy is measured downwards.
g
X\
Chapter 3
Problems
83
3.18 In the band diagram shown below, which carrier has the larger kinetic energy?
Explain the reason for your answer.
1I
leV
2eV
4eV
chapter 4
CARRIER PROCESSES:
DRIFT, DIFFUSION AND
GENERATIONRECOMBINATION
4.0 I N T R O D U C T I O N
In the last chapter, we identified electrons and holes as the current carriers in a
semiconductor. T h e electrons exist in t h e conduction b a n d and t h e holes are in the
valence band. In an intrinsic semiconductor, at t h e r m a l equilibrium, the density of
electrons is equal to the density of holes because the r a t e of generation of electronhole pairs is balanced by the r a t e of recombination of electrons and holes.
The addition of small traces of controlled impurities increases the density of
one carrier while decreasing the density of the other. The density of electrons is
e n h a n c e d by the addition of d o n o r impurities while the addition of acceptor impuri
ties increases t h e density of holes.
In o r d e r to d e t e r m i n e expressions for t h e carrier currents in terms of an
applied voltage and other factors, it is necessary to discuss and define the processes
by which carriers are caused to move.
Carrier m o t i o n is caused by two conditions; the application of an electric field
whose force accelerates the carriers and, second, a difference of carrier concentra
tion b e t w e e n two points causes carriers to move by diffusion from regions of high
concentration to regions of low concentration. F u r t h e r m o r e , t h e processes of carrier
generation and recombination affect t h e resulting densities of t h e carriers.
It is by considering t h e combined effects of carrier motion, generation, and
recombination, o n t h e distribution and motion of the carriers, that we can proceed
to set u p t h e expressions from which we d e t e r m i n e the relations describing the
o p e r a t i o n of semiconductor devices.
84
Section 4.2
Thermal Velocity
85
4.1 V E L O C I T Y LIMITATIONS
In the last chapter, we derived relationships for t h e densities of carriers in an extrin
sic semiconductor. W h e n we consider a bar of semiconductor, to which an electric
field is applied, t h e current in the bar will d e p e n d on t h e density of the carriers, the
crosssectional area of the bar and t h e velocity with which the carriers move. For an
/Vtype semiconductor, w h e r e n p , the current is given by
Q
mpS =
couls)
= H
/electro)
^) {^rj
/ couls \
lele^irj
/cm)
) *
,
A
9 X
}
4.2 T H E R M A L V E L O C I T Y
Electrons and holes in semiconductors are in constant motion because of t h e ther
mal energy they receive. Since they are in motion, they are not associated with any
particular lattice position. A t any o n e time, the electrons and holes m o v e in r a n d o m
directions with a m e a n r a n d o m velocity. This velocity, k n o w n as the thermal velocity,
is of t h e o r d e r of 1 0 c m / s for electrons in silicon. Because of the motion in ran
d o m directions, the current resulting from t h e m o t i o n of all t h e carriers in any one
direction is zero.
7
86
Chapter 4
(4.1)
w h e r e m* is the conductivity effective mass of t h e free electron, v its m e a n t h e r m a l
velocity, k is Boltzman's constant, and is the t e m p e r a t u r e in degrees Kelvin. For
silicon at = 300K, this velocity has b e e n calculated to be approximately 2.3 X
1 0 c m / s . T h e m e a n kinetic energy for all the electrons at t h e r m a l equilibrium is
E = (3/2)kT, and this translates into 0.04eV at 300K, which is slightly above the
conduction b a n d edge E , and for silicon represents approximately 1/25 of the b a n d
gap energy E .
Electrons traveling in a solid u n d e r the influence of a small applied electric
field collide with the lattice, exchange energy with the lattice, and start all over.
D e p e n d i n g u p o n the magnitude of the field, the electron gives u p a certain a m o u n t
of heat to the lattice.
th
4.3 C O L L I S I O N S A N D SCATTERING
W h e n a relatively lowintensity electric field is applied to a metal or a semiconduc
tor, electrons acquire a m e a n velocity in accordance with O h m ' s law. This new
velocity is superimposed on t h e t h e r m a l velocity and its magnitude is m u c h smaller
than t h e m e a n t h e r m a l velocity. This new velocity c o m p o n e n t refers to the average
rate of m o t i o n of the electron population in the direction of the force of the electric
field. This velocity is k n o w n as the drift velocity, v .
Sketches comparing the r a n d o m fieldfree motion of electrons with the fielddirected motion are shown in Fig. 4.1.
It is i m p o r t a n t to point out that since the drift velocity is, in most devices, much
smaller t h a n t h e t h e r m a l velocity, t h e drift velocity can be considered a p e r t u r b a t i o n
of the t h e r m a l velocity.
The electron that is moving in a solid, subjected to an electric field, collides
with the atoms and with t h e nonideal lattice structure, loses most of its velocity, and
then r e p e a t s the process. This process of collision is m o r e accurately k n o w n as scat
tering.
It is important to point out that quantummechanical calculations indicate
that scattering does not take place, and h e n c e n o energy exchange occurs, w h e n an
electron is traveling in a perfectly periodic and stationary lattice. A nonstationary
lattice results w h e n the atoms vibrate by acquiring heat.
Scattering does not refer only to a direct collision b e t w e e n the moving elecd
*From Halliday, Resnick, and Walker, Fundamentals of Physics, p. 582; copyright Wiley
(1993). Reprinted by permission of John Wiley & Sons, Inc.
Section 4.3
87
88
Chapter 4
4.4 C O L L I S I O N S EFFECTS
Drift Velocity
T h e effect of t h e scattering, or the collisions, on the motion of a carrier is equivalent
to a frictional resistive force that controls the acceleration of the particle and limits
its velocity to t h e drift velocity.
The m o t i o n of an electron that is subjected to an electric field, in the xdirection, obeys t h e equation,
d
m
dx

(42)
where % is t h e electric field intensity, m* is the electron effective mass, q is the elec
tronic charge, and the second t e r m on t h e right h a n d side of E q . (4.2) is the frictional
resistive force where w is a constant d e p e n d e n t on the solid.
The electric field does not accelerate the electron continuously because the
resistive force increases as t h e velocity increases until such time that t h e resistive
force balances the force of t h e electric field until t h e steadystate, w h e n t h e velocity
becomes the drift velocity. T h e drift velocity can b e found by setting t h e impulse
(force X time), applied to an electron during the time T b e t w e e n collisions, equal to
the m o m e n t u m gained by t h e electron in that period. In t h e steady state, t h e drift
velocity is d e t e r m i n e d from
x
<f*>xtc
> d
(43)
"<c
We digress briefly to consider the transient behavior of the electron velocity.
For an electron starting from rest at / = 0, t h e solution to E q . (4.2) is
,
dx
v(t) = =
dt
cf&

w
v
exp
wt
v[t) =
(4.4)
so that E q . (4.4)
(4.5)
Thus, after a certain time t > , the particle velocity will have a steadystate
value given by t h e drift velocity, the negative sign indicating that the velocity is in a
direction opposite to that of the electric field.
Suppose that, as t h e electron is being accelerated, the electric field is r e m o v e d
and at this time assume its velocity is v. For this condition, the velocity in accor
dance with the solution to E q . (4.2) b e c o m e s
(i) = exp(r/T )
0
(4.6)
Section 4.4
Collisions Effects
89
The time constant T may also b e interpreted as the factor that controls the
rate at which t h e velocity and t h e current decay to zero after the field is removed.
This time is k n o w n as the relaxation time and it is of the o r d e r of l ( F s .
The steadystate value of the drift velocity in t h e presence of the electric field
is also found by setting t > > , in E q . (4.4).This time has also b e e n r e p r e s e n t e d to
refer to a m e a s u r e of the time it takes for carriers in a semiconductor to move and
neutralize charges.
The r e a d e r may have concluded that w h e r e we speak of the velocity of an
electron, we are implying that it is possible to isolate the particle. In fact, t h e velocity
h e r e refers to the m e a n of t h e velocities of all the electrons that are free.
We defined earlier scattering in t h e presence of an electric field, as the process
of collision b e t w e e n the fielddirected electron and t h e lattice structure. We will
now illustrate this operation on an energy b a n d diagram.
c
15
Collisions a n d E n e r g y Exchanges
In Fig. 4.2(a), we show an Ntype semiconductor with metal contacts attached to
both ends. A t point A , a voltage + V is applied with respect to contact B, which is at
g r o u n d potential. We will study the effect on the energy of an electron released from
rest at point B. Before we do that, we will review t h e energy relations w h e n t h e bar
is replaced by a vacuum enclosure. A t terminal of the vacuum enclosure, an elec
tron is released from rest, and since there are n o obstacles and hence n o collisions
along the way, t h e electron accelerates and strikes plate A . A t point B, t h e electron
lost potential energy qV; at impact with terminal A , all the potential energy is con
verted to kinetic energy and hence heat at point A.
In the energy b a n d diagram of the semiconductor bar of Fig. 4.2(b), an elec
tron held at point has potential energy qV and its potential energy w h e n it arrives
(b)
Figure 4.2 Illustrating conduction in a bar and by use of the energy band diagram
(a) semiconductor bar with a voltage applied; (b) motion of electron on energy
band diagram.
90
Chapter 4
4.5 M O B I L I T Y
Referring to E q . (4.3), we observe that the drift velocity of electrons (or holes) is
given by the p r o d u c t of t h e electric field intensity and a factor that d e p e n d s on the
electronic charge, the effective mass, and t h e relaxation time of t h e semiconductor.
We label that factor the mobility of a carrier, , so that for an electron we have
< 4
8 )
(49)
 %
410
< >
Section 4.5
Mobility
91
GaAs electrons
I0
1(P
10
LO
io
The above t r e a t m e n t assumes that the time interval between collisions is inde
p e n d e n t of the magnitude of the applied electric field intensity. This is valid p r o
vided the drift velocity is m u c h smaller t h a n t h e t h e r m a l saturation velocity, v ,
which is approximately 10 c m s / s for electrons in silicon.
E x p e r i m e n t a l results of t h e m e a s u r e m e n t s of t h e drift velocity of electrons
and holes are shown in Fig. 4.3. It is quite evident that at very large values of electric
field intensity, the drift velocity approaches the saturation velocity a n d h e n c e the
mobility decreases with increasing electric field intensity.
We also note, in Fig. 4.3, the d e p e n d e n c e of the drift velocity, and hence the
mobility in the linear region, on the material through its d e p e n d e n c e on the effec
tive masses of the electrons and the holes.
&
Effects of I m p u r i t y C o n c e n t r a t i o n a n d T e m p e r a t u r e o n
Mobility
92
Chapter 4
50 I
100
200
I
500
. . I
1000
T(K)
Figure 4.4 Variation of electron mobility with temperature and doping in silicon.
Source: S.M. Sze, Semiconductors: Devices, Physics and Technology, p. 33; copyright
Wiley (1985). Reproduced by permission of John Wiley & Sons, Inc.
Section 4.5
Mobility
93
Expressions f o r t h e M o b i l i t y
W h e n two scattering processes occur at t h e same time, the combined effect on the
mobility is d e t e r m i n e d from the probabilities of the two processes. If in an incre
mental time, dt, t h e probability of scattering by process 1 is dt/j
where , is the
time intervals b e t w e e n two collisions and the probability of scattering by process 2
is / , t h e n t h e total probability dt/j is the sum of the two probabilities so that we
can write
v
dt/
= dt/j
(4.11)
or
(4.12)
312
3I2
(N
N ).
A
 3
7.4 1 0 "
m = 887y
1 +
3 3
17
(1.26 10 )
8
(4.13)
0.88
 2
1.36 1 0 
54.37
.(2.35
2 3
(4.14)
0.887
i o
0.146
0.146
"
(4.15)
17
1 + 0.698 X 10 7V
54.3 +
407
1 + 0.374
17
10~
(4.16)
Chapter 4
Drift
Section 4.6
95
4.6 DRIFT C U R R E N T A N D C O N D U C T I V I T Y
W h e n an electric field is applied to a semiconductor, the electrons and holes acquire
drift velocities and the c o n s e q u e n t m o t i o n of carriers results in an electric current.
This current is k n o w n as drift current. The electron drift current is given by:
I =Aqnv
n
(4.17)
l =Aqn^ %
n
(4.18)
The current due to electron motion is thus in t h e same direction as the electric field
intensity. The electron current density in a m p s / c m is
2
so that
J = %
(4.19)
p%
= %
qVlp
(4.20)
J = J +J
t
=%+%
(4.21)
'
= <? +
(422)
E X A M P L E 4.1
a) Determine the conductivity of intrinsic silicon at 300K.
b) Repeat part (a) for a sample of silicon doped with 10 cm~ of phosphorus.
17
96
Chapter 4
Solution
10
a) For intrinsic silicon at 300K, = = n = 10 cnr , the mobilities are calculated using Eqs.
(4.15) and (4.16) to be = 1340cm and  461.3cm '.The intrinsic conductivity becomes
i
, = q .( + ) = 2.88 X lO^ohmcmF
;)
17
b) The densities of electrons and holes are calculated to be 10 cm~ and 10 cm respectively and
the mobilities are calculated using Eqs. (4.15) and (4.16) to be = 825cm /Vs and =
350cm /Vs. The extrinsic conductivity becomes
2
Since the density of silicon atoms is approximately 10 cm" , the addition of one part of phos
phorus to a million parts of silicon increased the conductivity of the sample by a factor of about one
million.
4.7 RESISTIVITY A N D R E S I S T A N C E
The reciprocal of t h e conductivity, t h e resistivity p, has units of ohmcm. In general,
t h e semiconductor is either /Vtype or Ptype. For a semiconductor in t h e r m a l equi
librium, t h e resistivity becomes
= Vi^ 0 +
n
Po)
23
( )
 3
as \/{qv N )
n
(4.24a)
For a Ftype semiconductor, the expression for the resistivity is given approximately
by
s
l/(q\L N )
p
(4.24b)
By using values of t h e mobility calculated from t h e linear region of Fig. 4.3, a n d cal
culating resistivity using Eqs. (4.24(a)) and (4.24(b)), o n e obtains a plot of t h e resis
tivity versus impurity density. The results a r e shown in Fig. 4.6.
REVIEW QUESTIONS
Q41 Identify the processes that cause scattering.
Q42
Q43
Q44
Q45
Section 4.7
2 3 456 8
1 0
14
1 0
2 3 456 8
15
1 0
2 3 456 8
16
1 0
2 3 456 8
17
1 0
2 3 456 8
18
1 n
l9
2 3 456 8
1 Q
97
2 3 456 8
20
1 n
2l
3
98
Chapter 4
The mobility is also influenced by the temperature and the doping of the semiconduc
tor. The higher the doping the lower the mobility.
The current resulting from an applied electric field, known as the drift current, is the
product of the electric field intensity, and the sum of the conductivities of holes and
electrons. Conductivity, the reciprocal of resistivity, is the product of the carrier den
sity, the charge of the carrier, and the mobility.
EXERCISES
E41
15
b)
15
 3
15
of phosphorus and 5 X 1 0 c n r of
a) = 4 X 10 ~ , = 2.5 X 10 cnr
2
E42
15
b)
b) = 0.16 (ohmcm)
E43
1
For the sample in E42 calculate a) the resistivity b) the drift current density /
Ans:
E44
a) = 3.12 ohmcm,
b) / = 32 A / c m
Determine the resistance of the sample in E42 if its length is 10 cm and its area
is 1 cm .
2
Ans:
R = 31.2
E X A M P L E 4.2
A sample of silicon at 300K has resistivity of 5 ohmcm. Assume that arsenic is the only dopant.
Determine the impurity concentration.
Solution
<7/V*
14
[r
14
Section 4.7
99
E X A M P L E 4.3
2
A sample of silicon at 300K of length 2.5cm and a crosssectional area of 2mm is doped with
10 cm~ of phosphorus and 9 x 10 c m of boron. Determine:
a) The conductivities of the sample due to electrons and that due to holes.
b) The resistance of the sample.
17
16
 3
Solution
a) The conductivities of electrons and holes are given respectively by,
+ N
and pn = nj
and
= 1 X 10 cnT
The values of the mobilities are found from Eqs. (4.15) and (4.16), using = N + N , to be
A
= 626cm /Vs
and
16
13
(ohmcm)"
= /='1/^5/(2 lfr^==
i2$otas
We will now briefly discuss and explain the ohmic resistance of metals and semiconductors in
the light of our discussion of the scattering phenomena.
Resistance of M e t a l s a n d Semiconductors
E q u a t i o n (4.6) indicates that the electron motion undergoes damping as it travels
t h r o u g h o u t t h e lattice structure. Eventually, its velocity becomes zero. T h e electrons
are constantly traveling with high velocities, and w h e n an electron collides with the
lattice or with an impurity atom, it loses s o m e of its energy and thus starts out with a
new velocity.
This damping of the electron wave is actually caused by any disturbance in the
periodic lattice. These irregularities in the structure scatter t h e electron waves, just
as light is scattered in the atmosphere. In getting scattered, the electron wave is con
verted from a plane wave traveling in a certain direction with a certain m o m e n t u m
into waves traveling in all directions and corresponding to n o m o m e n t u m at all. The
electron has then lost its m o m e n t u m .
W h e n we consider a specific material, we find that its electrical resistance is
due to both perturbations in the solid lattice and due to the t e m p e r a t u r e . Thus, the
100
Chapter 4
4.8 PARTICLE D I F F U S I O N A N D D I F F U S I O N C U R R E N T
Diffusion
Carrier m o t i o n in semiconductors, which results in currents, is either caused by the
application of an electric field, which causes a drift current, or by t h e process of dif
fusion of carriers.
The process of diffusion consists of the m o t i o n of carriers away from regions
of high concentrations of carriers to regions of low concentration of carriers. This
process is well illustrated w h e n a light is flashed at a small region n e a r the center of
a long thin bar of semiconductor material. Electronhole pairs, in addition to those
available at t h e r m a l equilibrium, are t h e n generated at the center because of the
light energy that is absorbed by the semiconductor. These excess carriers will move
randomly to the right and to t h e left of t h e generation zone, resulting in an outwards
flux of particles away from the center. These carriers are said to diffuse (just as a
d r o p of ink in a glass of water quickly mixes with water).
We illustrate this process by first considering Fig. 4.7(a), which shows a distrib
ution of particles per unit volume. We are interested in the rate of motion of these
particles per unit area and unit time across a plane at x The particles are in con
stant t h e r m a l motion from left to right, and from right to left at x Since t h e density
of t h e particles is greater to the left of x than it is to t h e right of x m o r e particles
y
Section 4.8
"
= 
101
0p
/"semiconductor
(a)
diffusion
(C)
Figure 4.7 Illustrating diffusion: (a) diffusion of particles at JC,; (b) electrons injected at = 0
cause electron density profile in Psemiconductor; (c) hole diffusion and hole diffusion
current; and (d) electron diffusion and electron diffusion current.
F=
D
dP
dx
in the
(4.25)
X=X\
where F is t h e flux per unit area, per unit time, is t h e volume density of particles,
dP/dx is the gradient at x and D is labeled the diffusion constant. For F to have
units of particles per unit area p e r unit time, D has to have the dimension of dis
tance squared per unit time. According to E q . (4.25), and since the derivative is neg
ative at x particle flow is in the positive xdirection. In t h r e e dimensions, E q . (4.25)
becomes
1
F = DVP
(4.26)
102
Chapter 4
strong source of light focused at = 0. We assume n o electric field exists in the sam
ple and that t h e density of excess electrons (excess over those available in the
semiconductor) decreases with increasing x.
Because t h e gradient is negative in the direction of electron diffusion, the flux
of electrons (motion per unit area p e r unit time) passing through t h e plane at x as
d e t e r m i n e d from E q . (4.26), is given by
v
dn'
~
Dn
(4.27)
dx
3
_ 1
_ 1
Diffusion Current
U p to this point, t h e charge of the carrier has not b e e n introduced. To convert the
expression to current density, we introduce the charge q coulombs for an electron
so that the expression in E q . (4.27) is written in terms of C c m ~ s , which has units
of current density as
2
dn
' =
dn
_ 1
,.
( 4
'
2 8 )
_ 1
D  (f),
(430)
Section 4.9
Carrier Currents
103
E X A M P L E 4.4
The region shown in Fig. 4.7(b) is a section of a silicon device that is at room temperature (T =
300K). This region is doped with 10 cm acceptor atoms. A stream of minority carriers is injected
at = 0 and the distribution of minority carriers in the sample is assumed to be linear, decreasing
from a value of lO^cnT at = 0 to the equilibrium value at = W. where W is 10 microns.
Determine the diffusion current density ol" electrons.
15
Solution
be:
The thermal equilibrium densities of majority and minority carriers are calculated to
15
Using the density of impurity atoms, we determine the mobility of electrons in the region
from Eq. (4.15) to be:
= 1331 cmV(Vs)
"
"dx
The diffusion constant is obtained from Einstein's relationship at = 300K,
D
"
~q "
259
is 
dx
'
1 3 3 1
3 4
'
4 8
m2//S
'
14
10 x 10~
] [ =  10 cm"
/
14
4.9 C A R R I E R C U R R E N T S
In this chapter, we have examined t h e processes that cause carrier currents
namely drift and diffusion. In the next section we will study the mechanisms of car
rier recombination and generation.
By combining the expressions for the current densities of holes and electrons
due to drift and diffusion, we have the following expressions for J , J , and / , repre
senting t h e electron current density, the hole current density, and the total current
density respectively as:
n
104
Chapter 4
J,
q\x n% + qD.,
n
q^ p%
p
/, = /
qD
(4.31)
+ /
I
The units of the symbols are: / in a m p s / c m , q in coulombs, in c m / V  s ,
and are in c u r , D and D in c m / s and dn/dx and dp/dx are in c m .
It is evident from Eqs. (4.31) that in o r d e r to derive relations for the currents
in a device that is subjected to a certain applied voltage in the steadystate, one
n e e d s to derive expressions for both t h e electric field intensity distribution as a
function of distance, and the spatial distributions of the electron and hole densities.
Fortunately, at the current levels at which most of the devices normally operate, the
drift current of minority carriers is neglected. O u r emphasis will b e o n the diffusion
current of minority carriers. In the following section, we will set u p t h e continuity
equation from which we obtain expressions for the distribution of minority carriers.
2
 4
4.10 R E C O M B I N A T I O N A N D G E N E R A T I O N
Rates of RG
So far, we have discussed the two processes of carrier motiondrift and diffusion. A
third category of carrier actions that involve a variation in t h e densities of electrons
and holes are the processes of generation and recombination. In fact, they are two
separate processes that t a k e place simultaneously and their rates are equal only at
thermal equilibrium. These two processes indirectly affect the currents by changing
the carrier densities involved in drift and diffusion.
Generation is the process of creating new carriers, holes, and electrons.
Recombination
is the inverse of generation, whereby an electron and hole dis
a p p e a r simultaneously.
In an intrinsic semiconductor at t h e r m a l equilibrium, electrons and holes are
continuously generated and continuously recombine. Because the intrinsic carrier
density is d e t e r m i n e d solely by the energy gap and t e m p e r a t u r e , thus constant
u n d e r given conditions, the rate of generation of electronhole pairs is balanced by
the equal r a t e of recombination. In extrinsic semiconductors, the carrier densities
are d e t e r m i n e d by both the impurity doping density and t h e intrinsic carrier density,
hence one can also state that in t h e r m a l equilibrium, t h e rates of generation and
recombination are equal. We remind the reader that t h e r m a l equilibrium is a condi
tion in which n o external forces (such as light) or an electric field are applied to the
semiconductor. For intrinsic semiconductors and extrinsic n o n d e g e n e r a t e semicon
ductors, the relation np = n is valid at t h e r m a l equilibrium.
2
The question is then, when are the generation and recombination rates not
equal? To answer that question, we point out that semiconductor devices operate
normally u n d e r nonequilibrium condition. This translates into np indicating
that external effects are influencing the carrier densities and carrier distributions.
2
Section 4.10
105
Direct G e n e r a t i o n  R e c o m b i n a t i o n
Electrons and holes in semiconductors are generated when an electron is displaced
directly from t h e valence b a n d to the conduction b a n d due to absorption of thermal
energy or exposure to light energy. This process is labeled direct generation w h e n
the electron is excited directly from the valence b a n d to the conduction band, and it
results in an additional electronhole pair. Direct recombination
occurs w h e n an
electron falls from the conduction b a n d directly into the valence band, eliminating
both the electron and a hole. Such direct generationrecombination processes are
very unlikely in silicon and germanium, as discussed in Section 2.4, because of the
shape of t h e Ek diagram of Fig. 2.12. Direct recombination generation is the com
m o n process in IIIV c o m p o u n d s such as gallium arsenide, gallium phosphide and
indium antimonide. This direct recombination results in light emission. Sketches
illustrating direct recombinationgeneration are shown in Fig. 4.8a.
Direct generation of electronhole pairs may b e caused by one or m o r e of the
following processes: exposure to thermal energy, exposure to light energy, or impact
ionization. Impact ionization occurs w h e n the collision of an accelerated electron
with an a t o m transfers enough energy to the a t o m to cause the generation of new
carriers. U n d e r favorable conditions an avalanche process may ensue that results in
excessive heat being generated, which destroys the device.
To g e n e r a t e an electronhole pair, the a m o u n t of energy imparted to t h e elec
tron in the valence b a n d must b e at least of t h e o r d e r of the energy gap . Similarly,
the recombination process is accompanied by a release of light or heat energy of the
o r d e r of the energy gap. In a semiconductor at thermal equilibrium, the processes of
generation and recombination take place continuously and their rates are equal.
Indirect G e n e r a t i o n  R e c o m b i n a t i o n
The principal recombinationgeneration ( R  G ) process in germanium and silicon is
of the indirect type where a third party, acting as a catalytic agent, is involved. This
third party is a localized state in the forbidden band, which serves as a "stepping
s t o n e " b e t w e e n t h e conduction and valence bands. Such states are at energy levels
106
Chapter 4
generation
recombination
generation
(a) direct
Figure 4.8
recombination
(b) indirect
in which silicon and g e r m a n i u m atoms cannot exist. They are, however, available
only for foreign atoms. Sketches showing indirect R  G processes are shown in Fig.
4.8(b). E a c h of the indirect generation and the recombination processes occur in
two steps, w h e r e t h e "stepping s t o n e " is a localized state having an energy level E
near the middle of the forbidden band.
The frequency and location of recombination are probabilistic in nature. The
energy level of an impurity recombination center is d e t e r m i n e d as follows: A n elec
tron transition from E to t h e recombination center E has a probability of occur
rence of P (E ) whereas the probability of a hole transition from E to E is (E ).
Since a recombination is completed by t h e transition of both particles, the probabil
ity of t h e occurrence of such an event is proportional to the product,
P (E )P (E ),
of the two probabilities. The m a x i m u m of such a product is found to occur midway
in the b a n d gap at (E +
E )/2.
T h e r e are four processes that m a k e u p the generation and recombination in
indirect semiconductors. These are shown in Fig. 4.9, where an acceptor localized
state exists slightly above the middle of t h e forbidden band. The first process is the
capture of an electron by the localized state, while the inverse, second process,
whereby an electron is m o v e d from the localized state to the conduction band, is
k n o w n as emission of an electron.
The third process, known as the hole capture, is described by the transfer of an
electron from the localized state to t h e valence band. Finally, the fourth process con
sists in the transition of an electron from t h e valence b a n d to t h e localized state,
leaving a hole behind and called hole emission.
T
Section 4.10
before
after
electron capture
after
before
hole capture
before
after
electron emission
107
before
after
hole emission
Figure 4.9 Interaction of free carriers with localized states in indirect generationrecombination. The RG center is an acceptor type.
L o w  L e v e l Injection a n d R e c o m b i n a t i o n
Consider an /Vtype semiconductor in t h e r m a l equilibrium w h e r e n
p . This is
now p e r t u r b e d by s o m e mechanism that causes p to increase by Ap. To preserve
neutrality, n must also increase by An and An = Ap.
We will assume that Ap
n . Such a perturbation, which caused t h e excess
minority carrier density (excess over equilibrium) to b e m u c h smaller than the
majority carrier density, is said to have caused lowlevel injection. This condition of
lowlevel injection is that in which most devices are normally o p e r a t e d w h e n excess
0
108
Chapter 4
15
10
10
A n a l y t i c a l Relations
Expressions are n e e d e d for the timevariation of carrier densities u n d e r p e r t u r b e d
nonequilibrium conditions for the typically m o r e c o m m o n indirect recombinationgeneration processes. Before determining the expressions, and partly as a reminder,
we define s o m e terms as follows:
3
Section 4.10
dp/dt = K^Ap
109
(4.32)
dn/dt = KJST^n
1
(4.33)
= 1/
(a)
= 1/K N
2
(b)
(4.34)
(a)
dn/dt = ~/
(b)
(4.35)
M i n o r i t y Carrier Lifetime
Since Eqs. (4.35) r e p r e s e n t the rate of decay of t h e excess minority carriers, Ap in re
type and An in Ptype semiconductors, it stands to reason that t h e time constants
and i represent a m e a n time interval during which all excess minority carriers have
recombined. Some will r e c o m b i n e at t = 0 after the perturbation is r e m o v e d and
others will take a longer time. In that case, we define and as the average time an
excess minority carrier will survive in an environment of a great n u m b e r of majority
carriers. These are labeled t h e lifetimes of minority
carriers.
The lifetimes cannot b e accurately predicted as they d e p e n d on t h e density of
traps N , which itself is not a predictable quantity since it varies from one sample
of semiconductor to a n o t h e r and is also subject to changes during fabrication of
devices.
However, it is possible to increase t h e density of traps and hence reduce t h e
lifetime by the addition of certain metals t o t h e semiconductor such as gold in sili
con.
A t certain gold concentrations, an increase of two orders of m a g n i t u d e of gold
density m a y serve to reduce the lifetime by three orders of magnitude. The lifetimes
in typical samples may vary from a microsecond to a nanosecond. The main purpose
of decreasing the lifetime of minority carriers, as we shall explain in a later chapter,
is to increase the switching speed of semiconductors or improve the highfrequency
response of devices.
110
Chapter 4
REVIEW QUESTIONS
Q47
Explain, in your own words, what is meant by diffusion in general and, as relating to
semiconductors, in particular.
Q48 If the electron density at one end of a semiconductor bar is increased relative to the
other end, does the hole density take on a similar distribution? Why?
Q49 Explain why the rate of generation of electrons in a semiconductor is given by /
where n is the equilibrium electron density and is the lifetime of the minority car
rier electrons.
Q410 What is the rate of recombination and what does it depend on?
Q411 Explain the following statement: "If the minority carrier density at one point in a
semiconductor exceeds the thermal equilibrium value, the recombination rate
exceeds the generation rate."
0
HIGHLIGHTS
The current in some electronic devices is strictly a drift current. In others, it is a combi
nation of drift and diffusion. Diffusion of carriers results when the carrier density in a
semiconductor varies with distance.
Diffusion current is directly proportional to the slope of the carrier density distribu
tion. Diffusion current is the product of the slope, the charge q, the crosssectional
area, and the diffusion constant.
Diffusion constant is a measure of the ease of diffusion of the carriers and is directly
proportional to the mobility.
While carriers are accelerated in a semiconductor, two other processes are also active;
they are generation and recombination. Electrons and holes disappear as carriers by
recombining and are continuously replenished by generation. Generation is the
process of the transfer of electrons from the valence band to the conduction band.
To determine drift and diffusion currents, it is necessary to know the distribution, with
distance, of the carriers. The continuity equation, which is derived in the next section,
is used, subject to appropriate boundary conditions, to determine the distribution of
the minority carriers in a semiconductor.
EXERCISES
16
15
E45 The electron density in a silicon sample decreases linearly from 10 cm~ to 10 cm~
over a distance of 10cm. The area of the sample is 6cm and the diffusion constant is
15cm /s. Determine the diffusion current of electrons.
2
A n s : / = 12.96mA
E46 The distribution of electrons in a semiconductor sample is given by
n(x) = 10 exp(x/L) c m . The length L = lOcms and the diffusion constant is
25cm /s. Determine the diffusion current density at = 0.
15
3
A n s : / =  4 X 10" A/cm
Section 4.11
111
4.11 T H E C O N T I N U I T Y E Q U A T I O N
We will now derive t h e continuity equation, which will be used to determine expres
sions for the distribution of minority carriers in an extrinsic semiconductor. After
having d e t e r m i n e d these expressions, then we can deal with the diffusion currents of
minority carriers.
We consider in Fig. 4.10, a section of an Ntype semiconductor into which
excess minority carrier holes have b e e n injected and are moving within t h e section,
and we focus our attention on the hole density in a very small slice, Ax, of the sec
tion. Since hole current is flowing, a certain n u m b e r of holes are brought in by the
hole current at and a certain n u m b e r is carried out by the current at (x + Ax). In
addition and within the slice, generation and recombination take place. We assume
lowlevel injection (Ap
n ) as was discussed in the preceding section. The hole
density in the slice is and the equilibrium hole density is p .
The n u m b e r of coulombs of holes per square centimeter per second that flow
into the slice is J (x). The n u m b e r of coulombs entering per second is J (x)A and the
n u m b e r of holes that enter t h e slice per second is J (x)A/q. Similarly, the n u m b e r of
holes that leave t h e slice per second is J (x + Ax)A/q. Therefore, the rate of change
(per second) of holes within the slice caused by the hole currents is t h e difference
b e t w e e n t h e two
Q
[J (x)  J (x + Ax)]A/q
p
holes/s
(p
p )AAx/i
Q
where is the lifetime of holes in the material and t h e negative sign is inserted to
indicate a decrease in as long as > p .
0
J (x)
P
] (x + Ax)
P
+ Ax
Figure 4.10
equation.
Chapter 4
light pulses
C
metallic contact
D
/ metallic contact
 L
jf
(a)
hole density
(b)
Figure 4.11 (a) HaynesShockley experimental setup; (b) the effects of drift,
diffusion, and recombination on the minority carrier hole pulse.
= = ^ ^ A A x
[ J p i x )
 "
{X
^
=  ^  qW
dt
(4.36)
dt
div/
(4.37)
Section 4.13
HaynesShockley Experiment
113
4.12 H A Y N E S  S H O C K L E Y E X P E R I M E N T
In 1951, two researchers at Bell Labs, H a y n e s and Shockley, r e p o r t e d an experiment
they performed that enabled t h e m to m e a s u r e mobility, diffusion constant, and life
time. It is i m p o r t a n t to emphasize again that, in determining currentvoltage rela
tions of semiconductor devices, it is t h e carrier processes of minority carriers that
will serve as t h e mechanisms of the derivations.
We will briefly explain the experiment without entering into t h e details of the
calculations that follow the results. The main purpose of describing the experiment
in this b o o k is to illustrate the combined t h r e e carrier processes that minority carri
ers undergo.
A s shown in Fig. 4.11(a), t h r e e contacts are m a d e to an semiconductor bar.
The two contacts at A and are used to connect a battery in series with a resistance
R The function of the battery is to establish an electric field that will cause carriers
to drift in the bar with holes moving from left to right. A t point C, we focus pulses of
a b e a m of light on the bar and the light energy moves m o r e electrons from the con
duction to t h e valence band, thus generating, at C, additional electrons and holes. A t
point D, a special contact is m a d e to the bar, which we will study in the following
chapter, that permits only minority carriers to b e picked up.
The light pulse b e a m generates a pulse of electrons and holes at C. The n u m
ber of electronhole pairs generated at C, because of lowinjection, is so small that
the n u m b e r of electrons is negligible c o m p a r e d to those in the bar. However, the
n u m b e r of holes is m u c h greater t h a n those (minority carriers) in t h e bar. The pulse
of holes generated drifts in the bar, is picked u p at D, and the current of holes causes
a voltage d r o p across R to ground. This voltage across R is connected to the Y ter
minal of t h e oscilloscope and appears as a pulse on the scope. The Xtime scale of
t h e scope is synchronized with t h e timing of the light pulse generator. In brief,
pulses of holes are lightgenerated at C and displayed on the scope after travelling
in the bar.
O u r interest h e r e is in the shape of the hole pulse as it travels from C to D. T h e
hole pulse is subjected to the three carrier processes as shown in Fig. 4.11(b). T h e
pulse moves along the bar as a result of drift, its amplitude decreases and the pulse
y
114
Chapter 4
PROBLEMS
5
15
17
14
4.5
b)
The energy in eV and in Joules delivered to the bar.
At = 300K, the intrinsic carrier density of a GaAs sample is 1.8 X 10 cm~ , =
8500, and = 400cm /Vs. An electric field applied to a bar of extrinsic GaAs causes
equal electron and hole current densities. Determine
a)
The equilibrium electron and hole densities.
b)
The net doping density.
A semiconductor has an intrinsic resistivity of 3 10 ohmcm. Donor and acceptor
atoms are added with densities of 10 cm~ and 5 X 1 0 c n r respectively. Given =
1600 and = 600cm /Vs, determine the current density if the applied electric field
is lOOmV/cm.
A bar of intrinsic semiconductor has a resistance of 5 ohms at 360K and 50 ohms at
330K. Assume that the change in resistance is a result of the change in only.
Calculate the band gap energy of the semiconductor.
A certain silicon sample has N = 5.01 X 1 0 c n r and N = 5.02 X 10 cm . Given
= 1200 and = 400cm /Vs, determine
6
4.6
14
12
4.7
4.8
16
16
3
Chapter 4
a)
b)
Problems
115
The conductivity.
Repeat part (a) if N = 10" and N = 1.1 X 10 cm . Use the mobilities of
part (a).
4.9 Determine the resistance of a bar of silicon that has the following properties: length L
= 0.8cm, area A = 1mm , N = 3 x 10 cm , N = 10 cm" , = 1000 and =
500cm /Vs.
4.10 A silicon bar has a length of 1cm, a height of 0.01cm, and a depth of 0.2cm. At =
300K, determine the resistance of the bar for the following conditions:
a)
Intrinsic.
b)
A donor doping of 10 cm~ .
c)
An acceptor doping of 10 cm~ .
4.11
a)
For minimum conductivity of a semiconductor sample, determine an expres
sion for the electron and hole density, at a given temperature, in terms of ,
, and the intrinsic carrier density,
b)
Use = 3900, = 1800cm /Vs, and n = 10 cm to calculate the hole
density and the maximum resistivity.
A
17
14
16
15
15
10
13
4.12 Given a Ge sample that has, at a given temperature, n = 2.5 x 10 cm~ , = 3900
and, = 1900cm /Vs, determine:
a)
The intrinsic conductivity.
b)
The minimum conductivity.
i
4.13 Given a semiconductor that has a mobility ratio that is independent of impurity den
sity given by = / and K> 1, determine an expression for the maximum resis
tivity in terms of and the intrinsic resistivity.
4.14 Determine all the possible values of hole and electron densities that cause the con
ductivity of a semiconductor to be equal to the intrinsic conductivity.
4.15 An electric field of lOV/cm is applied to an intrinsic silicon sample. If the carriers drift
lcm in 100, determine, at = 300K:
4.16
a)
The drift velocity.
b)
The diffusion constant.
c)
The conductivity.
A bar of heavily doped Psilicon, to which an electric field is applied, has a drift cur
rent density of 50A/cm . The hole drift velocity is 50cm/s. Determine the hole den
sity.
An Ntype Ge bar has a resistivity of 5 ohmcm. Determine the time it takes an elec
tron to travel 5 X 10 cm if the current density is 0.1A/cm . Use the plots in the text.
A voltage is applied between two contacts that establishes an electric field of
lOOV/cm in the space separating the contacts. For an electron starting from rest at the
first contact, determine:
a)
The drift velocity of an electron if the space between the contacts is occupied
by an Nsemiconductor having = 3900 cm /Vs
b)
The velocity of an electron at a distance of lcm if the space between the con
tacts is a vacuum.
A bar of Ntype silicon lcm long is doped with 10 cm~ of donor atoms at = 300K
and a voltage of 5V is applied to the ends of the bar. Determine:
a)
The hole drift current density.
2
4.17
_3
4.18
4.19
15
116
Chapter 4
4.20 A hole current of 10" A/cm is injected into the side (x = 0) of a long Nsilicon bar.
Assuming that the holes flow only by diffusion, and that at very large values of x, the
distribution of excess holes decays to zero. Determine:
a)
The steadystate excess hole density at = 0.
b)
Repeat part (a) at = .
Given = 400cm /Vs, = 1600cm /Vs, and the lifetime of holes is 25 .
4.21 The bar of Prob. 4.20 has N = 10 ~ , determine:
a)
The rate of generation of electronhole pairs at = .
b)
The rate of recombination of electronhole pairs at = .
2
15
chapter 5
THE PN JUNCTION DIODE
5.0 I N T R O D U C T I O N
In earlier chapters, we have studied the properties of semiconductors u n d e r equilib
rium and nonequilibrium conditions. It was indicated that p u r e intrinsic semicon
ductors are of very limited use. Semiconductors that are d o p e d with impurities form
the basis of the devices we are about to study.
A semiconductor that has b e e n d o p e d with acceptor impurities and into the
surface of which d o n o r atoms are diffused forms an extremely interesting junction
k n o w n as the P N junction diode. The currentvoltage characteristic of a typical
diode is shown in Fig. 5.1.
In closely studying this characteristic, we observe the two most important
properties of diodes. In t h e first place, w h e r e the voltage is positive, very small volt
ages, less t h a n one volt, cause large currents, whereas w h e n t h e voltage is negative
the current is extremely small. Second, t h e slope of the characteristic in t h e first
q u a d r a n t is extremely large and the slope in t h e third quadrant, preceding the sud
den drop, is very small.
The first property points to t h e use of the diode as a rectifier in converting
alternating (AC) voltages to undirectional voltages and eventually through filtering
to direct voltages ( D C ) . The second important application of t h e diode is its use as a
switch from an almost short circuit (low V / I ) in the first q u a d r a n t to an o p e n circuit
(high V / I ) in t h e third quadrant.
T h e P N junction diode provides an essential background to the study of t h e
bipolar and junction fieldeffect transistors.
a
117
118
Chapter 5
10
6
4
2
0
2
4
6
8
10
40
20
I
0
K(V)
20
40
(a)
12
10
8
6
4
2
0.2
0.4
0.6
0.8
1.0
V (V)
a
(c)
(b)
Figure 5.1 (a) Currentvoltage characteristic of diode; (b) Currentvoltage
characteristic in the positive I and V regions; (c) Circuit from which (b) is obtained
showing diode symbol.
5.1
SPACECHARGE REGION
F o r m a t i o n of R e g i o n
Let us assume that a slab of material and a slab of material are brought together
in a m a n n e r in which their structures line u p and a single crystal is formed, as shown
in Fig. 5.2. For t h e sake of simplicity, we assume that t h e (metallurgical) b o u n d a r y
b e t w e e n the and regions represents a step junction or an abrupt junction. It is
assumed, in an abrupt junction, that t h e transition from t h e to the region takes
Metal top contact
Figure 5.2
Section 5.1
SpaceCharge Region
119
Qn
Qp
Qn
Qp
Qn
'
POp
'
nf
"~ N
n =N
0n
(a)
Po
nop
POn
(b)
Figure 5.3 (a) Dopant and carrier densities before contact; (b) Carrier densities
after contact showing diffusion of carriers.
120
Chapter 5
qN
p0
qN 
metallurgical
junction
hi
bi
Barrier V o l t a g e a n d E n e r g y B a n d s
T h e charges in the depletion region cause an electric field, which results in a voltage
across the depletion region labeled t h e builtin voltage, V , and shown in Fig. 5.4.
The builtin voltage, V , is accompanied by bending of the energy bands, as
shown in Fig. 5.5. T h e bending is explained as follows: Voltage b e t w e e n two points a
and b is defined as the energy e x p e n d e d or acquired in moving a unit positive
charge from a to b. If t h e positive charge is at a and b is at a higher voltage V, then
work is e x p e n d e d in moving the charge. After arriving at b, the positive charge has
gained potential energy equal to t h e energy e x p e n d e d that is given by (qV), w h e r e q
is the charge in coulombs of the electron.
If an electron is m o v e d from a to b, with b at a higher voltage, V, t h e electron
loses potential energy so that at b the electron is said to have lower potential energy
than at a. This energy at b is equal to (qV).
bi
hj
Section 5.1
SpaceCharge Region
121
distance
Figure 5.5
n0
bi
bi
122
Chapter 5
5.2 A N A L Y T I C A L R E L A T I O N S AT E Q U I L I B R I U M
In this section, we will establish analytical relations for the conditions in the diode at
t h e r m a l equilibrium. After reviewing some basic concepts in electrostatics, we will
d e t e r m i n e the location of the Fermi level t h r o u g h o u t t h e diode. We will then derive
an expression for the builtin voltage at the junction. This is followed by the deriva
tion of analytical relations for the electric field, the potential distributions, and the
concomitant bending of the energy b a n d s in the diode.
Electrostatics of t h e S p a c e C h a r g e R e g i o n
The electric field is t h e force exerted on a unit positive charge. The force on an elec
tron is q%, where q is the m a g n i t u d e of the electronic charge.
We also k n o w that force is the negative of the gradient of potential energy so
that t h e force acting on an electron becomes
q% =  (gradient of potential energy of the electron)
The expression for the electric field becomes
^=grad
q
(5.1)
q dx
(5.2)
dx
(5.3)
(5.4)
q
Finally, we use Poisson's relation in E q . (5.5) to relate the potential, the poten
tial energy, and the electric field to the diode constants. Poisson's equation in one
dimension is
Section 5.2
123
d*<j>
dx
~P/
5 5
()
(5
*  7
6)
Constancy of t h e Fermi L e v e l
In the previous section, we concluded that, at thermal equilibrium, the net current
of both electrons and holes across the diode junction is zero. We refer to E q . (4.31b),
and r e p e a t e d here, for the hole current density in one dimension
dp
J =
(5.8)
qD
?Tx
T h e expressions for t h e hole density, given by E q . (3.41), and its derivative are
= . exp [(E  E )/kT]
i
(a)
(5.9)
*>*[jy/HH
dE,
dE
dx
dx
(b)
kT/q, %
dE,
dx
n exp (E  E )/kT
i
= [n exp (E j
E )/kT]
F
dA
dx
dE
dx
(5.10)
dE,
E q u a t i o n (5.10) is simplified so that~ ^= 0.
dx
This result indicates that t h e Fermi level is constant as we m o v e from t h e
region to the region. T h e identical result is obtained by using expressions for the
electron density.
We conclude that in equilibrium, the Fermi level must be constant t h r o u g h o u t
the semiconductor, as shown in Fig. 5.6.
JL
124
Chapter 5
distance
Figure 5.6 Energy distributions, the Fermi level, E , and the Fermi potentials at
equilibrium.
f
and E are bent, the Fermi level E is flat t h r o u g h o u t and not subject to any bend
ing. The question is then: By h o w m u c h energy are the b a n d s bent or what is the
m a g n i t u d e of the potential across the junction?
The potential is developed across the space charge region, which we have
assumed consists entirely of ionized donors and acceptors. We explained this
assumption by acknowledging that t h e r e are some carriers that are continuously
crossing the depletion layer. The exception is that their density is small in compari
son to the densities of the ionized impurities.
Since the intrinsic energy levels, E in the region and E in the region, are
b e n t and t h e Fermi level E is constant across t h e spacecharge region, we define
Fermi potentials in each of the charge neutral and regions, cb and ^ , as
v
jp
EF Ejp
(PFp ~
EF
<PFn
"
PJlj
q
q
We remind the r e a d e r that in the region, E is higher than E so that is
positive. In the region, E is above E so that <> is negative. Since E is constant
throughout, t h e total potential separation b e t w e e n E and E is given by ( ^ +
^), which we label t h e builtin voltage V . The bands of E , E and E are each
bent by qV .
A sketch of the energy bands at equilibrium showing the Fermi potentials is
shown in Fig. 5.6.
ip
in
ip
bj
in
bi
Section 5.2
125
\dp
or
%
*1<%
(5 13)
n0
Qn
kT
I " dp
P
%dx
(5.14)
where the first subscript in p, 0, refers t o the equilibrium value a n d the second sub
script refers to the relevant or regions.
The t e r m on the righthand side of E q . (5.14) is the negative of the builtin
voltage, V , across the depletion layer so that after integrating we have
u
b i
= in ^
Q
(5.15)
0n
Qn
0n
5 16
( )
(5.17)
. ' ^
bi
bi
126
Chapter 5
fJfo+
(*>
n0
n0
p0
n0
p0
nQ
The whole diode, t a k e n as one unit, is charge neutral and the total charge per
unit area on one side of = 0 is equal in m a g n i t u d e and opposite in sign to the
charge on the other side of = 0 as shown in E q . (5.19).
<l A x
p0
= qN x
D
(5.19)
na
where N and are the densities of the ionized impurities on the side and side
of = 0 in the depletion region respectively.
The doping is not symmetrical, and ratios of N to N or N to N in practical
diodes have values of about 100 so that we can conclude from E q . (5.19) that the
depletion layer extends d e e p e r into the lightly d o p e d region. The expression for
%(x) in t h e side of the spacecharge region, subject to the assumptions of the
depletion approximation, is obtained by integrating E q . (5.18) as
A
%{x) = x
A t = x ,%(x)
nQ
+ C, for 0 =s =s
=  ^ x
Section 5.2
%(x) =&&(
For x
pQ
=s
for 0
127
(5.20)
nQ
=s = 0
(5.21)
=  *
= ,
'
at
x = 0
(5.22)
A section of the diode, t h e charge distribution, the electric field intensity, t h e poten
tial and the potential energy distribution are shown in Fig. 5.7 for N
N.
The expression for
from E q . (5.22) is used in Eqs. (5.20) and (5.21) so that they
are rewritten as
A
%(x) = %(l~)iov0^x^x
n0
(a)
Figure 5.7 Distributions of (a) charge, (b) electric field, (c) potential, and (d)
energy at equilibrium.
128
Chapter 5
%{x) = % ( \
+ f]
for  x ^ ^ O
(5.23)
= J %{x)dx for x
(b)
= =s 0
pQ
()
= % [x
max
+ ^ )
for x
^O^x
p0
(5.24a)
() = % (x
+  ^ ~ )
maK
(5.24b)
A l t h o u g h the above equation is only valid for the limits shown, the zero level for
( ) is at = x The total voltage across the spacecharge region is given by set
ting = x in the expression for so that
p0
nQ
bl
= ~^(x
+ xj
m a x
bj
W i d t h of t h e Space C h a r g e R e g i o n
We will now obtain an expression for the total depletion width in terms of the phys
ical properties of the diode and t h e builtin voltage.
By using t h e expression for > from E q . (5.22) in Eq. (5.25), and realizing
that (x + x ) is the width of the depletion region, we have
<
mla
pQ
nQ
bi
= ^ ~ ( ^
+ *o)
(526)
where = e s .
The expression for x from E q . (5.19) is used in E q . (5.26) and the resulting
equation is solved for x as
r
n0
Section 5.2
/
X
"
2sV N
hl
\qN (N
A
129
\\
+ N )l
The width of t h e depletion layer. W, becomes, by using Eqs. (5.19) and (5.27),
(2 (N
iy
p
gv ^ L / V / V" '
+ TV,,),
(5.28)
EXAMPLE 5.1
I7
is
:i
An abrupt (step) junction diode made of silicon has N = 1 0 c n r and N = 10 cm~ . The diode
is at 300K and has area = 10 cm . The relative permittivity of silicon is 11.8.
Calculate:
a) The builtin voltage, V .
b) The depletion widths x and .
c) The maximum value of the electric field intensity.
d) The charge stored in each of the depletion regions.
A
hi
n0
Solution
From Eq. (5.17),
a)
We determine x
p0
V = 0.0259 in
10
15
hl
 (
{
1 1
8 5 4
= 0.715V
1 Q 2 0
n0
X 10"
1 0
1 4
1 0
"
"^'
1.6 X 1 0 " " X 10 (1.01 X 10 )
17
.,
17
0.0961 microns
pa
*
x
0.961 microns
W = x + x
= 0.970 microns
The maximum value of the electric field is found from Eq. (5.22)
ZV x
 1 . 6 X 10" X 96.1 X 10~
11.8 X 8.854 10~
19
15
14
 1 4 8 10 V/cm
Q* (depletion layer) =
d)
Q = 1.6
19
10
15
qNpX^A
6
X 96.1 10~ 1 0
130
Chapter 5
Q = 153.3 X 10" C
It is worth noting that the depletion layer extends much deeper into the region with the lighter
doping as shown by Eq. (5.19) and as calculated in the preceding example.
5.3
C O N D I T I O N S IN T H E D I O D E W I T H V O L T A G E A P P L I E D
Biasing a diode is the process of connecting a voltage source (or a current source)
b e t w e e n the metallic contacts of the and regions. The biasing voltage, depend
ing u p o n its direction, causes the diode to conduct either in the forward direction, as
in t h e first q u a d r a n t of the characteristics of Fig. 5.1, or in t h e reverse direction. In
the forward direction, a small voltage causes a large current, whereas in the reverse
direction, the current is negligibly small unless the voltage is so high that b r e a k d o w n
occurs, as shown in the large increase in the current at a certain fixed voltage in the
third quadrant.
Forward bias is achieved by connecting the positive lead of the voltage source
to t h e region contact and the negative lead to the region contact. Obviously, a
diode is said to be reversebiased when the opposite connections are made. We will
d e t e r m i n e t h e effect of t h e biasing on the characteristics of the spacecharge region
of Fig. 5.4.
W h e n a forward bias is applied to t h e diode by making positive with respect
to N, as shown in Fig. 5.8, the question is: H o w is this voltage distributed across the
diode? T h e r e are five places that can share the voltage as d e t e r m i n e d by the ohmic
drops. T h e two aluminum contacts to each of the and regions have very low
resistivity and hence low resistance so that there are negligible voltage drops across
t h e m . O n comparing the conductivities of the three remaining regions, we find that
t h e neutral regions have m u c h higher conductivities, hence lower resistance, com
p a r e d to the depletion region. The neutral regions have an a b u n d a n c e of carriers
Section 5.3
131
and the spacecharge region is depleted of carriers, causing it to have a high resis
tance. Thus, the applied voltage, at n o r m a l current levels, may be assumed to a p p e a r
totally across t h e depletion region.
The applied voltage V , opposes the builtin voltage, V , when V is connected,
to m a k e positive with respect to and V aids the builtin voltage w h e n it is con
nected to m a k e positive with respect to P. This is shown in Fig. 5.8(a), w h e r e the
former connection is labeled forward bias, and shown in Fig. 5.8(b), where t h e latter
connection is labeled reverse bias.
Because t h e applied voltage adds to or subtracts from the builtin voltage, the
electric field intensity must change. This necessitates a change in the a m o u n t of
charge in each side of depletion layer. T h e dopings N and N cannot change, thus a
change in t h e width of t h e depletion layer must accompany t h e application of a volt
age to the diode. Applying a forward bias reduces the electric field intensity, thus
reducing the charge and reducing the width of the depletion layer, whereas a
reverse bias increases that width. The effects of biasing on charge distributions, on
the electric field, on t h e potential distribution, and on t h e energy b a n d s are shown in
Figs. 5.9 and 5.10.
Since we concluded that all the applied voltage, at least at t h e n o r m a l current
levels, appears across the depletion layer, t h e voltage, V across this layer becomes
a
bi
v, * *
29
( >
where F is the barrier voltage and V is the applied voltage (with V being positive
for forward bias and negative for reverse bias).
T h e expression for the depletion layer width in E q . (5.28) is modified by
replacing V" by V. as
b i
bi
W =
2s(N
q\
NN
A
)>
(5.30)
REVIEW QUESTIONS
Q51 For a PN junction diode in thermal equilibrium, explain why each of the electron and
hole currents are zero.
Q52 Briefly explain why the section separating the and regions is labeled (a) a space
charge region (b) a depletion region.
Q53 A voltage appears across the depletion region of a PN junction diode in thermal equi
librium. Can this voltage be measured by connecting a voltmeter to the metal terminals
of the diode?
Q54 Since a voltage appears across the depletion region, which region has the higher volt
age, the or the region?
Q55 The depletion region extends deeper into the lightly doped (N) region of the PN diode.
Explain why.
132
Chapter 5
Forward bias
metallurgical junction
metallurgical junction
x=0
x=0
Xy.
x
p0
x" = 0
'=0
^
qN
v = o
v >
qN
qN A
(a)
V > 0
(b)
Vt
V >0
V = 0
14 V
v
bl
(C)
M bi~ a)
E:
v
(d)
Figure 5.9 Distributions of (a) charge density, (b) electric field, (c) potential, and (d) energy
at equilibrium and at forward bias (V > 0) for JV. N .
a
Section 5.3
Equilibrium
133
Reverse bias
metallurgical junction
metallurgical junction
I =0
TV
Xp
x" = 0
n0
x' = 0
<b
qN
v =o
<
 q N
qN
(a)
K, = 0
(b)
V)
V<0
X
(c)
qv
bl
q(v K)
bi
(d)
Figure 5.10 Distributions of (a) charge density, (b) electric field, (c) potential, and
(d) energy at equilibrium and at reverse bias (V < 0) for N
N.
a
134
Chapter 5
Q56 In Fig. 5.3(a), the change from the side of the depletion layer to the side looks per
fectly abrupt. This is not realistic. Why?
HIGHLIGHTS
nQ
EXERCISES
ES1
16
E52
15
5.4
C U R R E N T S IN D I O D E
M o t i o n of Carriers w i t h Bias A p p l i e d
In this section, we will derive the expression for the currentvoltage characteristic of
the P N junction diode. First, we will analyze qualitatively the currents across the
junction b o t h at equilibrium and with an applied voltage. Let us establish s o m e rules
concerning the location and m o t i o n of the carriers:
Section 5.4
Currents in Diode
135
3. Electrons in and holes in are minority carriers. They move across t h e junc
tion because of the electric field at the junction with electrons moving from to
and holes from to P. These are drift currents.
4. Majority carriers move across the junction, electrons from to and holes
from to N, because of the large concentration gradients at the junction. These
gradients result from the large concentration of electrons in and holes in P,
and the small density of holes in and electrons in P. Currents resulting from
these gradients are diffusion currents.
5. W h e n electron and holes move from one region to another, across the depletion
region, their m o t i o n occurs at a fixed energy level E. In this motion, while the
total energy is constant, the energy of the carrier changes from potential to
kinetic or vice versa.
A t t h e r m a l equilibrium, each of the electron and hole currents is zero. The
electron diffusion current caused by electrons moving from to is equal and
opposite to the electron drift current caused by electrons from to N. The same
analogous processes occur for holes. A schematic diagram to illustrate the motions
is shown in Fig. 5.11.
Two properties of t h e figures in the following pages require mentioning here.
First, we observe in Fig. 5.10 that the magnitude and sign of the slope of the energy
b a n d diagram in the depletion layer are a m e a s u r e of t h e intensity and direction of
t h e electric field respectively. We n o t e the electric field is directed from right to left.
A t forward bias, and as shown in Fig. 5.12, t h e slope has decreased, indicating a
decrease in the m a g n i t u d e of the electric field intensity as c o m p a r e d to the condi
tion at equilibrium.
electrons
drift ^ ^
<
/
drift
"
<S
/
electrons
diffusion ^ ^ diffuse
1^1 _.
*
___ __
,holes>
,
diffuse
< holes
,,
*
J
"
diffusion
Figure 5.11
d r i f t
distance
136
Chapter 5
Figure 5.12 Energy band diagram with forward bias. Note the negative slopes of
the energy levels in the spacecharge region. The slope decreases at forward bias,
reflecting a decrease in the electric field.
The second item is the use of filled and hollow dots to represent electrons and
holes arranged in pyramidlike structures in the conduction b a n d for electrons and
in t h e valence b a n d for holes. Pictorially, these represent the energy distributions of
the carrier densities that we have previously d e t e r m i n e d analytically from t h e prod
uct of the density of states function and the FermiDirac distribution. The distribu
tions of carrier densities represent an exponentially decreasing function of energy,
as m e a s u r e d upwards from the b o t t o m of the conduction b a n d for electrons and as
m e a s u r e d downwards from t h e t o p of the valence b a n d for holes.
W h e n a bias is applied to the P N junction diode, we assume that the region is
fixed in energy and we allow t h e Nregion energy levels to move u p for forward bias
and down for reverse bias. Since t h e Fermi levels in t h e neutral regions are fixed
with respect to E and E , t h e relations for t h e energy distributions of t h e carriers
apply and the n u m b e r of carriers available for transfer from one region to another
varies as the energy levels m o v e u p or down.
c
Conditions w i t h F o r w a r d Bias
W h e n a forward bias is applied to the diode, the electric field at the junction is
reduced and so is t h e voltage across the junction. If we assume that the energy levels
in the region remain fixed, t h e n the energy levels in the region are raised by
qV , as shown by Fig. 5.12.
The n u m b e r of electrons, minority carriers in the region, above E , is t h e
same as that at equilibrium. They constitute t h e electron drift current. The n u m b e r
of electrons in that have energies above the E of t h e region is considerably
greater than that at equilibrium. These electrons will diffuse to P. Thus, we have an
electron diffusion current across t h e junction that is greater than the drift current
caused by electrons moving from to N. Analogously, holes will diffuse from to
and this diffusion current is m u c h greater than t h e drift current due to holes from
to P.
A n illustration of the carrier motions on the energy b a n d diagram is shown in
Fig. 5.13.
a
Section 5.4
Fp
Ev
Currents in Diode
137
drift
ogo
Figure 5.13 Carrier motions with forward bias (V > 0). Note the larger number
of holes in and electrons in that are available for diffusion (compared to
equilibrium).
Conditions W i t h R e v e r s e Bias
W h e n a reverse bias is applied, the drift currents do not change since the density of
minority carriers has not changed. The slope of the majority carrier density distribu
tion across t h e junction is still very large but the currents due to diffusion are small
because of t h e smaller n u m b e r of these carriers, electrons in and holes in P, that
have energies that are greater than the E of and E of respectively. This is illus
trated in Fig. 5.14. Stated otherwise, electrons in and holes in have a higher
energy barrier to overcome, c o m p a r e d to the conditions at equilibrium, so that the
net current is negative. This current is negative because it is composed of a larger
c
%
P
J drift
n
/ diffusion
/ diffusion
jy
J drift
distance
Figure 5.14 Motion of carriers at reverse bias ( V < 0). Note the smaller number
of holes in and electrons in that are available for diffusion (compared to
equilibrium).
138
Chapter 5
Section 5.4
Currents in Diode
139
S o l u t i o n of C o n t i n u i t y E q u a t i o n
We refer to t h e continuity equation, E q . (4.36) for holes in the region, in
one dimension, in the steady state and replace J by its diffusion c o m p o n e n t ,
qD dp/dx, so that
p
Qn
SO'
2
dx
p'
=
Dr
p
5 32
( )
>
nQ
p' = B e x p ( ^ ) + B
1
p0
exp^J
(5.33)
L = VD^
(5.34)
The diffusion length represents t h e average distance that excess minority car
riers, holes in this case, diffuse before they recombine. To d e t e r m i n e and B , we
need two b o u n d a r y conditions. They are
p
f>'(0) =
p'(W )
n
ft.
= 0
plf I1
atx'
= 0
atx'
= W
(5.35)
The voltage V is the applied voltage and W is the distance in t h e region from the
edge of the depletion layer to the ohmic contact.
We shall indicate in a subsequent section* how we arrive at the first b o u n d a r y
condition. The second b o u n d a r y condition indicates that at the end of the region,
the excess carrier density (p p ) is zero. In other words, t h e hole density takes on
its equilibrium value. This is not due to recombination in the bulk of t h e region
but r a t h e r to the ohmic contact at the end of the region. E v e n if the width of the
region is so small that n o recombination occurs, the hole density at t h e ohmic con
tact will b e p . The ohmic contact is a surface of very high recombination r a t e
because of the great a b u n d a n c e of electrons.
W h e n we use the conditions of Eqs. (5.35) in E q . (5.33), we obtain
0n
0n
140
Chapter 5
OK
exp
Pan
exp
kT
exp
L,
(5.36)
exp
Before we p r o c e e d further, let us distinguish two cases: W
L and W
=10L
andW
ssO.lL
exp
~ =
qVa
1
exp
ioxW
(a)
(5.37)
kT
1
exp
1 
for W L
(b)
kT
The excess electron density
in the region, for the condition that the width of
the region is m u c h greater than L , is analogously given by
n
fx
l
0p
0p
exp
kT
X,
(5.38)
exp 
w h e r e n' is the excess electron density in the region, is the electron density in
the region, and n is t h e equilibrium value of t h e electron density in the region.
Sketches of the carrier distributions are shown in Fig. 5.15, w h e r e p (0) is the
value of at = (' = 0) and n^(0) is the value of at = x (x'' = 0). The rela
tions in Eqs. (5.37) for the minority carrier densities are valid at equilibrium as well
as for forward and reverse bias. For reverse bias, V is negative and from E q . (5.37)
the value of at = is zero. Also, zero is the value of at = .
Q
'
Currents Crossing J u n c t i o n
Since we have concluded that minority carriers cause only a diffusion current, the
expression for the hole current density in b e c o m e s
dp
n
d~x
dp'
(5.39)
qD
 >;
qVa
kT
exp
exp
forW
(a)
(5.40)
p\
'
qVa
1
exp
kT
forVy
(b)
JJx
qTJ p
p
0n
exp
forW
(5.41)
Section 5.4
Currents in Diode
141
Figure 5.15 Distributions of carrier densities at (a) equilibrium, (b) forward bias
for W L , and (c) reverse bias. Note the change in the width of the depletion
layer with bias (highly exaggerated).
n
w h e r e D ,p
and L refer t o holes in t h e region and W is t h e length of t h e neu
tral region.
A n analogous derivation for t h e electron current density at t h e edge of t h e
transition layer in t h e region gives
p
0n
142
Chapter 5
J (x )
n
0p
exp
1
kT
for W
(5.42)
Qp
[J (x) +
J(* )l
T h e diode current b e c o m e s
P
I =
I = A
qDpPon
qD nop
JlVa
expkT
for W
and W
(5.43)
n
for
W.
and
L
(5.44)
Qp
0n
E X A M P L E 5.2
Given a silicon diode that has the following properties:
N in = 2 X 10 cm"
7V inN = 2 X 10 cmJunction area = 2 x 10~ cm
l7
16
Currents in Diode
Section 5.4
Figure 5.16
143
Solution
The intrinsic carrier density for silicon at 300K is 1 X 10 cm~ . Because N and N
greater than , n N = 2 X 1 0 c u r and p N = 2 X 10 cm " . Therefore,
10
16
0n
17
0p
are much
_ nj_ _ 1 10 "
= 5 X 10 cm^
~ n ~ 2 10
3
16
Ik.
5 io cm~
Pop
1/2
1/2
= 11.4 X 10~ cm
1.6
x I P ' X 13 X 5 x 10
1 9
11.4 "
J { ) 0.1064 A / c m
/0.66
L \0259
eXP
and
J ( ) = 0.00774A/cm
l
144
Chapter 5
A [J (x ) + / , (  * , ) ] = 0.228mA
n
It is worthwhile noting that the much larger hole component of the current is a result of the
much higher doping of the region as compared to that of the region.
T h e Current L o o p
In the analysis of t h e currents in the diode, we have concentrated on the diffusion of
minority carriers w h e n a forward bias is applied to t h e diode. Naturally, the inquisi
tive student will ask: W h a t h a p p e n s to t h e carriers once outside the depletion
region, in t h e neutral regions, at the metallic contacts, and in t h e outside circuit?
Consider first a forwardbiased diode with t h e positive terminal of t h e battery
connected to the metallic contact of the region and t h e negative terminal con
nected to the metallic contact of the region. The metallic contact is an a b u n d a n t
source of electrons and thus the metalsemiconductor contact is a mechanism for
exchanging electrons and holes. A sketch of t h e diode showing carrier motion is
shown in Fig. 5.17.
A t t h e side, electrons are drawn by the positive terminal of the battery to b e
circulated in t h e wires external to t h e diode. The electrons come from t h e metal con
tact and these are replenished by other electrons from the semiconductor. Those
electrons that were carried away from t h e semiconductor contact created holes
behind t h e m . These holes travel to the depletion layer edge where they diffuse to
the region. In their quick j o u r n e y in the region, some of these holes recombine
with t h e minority carrier electrons that have diffused from the region across the
depletion layer.
T h e holes that diffuse from to r e c o m b i n e with electrons supplied from the
metalN semiconductor contact that had b e e n extracted from t h e metallic contact at
the region. Since electrons are majority carriers in the region, only a few of
t h e m will r e c o m b i n e with holes. The remaining will proceed to t h e depletion layer
to diffuse to the region.
In Fig. 5.16, we identified J in and / in as minority carrier diffusion cur
rents. We obtained the variation of J in and / in by an indirect m e t h o d . The
n
Section 5.4
electrons drawn by V from contact
>
holes drift in
< >
>
holes drift in
>
o
o>
holes drift in
>
holes diffuse in
(
holes diffuse in
>3
4
<*
electrons diffuse in
}
*~(oW)
PbyV
electrons drift in
o>
P region 
145
o>
Currents in Diode
*
depletion
region
electrons drift in
<
Nregion
metallic contact
metallic contact
electrons moved by V
.
electrons moved by V
electron
hole
( Q t ) electronhole
recombination
Figure 5.17
The minority carriers, in reverse bias, drift across the depletion layer (they all fall
down the very steep barrier) because of the large electric field that is established.
Of the electronhole pairs g e n e r a t e d on the side of the depletion layer, elec
trons fall into the side and holes migrate towards the metal semiconductor con
tact to m e e t t h e electrons that were transported there by the applied bias. They
recombine there and the resulting hole density is t h e equilibrium value, whereas at
the edge of the depletion layer, the hole density is zero. The electrons g e n e r a t e d on
the side drift to the region, w h e r e they migrate to the metallic contact and are
attracted to t h e outside circuit by the positive terminal of the battery, which is now
connected to t h e region.
Similarly, for t h e electronhole pairs g e n e r a t e d in the side of t h e depletion
layer, holes drift down the potential hill across the depletion layer to the contact in
and electrons migrate towards the semiconductorcontact to enter the external
circuit and m o v e towards the metalP semiconductor contact.
146
Chapter 5
S a t u r a t i o n Current
The currentvoltage characteristics of diodes in the forward direction, subject to the
assumption of low injection, and for a fixed forward bias, are d e t e r m i n e d by the
m a g n i t u d e of the reverse saturation current given by Eq.(5.44). For two diodes hav
ing identical doping densities and identical minority carrier lifetimes, t h e current
d e p e n d s on the minority carrier densities and o n the mobilities of minority carriers
t h r o u g h the diffusion constants. For two such diodes, t h e current depends o n the
semiconductor b a n d gap, the mobility of minority carriers, and on t h e temperature.
A comparison of the characteristics of silicon, germanium, and gallium
arsenide diodes having the same areas, t h e same lifetimes, and the same dopings is
shown in Fig. 5.18.
T h e smaller b a n d gap of g e r m a n i u m results in a higher intrinsic carrier den
sity, thus higher minority carrier d e n s i t i e s , p and n . T h i s causes an increase in the
saturation current and hence an increase in the forward and reverse currents.
A n o t h e r reason for the higher current of t h e germanium diode is the higher mobil
ity of carriers in g e r m a n i u m w h e n c o m p a r e d to that in silicon, which is evidenced in
t h e higher diffusion constant, D, which appears in Eqs. (5.44).
In contrast to g e r m a n i u m and silicon, gallium arsenide has a smaller value of
intrinsic carrier density, thus lower minority carrier densities while having a m u c h
higher mobility of electrons. It is t h e difference in the b a n d gaps of the t h r e e materi
als that mainly accounts for t h e differences in the shapes of the currentvoltage
characteristics. A t r o o m t e m p e r a t u r e , the b a n d gap of germanium is 0.67eV and of
silicon is 1.12eV, whereas it is 1.42eV for gallium arsenide.
A n o t h e r important factor that influences the diode current is t e m p e r a t u r e . A n
increase of the operating t e m p e r a t u r e exerts t h e following effects:
0n
0/)
Ge
GaAs
Si
10 
6
"3
2 
0.2
0.4
0.6
V (V)
a
0.8
1.0
Section 5.4
Currents in Diode
147
The first two effects cause an increase in the current, whereas the decrease of
mobility decreases t h e current. T h e increase of t e m p e r a t u r e decreases the value of
the exponential in E q . (5.43), resulting in a decrease of t h e current. The net result
of all the factors is such that an increase of t e m p e r a t u r e tends to increase t h e cur
rent. This increase of current could result in instability since the power dissipated in
the diode acts, in turn, to increase t h e t e m p e r a t u r e further.
B o u n d a r y Condition a t J u n c t i o n
T h e expression for the builtin voltage (barrier voltage), V , given by Eq.(5.15)
relates the equilibrium hole densities on b o t h sides of the spacecharge region. This
d e p e n d e n c e is also obtained in terms of t h e relevant electron densities. E q u a t i o n
(5.15) can b e rewritten as
bi
POP
VSs.
qV
kT
"Op
( 5
4 5 )
This expression was derived for equilibrium conditions where each of the hole
and electron currents across the junction are zero, making each carrier's diffusion
current equal and opposite to that carrier's drift current across the junction.
It is of interest to calculate using t h e data of E x a m p l e 5.2, the magnitudes of
the diffusion currents at equilibrium. The magnitudes of the hole and electron diffu
sion current densities across the depletion region, assuming linear distributions of
holes and electrons, are
n\ = qD r>
P
J(x/ * = 0)
r> PP ^~
P
= qD
i(a)\
J (x"
= qDp*=f
(b)
= 0) = qD ^
n
(5.46)
w h e r e W is t h e width of the spacecharge region and the values of the carrier densi
ties are those at the edges of the region.
REVIEW QUESTIONS
Q57
Q58
Q59
Q510
Q511
Q512
Q513
148
Chapter 5
HIGHLIGHTS
When a voltage is applied that makes positive with respect to N, the charge in the
depletion layer is reduced, the electric field is reduced, and so is the voltage across it.
The diode is said to be forwardbiased.
At forward bias, the hole density in at the edge of the depletion layer is increased
and a hole gradient is established throughout the region. Analogously, electrons in
have a similar distribution. As a result of the gradients of holes in and electrons in P,
holes diffuse in and electrons diffuse in P.
To determine the analytical relationship for the current in a diode, expressions for the
hole and electron diffusion currents are required at the edges of the and regions
respectively. Expressions for the diffusion currents are obtained from a knowledge of
the distributions of minority carriers in both regions and these distributions are deter
mined from a solution of the continuity equations for these carriers.
The sum of the hole diffusion current at the edge of the depletion layer in and the
electron diffusion current at the edge of the depletion layer in results in the expo
nential IV equation.
When the voltage that is applied to the diode is connected to make positive with
respect to (reverse bias), the voltage across the depletion layer increases, the electric
field increases, and the energy barrier increases, making it more difficult for holes to
go from to and electrons to cross from to P. It becomes easy for electrons in
and holes in to cross to the other region. Thus, a very small current results, opposite
in direction, to the current caused by the forward bias.
EXERCISES
E53
Assume the distribution of excess hole diffusion density in the regions is given
p' = p'(Q) (1  xJWJ, where p'(0) is the excess hole density at x = 0 and W is
the width of the region.
Derive an expression for the hole diffusion current in N.
An N+P silicon diode has N = 1 0 c n r and N = 1 0 c u r with W L . Given
= = l a s and/I = 10~ cm .
n
E54
18
16
Ans: a) I = 0.107mA
EXAMPLE 5.3
Calculate, for example 5.2, the hole and electron diffusion current densities at thermal equilib
rium.
Solution The expressions for the width of the spacecharge region and the barrier voltage are
given by Eqs. (5.28) and (5.17) respectively as
W =
2s(N + N )
qN N
A
1/2
and
kT
NN
V . = in
~
q
nf
r
AA
D
D
Section 5.4
Currents in Diode
149
The value of V is calculated to be 0.81TV and W is 2.41 X 10 cm so that, at equilibrium, the hole
and electron current densities at the edges of the depletion region are calculated using Eq. (5.46) to
be
3
, , .
1.6 X 10
J =
19
17
13 X (2 10  5 10 )
,
5
' = 1726 x 10 A/cm
3
1 Q
16
, ,
,
(2 X 10  5 X 10 )
, ; ,
J (xJ
= 1.6 x 10 X 34 X
= 4.51 x 10 A/cm
2.41 X 10
1 9Q
p U
These values are much greater than those calculated for forward bias in Example 5.2.
(>'>)=
where p'(0) = (p(0)  p ) is the excess hole density in at the edge of the spacecharge region.
The assumption of lowlevel injection has resulted in a hole current across the
junction that is the difference b e t w e e n two large hole current components at equi
librium, namely J (drift) and / (diffusion).
0n
bi
so that
p(x )
p
= M*)(exp ^ )
p ( ^ )
4 8
(5 )
rf^jWgP^)
(5.49)
The value of t h e hole density in at the edge of the depletion layer is, for all
practical p u r p o s e s , / ? at low injection. So,p(x )
= p and from E q . (5.49)
0p
p(* )=P(0)=
n
0p
A ) n
exp^
(5.50)
150
Chapter 5
is p(0) p ,
so we can write
0n
exp
(5.51)
kT
d p'
p'
(5.52)
dx
= (D )
V
'
where
B.
exp
P ' ( 0 ) = P exp
0
(5.53)
exp
(5.54)
= B, + B
(5.55)
= 0
p'(W)
(5.56)
n\
(5.57)
p'(0)
p'(0)
1
2W,
exp
exp (2W
(5.58)
exp
p'(0)
exp
exp
[) ~ l2W x'Y\
n
(2W \
n
(5.59)
Problems
151
L . E q . (5.59) becomes:
p
p'(0)
exp
(z3
p'(0)
(5.60)
exp
exp
and this can in turn be rewritten as
p'(o)
P' =
p'(o)
exp
L,
exp
2(W
L,

(5.61)
x')
exp
For values of (W
x'
P'(0)
(5.62)
exp
Case II: W L . For this case, we shall expand the exponential in E q . (5.59)
into a series, so that we have
n
2W
P'(0) 1 +
_ x'
1 +
(5.63)
2%
p'(0)
X'
 +
2p'(0) ' 
L . We rewrite
Ep_
2W
L
2W,
or
' = ' ( 
(5.64)
0p
0n
152
Chapter 5
15
bi
n0
pQ
17
14
13
bi
a)
N /N .
b)
The ratio of the conductivity of to that of N.
Assume = . , and = 2.5 .
A
'
19
16
18
16
16
14
20
6
18
_2
13
3
16
3
Chapters
Problems
153
e)
The width, W, of the depletion layer when a reverse bias of 10V is applied.
5.10 A PN junction diode has a reverse saturation current of at = 300K. Determine
the applied voltage for currents of
a)
1mA.
b)
10mA.
5.11 We will repeat Prob. 5.10 except that we will account for the IR drops. A PN junction
diode has a reverse saturation current of . The resistivity of the region 0.05
ohmcm and that of the region is 0.2 ohmcm. Each region is 1mm long and has an
area A = 0.5mm . Include the IR drops in the and regions and find the forward
voltage, V, for
2
a)
/ = 1mA.
b)
I = 10mA.
5.12 A PN junction silicon diode has resistivities for the and regions, p = 0.2 ohmcm
and p = cm. Given that the lifetime of minority carriers in is 10~ sec. the life
time of holes in is 10~ sec and A = 10~ cm :
n
a)
Calculate the density of minority carriers at the edge of the depletion region
in when the applied forward voltage is 0.6V.
b)
Plot the values of the majority and minority carrier currents as functions of
the distance from the junction, on both sides of the depletion layer, for V =
0.6V.
a
c)
Locate the plane in at which the majoritycarrier current equals the minor
itycarrier current for V = 0.6V.
a
17
15
5.13 An abrupt junction silicon diode has N = 10 cm~ and N = 10 cm~ . For electrons
in , = 800cm /Vs and = 0.1 , and for holes in , = 480cm /Vs and =
. Use = 10~ cm :
A
a)
b)
Assume that the mobility and lifetime do not change with temperature.
Repeat part (a) for = 400K.
5.14 For a P N diode, use the results obtained in the text to determine:
a)
An expression for the excess minority charge Q stored in the region
assuming W L .
b)
An expression for the hole current at the edge of the depletion layer in N, at
= x , in terms of Q .
+
n0
5.15 The limit of lowlevel injection is normally assumed to be when the minority carrier
density at the edge of the depletion layer in the lower doped region becomes equal to
one tenth the majority carrier density in that region. For a silicon diode having N =
10 cnT and N = 10 cm~ , determine the value of the applied voltage at which the
limit of lowlevel injection is reached.
A
17
15
18
3
= 1 0 c m and N
15
= 10 cm" . Determine:
a)
b)
n0
5.17 Certain PN junctions have a doping profile that is known as linearly graded, as shown
in the figure, such that (N N ) = ax in the depletion region. Assume symmetrical
doping so t h a t x = W/2 andx^ = W/2.
D
154
Chapter 5
Determine:
a)
An expression for the electric field distribution in the depletion layer.
b)
An expression for V .
u
c)
5.18 A P N
Given
a)
b)
17
15
bi
bi
16
13
bi
I=
CPexp(E /kT)
5.21
a)
1 dE
L dT
__
3
E
kT
Chapter 5
b)
5.22
a)
Problems
155
b)
5.23
a)
b)
18
c)
18
15
chapter 6
FABRICATION
TECHNOLOGY
6.0
INTRODUCTION
The use of semiconductors has had a profound impact on the electronics industry
and t h e consequent introduction of integrated circuits has had a major impact on
everyday life. The microminiaturization of electronics circuits and systems and their
concomitant application to computers and communications represent major inno
vations of the twentieth century. These have led to t h e introduction of new applica
tions that were not possible with discrete devices.
The simultaneous formation of m a n y integrated circuits on a single silicon
wafer followed by t h e increase of t h e size of t h e wafer to a c c o m m o d a t e m a n y m o r e
such circuits served to significantly reduce the costs while increasing the reliability
of these circuits.
W h e r e a s t h e electronics engineer was previously concerned with the design of
circuits using discrete elements, the engineer is now involved with the ubiquitous
interaction b e t w e e n t h e circuit and t h e fabrication process, which itself influences
the circuit design, thus forming an integral design feedback loop. Design engineers
are n o w r e q u i r e d to design the systems, the logic, the circuits, and t h e layout of the
integrated circuits on a wafer. T h e ingenuity in the design of economically competi
tive circuits that m e e t t h e r e q u i r e m e n t s of b o t h speed and power dissipation is
partly a result of the engineer's expertise in the allocation of space on a silicon chip.
R e a l estate on t h e wafer and on the chip has b e c o m e a p r i m e commodity.
Before the student proceeds to the study of the operation and characteristics
of o t h e r semiconductor devices, he or she should be aware of, first, the materials and
t h e processes used in the fabrication of integrated circuits and devices, second, of
the layout of t h e devices on a silicon chip, and, third, of the dimensions involved, in
156
Section 6.1
Figure 6.1
Why Silicon?
157
particular, in integrated circuits. T h e main reason for introducing the subject of fab
rication early on is that while we use simple sketches to describe the o p e r a t i o n of
the device in later chapters, the student will be cognizant of, and can visualize, the
actual construction of the device so that she or he may better u n d e r s t a n d the p r o p
erties and t h e limitations of these devices. A p h o t o g r a p h of a wafer containing hun
dreds of dice (chips) and a drawing of a chip are shown in Fig. 6.1. The identical
chips, each of which may vary in area from 10 to over 100 m m , may contain u p to
several million devices.
In this chapter, we will first describe the process by which boules or ingots of
silicon are obtained. This process results in the formation of solid cylindricalshaped
ingots. It is from these that wafers are sawed off.
Following this, we describe the various processes that are used in the fabrica
tion of devices. Since we have already studied the P N junction diode in the previous
chapter, we end this chapter by applying the processes to illustrate the fabrication of
a diode, a capacitor, and a resistor for an integrated circuit.
In C h a p t e r s 8 , 1 1 , and 12, wherein the other major devices are studied, we will
apply these processes to the fabrication of a bipolar transistor, a metalsemiconduc
tor fieldeffect transistor, and a metaloxidesemiconductor fieldeffect transistor.
2
6.1 W H Y S I L I C O N ?
Semiconductor devices are m a d e in one of two forms: either as single discrete units,
such as a diode or a transistor, or in conjunction with other circuit elements making
u p an integrated circuit. Integrated circuits may be monolithic, w h e r e b y transistors,
diodes, resistors, and capacitors are fabricated and interconnected on the same sili
con chip or they may b e hybrid. In hybrid circuits, some of the circuit elements are
in discrete form and others are interconnected on a chip with the discrete elements
connected externally to those formed on the chip.
158
Chapter 6
Fabrication Technology
15
6.2 T H E P U R I T Y O F SILICON
The starting form of silicon, which manufacturers of devices and integrated circuits
use, is a circular slice k n o w n as a wafer. T h e wafers are cut from single cylindrical
ingots of silicon, with wafer diameters, varying from 10 to 20cm and expected to
reach 30cm in the not too distant future. The wafer thickness is of the order of sev
eral h u n d r e d microns. L a r g e d i a m e t e r wafers are very costeffective because of t h e
larger n u m b e r of integrated circuits they accommodate.
Silicon is found in a b u n d a n c e in n a t u r e as an oxide in sand and quartz. A num
ber of processes are required to convert the sand into silicon wafers. To b e useful in
Section 6.2
159
1 9
 3
15
13
22
Silicon From S a n d
Since we n e e d silicon in crystal form for integrated circuit fabrication, the question
is: H o w d o we find it or obtain it?
Silicon, as the element, is not found in nature. It is, however, found abundantly
in n a t u r e in the form of silicon dioxide, which constitutes about 20 percent of t h e
earth's crust. Silicon is commonly found as quartz or sand. Therefore, we have to
first convert silicon dioxide into silicon.
O n e might consider reducing S i O by the addition of hydrogen, but this is not
possible because S i O is a very stable c o m p o u n d . T h e r e are several m e t h o d s avail
able for obtaining silicon and t h e most c o m m o n is to first refine silicon dioxide
chemically with carbon in an arc furnace at very high t e m p e r a t u r e s resulting in Si
and C 0 . The C O evaporates as a gas leaving impure silicon, which is 9099 percent
p u r e and is k n o w n as metallurgical grade silicon. The next step is to purify the sili
con.
To purify silicon, a c o m m o n m e t h o d is to first p r o d u c e silicon tetrachloride by
burning the silicon:
z
Si + 2 C  SiC(
2
160
Chapter 6
Fabrication Technology
6.3 T H E C Z O C H R A L S K I G R O W I N G P R O C E S S
The Melt and the Dopant
The e q u i p m e n t setup for this process is shown in Fig. 6.2. To grow crystals, one starts
with very p u r e semiconductor grade silicon, which is melted in a quartzlined
graphite crucible. The melt is held at a t e m p e r a t u r e of 1690K, which is slightly
greater than the melting point (1685K) of silicon. The surrounding heaters and heat
shield establish a carefully controlled t e m p e r a t u r e with the center of the melt being
t h e coolest.
Pull
direction
C__J5 Pull and rotate
Seed holder
Seed
Solidliquid interface
RF heating coil
Section 6.3
161
A precisely controlled quantity of the dopant is added to the melt; added boron
makes silicon; a d d e d p h o s p h o r o u s m a k e s silicon. O n e assumes that t h e density
of impurities to be added is determined accurately by the desired resulting conduc
tivity. However, the p r o b l e m is not as simple as it m a y seem. W h e n a material freezes,
the concentration of impurities incorporated in the solid is usually smaller than the
concentration in the liquid. T h e ratio of the concentration of impurities in the solid,
C , t o that in the liquid, C , is k n o w n as the equilibrium segregation coefficient k ,
0
= C /C
23
Solution
a)
b)
)6
16
s s > x
a l o m
20
6.023 10 atoms/mole
We observe that a very small amount of phosphorus, ~ 0.03g. is needed to dope 50Kg of
silicon.
162
Chapter 6
Fabrication Technology
while, at the same time, the crucible is r o t a t e d in the opposite direction. The molten
semiconductor attaches itself to the seed and it b e c o m e s identical to the seed in
structure and orientation.
A s the seed is pulled up, the melted material that is attached to the seed solid
ifies (freezes). Its crystal structure b e c o m e s the same as that of t h e seed and a larger
crystal is formed. By this m e t h o d , cylindrical single crystal bars of silicon are p r o
duced. A s the molten silicon solidifies o n the seed, the purity of the silicon is
improved as most of the impurities t e n d e d to remain in the liquid and melt as the
melted silicon gradually solidifies.
The desired silicon bar diameter is obtained by controlling b o t h the tempera
ture and the pulling speed. In the final process, when the bulk of the melt has b e e n
grown, the crystal diameter is decreased until t h e r e is a point contact with the melt.
The resulting ingot is cooled and r e m o v e d to be m a d e into wafers. The ingots have
diameters as large as 200mm, with latest ones approaching 300mm. The ingot length
is of the o r d e r of 100cm.
I n g o t Slicing a n d W a f e r Preparation
The ingot surface is ground t h r o u g h o u t to an exact diameter and t h e t o p and bot
t o m portions are cut off. Following this, circular wafers are sliced off t h e ingot with a
high speed d i a m o n d saw. T h e wafer thicknesses vary from 0.4 to 1.0mm.
Slicing t h e wafers to be used in the fabrication of integrated circuits is a proce
d u r e that requires precision equipment. The object is to p r o d u c e slices that are per
fectly flat and as s m o o t h as possible, with n o d a m a g e to the crystal structure. The
wafers n e e d to b e subjected t o a n u m b e r of steps k n o w n as lapping, polishing, and
chemical etching. T h e wafers are first lapped with a suitable abrasive, such as dia
m o n d , to r e m o v e t h e irregularities introduced by the sawing. They are also chemi
cally etched t o p r o d u c e flat and parallel surfaces and finally polished to a
mirrorlike finish.
The wafers are cleaned, rinsed, and dried for use in t h e fabrication of discrete
devices and integrated circuits. It is interesting t o n o t e that the final wafer thickness
is about o n e third less t h a n that after the sawing.
The growth of G a A s crystals is m u c h m o r e complex than that of silicon. The
largest commercially available wafers are about 10cm in diameter. O n e reason for
this is that the wafers are brittle and m a y crack. F u r t h e r m o r e , G a A s crystals contain
a high concentration of crystal defects that can d e g r a d e the device yield signifi
cantly.
In the next several pages, we will explain and illustrate with sketches, w h e r e
necessary and relevant, the major operations required for the fabrication of circuits
and devices o n a wafer of silicon. Having d o n e that, we will apply t h e knowledge
gained to t h e fabrication of a P N junction diode. A s we study the major devices in
the chapters following, we will refer to the operations involved in their fabrication
and illustrate the fabrication of a typical device.
Section 6.4
6.4
Fabrication Processes
163
FABRICATION P R O C E S S E S
The category of processes that are used in the fabrication of devices and integrated
circuits are t h e following:
Oxidation
Diffusion
I o n Implantation
Photolithography
Epitaxy
We will now consider each process separately and apply some of these to t h e
formation of a diode, in this chapter, and to the fabrication of transistors, in later
chapters.
T h e basic fabrication process is k n o w n as t h e planar process, in which the
introduction of impurities and metallic interconnections is carried out from the t o p
of t h e wafer. A major advantage of the planar process is that each fabrication step is
applied to all identical circuits and devices on each of the m a n y wafers at t h e same
time.
It is i m p o r t a n t to initially emphasize that the fabrication requires an
extremely clean environment in addition to the precise control of t e m p e r a t u r e and
humidity.
Thermal Oxidation
The process of oxidation consists of growing a thin film of silicon dioxide on t h e sur
face of t h e silicon wafer. In the planar process, all operations are carried out from
the t o p surface. It b e c o m e s necessary to shield certain regions of the surface so that
d o p a n t atoms, by diffusion or ion implantation, may be driven into o t h e r selected
regions. T h e formation of a silicon dioxide layer is shown in Fig. 6.3 and its shielding
effect is illustrated in Fig. 6.6. Silicon dioxide, as we shall see later, plays an impor
tant role in making this possible. F u r t h e r m o r e , an S i O layer serves as a passivating
or protective layer o n the silicon surface to protect the devices during subsequent
processing.
The commonly used silicon dopants, such as boron, phosphorous, arsenic, and
antimony, have very low diffusion coefficients (diffuse with great difficulty) in S i 0 .
Because of this, S i O is used as a shield against infiltration of these dopants. O n the
other hand, these dopants diffuse very easily if the surface is silicon.
Oxidation is accomplished by placing t h e silicon wafers vertically into a quartz
b o a t in a quartz tube, which is slowly passed through a resistanceheated furnace, in
t h e presence of oxygen, operating at a t e m p e r a t u r e of about 1000C. The oxidizing
agent m a y b e dry using dry oxygen or wet using a mixture of water vapor and oxy
gen. The oxide growth rate in the dry process is m u c h slower but it produces an oxi
dized layer that has excellent electrical properties. T h e whole operation is
z
164
Chapter 6
Fabrication Technology
(b)
Figure 6.3 (a) Thermal oxidation system and (b) growth of Si0 .
2
E X A M P L E 6.2
Determine the thickness of silicon that is consumed when a silicon dioxide layer of thickness x
is grown on the surface by thermal oxidation.
Solution The volume of one mole of Si or S i 0 is the ratio of its molecular weight to its density,
determined as,
2
60.08g/mole
,
,, ,
, = 27.23 cm
2.21g/cm
3
,  = 12.055 cm
2.33g/cm
Since 1 mole of SiO uses up 1 mole of silicon over the same area, we have
z
Section 6.4
thickness of Si
Fabrication Processes
165
thickness of Si X area
thickness of S i 0 X area
2
12.055
= 0.44
"27.23
Etching Techniques
Etching is the process of selective removal of regions of a semiconductor, metal, or
silicon dioxide.
T h e r e are two types of etchings: wet and dry. In wet etching, the wafers are
immersed in a chemical solution at a p r e d e t e r m i n e d t e m p e r a t u r e . In this process,
the material to be etched is r e m o v e d equally in all directions so that some material
is etched from regions where it is to be left. This becomes a serious p r o b l e m when
dealing with small dimensions.
In dry (or plasma) etching, the wafers are immersed in a gaseous plasma cre
ated by a radiofrequency electric field applied to a gas such as argon. T h e gas
breaks d o w n and becomes ionized. Electrons are initially released by field emission
from an electrode. These electrodes gain kinetic energy from the field, collide with,
and transfer energy to the gas molecules, which results in generating ions and elec
trons. The newly g e n e r a t e d electrons collide with other gas molecules and the
avalanche process continues t h r o u g h o u t the gas, forming a plasma. The wafer to be
etched is placed on an electrode and is subjected to the b o m b a r d m e n t of its surface
by gas ions. A s a result, atoms at or near the surface to be etched are r e m o v e d by the
transfer of m o m e n t u m from the ions to the atoms.
Diffusion
This process consists of t h e introduction of a few tenths to several micrometers of
impurities by the solidstate diffusion of dopants into selected regions of a wafer to
form junctions. Most of these diffusion processes occur in two steps: t h e predeposi
tion and the drivein diffusion. In the predeposition step, a high concentration of
d o p a n t atoms are introduced at the silicon surface by a vapor that contains the
d o p a n t at a t e m p e r a t u r e of about 1000C. M o r e recently, a m o r e accurate m e t h o d of
predeposition, to b e explained later, and k n o w n as ion implantation, is used.
A t t h e t e m p e r a t u r e of 1000C, silicon atoms move out of their lattice sites cre
ating a high density of vacancies and breaking the b o n d with the neighboring atoms.
The impurity atoms, which are incident on the surface, move into the silicon because
166
Chapter 6
Fabrication Technology
substrate
Section 6.4
Fabrication Processes
167
resistance heaters
quartz furnace tube
quartz tube
Vent
end caps
Liquid impurity source
Figure 6.5
silicon wafers
Nitrogen and oxygen gas are m a d e to pass over the container. These gases
carry the d o p a n t vapor into the furnace, where t h e gases are deposited o n t h e sur
face of t h e wafers. These gases react with the silicon, forming a layer on the surface
of t h e wafer that contains silicon, oxygen, and phosphorus. A t the high t e m p e r a t u r e
of the furnace, p h o s p h o r u s diffuses easily into t h e silicon.
So that the d o p a n t may b e diffused d e e p e r into t h e silicon, the drivein step
follows. This is d o n e at a higher t e m p e r a t u r e of about 1100C inside a furnace, simi
lar to that used for predeposition, except that n o d o p a n t is introduced into the fur
nace. T h e higher t e m p e r a t u r e causes t h e d o p a n t atoms to m o v e into t h e silicon
m o r e quickly. Diffusion d e p t h is controlled by t h e time and t e m p e r a t u r e of the
drivein process. By precise control of the time and t e m p e r a t u r e (to within 0.25C),
accurate junction depths of fraction of a micron can b e obtained. Diffusion of
d o p a n t into silicon is illustrated in Fig. 6.6.
phosphorus
dopant
atoms
deposited
silicon dioxide
semiconductor silicon
SiO,
dopant atoms diffuse in
silicon but not in S i 0
2
Figure 6.6
168
Chapter 6
Fabrication Technology
klT
k is a constant.
Vz)i.
L = diffusion length =
C(x,t)
= >0
(6.1)
QOcii) = C, erfc
(x/V4D t )
(6.2)
i l
(6.3)
and
erfc(x) = 1  e r f ( )
(6.4)
dx
j eric(x)dx
Jo
= 1/VTT
= C (2/JT
s
\fDytjD t )
2 2
e* '^
(6.5)
Sample calculations and plots are shown in Fig. 6.7 for p h o s p h o r u s predeposition at
1000C for 8 minutes followed by a drivein diffusion at 1250C for 32 minutes. T h e
Section 6.4
0.5
1.5
2.5
Fabrication Processes
3.5
4.5
169
depth in
Figure 6.7
12
21
1 4
cm /s
Ion Implantation
This is a process of introducing d o p a n t s into selected areas of the surface of the
wafer by b o m b a r d i n g the surface with highenergy ions of the particular dopant.
To generate ions, such as those of phosphorus, an arc discharge is m a d e to
occur in a gas, such as phosphine ( P H ) , that contains t h e dopant. T h e ions are then
accelerated in an electric field so that they acquire an energy of about 20keV and
are passed through a strong magnetic field. Because during t h e arc discharge
u n w a n t e d impurities may have b e e n generated, the magnetic field acts to separate
these impurities from t h e d o p a n t ions based on the fact that the a m o u n t of deflec
tion of a particle in a magnetic field d e p e n d s on its mass.
3
Following the action of the magnetic field, the ions are further accelerated so
that their energy reaches several h u n d r e d keV, w h e r e u p o n they are focused on and
strike the surface of the silicon wafer. A t this time, t h e ion current is of the order of
1mA.
A s is the case with diffusion, the ion b e a m is m a d e to p e n e t r a t e only into
selected regions of the wafer by a process of masking, which will be discussed later.
170
Chapter 6
Fabrication Technology
O n entering the wafer, the ions collide with electrons and with nuclei of silicon
atoms, and lose their energy. The d e p t h of p e n e t r a t i o n is about 0.1 to 1 micron. The
higher the energy of t h e ions and t h e smaller their mass, the greater is the depth of
penetration.
Ion implantation has t h e following advantages:
Photomask Generation
The whole process of integrated circuit fabrication consists of identifying selected
regions of each circuit (or dies) of t h e wafer surface into which identical d o p a n t or
metallic interconnections are m a d e , while protecting the other regions of the wafer
surface. To carry out one of t h e m a n y processes of oxidation, diffusion, ion implanta
tion, or epitaxy, a separate mask or mini mask is required for each operation whose
Section 6.4
Fabrication Processes
171
Photolithography
In this process, the image on the reticle is transferred to t h e surface of t h e wafer.
This is d o n e to o p e n identical windows so that t h e diffusion process, for example,
may t a k e place in all identical regions of t h e same I C and for all ICs o n the wafer.
A s an illustration, we assume that t h e first reticle is used over an oxidized surface.
To transfer t h e pattern, the wafer is coated with a lightsensitive p h o t o e m u l sion, k n o w n as photoresist. By applying about 1cm of the liquid to the wafer surface
and spinning the wafer very rapidly, a uniform film, about 1 micrometer thick, of
photoresist is formed over t h e oxidized surface of the wafer. After this, the following
steps, also shown in Fig. 6.8, are t a k e n to o p e n a window on the wafer:
3
172
Chapter 6
Fabrication Technology
SiO,
silicon substrate
and oxide film
substrate
'/////////////////////.
*
positive photoresist
SiO,
photoresist applied
substrate
1 II II II II It 1
mask
'////////////////////,
SiO,
substrate
photoresist etched
away under transparent
regions of mask
V////////A
SiO,
substrate
//////////
'//////////
SiO etched
away
z
SiO,
substrate
remaining
photoresist etched
away
Figure 6.8
SiO,
substrate
Steps in oxidation and window opening.
Section 6.4
Fabrication Processes
173
We indicated earlier that the reticle or mask is placed in direct contact with
the wafer. It may b e possible that t h e wafer has few irregular particles at the surface
of the crystal or it may have some dust particles. These particles stick to the mask
and cause defects in the surface of each successive operation. This p r o b l e m is cured
by a process k n o w n as proximity printing, wherein the mask is separated from the
wafer by a distance of about 1020 micrometers.
Epitaxial G r o w t h
Epitaxy is the process of the controlled growth of a crystalline d o p e d layer of silicon
on a single crystal substrate. T h e processes of diffusion and ion implantation, which
w e r e earlier described, p r o d u c e a layer at the surface that is of higher doping den
sity t h a n that which existed before the d o p a n t was added. It is not possible by these
m e t h o d s to produce, at the surface, a layer of lower concentration than exists there.
This can, however, b e accomplished by the m e t h o d of epitaxy. In the processes of
diffusion and ion implantation, a d o p a n t is driven into a substrate of d o p e d silicon.
In epitaxy, a layer of d o p e d silicon is deposited on t o p of the surface of the sub
strate. Normally, this single crystal layer has different type doping from that of the
substrate.
Epitaxy is used to deposit on N silicon, which is impossible to accomplish
by diffusion. It is also used in isolation b e t w e e n bipolar transistors wherein N~ is
deposited on P. It may also b e used to improve the surface quality of an substrate
by depositing material over it. The system for growing an epitaxial layer is shown
in Fig. 6.9.
In this system, silicon wafers are placed in a long boatshaped crucible m a d e of
graphite. The b o a t is placed in a long cylindrical quartz tube, which has inlets and
+
RF heating coil
so
cl
 temperature bath
Figure 6.9
SiCI
174
Chapter 6
Fabrication Technology
7. I n t r o d u c e dopant.
A n u m b e r of different chemical reactions can b e used to deposit t h e epitaxial
layer. Silane ( S i H ) or SiCt? can b e used with t h e following reactions:
4
S i H > Si + 2 H
4
SiC + H > Si + 4 H C
4
A silicon layer can b e p r o d u c e d from silane by the addition of heat, while sili
con tetrachloride requires a reduction by hydrogen.
To grow a layer of Ntype silicon, very small a m o u n t s of impurities, such as
P H , A s H , or S b H , are introduced simultaneously with the gases. D i b o r a n e ( B H )
is used to form a Player of silicon. D u r i n g the epitaxial layer deposition, t h e d o p a n t
atoms d e c o m p o s e and they b e c o m e part of the layer.
Thus, epitaxy provides a m e a n s for accurately controlling t h e doping profile in
o r d e r to optimize t h e performance of devices and circuits.
It is extremely important to r e p e a t o n e major consideration in t h e p r e p a r a t i o n
of processes and devices, namely that t h e wafers must b e very smooth and clean.
Dirt particles and irregularities can have damaging effects on t h e properties of the
devices. A s indicated earlier, and in all these operations, t h e wafers are thoroughly
cleaned before they are placed in a boat. O n c e inside t h e quartz tube, the cleaning
process is carried out by using nitrogen to flush the air out and hydrochloric acid is
m a d e to pass over the wafers in o r d e r to etch away a very thin layer of the surface.
3
Section 6.4
Fabrication Processes
175
M e t a l l i z a t i o n a n d Interconnections
After all semiconductor fabrication steps of a device or of an integrated circuit are
completed, it b e c o m e s necessary to provide metallic interconnections for the inte
grated circuit and for external connections to b o t h the device and to the IC. The
r e q u i r e m e n t that must b e m e t by t h e interconnections is that they have low resis
tance to minimize b o t h the voltage drops on t h e lines as well as the capacitances
b e t w e e n t h e lines so as to reduce delay times. The connections must also m a k e
ohmic contacts to semiconductors in t h e devices such as the and regions of a P N
junction diode. A n ohmic contact is one that exhibits a very low resistance, allowing
currents to pass easily in b o t h directions through t h e contact.
The high conductivity of aluminum m a k e s it t h e metal of obvious choice, par
ticularly in siliconbased devices. It also has the following advantages:
1. easy to e v a p o r a t e
2. can be easily etched
3. not expensive
4. adheres well to silicon dioxide
T h e r e are a variety of m e t h o d s for depositing aluminum on silicon substrates,
and we shall briefly present t h r e e c o m m o n methods, which are: resistance heating,
electron b e a m heating, and sputtering. In resistance heating, the source of the
h e a t e d e l e m e n t and the silicon substrate are located in an evacuated chamber. The
source is a small piece of aluminum attached to a coil of tungsten, which serves as
the heater. The h e a t e d filament with a high melting point remains solid while the
a l u m i n u m (with a small addition of silver or copper) is vaporized. The aluminum
molecules travel to the substrate w h e r e they condense, depositing an aluminum
layer o n the surface of the silicon. A photolithographic masking and etching
m e t h o d , using a phosphoric acid ( H P 0 ) solution or a dry etching technique, is
used to r e m o v e the metal from regions w h e r e it is not wanted. A typical intercon
nection b e t w e e n two diffused layers is shown in Fig. 6.10.
A n o t h e r m e t h o d of generating vaporized aluminum is to place the aluminum
in a crucible into a v a c u u m c h a m b e r together with t h e substrate. T h e aluminum is
3
diffusions
Figure 6.10
176
Chapter 6
Fabrication Technology
leads
Section 6.5
177
6.5 P L A N A R P N J U N C T I O N D I O D E F A B R I C A T I O N
To illustrate t h e various steps in the fabrication of a discrete P N junction diode, we
show in Fig. 6.12 a series of drawings that include most of the processes discussed
earlier. We list below t h e various steps in t h e fabrication of a P N junction diode. It is
important to indicate here that in integrated circuits, where all interconnections
and
device terminals are made at the surface, a diode is formed from a bipolar
transistor
by placing a shortcircuit between two of the three terminals of the transistor (collec
tor to base).
Figure
Process Description
+
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
178
Chapter 6
Fabrication Technology
/ / / / / / / / /
//
///777[
SiO,
N substrate
(a&b)
(C)
\ t I \ \ t t \ \ \ \
positive
photoresist
positive
photoresist
A'
(e)
(d)
SiO,
K///////I
V/////A
TV
(g)
(f)
SiO,
N+
N+
(h)
(i&j)
(j)
6.6
F A B R I C A T I O N O F R E S I S T O R S A N D C A P A C I T O R S IN ICs
Resistors
In integrated circuits, resistors are usually m a d e of impurities that are diffused into
a semiconductor, which is of opposite polarity. They are m a d e by and from t h e
processes that are used to form devices. Figure 6.13(a) shows a resistor m a d e of a
region diffused into an Nepitaxial layer and to t h e ends of which metallic contacts
are made. The section of the resistor, as dictated by the diffusion, is very nearly rec
tangular in shape, as shown in Fig. 6.13(b).
T h e resistance of the layer is given by
179
(b)
(c)
Figure 6.13 Resistors in integrated circuits (a) crosssection, (b) rectangular
dimensions, and (c) meander pattern.
R =
pL/Wd
180
Chapter 6
Fabrication Technology
V/////////.
'V/////////////A
Figure 6.14
PROBLEMS A N D QUESTIONS
6.1 A silicon crystal is to be grown by the Czochralski process and is to have in the melt
an arsenic concentration of 5 10 atoms/cm . The segregation coefficient of arsenic
in silicon is 0.3. Determine the initial arsenic concentration in the crystal.
6.2 A silicon crystal is to be grown by the Czochralski process and is to contain 5 10
boron atoms/cm . Given k for boron is 0.8:
a)
Determine the initial concentration of boron atoms in the melt to produce
the required density.
b)
If the initial amount of silicon in the crucible is 50kg, how many grams of
boron should be added?
6.3 A crystal of silicon is to be grown using the Czochralski process. The melt contains
10kg of silicon to which is added lmg of phosphorus. Given k (phosphorus) = 0.35,
atomic weight of silicon = 28.09, atomic weight of phosphorus = 30.97, and density of
phosphorus = 0.35g/cm , determine the initial dopant concentration in the solid at
the beginning of the growth if the atomic density of silicon is 5 10 c m .
6.4 In the diffusion process, Q is defined as the total number of atoms per unit area of the
semiconductor.
15
15
22
3
15
14
20
Chapter 6
Problems
181
b)
The surface concentration C after this step.
c)
The number of boron atoms/cm , Q, after this step.
Note that the dopant density cannot exceed the solid solubility in silicon.
6.6 The predeposition step of Problem 6.5 is followed by a drivein diffusion step at
1050C for 4 hours. Given the diffusion coefficient of boron in silicon at 1050C is 17.3
X 10" cm /s, determine:
S
14
a)
b)
c)
chapter 7
LIMITATIONS TO IDEAL
DIODE THEORY
7.0 I N T R O D U C T I O N
In the previous chapter, we explained the operation of t h e P N junction diode and
d e t e r m i n e d the ideal diode expression relating the diode current to t h e voltage that
is applied.
U p o n comparing experimentally d e t e r m i n e d characteristics with those pre
dicted by the currentvoltage exponential relationship, we observe deviations
b e t w e e n t h e two. These deviations occur mainly at the lowcurrent and t h e highcur
rent ends for forward bias, and in the reversebias regions of the characteristics. We
also observe a sudden increase of current in the reverse bias region of the experi
m e n t a l characteristic, labeled a b r e a k d o w n p h e n o m e n o n , that is not predicted by
our simple theory.
O u r emphasis in this chapter is on t h e explanation of the physical processes
that cause the actual characteristics to deviate from t h e exponential theoretical
characteristics. The major differences arise because of t h e assumptions that we
m a d e in deriving the relations b e t w e e n the current in the diode and the voltage
applied.
We will also study the switching properties of t h e diode and the relations
b e t w e e n these properties and the capacitances inherent within.
7.1 D E V I A T I O N S IN T H E F O R W A R D R E G I O N O F THE C H A R A C T E R I S T I C
R e v i e w of A s s u m p t i o n s
In deriving the ideal currentvoltage characteristic, and in the earlier subsection,
"Assumptions for Ideal D i o d e Equation," we m a d e the following major assumptions:
182
Section 7.1
183
184
Chapter 7
(a)
N
2
W(V )
a
metallic contact
electron
hole
(o) recombination
(b)
Figure 7.1 (a) A forwardbiased diode and (b) diode currents made up of
recombination currents.
exp
(7.1)
kT
Aqn]
l(L N )
p
by n /N
0jj
D
{L N )\
n
and n
Qp
by n /N
i
as
Section 7.1
185
(7.2)
The t e r m I
is given by 'Aqn/W/2T )
w h e r e T is assumed to represent the lifetime
of holes and electrons in the depletion layer and W is the width of t h e depletion
layer, which is a function of the applied voltage.
Interestingly enough, E q . (7.2) has b e e n shown to represent also t h e genera
tion current, I , in the depletion layer w h e n t h e diode is reversebiased. For that,
RO
^GEN ~~
IRO
(7.3)
R 0
).
*From Tyagi, Introduction to Semiconductor Materials and Devices, p. 198, copyright Wiley,
(1991). Reprinted by permission of John Wiley & Sons, Inc.
186
Chapter 7
RO
/ cc exp
*From S. M. Sze, Physics of Semiconductor Devices, p. 92, copyright Wiley (1981). Reprinted
by permission of John Wiley & Sons, Inc.
(7.4)
Section 7.1
187
where = 1 for diffusion current only and = 2 for recombination current only.
W h e n b o t h currents contribute to the total current, varies b e t w e e n 1 and 2.
High Injection
A t values of forward bias greater t h a n about 0.7V in silicon, the diode IV charac
teristic, shown in Fig. 7.2, departs from being a straight line with slope of q/kT to a
curve indicating a decrease in t h e current for an increase of voltage when c o m p a r e d
to that given by the ideal diode equation. The characteristic in this region is influ
enced by two p h e n o m e n a : highlevel injection and bulk resistance.
Highlevel injection is assumed to occur w h e n t h e minority carrier density at
t h e edge of t h e depletion layer is comparable to the equilibrium majority carrier
density.
The relative carrier densities at lowlevel injection and highlevel injection for
a P N diode are shown in Fig. 7.3.
In low injection regime, t h e minority carrier density at t h e edge of the deple
tion layer is about six orders of m a g n i t u d e smaller than the equilibrium majority
carrier density. A t high injection, the difference b e c o m e s negligible, as we illustrate
for holes in N, w h e r e N = 1 0 c m ~ :
+
16
A t equilibrium:
1 6
Qn
= 10 cmr ,p
10
= 1 X 10 cm
4
0 n
 3
16
so that,
16
16
= 1.1 X 1 0 c m "
(0)
21
16
Ap (0)
n
p,n
p,n
Po
"On
POn
"Op
"Op
low injection
Figure 7.3
high injection
188
Chapter 7
7.2 R E A L D I O D E C H A R A C T E R I S T I C S IN T H E R E V E R S E DIRECTION
A comparison of t h e experimentally observed reverse characteristic with that pre
dicted by the ideal diode equation is shown in Fig. 7.4.
The actual reverse characteristic exhibits two major deviations from that pre
dicted by the ideal diode equation. First, the equation predicts that for a reverse bias
*From M.S. Tyagi, Introduction to Semiconductor Materials and Devices, p. 205, copyright
Wiley, (1991). Reprinted by permission of John Wiley & Sons, Inc.
Section 7.2
ideal diode
equation
Vbr
189
br
0n
bi
/=
 VS
Ro)
(7.6)
where I and I
are defined following Eqs. (7.1) and (7.2). We note that while I is
i n d e p e n d e n t of voltage, I
increases with an increase of the reverse bias. T h e
increased reverse bias increased the width of the depletion region W, resulting in an
increase of the reverse current. The t e r m I
accounts for the increased recombina
tion current at low values of forward voltage and also accounts for an increase in the
generation current with increasing reverse bias.
s
RO
RO
R0
190
Chapter 7
Junction Breakdown
A s the reverse bias applied to a diode increases, both t h e width of the depletion
region and t h e electric field in that region increase. In general, t h e u p p e r limit on
the reverse voltage is placed by the p h e n o m e n o n of b r e a k d o w n that occurs at a cer
tain critical value of the electric field. A t b r e a k d o w n the reverse current increases
very rapidly with miniscule increases of the reverse voltage, as shown in Fig. 7.4.
T h e r e are two types of b r e a k d o w n : Zener b r e a k d o w n and avalanche break
down. Z e n e r b r e a k d o w n results from the tunnelling of a large n u m b e r of electrons
through t h e energy barrier, causing a large current. Avalanche b r e a k d o w n results
from impact ionization of a t o m s by electrons that have acquired a high value of
kinetic energy from the high electric field in t h e depletion region. We will discuss
avalanche b r e a k d o w n first.
Avalanche Breakdown
A t a critical value of the electric field in t h e depletion region, a sudden increase of
t h e reverse current caused by impact ionization occurs. The avalanche or snow
balling effect that occurs in the depletion region and resulting from impact ioniza
tion is illustrated in Fig. 7.5, w h e r e a large reverse bias has b e e n applied to the
diode. While b o t h electrons and holes take part in impact ionization, to simplify the
explanation, we consider t h e m o t i o n of electrons in the depletion region. We recall
that the kinetic energy of an electron is m e a s u r e d by its energy separation above the
b o t t o m of the conduction b a n d E . A n electron at E has zero kinetic energy.
W h e n an electron in the depletion layer is accelerated by t h e high electric field
caused by a large reverse bias, it gains kinetic energy. If t h e electron gains kinetic
energy equal to or greater than the b a n d gap energy, , and it collides with t h e lat
tice, a covalent b o n d is b r o k e n . The breaking of a covalent bond, which is equal
to t h e elevation of an electron from t h e valence b a n d to the conduction band,
results in the generation of an electronhole pair, as shown in Fig. 7.5 for electron
N u m b e r 1.
c
Figure 7.5 Illustration of impact ionization for electrons only. Each electron
generates a hole and an electron. Generation of electronhole pairs by hole
ionization is not shown.
Section 7.2
191
T h e two electrons, electron N u m b e r 1 and the one resulting from the collision,
are in turn accelerated by the field, gain kinetic energy KE > , collide with the
lattice, and generate two additional electronhole pairs. While we have shown the
consequences of electrons colliding only with t h e lattice, it is important to recognize
that the holes behave in an analogous m a n n e r in causing impact ionization. The
avalanche process of carrier generation by collision results in a very large n u m b e r of
carriers and hence a large increase in t h e current.
Because of t h e avalanche process, the current entering the depletion layer is
multiplied by a factor M, k n o w n as t h e multiplication factor, as it crosses the layer. If
without any avalanche process the reverse current is I , then the actual reverse
current, I , is d e t e r m i n e d from
RO
= I
o u t
/I, = I / I
n
(77)
R O
where /
and /. refer to exit from and entry into t h e depletion region. A n expres
sion for M, as a function of an empirically d e t e r m i n e d exponent n, is given as
1
(7.8)
1
where V is the applied reverse voltage and V is the b r e a k d o w n voltage of the
junction.
The use of the word b r e a k d o w n is unfortunate as it might convey the impres
sion that a destruction of the diode has t a k e n place. The diode is destroyed only
when its r a t e d p o w e r dissipation is exceeded. Therefore, this will d e p e n d on the cur
rent at b r e a k d o w n . The current is normally restricted to be less than the m a x i m u m
allowable value by t h e addition of a series resistor. The product of t h e m a x i m u m
allowable value of the current and t h e voltage at b r e a k d o w n must be less than the
rated power dissipation of the diode.
R
br
Expression f o r t h e B r e a k d o w n V o l t a g e
It is evident from Fig. 7.4 that an increase in the reverse bias increases the current
and causes b r e a k d o w n to t a k e place. T h e applied voltage causes a high electric field,
which causes the avalanche b r e a k d o w n as it accelerates the carriers, which in turn
p r o d u c e new carriers.
It appears then that there is a critical value of the electric field, % , which
causes the b r e a k d o w n . The critical field is usually assumed to be a constant for a
certain semiconductor, although it varies slightly with the doping.
We will now relate t h e m a g n i t u d e of the b r e a k d o w n voltage, V , to t h e diode
constants, assuming that t h e critical value of the field is a constant.
We n o t e from Fig. (5.10) that t h e electric field m a x i m u m occurs at = 0 and
has value, at equilibrium, d e t e r m i n e d from E q . (5.25) as 2 V / W a n d W i s given by
E q . (5.28). For biasing conditions, we replace V by V. (V  VJ, W by E q . (5.28),
and solve (5.25) for ^
as
cr
br
bi
bi
m a x
m a x
= " [ 2 V, q (N N )/s(N
A
112
+ N )]
D
(7.9)
192
Chapter 7
For
2qV,N
1/2
(7.10)
where %
is d e t e r m i n e d by the region with the lower doping density.
We assume that b r e a k d o w n occurs at  V , where %
and also that V.
I V \, so that V = V . Solving for V , we have
hr
hr
br
V,br
2qN
(7.11)
n
E X A M P L E 7.1
3
The doping densities of an abruptjunction silicon PN diode are N = 10" atoms/cm and N =
8 10 atoms/cm^ . Determine the breakdown voltage if the critical field is 3 X KfiV/cm.
Avalanche breakdown takes place when the maximum electric field intensity in the depletion
region is equal to the critical value.
A
15
Solution
(7.9) as
The expression for the maximum value of the electric field intensity is given by Eq.
0.5
2gN N V,
s(N
N )}
A
14
3 10
X 1.6 '" 10
1.04 10
12
V,
10
17
15
+ 8 10 )
= 39.48V
The voltage that must be applied is calculated after V is found from Eq. (5.17):
bi
Thus, a reverse voltage having magnitude (39.48  0.77 = 38.7) will cause breakdown in the
diode.
Section 7.2
193
E X A M P L E 7.2
In the absence of any multiplication of carriers when a reverse bias is applied to a diode, the
magnitude of the reverse current is the saturation current / . To obtain an estimate of the effect of
multiplication, calculate for the diode of Example 7.1, whose breakdown voltage, V , is 38.7V, the
values of the multiplication factors for reverse voltages of 10,20,30,38.3 and 38.6. Let  3 in the
expression for M.
?
hr
Solution
=
1 where For V
V
V
V
V.
a
=
=
=
=
=
10V,
20V,
30V,
38.3V,
38.6 V,
=
=
=
=
=
1.0175
1.1601
1.872
32.58
129.33
It is important to note the dramatic sudden increase of the multiplication factor, which is the
ratio of the reverse current leaving the depletion layer to that entering, as the breakdown voltage is
approached. Increasing the applied voltage by 0.3V, from 38.3V to 38.6V, increased the current ratio
from 32.58 to 129.33. A smaller value of causes sharper increases in as  1  , _ is approached. For
n = 2 a n d V = 38.6V, has a value of 193.
Zener B r e a k d o w n
The physical mechanisms that result in Z e n e r b r e a k d o w n are completely different
from those causing avalanche b r e a k d o w n . The only similarity b e t w e e n the two is the
general shape of the reverse characteristic of the diode. They have two major differ
ences. First, Z e n e r b r e a k d o w n in a silicon diode takes place at reverse voltages of
the o r d e r of 5 volts or less. Second, Z e n e r b r e a k d o w n characteristic exhibits a m o r e
a b r u p t rise in current at b r e a k d o w n . The difference in t h e shape of t h e characteristic
is shown in Fig. 7.6.
Figure 7.6
194
Chapter 7
Figure 7.7
Section 7.2
195
REVIEW QUESTIONS
Q71 Briefly explain the mechanism by which the recombination current, in a certain voltage
Q72
Q73
Q74
Q75
HIGHLIGHTS
Actual diode IV characteristics deviate from those predicted by the ideal diode equa
tion in both the forward and reverse directions.
In the forward direction, the actual characteristics establish a higher current at low
forward voltages (=s 0.3V) because of recombination in the depletion layer. At the
high end of the forward region (& 0.7V), the actual current is lower because of the
effects of highlevel injection and voltage drops in the neutral regions.
The ideal relations show a very small constant current in the reverse directions,
whereas the actual characteristics indicate both an initial increase of current with
increase of voltage and a rather sudden jump to an extremely large current. This large
current can only be limited by a series resistance in the circuit. This sudden large cur
rent leads to breakdown.
There are two types of breakdown. In silicon diodes the breakdown that takes place at
reverse voltages of 5V or less is known as Zener breakdown. At higher voltages,
breakdown is of the avalanche type. Zener breakdown results from the tunnelling of
carriers across the junction.
The probability of the transition of an electron through a barrier (tunnelling) is a
strong function of the barrier thickness. Tunnelling is only significant in highly doped
semiconductors where the depletion region is very narrow and the electric field is very
high.
EXERCISES
E71
15
10
10
A diode has a breakdown voltage of 20V. Use = 3 in Eq. (7.8) to determine the
multiplication factor at a reverse voltage of 19V.
Ans: = 1
E73
The critical field for breakdown in silicon is 3 X 10 V/cm. Calculate the doping of
the side of a P N diode if breakdown is to occur at 20V.
+
16
A n s : N = 1.47 X 10 cmr .
n
196
7.3
Chapter 7
C
(7.X2)
The transition capacitance exists w h e n a diode is reversebiased or forwardbiased, whereas t h e storage capacitance occurs only w h e n the diode is forwardbiased.
T h e Transition ( J u n c t i o n ) Capacitance
The depletion layer of a diode consists of positively ionized donors and negatively
ionized acceptor atoms. We refer to Figs. 5.9 and 5.10, which show the distribution of
the ions in t h e depletion layer. T h e depletion layer extends d e e p e r into the lightly
d o p e d region, so that for N N , x x . H o w e v e r , t h e charges contained
within b o t h parts of the depletion layer are equal as indicated by E q . (5.19).
The change in the voltage, AV , applied to t h e diode causes the width of the
depletion layer t o change, as shown in Fig. 7.8, resulting in a change of charge AQ.
Because the relation b e t w e e n the charge and the applied voltage, as we shall see, is
nonlinear, we define a transition capacitance or junction capacitance C as
A
<
>
bi
Q = AqN x
s
bi
= AqN x
A
(7.14)
(V
bj
Section 7.3
197
depletion layer (V ) a
 depletion layer
\(V + AV )
a
x=0
Q
+Q
Qs
Figure 7.8 Storage charge change and a decrease in the depletion layer width
result from positive AV bias. The distances x and x represent the width of each of
the depletion layers in and respectively.
a
For N
0.5i/v; 0.5
q N
N
dQs
= A
dV
2(N + N )
D
(7.16)
N , Cj is simplified to
D
qeN
1/2
2(V
(7.17)
 V.)
bi
It is to be noted that with a larger V. at reverse bias, C. is smaller than the value
at forward bias.
W h e n V in E q . (5.28) is replaced by V., for N N and x x t h e deple
tion layer width, x , is
bi
2 Vf\i/2
qN
For biased conditions, x a n d x replaced x
n
n0
and x
p0
V,
(718)
 MR
(7 19)
(7.20)
198
Chapter 7
1 v {V)
Figure 7.9
Plot of l / C vs. V .
a
hi
E X A M P L E 7.3
17
Solution
The expression for the junction capacitance of an abrupt junction is given by Eq. (7.16) as
*v*r
where V = V  V . For zero bias, V. = V , and C = C*.
bi
'
a )
* =
1 0
J1.6
" (
bi
 1 9
1.04 X 10~ x
2(1.08 X 10")
12
17
b!
;0
q=
c) At V = 0.7V, V = 0.77  0.7 = 0.07V
t
KFN
1
)
m
8 10
.0259 n  ^ 5  = 0.77V
C = 0.56pF
32
V=
0.188pF
Section 7.3
199
UJ
/0.77\
C = 0.56
pF = 1.85pF.
'o.()7/.
We conclude thai for the same diode, the transition capacitance in forward bias is much larger
that in reverse bias. The reason is that the width of the depletion region is much smaller when
iode is forwardbiased.
;
S t o r a g e Capacitance
The application of a forward bias, V , t o a diode has the following consequences: a
reduction in t h e barrier height, a reduction in the width of the spacecharge layer,
and an injection of majority carriers across t h e depletion layer into the opposite
region where they are stored as excess minority carriers. T h e density of the excess
stored minority carriers increases with an increase of the forward bias.
In E q . (5.37(a)) we d e t e r m i n e d an expression for the excess hole density dis
tribution in t h e region. By relocating the origin at x' = 0, (x = x ,) the expression
of E q . (5.37(a)) for the excess hole density in N, for W L , becomes
a
P;=Po(expl)^'
"
(7.21)
= qA\
Jo
p'dx'
= ^ L ^ e x p ^  l )
(7.22)
= d
Po
exp
w ^r ? "
q V J k T
(7 23)
We observe that the storage capacitance increases exponentially with the for
ward bias. I n addition to the storage capacitance, a n d as we found earlier, a forwardbiased diode has a transition capacitance as well. T h e transition capacitance, in
accordance with E q . (7.20), of a forwardbiased diode has a larger value t h a n that of
a reversebiased diode since the depletion layer is narrower. However, for a for
wardbiased diode the storage capacitance is significantly larger than the transition
capacitance.
It is essential t o point out that the charge storage we have referred t o is that of
excess minority carriers. T h e and regions are neutral at low injection because of
the presence of excess majority carriers drawn from the metal contact. T h e r e is n o
actual net space charge in t h e neutral regions as exists in the depletion region.
The storage capacitance is a m e a s u r e of the change of the area u n d e r the
minority carrier distribution as the voltage changes. This requires a time delay as
m e a s u r e d by the capacitance.
200
Chapter 7
E X A M P L E 7.4
17
15
Solution
as
Since N >5> N , the storage capacitance is given by the expression for C in the region
A
(Aq L p ,
p
qV \
0l
exp
^V~kf W
10
20
f ,
()
a 5
= 126 X 10~ cm
a) At V = 0.6V, C becomes
a
C = 179r>F
s
b) At V = 0.65V, C becomes
C\ = 1234.3pF
The results of the above example and those of example 7.3 indicate that, in the forward direction,
the storage capacitance is much larger than the transition capacitance.
7.4 S M A L L  S I G N A L E Q U I V A L E N T CIRCUIT
We define the incremental resistance of the diode as
(7.24)
By using the exponential relationship of Eq. (5.43) for the current of the
diode, and assuming N N so that n p and V' kT/q, t h e expression
becomes
A
op
/ = (qADyPjLJ
exp qVjkT
(7.25)
"
(dVJ
'
[qlkT)qA
D p^
p
exp {qVJkT)
( ?
'
2 6 )
By taking the product of the expressions for the incremental resistance and
the storage capacitance, using E q s (7.23) and (7.26) we have
Section 7.4
201
We n o t e that this time constant is equal to the lifetime of holes in the region.
The following example illustrates the dramatic decrease of the incremental
resistance as the forward bias is increased.
EXAMPLE 7.5
For the diode of Example 7.4, calculate the incremental resistance at a) 0.6V and b) 0.65V.
Solution The simple expression that can be used to calculate the incremental resistance of a
diode, where N N , is given by Eq. (7.27) as
A
' 
Using the values for C determined in Example 7.4 and substituting the value of 0.1 for the
lifetime, we have
s
for V = 0.6V,
a
r = 558.6ohms
rf
81 ohms.
E q u i v a l e n t Circuit of t h e D i o d e
A circuit containing t h e capacitances and incremental resistance can be used to
replace t h e diode for incremental variations of the voltage or current. W h e n , for
example, an incremental change in the applied voltage is m a d e , by circuit analysis,
one can t h e n d e t e r m i n e the corresponding change in t h e current but can also deter
mine the time response of t h e diode, such as h o w long it takes the diode current to
reach a certain fraction of its final value. The response to a sinusoidal voltage or cur
rent can b e d e t e r m i n e d using circuit analysis. A s we have just seen, t h e values of the
circuit elements d e p e n d u p o n the operating D C voltage which is applied to the
diode. Such a circuit is k n o w n as a smallsignal equivalent
circuit.
Small signal equivalent circuits for the diode in the reverse and forward direc
tions are shown in Fig. 7.10. The circuit for the reversebiased diode includes the
WW
(a)
Figure 7.10
(b)
202
Chapter 7
"= / " V
28)
7.5 P R O P E R T I E S O F T H E S H O R T  B A S E D I O D E
We will label the base of a diode as the region that has the w e a k e r doping. It is the
region of a P N junction diode. A shortbase P N diode is one in which the
base width, W , is m u c h smaller than the diffusion length, L , of minority carriers,
W L .
+
w
kT
n
A sketch of this equation is shown in Fig. 7.11, indicating a straight line distribution.
The expression for t h e current of a shortbase diode is given by t h e diffusion
current anywhere in t h e region, since the slope of the excess carriers is constant
t h r o u g h o u t the base. A constant current implies zero recombination, which is
expected since we have assumed that W is m u c h smaller than a diffusion length.
The forward diffusion current, for N N , b e c o m e s
A
dp' _
qAD
dx'
exp
qVa
(7.30)
kT
Section 7.5
203
=0
Figure 7.11 Excess hole density distribution p'(x') in the base of a shortbase
diode (W L ).
Qs = 9A(^J
XW ^
qApJ^
( e x p ^ l )
(7.31)
_ q A W
n P o n
qV
The p r o d u c t r C
d
<
W.
w ~
(7.33)
(^W^expf
becomes
rC
d
(7.34)
/(short) = ^ / ( l o n g )
C (short) = (Wj2L )
p
(a)
C (long)
(b)
(7.35)
204
Chapter 7
Therefore, the shortbase diode has a higher current and a smaller storage
capacitance t h a n the longbase diode.
U p o n comparing E q . (7.23) for the storage capacitance of the longbase diode,
E q . (7.32) for t h e capacitance of the shortbase diode, and the conditions of Eqs.
(7.35), we observe that t h e capacitance of the shortbase diode is much smaller than
that of the longbase diode. This is a result of the fact that m u c h less excess charge is
stored in t h e base of t h e shortbase diode, as shown in Eqs. (7.22) and (7.31). This
excess charge of minority carriers accumulates w h e n the diode is conducting in the
forward direction.
For the same forward voltage, V , and hence for the same forward current, less
charge is stored in t h e base of the shortbase diode because the width of the base is
much smaller t h a n a diffusion length. We recall that t h e longbase diode is defined
as one w h e r e t h e width of t h e base is equal m a n y diffusion lengths.
In the next section, we will consider the time it takes a diode to switch from a
reverse state to forward and back to reverse. We will conclude that this time is m a d e
u p mainly of the time t a k e n to turn the diode O F F from an O N position. This turnoff time d e p e n d s directly on the a m o u n t of excess charge that is stored in the base
while the diode is in the forward state, and hence on t h e storage capacitance.
a
7.6 D I O D E S W I T C H I N G C H A R A C T E R I S T I C S
Junction diodes are used as switching elements in a variety of applications. W h e n
used as a switch, the diode is said to b e O N w h e n it is operating in the forward direc
tion and O F F w h e n it is reversebiased in t h e region of the characteristic that pre
cedes b r e a k d o w n . The performance of a diode as a lowcurrent switch is m e a s u r e d
mainly by the time increment it takes to switch from o n e state to another.
F u r t h e r m o r e , an ideal switch is required to have zero resistance in t h e forward O N
state and infinite resistance in t h e O F F state. This resistance is the ratio of the
steadystate voltage to t h e steadystate current.
We will determine, in this section, the physical p a r a m e t e r s that d e t e r m i n e the
switching times.
In our analysis, we assume a P N junction diode so that the major d e t e r m i n a n t
of excess carrier storage is t h e excess hole density storage in the region. We there
fore neglect the operations in t h e region.
+
TurnON T i m e
To d e t e r m i n e the O N time, we consider a longbase diode, W L , in t h e circuit
of Fig. 7.12(a) that is in t h e reversebiased state having a V = V and a negligibly
small reverse current I . The p r o c e d u r e we will follow consists in determining the
relation b e t w e e n the excess hole charge stored in and t h e forward current, I , that
will b e applied at t = 0. To turn the diode ON, we can apply a positive voltage pulse
or a positive current pulse. Because of its simplicity, we will use a current pulse
formed in t h e circuit of Fig. 7.12(a) by moving t h e switch to position A , w h e r e a for
ward voltage V ( ~ 2 0 V ) is in series with a large resistance R .The positive current I
V /R for all t > 0,  V j V , and V is V at large t.
D
Section 7.6
h
Wv\
WA
V
A
205
\LV
Rf
(A)
Po(e^  1 )
(b)
(c)
Figure 7.12 (a) Switching circuit of the diode, (b) excess hole buildup in the
region, and (c) diode voltage buildup. The voltage V is the steadystate value and
V is the instantaneous value of the diode voltage.
a
10
Qn
J = I /A
F
= qD
pjL
[exp (q VjkT)
 1]
(a)
206
Chapter 7
= qAL
[exp (q VjkT)
Qn
 1]
(b)
(7.36)
= (L /D )l =7 I
(7.37)
p F
= I W J2D
F
(7.38)
TurnOFF T i m e
To turn a diode OFF, it is necessary to r e m o v e t h e excess charges stored by t h e for
ward current. The diode will b e O F F w h e n t h e current through it is I .
A t t = 0, the switch in Fig. 7.12(a) is almost instantaneously m o v e d to position
so that a reverse bias is applied to t h e diode. T h e instant before moving the
switch, at t = 0", t h e diode is in a forward state, with current I , and V = V as seen
in Fig. 7.12c, which is m u c h smaller t h a n V . T h e current is instantly changed from
its forward value I to a constant reverse value, as seen in Fig 7.13b, and given by
s
 ^
K
(7.39)
R
0n
Section 7.6
207
n(x )
v +v
R
Rn
<P(x)
RR
t = QPOn
=0
=0
(a)
(b)
Figure 7.13 (a) Decay of minority carriers in base and (b) the current variation
corresponding to the hole decay.
t = n
s
1+
(7.40)
A
(7.41)
*From M. S. Tyagi, Introduction to Semiconductor Materials and Devices, p. 257, copyright
Wiley (1991). Reprinted by permission of John Wiley & Sons, Inc.
208
Chapter 7
REVIEW QUESTIONS
Q76
Q77
HIGHLIGHTS
Chapter 7
Problems
209
made up of the transition and storage capacitances and the dynamic resistance dV/dl
of the diode.
When the diode is used as a switch, a low turnON time is favored by a small lifetime
for a longbase diode and by a small transit time for a shortbase diode.
EXERCISES
E74
E75
a)
b)
'
10
10
PROBLEMS
7.1 The effect of recombination in the depletion layer of a diode is studied by comparing
the magnitudes of the recombination current and the diffusion current. A P N silicon
diode, at = 300K, has N = 10 cm , N = 5 X 10 cm~ , and A = 10" cm . The
lifetimes in both and are assumed to be .. Determine the ideal and total diode
current for,
+
16
15
a)
V =0.1V
V = 0.5V
7.2 The lowvoltage characteristics of the PN junction diode can be modeled by two
diodes in parallel, as shown in the figure.
Plot on semilog paper / vs. V, at = 300K, as V increases from 0.01V to 0.75V.
Assume I = 1 0  , 7 = 10~ A, and that I does not vary with V .
b)
15
13
RQ
v.
210
Chapter 7
7.3 Use the diode parameters given in Prob. 7.1 to determine; at = 300K,
(a)
the excess hole density in the base at the emitter junction for an applied volt
age of 0.6V.
(b)
the voltage at which high injection is reached.
7.4 A PN silicon diode has N = 10 cm" , N = 5 X 1 0 c n r , = = 0.1 , and A =
10" cm . Determine, at = 300K:
(a)
the junction capacitance at zero applied voltage.
(b)
the storage capacitance at V = 0.5V.
(c)
the junction capacitance at V' = 10V.
16
15
7.5 A silicon P N junction diode has A = 10" cm . The relation between the junction
capacitance and the voltage is given by
+
18
Determine, at = 300K,
(a)
the builtin voltage V .
(b)
the doping density in N.
hj
(c)
the doping density in P.
7.6 Measurements on a silicon P N junction diode operating at = 300K yield the fol
lowing results:
+
At V =  4 . 3 V
C. = 20pF
At V = 0.55
C. = 40pF
Given that N
(a)
18
(b)
N.
7.7 A P N step junction diode has a breakdown voltage of 500V. The critical electric field
is 3 X 10 V/cm. Determine, at = 300K;
D
(a)
(b)
N.
D
15
3
fl
18
16
(a)
(b)
%(x = 0)
C
(c)
C,
(d)
T inN
(e)
in
(f)
depletion layer thickness
Briefly explain your reasoning.
p
Chapter 7
7.11
(a)
Problems
211
14
18
c =\
A
16
13/
3
chapter 8
BIPOLAR TRANSISTORS I:
CHARACTERISTICS AND
FIRSTORDER MODEL
8.0
INTRODUCTION
A transistor is a device m a d e from semiconductor material and has three terminals.
C u r r e n t is m a d e to flow t h r o u g h the semiconductor from one terminal to a n o t h e r
terminal, while t h e third terminal controls t h e flow of this current.
T h e r e are two types of transistors: the bipolar junction transistor and the fieldeffect transistor. The major difference b e t w e e n t h e two lies in t h e mechanism by
which control of t h e current is achieved.
The two major applications of transistors are amplification and switching.
Current, voltage and p o w e r can b e amplified using transistors in the p r o p e r circuit.
For digital applications, a transistor can b e m a d e to switch b e t w e e n t h e O N and the
O F F states in a very short duration of time.
We will study the structure, operation, and characteristics of bipolar junction
transistors in this chapter.
8.1 S T R U C T U R E A N D B A S I C O P E R A T I O N
The essential physical structure of a P N P transistor is shown in Fig. 8.1(a), with the
three regions and their corresponding depletion layers at the two junctions. While
this m o d e l will b e used to analyze t h e operation and characteristics of t h e bipolar
junction transistor (BJT), actual cross sections of devices are shown in Figs. 8.1(b)
and 8.1(c).
The B J T is composed of three d o p e d regions of a semiconductor in contrast to
the two regions of a P N junction diode. The middle region, which is sandwiched
212
Section 8.1
emitter
base
collector
)
B
213
214
Chapter 8
b e t w e e n two regions having the same type of doping, has the opposite type of dop
ing so that a transistor may b e of P N P or N P N type.
The B J T is labeled as bipolar, in reference to t h e two types of carriers, elec
trons and holes, which t a k e part in the conduction process. Conduction in t h e fieldeffect transistor is carried out by means of only one type of carrier; hence, it is
k n o w n as a unipolar device.
The t h r e e regions of the B J T are the emitter, the base, and t h e collector. These
three regions form two interacting P N junction diodes. In the most basic operation
of t h e device, k n o w n as t h e n o r m a l or forward active operation, t h e emitterbase
junction is forwardbiased and the collectorbase junction is reversebiased. In this
state, majority carriers from the emitter diffuse into t h e base, w h e r e a very small
percentage of t h e m r e c o m b i n e (because the diffusion length of these carriers is
m u c h greater t h a n t h e width of t h e base) and t h e rest are swept into the collector by
the reversebiased junction. If the width of the base is m u c h greater than the diffu
sion length, t h e n all the carriers would r e c o m b i n e in the base and the transistor
effectively b e c o m e s a combination of two diodes back to back and the collector cur
rent effectively becomes the reverse saturation current of one of t h e diodes.
The word transistor is an acronym for transferresistor, w h e r e transfer refers to
the relation of the output to t h e input. A low resistance is seen at t h e forwardbiased
diode w h e r e the ratio of voltage to current is relatively low, while the second diode
exhibits high resistance as obtained from the ratio of voltage to current.
Figure 8.1(b) represents a cross section of a P N P discrete transistor. We use
the t e r m discrete for a transistor that is separately packaged, whose t h r e e terminals
are accessible, in contrast to an integrated circuit transistor. The drawing in Fig.
8.1(a) represents the section shown by t h e dotted lines in Fig. 8.1(b), with the orien
tation shifted by 90 degrees.
In t h e majority of applications, B J T s are incorporated into an integrated circuit
(IC), also k n o w n as a microchip. The I C incorporates combinations of transistors,
resistors, and capacitors, all fabricated simultaneously on a single chip of crystal sili
con. H u n d r e d s of thousands, and m o r e recently several millions, of devices are fabri
cated on a single silicon chip. We show in Fig. 8.1(c) a cross section of an integrated
circuit N P N BJT. All connections on the chip are m a d e at the surface of the I C and
consequently all currents are confined to a very thin region at the surface.
The n o r m a l direction of current is shown by the dotted arrows in Fig. 8.1(b). In
the discrete transistor, the emitter and base contacts are located at the top, t h e col
lector contact is shown at the b o t t o m , and t h e carriers cross part of t h e base region,
which is n o r m a l to the direction of flow. The purpose of the two base contacts is to
reduce the ohmic base resistance.
While we will use the single m o d e l of Fig. 8.1(a) in our discussion, the actual
construction of a B J T in an I C is t h e one to k e e p in mind. We observe that all three
contacts to t h e transistor in an integrated circuit are m a d e from the top. O n e obvi
ous reason for this is that circuit connections to other elements in t h e I C chip need
to b e m a d e in one plane.
Section 8.2
215
8.2 F A B R I C A T I O N O F T H E B I P O L A R I N T E G R A T E D CIRCUIT
TRANSISTOR
We n o w consider the fabrication of an N P N silicon B J T on an integrated circuit
using t h e processes that we discussed in C h a p t e r 6. While our discussion will be
focused on the BJT, it is u n d e r s t o o d that the surface of a whole wafer is being
processed.
The starting material is a b o r o n  d o p e d wafer on a very small area w h e r e we
will form a BJT. T h e base o n which t h e transistor is m a d e is k n o w n as the substrate
and its function is to act as t h e mechanical support for the device. The reason for the
use of a substrate for the N P N will be clarified w h e n the t e r m isolation is dis
cussed. This substrate has a resistivity of 310 ohmcm with a thickness b e t w e e n 250
and 400 for wafers having diameters over 100mm.
The fabrication steps u p to, and including, the metal contacts to the three
regions are illustrated in Fig. 8.2.
First, a layer of S i 0 , about 5000A thick, is deposited on the surface of the sub
strate by thermal oxidation. Using the first mask and the p h o t o lithographic process,
windows are o p e n e d in t h e oxide for the buried layer. The N buried layer is dif
fused to a d e p t h of about 3, followed by oxide removal. This layer serves to col
lect the carriers that have crossed the base on their way to t h e collector terminal, as
shown in Fig. 8.1(c). It serves as a subcollector and is used to reduce the collector
ohmic resistance, as we will explain later.
After the buried layer is diffused, the wafer is stripped of all oxide to permit
t h e next deposition. It is to b e noted that during the subsequent hightemperature
processes, the buried layer tends to diffuse out.
The second operation is the deposition of a p h o s p h o r u s  d o p e d Nepitaxial
layer on the whole wafer and in which all devices are m a d e . This layer has a resistiv
ity of 0.1 to 1 ohmcm, with a thickness of 0.5 to 5 for highspeed digital circuit
applications and 1020 for linear analog circuits. A layer of silicon dioxide about
5000 to , thick is grown thermally on t h e surface of the epilayer.
Since the collector of the N P N is type, and so are the collectors of adjacent
transistors, there is an obvious n e e d to isolate t h e collectors from each other. This is
accomplished by what is k n o w n as isolation. To start with, each transistor is placed
in an island, or tub, as shown in Fig. 8.3, with a Pisolation region enveloping each
t u b and extending from the surface of the wafer all the way to t h e substrate. To
guarantee isolation of N P N translators, the substrate is connected to the most nega
tive point on the circuit, so that reversebiased junctions are formed b e t w e e n collec
tors of adjacent transistors, since the collectors are normally at a positive potential.
The second mask is used to etch windows for the isolation regions, which are
formed by the subsequent diffusion of b o r o n extending from the surface down to
t h e substrate. A n Nepitaxial layer separates the isolation regions, thus serving as
t h e tub in which each transistor is formed. The diffusion of the isolation region is
followed by oxidation of the wafer surface.
2
Figure 8.2
Section 8.2
217
///////////W//M //1kY//////////////////////////
TV
Figure 8.3
Transistor tubs.
Mask number three is used to o p e n a window for the Ptype base of the transis
tor. Ptype diffusion or ion implantation is driven to form the base to a d e p t h of
about 2  3 . This is followed by an oxide layer.
The fourth mask is used to o p e n windows in the oxide for the N emitter and
the collector contacts. T h e p h o s p h o r u s or arsenic diffusion is driven to a d e p t h of
about 2. T h e n e e d for an N collector contact is to form a good ohmic contact.
A s we will see in t h e discussion of metalsemiconductor contacts, these contacts
may b e rectifying or ohmic. The ohmic contact permits easy current flow in both
directions. To form a good ohmic contact to an material, an N region is n e e d e d
b e t w e e n t h e metal on t o p and the region. Following t h e N diffusion, an oxida
tion layer is formed over the entire wafer surface.
Mask number five is used to o p e n windows for the formation of metallic con
tacts to t h e transistor terminals. Then, an aluminum thin film 0.5 to thick is
deposited, by evaporation or sputtering, on t h e t o p surface of the I C wafer.
Assuming that complete circuits are to b e formed on the surface of the wafer,
the sixth mask is used to define t h e interconnection
pattern in the circuits. These
interconnections are etched into t h e metal that has b e e n deposited on the surface.
To protect t h e surface of t h e wafer from moisture and chemical contamina
tion, a passivation layer o n t h e surface is required. The material is a phosphorusd o p e d oxide, which is deposited o n t h e surface.
Contacts to the integrated circuits are m a d e on pads that are located on the
periphery of the I C chip. Since the I C chip will b e b o n d e d to an I C package, connec
tions are to b e m a d e from the package leads to the bonding pads on the I C chip.
Mask number seven is used to define the bonding holes over t h e aluminum pads for
external connections.
Following t h e seven masks, the circuits are tested by a computercontrolled
system and all faulty chips are identified and m a r k e d . The wafer is t h e n sawed into
chips, which are b o n d e d on I C packages. G o l d wires about 25 in diameter are
used to connect the package leads to t h e bonding pads on t h e chip.
Additional detail on the buried layer is illustrated in Fig. 8.4. In Fig. 8.4(a), we
show the p a t h t a k e n by t h e carriers on their way from t h e emitter to the base and to
the collector. This p a t h is considerably longer than the p a t h in a discrete BJT, shown
in Fig. 8.16, and because of this the collector series resistance, labeled the parasitic
resistance, is quite large and is of the order of h u n d r e d s of ohms. To reduce this
+
218
Chapter 8
(b)
Figure 8.4 Path taken by electrons from emitter to collector (a) without a buried
layer through Nepilayer and (b) with a low resistivity buried layer.
resistance, we have placed t h e low resistivity buried layer in Fig. 8.5(b) into the path
of the carriers, which acts as a subcollector. The use of the buried layer reduces the
collector resistance by as m u c h as a factor of 20. The result of this is to improve the
quality of t h e transistor, defined by a t e r m k n o w n as t h e figure of merit, or gainbandwidth product, by the same factor.
8.3 T E R M I N O L O G Y , S Y M B O L S , A N D R E G I O N S O F O P E R A T I O N
Terminology a n d S y m b o l s
The symbols for a P N P and an N P N transistor are shown in Fig. 8.5. The arrow on
the emitter lead identifies the actual direction of t h e current. In a P N P transistor,
holes are normally m a d e to move from the emitter to the collector, hence the arrow
on t h e emitter terminal points into t h e lead labeled E. The arrow on the N P N emit
ter lead points outward, indicating that t h e actual direction of t h e current is out of
the emitter as a result of the flow of electrons from emitter to collector.
In this book, we will use t h e P N P structure in the analysis of t h e operation and
in the d e v e l o p m e n t of t h e analytical relations for t h e BJT. The main reason for this
is that the P N P follows directly from the P N junction diode discussions and the
equations pertaining to it developed in the previous two chapters. To apply the
information of this chapter to t h e N P N transistor, t h e following is n e e d e d : First,
replace the references to carriers such that the terms, holes and electrons, and their
Section 8.3
Figure 8.5
219
(a)
(b)
relevant constants are replaced by electrons and holes and their constants respec
tively. Second, reverse polarities of all voltages and the directions of all currents.
M o d e s of O p e r a t i o n
The bipolar junction transistor has two P N junctions: t h e emitterbase and collectorbase junctions. T h e r e are, therefore, four possible combinations of biasing of these
junctions, as shown in Fig. 8.6: b o t h forwardbiased junctions, b o t h reversebiased
junctions, and o n e forwardbiased with a n o t h e r reversebiased junction.
Different sets of currents of the transistor will result for each set of biasing.
These combinations of currents and voltages, corresponding to each q u a d r a n t of
Fig. 8.6, will be labeled as modes of operation.These
m o d e s are named: active, satu
ration, cutoff, and inverse active.
The active mode is most commonly used in analog circuits. This corresponds to
the fourth q u a d r a n t of Fig. 8.6, where t h e emitterbase junction is forwardbiased
Inverse active
Saturation
Cutoff
Active
Figure 8.6 Modes of Operation for a
PNP transistor. For NPN transistor, V
replaces V and V replaces V .
EB
BC
CB
B1
220
Chapter 8
CB
BE
BC
8.4 CIRCUIT A R R A N G E M E N T S
In circuit applications, the transistor is connected so that it presents to the rest of the
circuit, two input terminals and two output terminals. This is d o n e by having one of
t h e three transistor terminals c o m m o n to the input and output circuits. A s a result,
there are t h r e e possible circuit connections for the bipolar transistor, as shown in
Fig. 8.7. They are the commonbase,
commonemitter,
and commoncollector
con
nections. T h e most commonly used connection is the commonemitter type, with t h e
commoncollector being the one least commonly used. In Fig. 8.7, actual directions
of currents are shown.
8.5 T R A N S I S T O R C U R R E N T S IN T H E A C T I V E R E G I O N
We will use the twodimensional representation of the Fig. 8.1(b), r e p e a t e d in Fig.
8.8, to analyze t h e sources of the currents in the P N P bipolar junction transistor.
There are three terminal currents and t h r e e voltages so that we may write
Kirchhoff s current and voltage laws as
I = I
E
+ I
(a)
Section 8.5
221
+
V
EB
"Si
EB
CB
(a)
(c)
(*>)
Figure 8.7 (a) Commonemitter, (b) commonbase, and (c) commoncollector transistor
connection for a PNP transistor. For an NPN transistor, all currents are in opposite directions
to those shown above and the subscripts to all voltage are interchanged.
E B
+ V
B C
+ V
C E
(b)
(8.1)
CB
EB
BE
CB
BC
CE
EC
Emitter Current
For a forwardbiased emitterbase junction, and just as we d e t e r m i n e d in the P N
junction diode, holes injected from the emitter diffuse into t h e base and electrons
injected from the base diffuse into the emitter region. The sum of these two currents
forms the emitter current I , as
E
= h
(82)
+ hn
where I represents the hole current c o m p o n e n t and I the electron current com
p o n e n t of the emitter current.
The electrons that cross from base to emitter and that form I
recombine
with holes in t h e emitter. For typical devices, the electron density at the metal emit
ter contact is the thermal equilibrium value, as shown in Fig. 8.8(b). A t the emitter
metallic contact, electrons are forced out of the contact by the D C voltage, V , into
the external circuit, thus freeing excess holes. Some of these holes, which are major
ity carriers in the emitter, recombine with the electrons arriving from t h e base.
E
En
En
BB
222
Chapter 8
Figure 8.8 (a) Twodimensional model of the PNP transistor in the active region
and (b) profiles of minority carrier distributions.
T h e rest are injected into the base as a result of the forwardbiased emitterbase
junction.
Collector Current
The holes that are injected into the base from t h e emitter face either of two possibil
ities: one, if the width of the base is m u c h greater than the diffusion length of holes
(which d e p e n d s o n t h e lifetime of holes as minority carriers in the base), t h e n all the
holes will recombine in the base. If, on the other hand, the width of the base is m u c h
smaller than the diffusion length of holes, then t h e great majority of the holes reach
the reversebiased collector junction. The former condition m a k e s the transistor two
P N junction diodes back to back, since n o interaction between t h e emitter and col
lector takes place. T h e latter conditions m a k e it possible for the holes that origi
n a t e d in the emitter to cross into the collector and contribute to t h e collector
current. W h y will t h e holes cross into the collector?
Section 8.5
Emitter
Base
223
Collector
(b)
Figure 8.9 Energylevel diagrams for electrons in the PNP transistor in (a)
thermal equilibrium (zero bias) and (b) active region. Because the diagrams are
drawn for electrons, electrons find it easy to roll down the steep hill.
Simultaneously, holes find it just as easy to bubble up the same steep hill. It is
assumed that in diagram (b) V (V )
is of the order of 0.5V and V ( V ) is
of the order of 5V.
EB
BE
CE
The answer is found in the energy level diagrams shown in Fig. 8.9, where
u n d e r reversebias the electric field in the collectorbase junction provides an easy
p a t h for minority carriers (for holes to cross from the base into the collector and
electrons to cross from collector to base).
The holes originated in the emitter and diffused in t h e base to b e injected into
the collector. The reverse bias across t h e collectorbase depletion layer has
decreased the density of electrons and holes to values that are lower than their ther
mal equilibrium. H e n c e , t h e r m a l generation exceeds recombination and electrons
224
Chapter 8
and holes are generated in the layer. The holes join those arriving from the base and
cross into the collector.
The electrons generated in the collectorbase depletion layer roll d o w n the
steep hill from collector to base, as shown in Fig. 8.9(b). This current of electrons,
which is of the o r d e r of picoamperes w h e n the collector current is in m A , is also
k n o w n as the leakage current.
We n o w have a collector current m a d e of two components where I is the
current of holes bubbling u p the hill from base to collector and I is the current of
the electrons that roll d o w n t h e steep hill from collector to base
c
Cn
Ic = Ic
(83)
Base Current
The base current, I , consists of three componentsall shown in Fig. 8.10.
The first component, I , is the l
of the emitter current, which consists of
electrons that diffuse from base to emitter, and which is directed out of the base ter
minal. The second c o m p o n e n t , I = I , is p r o d u c e d by the motion of electrons
from t h e base lead that cross into t h e base in o r d e r to recombine with some of the
holes that are diffusing from the emitter to the collector. The direction of this cur
rent is also out of the base terminal. The current of electrons from collector to base,
I , represents t h e third c o m p o n e n t of the base current, so that the total base cur
rent becomes
B
B1
En
B2
tec
h = hx + hi ~ B3
J
Figure 8.10
(84)
Section 8.6
225
In reviewing this section, we have not highlighted the reason for our interest in
this transistor. This is the subject of our next section.
8.6 T H E B J T A S A C U R R E N T A M P L I F I E R
A p p r o x i m a t i o n s t o Base Current
By making assumptions that are quite valid in modernday silicon BJTs, we will
d e m o n s t r a t e that a B J T in the commonemitter connection is in fact a linear current
amplifier.
We have concluded earlier that carrier motion in a BJT, which is operating in
the active m o d e , results in five distinct currents. For the P N P transistor t h e r e is, first,
a current due to holes that diffuse from the emitter to the base, which m a k e s up the
overwhelming part of the emitter current. Second, t h e r e is a current of holes that
diffuse in the base and this current m a k e s u p practically all of the collector current.
Lastly, there are t h r e e c o m p o n e n t s of the base current: the current due to electrons
injected from the base to t h e emitter labeled I ; the current I , which results from
electrons supplied by the base contact to recombine with some of the holes that are
traversing t h e base to the collector; and the current I , which consists of a genera
tion current originating at the reversebiased collectorbase junction (which sends
electrons into t h e base). Simultaneous to the generation electrons going into the
base, generation holes are driven by the electric field into the collector. However,
this current due to holes is negligible c o m p a r e d to the current of holes that have dif
fused into the base and are collected by the collector. The current of electrons, I
may b e l u m p e d partly with the electrons that recombine in the base and partly with
the electrons that are injected into t h e emitter.
We have therefore reduced t h e base current to two currents: one of electrons
injected from the base that diffuse in the emitter and the second a current of elec
trons that recombine with holes in t h e base.
In the BJTs of the 1960s, the recombinations c o m p o n e n t of the base current
was considered to m a k e u p the larger portion of t h e base current because of t h e
combination of large basewidths, of the order of 10 microns, and the low lifetime of
minority carriers in t h e base. In today's transistors with basewidth of less than one
micron and with long lifetimes, t h e recombination current can b e safely neglected
for a very b r o a d range of transistors.
Therefore, we have limited the effective base current to the current of elec
trons that are injected into the emitter. While this is valid for t h e purpose of our dis
cussion here, in a later section we will derive complete expressions for I , I , and I ,
in terms of the physical properties of the transistor.
m
B2
B3
By
226
Chapter 8
Let us establish why we label I as the control current. We can safely assume
that in the great majority of transistors, t h e width of the base is much smaller than
the diffusion length of holes in the base of the P N P device, so that the P N emitterbase junction has properties that are similar to the shortbase diode discussed in the
last chapter. Therefore, the profile of the hole density distribution in the base is a
straight line having slope p(0)/W ,
so that the hole current at the collector junc
tion in a C E device is given by
m
= {qA D )dp/dx
E
= qA p(0)/W
Ep
(8.5)
pPa
(8.6)
EB
where A is the emitter junction a r e a , p is the equilibrium hole density in the base,
p ( 0 ) is the hole density in the base at the emitterjunction, W is the effective width
of the base, and D is the diffusion constant of holes in the base.
E
nE
nE
nE
(qA D njL )
E
nE
nE
exp (qVJkT)
nE
(8.7)
where n (0) is the electron density at the emitter side of the junction, n is the equi
librium electron density in the emitter, L is the diffusion length of electrons in the
emitter, D is the diffusion constant of electrons in t h e emitter, and A is the emit
ter area.
F r o m Eqs. (8.6) and (8.7), we conclude that by fixing the base (input) current
to a C E B J T circuit, V is fixed. If V is fixed, t h e collector current, I , is also fixed,
so that t h e ratio of I to I is given by
E
QE
nE
nE
EB
EB
h / h
T^f"
(88)
EB
BE
EB
EB
EB
EB
EB
EB
BB
Section 8.7
Transistor Parameters
BE
227
BB
EB
8.7 T R A N S I S T O R P A R A M E T E R S
The terminal currents of the P N P transistor operating in the active m o d e are sum
marized below as
I
= C
= hn
En
E
I
()
+ Cn = C
Kec ~
+ CBO
hBO
00
(89)
C
()
We will identify t h e t e r m I
by its subscripts as follows: T h e first two sub
scripts refer to the two terminals b e t w e e n which the current is m e a s u r e d and the
third subscript refers to the state of t h e third terminal ( O for o p e n ) . The current,
l
, is therefore t h e collectortobase current with the emitter open. It includes the
electrons and holes that are generated in t h e depletion region and swept into the
base and the collector respectively.
CB0
C B 0
228
Chapter 8
is
CBO
REC
(810)
= = I /I
7
CP
(8.11)
common
(8.12)
It is obvious that in a good BJT, t h e value of a is very close t o unity and varies
usually from 0.99 t o 1.
By incorporating E q . (8.12) in E q . (8.9(b)), we have
Ic = *I
+ ICBO
(813)
that
a
I CBO
= z
I + f ^
1 a
1 a
where a n d I
(8.14)
C E O
, so that E q . (8.14) b e c o m e s
1 5
+ ICEO
(8 )
a r e given by
CE0
= /(1  )
(a)
(8.16)
ICEO
IcBo/i
~ 0
()
The symbol I
refers to the collectortoemitter current with the base opencircuited. For I = 0, I = I This current is larger than I
because t h e electrons
generated at t h e reversebiased CB junction are swept into t h e base w h e r e u p o n
they diffuse into t h e emitter causing a larger diffusion of holes from t h e emitter into
the base to p r o c e e d to t h e collector. It is as if these holes represent a magnified cur
rent c o m p a r e d t o t h e electron current. T h e E  B junction is slightly forwardbiased
by V , causing t h e diffusion of electrons a n d holes. Since is almost unity, is a
large n u m b e r k n o w n as t h e DC commonemitter
current gain. This number, for a
good BJT, is 100 or more. T h e relations in Eqs. (8.1) through (8.16) apply as well t o
the N P N transistor provided that actual directions for currents a r e used.
In t h e following example, we will carry out calculations for transistor currents
and transistor parameters.
CEO
EC
CEO
CBO
Section 8.8
229
E X A M P L E 8.1
Given a PNP transistor that has the following current components: l = 2mA, l 0.01mA,
I = 1.98mA, and I = 0.001mA, determine: a) the base transport factor, b) the injection effi
ciency, c) and , d) I , l , and I , and e) repeat part (c) for 7 , = 1,99mA.
Ep
Bn
Cn
CBO
CEO
Solution
a) The base transport factor is
= I /I
Cp
= 1.98/2 = 0.99
Ep
= hJihj
+Q
= /(
01
+  ) = 0.995
c) a = 7 = 0.985
= /(1  ) = 65.67
D
>
"
En
CBO
'CEO
c~
115
8.8 G R A P H I C A L C H A R A C T E R I S T I C S A N D M O D E S O F O P E R A T I O N
In Fig. 8.12, we display t h e o u t p u t graphical characteristics for the P N P B J T
together with the relevant circuits of the connections. These are the characteristics
of interest w h e n the B J T is used in amplification or switching. C o m p l e t e analytical
dependencies for both t h e input and output variables in the P N P c o m m o n  e m i t t e r
and commonbase connections are represented by the following relations:
C E input, V
= f{J , y*c)
FB
o u t p u t , I = f(l ,
c
CB input, V
EB
V )
EC
= f(I ,
V )
BC
(a)
(b)
( S
'
(a)
(8.18)
output, 1 = f{J V )
C
BC
(b)
230
Chapter 8
Cutoff
CB0
Figure 8.12 (a) Commonemitter circuit and output characteristics of the PNP transistor
(b) Commonbase circuit and output characteristics of the PNP transistor.
Section 8.8
231
CB
Inverse active
Saturation
Cutoff
Active
Figure 8.13 Distributions of minority carriers in the four operation modes of the
PNP BJT.
EB
(a)
(8.19)
= n
CB
0 E
(b)
(a)
(b)
EB
(8.20)
In the collector at x" = 0, n ( 0 ) = n
c
QC
CB
EB
232
Chapter 8
this is equal to the collector current, l . A s long as t h e E  B junction is forwardbiased and t h e CB junction is reverse biased (V
< 0 or, V > 0), I and I are
given approximately by Eqs. (8.6) and (8.7). For a constant I , I is fixed so long as
V
has a value that makes V > 0, as d e t e r m i n e d by
Q
CB
BC
EC
BC
EC
= Veb + V
BC
= V
 V
(8.21)
EB
BC
BC
BC
CE S a t u r a t i o n M o d e
In t h e saturation m o d e , t h e CB junction is forwardbiased causing an increase in
the hole density at = W , which for a fixed I and V
reduces t h e slope of t h e
hole density distribution, thus decreasing t h e collector current, as shown in Fig.
8.12(a). A s V
is decreased further, for a fixed I , V
becomes m o r e positive
(more forward bias o n CB junction) and t h e slope of t h e hole density in the base
decreases, thus I decreases.
A t larger values of I , h e n c e larger values of V , t h e dividing line between the
active and saturation modes, which occurs at V
= 0, moves t o lower values of V .
This m o d e represents t h e O N position w h e n t h e B J T is used as a switch.
B
EC
EB
CB
EB
BC
EC
CE Cutoff M o d e
W h e n b o t h junctions a r e reversebiased, t h e base current b e c o m e s negative. We
consider first t h e condition at t h e b o u n d a r y b e t w e e n positive a n d negative I , at
I = 0. With t h e applied V , t h e emitter is positive with respect t o t h e collector. In
this condition, V is positive, t h e CB junction is reversebiased, a n d electrons are
injected from collector to base. Since t h e electrons cannot exit through t h e base,
they neutralize some of t h e d o n o r ions at t h e E  B junction, attracting holes from t h e
emitter in o r d e r to neutralize acceptor ions. The emitterbase junction is, therefore,
slightly forwardbiased, causing a gradient of holes in t h e base that diffuse t o t h e
collector. Because of t h e high doping of t h e emitter, t h e relatively smaller n u m b e r
of electrons that left t h e collector cause a larger n u m b e r of holes to b e injected into
the base and diffuse to t h e collector. This collector current, which is an amplified
form of t h e electron current, is labeled I  T h e collector t o emitter current with
the base o p e n is t h e collector current at I = 0 a n d represents t h e edge of cutoff.
For I < I , b o t h junctions are reverse biased and t h e collector current is
very small. This m o d e represents t h e O F F position w h e n t h e transistor is used as a
switch.
B
EC
BC
CEO
CEO
CE Inverse A c t i v e M o d e
The inverse active m o d e occurs w h e n t h e emitterbase junction is reversebiased
and t h e collectorbase junction is forwardbiased. T h e roles of t h e emitter and col
lector are thus interchanged, with the former collector emitting t h e holes (in the
P N P B J T ) a n d t h e former emitter ( n e w collector) collecting t h e m . T h e general
Section 8.8
233
(inverse active) =
AED poL
AD n W
p
nC
0C
(8.22)
nC
AE
AC
QE
CB A c t i v e M o d e
In the C B active m o d e , I is the relevant input current and is m u c h larger t h a n the
input current to the C E mode. The base current consists of electrons injected from
the base into t h e emitter and the collector current consists of holes diffusing from
the emitter into the base. T h e portion of t h e emitter current due to holes diffusing
in t h e base is almost equal to the collector current, as shown in t h e characteristics of
Fig. 8.12(b). A n increase in I brought about by an increase in t h e slope of the hole
density in t h e base results, for a fixed V , from an increase of I . B o t h of these cor
respond to a linear increase in I that is almost equal that of I .
A t higher values of I , for a fixed V , a higher I is required, which results in a
larger I with l = I .
E
BC
BC
CB S a t u r a t i o n M o d e
To the left of V
= 0, the collector junction is forwardbiased and the transistor is
in saturation [V
< 0). H o l e s are injected from the collector to the base and these
are, for (V
> 0, I > 0) in a direction that is opposite to the direction of holes that
originate in the emitter and end u p in the collector. A s a result, the collector current
is t h e difference b e t w e e n t h e two hole currents crossing t h e collector junction and is
smaller as V b e c o m e s m o r e negative. A t this time, the emitter current is constant.
The larger the forward bias on the CB junction, the larger the injection of holes
from collector to base. For the same I , I is smaller. A t larger values of I , which
indicates m o r e holes injected from base to collector, and to obtain t h e same I , m o r e
holes have to be injected from collector to base, requiring a larger forward bias on
the CB junction (larger V ).
BC
BC
EB
BC
CB
CB Cutoff
In the CB cutoff m o d e , I = 0 represents the edge of cutoff. The collector current is
labeled I  With the emitter opencircuited and V
> 0, I becomes the reverse
saturation current of the CB junction. This current is smaller t h a n the reverse cur
rent of a P N diode because for I = 0, t h e hole gradient in t h e base at t h e emitter
E
CBO
BC
234
Chapter 8
EB
REVIEW QUESTIONS
Q81 Identify the junction voltages and their signs in each of the four modes of operation for
an NPN transistor.
Q82 Identify the components of each of the three terminal currents for an NPN transistor.
Q83 What is meant by emitter efficiency?
Q84 What is meant by base transport factor?
Q8S State, in equation form, the relationship between the three voltages for a PNP transis
tor.
Q86 Why is the collector current in a PNP transistor in the inverse active region much
smaller than in the active region?
Q87 Why does the collector current decrease when the transistor is operated in the satura
tion region?
Q88 What is the function of the buried layer?
Q89 Why are the devices placed in tubs?
HIGHLIGHTS
CE
CB
EXERCISES
E81 The following components of the currents have been determined as: I = 2.712 x
10" A, I = 0.678mA, I = 9.4 X 10 A, and I = 0.6779mA.
Determine: a) the injection efficiency, b) the transport factor, c) alpha, d) beta, and e)
En
15
Ep
Cn
Cp
^CEO
Section 8.9
235
Lowlevel injection.
T h e widths of the emitter and collector regions are m u c h greater t h a n the dif
fusion length of minority carriers so that the minority carrier densities have
their equilibrium values at t h e contacts.
The collector area is m u c h larger t h a n t h e emitter area so as to collect all
holes crossing t h e collector junction.
For the P N P cross section shown in Fig. 8.14, we will use t h e symbols shown
below the figure. The p r o c e d u r e we will follow in determining expressions for the
currents in terms of V
and V
is identical to the one we used for t h e diode in
C h a p t e r 5. Again, we d e t e r m i n e expressions for t h e minority carrier distributions in
the emitter, base, and collector. Using these distributions, we will then derive
expressions for the emitter and collector currents, as d e t e r m i n e d from the minority
carrier currents at b o t h ends of each depletion layer.
We draw t h e attention of the reader to t h e use of p r i m e d symbols to refer to
the excess density, over that of equilibrium, of minority carriers. We also n o t e the
zero axes for distances are = 0 in t h e base, x' = 0 for the emitter, and " = 0 for
the collector.
EB
CB
Emitter Current
Just as in t h e case of the diode, we will determine the expression for the emitter cur
rent from the distributions of t h e minority carrier currents at the edges of t h e emit
terbase depletion layer.
We start with the continuity equation for holes in the steadystate in t h e base
of a P N P transistor, given by E q . (4.36), in one dimension, as
_ P ^ P o _ l ^
q dx
where is t h e hole density anywhere in the base, p is the equilibrium hole density
in the base, is the lifetime of holes in the base, and J is t h e hole current density
0
236
Chapter 8
Emitter
Collector
Base
For
r electrons
in
Dp
aE
For
> holes
in
For
electrons
in
D aC
(a)
x' = 0
=0
x" = 0
(b)
Figure 8.14 (a) Section of transistor together with notation for minority carriers, (b)
Minority carrier distributions for active mode operation. The vertical scale represents the
relevant carrier and the subscripts and C refer to the emitter and collector respectively
where the hole density in the base is represented without subscript. The equilibrium values in
the emitter and collector are n and , and p in the base.
QE
o c
d p'
2 dx
2
2
L
= / ,
so t h a t
(825)
p> = B e
1
x/L
+ B e2
(8.26)
T h e lowlevel b o u n d a r y conditions a r e
p'(0)=p [ex (qV /kT)l]
0
EB
(a)
Section 8.9
P'(W )
[exp (qVjkT)
P o
 1]
(b)
237
(8.27)
=Po+
exp
Po
sinh (W
 1
{ kT
sinh
x)/L
(W /L )
B
(8.28)
+ p [exp (qV /kT)
0
smh(x/L
 l]
CB
forO
nnh{W /Lp)
B
is found from
(~qD
Ep
A)(dp/dx)
at = 0
so that
l
Ep
[ ^ (qV /kT)
 l][coth
EB
(W /L)]
B
(8.29)
 l]/[sinh ( W / L ) ] j
CB
f l
d n'
n'
w h e r e n' = n n
and n , as shown in Fig. 8.11, is the equilibrium value of the
electron density in the emitter. The emitter width is assumed to b e much greater
than L so that the b o u n d a r y conditions are:
E
QE
OE
nE
exp
OE
(1 EB\
1 , at x'
kT
n' = 0 at x'
E
=0
(a)
(8.31)
(b)
= OO
n' = ()
e x
(8.32)
(x'/L )
nE
( a t %
'
0 )
 l]
(8.33)
 1]
EB
^nE
+
hp
h
Ln
= [(q A 7 ) / L ) ( c o t h (W /L ))
A
+ (qA D njL )}
nE
CB
nE
 1]
EB
(8.34)
238
Chapter 8
Collector Current
T h e collector current is m a d e of two components: t h e hole current crossing t h e col
lector junction from base t o collector I a n d t h e current of t h e electrons that cross
this junction from collector to base I . Therefore, I = I
+ I , where I
is
d e t e r m i n e d by using t h e hole density distribution in t h e base at t h e edge of t h e CB
junction from E q . (8.28).
cp
CN
I =
q A D
Cp
dp
~dx
cp
CN
CP
Wn
at
(W /L.)
B
 \exp(q V /kf)
 1] coth (W /L )
CB
(8.35)
d n'
(8.36)
dx'ji2
w h e r e n' = n  n
and n
is t h e equilibrium electron density in t h e collector.
The b o u n d a r y conditions for t h e electron distribution in t h e collector are
c
= n
[exp(q V /kf)
 l ] at x" = 0
CB
(a)
n^ = 0 a t x " = OO
(8.37)
(b)
(8.38)
nC
Cn
nC
= 0) =
oc
L nC
[exp (q V /kf)
CB
 1] (8.39)
c =
D po
p
L inh(W /L )
pS
[cxp(q V /kf)
^^ oth(W /L )
C
 1]
EB
qA
+
f"
cnoc
[exp(q V /kT)
CB
 l]
(8.40)
'c
(841)
Section 8.9
Figure 8.15
239
third current equal to t h e difference b e t w e e n the hole current crossing the emitter
junction and the hole current crossing t h e collector junction. The third t e r m is the
recombination current in the base and results from the electrons that t h e base lead
must supply to r e c o m b i n e with holes. The transistor currents and their c o m p o n e n t s
are shown in Fig. 8.15.
It is i m p o r t a n t to emphasize that, subject to t h e assumptions that were m a d e
at t h e beginning of the section Assumptions
and Procedure, the equations that we
have derived for t h e transistor currents are applicable in all four regions of opera
tion. Later in this chapter, we will simplify these relations and illustrate their appli
cations through examples.
We show in Fig. 8.16 the distributions of minority carriers in the emitter, base,
and collector for t h e active, saturation, and cutoff m o d e .
In a subsequent section, we will further investigate the recombination current
in t h e base.
So far, the relations derived for the currents in the P N P transistor can b e m a d e
applicable to t h e N P N transistor w h e n t h e following changes are m a d e :
REPLACE
D by D ,p
p
nE
D
P
by D
n c
EB
0E
W '
by p ,
pa
by yBE
by n , and L by L
oc
a n d
oc
CB
by
nE
and L
nC
by
pE
pC
BC
We will n o w use Eqs. (8.34) and (8.40) to calculate t h e values of currents and
their corresponding components. We have two objectives. First, we present the
reader with an o r d e r of m a g n i t u d e of the values of the terms in t h e equations that
we have derived. Second, we want to highlight t h e effect of a decrease of t h e doping
of t h e emitter, and later, of an increase in t h e width of the base, on the current gain
of the transistor, and on the magnitudes of the currents.
EXAMPLE 8.2
To illustrate the application of the relations derived so far, we consider a PNP silicon transistor
having N = 10 cm~\
= 10%~ , and N = 10 cm~ . Assume an effective cross section
17
AE
15
AC
240
Chapter 8
0
(b)
x'
(c)
Figure 8.16 Distribution of minority carriers in the bulk of the PNP transistor for
(a) active, (b) saturation, and (c) cutoff operation. The width of the base is assumed
here to remain constant at W in all modes of operation.
B
area A = 10~ cm , a base width W = , and minority carrier lifetimes in each of three regions
at = 10~ s.
For active mode operation, and at V = 0.63V, calculate:
a) I , 7 , a n d 7
6
EB
& )
b) I ,I ,
and 7
c) The three components of /
d) The beta of the transistor
Cp
Cn
20
Solution We determine the minority carrier densities in the three regions by using nj = 10 cm~
for silicon:
n
oE
10 cm~ ,
= 10 cm~ ,
and
= 10 cm"
The mobilities are calculated from the relations in Eqs. (4.15) and (4.16), the diffusion con
stants are then determined using the Finstein relation and the diffusion lengths are calculated using
L = (7JT) '
1 2
,
D
L
nE
:/
= 826cm /Vsec,
= 21.4cm /sec,
= 4.62 x 10~ cm,
2
, = 447cm /Vsec,
D  11.6cm /sec,
L = 3.4 X 10^ cm,
2
nC
Section 8.9
241
The factors multiplying the exponentials in Eqs. (8.34) and (8.40) are calculated to be
q
 ^ O E
10"
For W /L
B
4 1
'^
.\
 ( )
4 5
,,
 0
A ;
 ( )
, ,
= 0.0294117:
sinh
= 34.0099
10
EB
cg
and I
respectively, we have
Pn
= .678mA
= 2.712 X 10 '
Ep
En
to be neg
/ , . = / , , + /,, = 0.681mA
b) The components of the collector current, I , are calculated by using Eqs. (8.35) and (8.39) as
c
L . = 0.678mA
Cp
15
= 9.4 X 10" A
Cn
/,.. = 0.678mA
c) The three components of the base current are:
I
=I = 2.712 X 10 "A
15
Cn
= h i
+ B2
!
"
d) = f = 225.7
It is important to compare the magnitudes of the three components of the base current. It is
obvious that the collectorbase leakage current, / , i s negligibly small. Furthermore, the recombina
tion current, for the ratio of W to L used is about one tenth of the base current. This is quite true
in modern silicon transistors where the dominant component of the base current is the one result
ing from majority carrier electrons crossing from base to emitter.
B3
E X A M P L E 8.3
Use the data of Example 8.2 except replace the doping of the emitter by N
to calculate the value of beta.
AE
16
= 5 X 10 cm"
Solution Assuming that the lifetime of minority carriers is unchanged, the relevant terms for
minority carriers in the emitter become
n
OE
= 2 X 10 cnr ,
= 1016cm /Vs, D
tlE
= 26.3cm /s
242
Chapter 8
nE
= 5.13 X l O ^ c m a n d
16
= 1.6405 X 10" A
I is unchanged at 0.678mA
c
107.5
We observe the dramatic decrease of beta as the emitter doping was reduced. The reduction of
the doping increased the equilibrium density of electrons in the emitter and consequently
increased I . The increase of I served to increase I and I but had no effect on the value of I .
En
EXAMPLE
En
8.4
Repeat parts (a) and (b) of Example 8.2 except for a base width W = 2.
B
The value of
= 0.0588574, coth (W /L )
B
= 17.01962
a) By using the above relations and the terms calculated in Example 8.2, we have
I = 0.342mA,
E
b) =
I = 0.3389mA
c
= 102.7
An increase in the width of the base caused a decrease in the slope of the hole density at the
base side of the depletion layer. This resulted in m decrease of both l and I and hence decreases
in the emitter and collector currents. Both the injection efficiency and the transport factor
decreased with a larger relative decrease in the injection efficiency. This caused a sizeable decrease
in the value of beta.
Ep
Cp
Relations f o r t h e N P N Transistor
By making the substitutions listed earlier in this section relevant, emitter and collec
tor current equations for the N P N transistor in the active region are determined
below and in accordance with t h e reference, as shown in Fig. 8.17. Applications of
these equations are illustrated in Example 8.4.
Relations for t h e N P N transistor corresponding to the P N P device equations,
given by Eqs. (8.29), (8.33), (8.35), and (8.39), are shown below
Section 8.9
Electron motion
emitter
243
Hole motion
base
collector
'Bp
DpE
DPC
Figure 8.17 Reference currents in NPN BJT and symbols identifying minority carriers in the
three regions.
qAD
L
in at  junction
BE
exp 
kT
coth
exp^f1
1
VBC
(8.42)
sinhf^
AD poE
pE
in at  junction
LpE
qAD p
pC
0C
in C at C  junction
= I
En
+ I
Ep
(8.43)
sinh^
Cn
in at C  junction
where I
q VBE
e x p  ^  l
and I
= I
Cn
pC
I .
Cp
kT
coth 
q VBC
e x p  1
(8.44)
(8.45)
244
Chapter 8
E X A M P L E 8.5
,7
16
15
AS
DC
20
BE
b) / . , /
C
Cp)
and/
HC
BE
d)
20
17
Emitter
3
cmcm /Vs
cm /s
cm
2
D
L
Base
3
OE
pE
p B
=W
= 350.5
= 9.078
= 3.013 "
15
and p
0 5
(DT).
oc
= ^/ , mobilities
Collector
4
= 10
, = 1258.3
D ~ 32.59
L = 5.7 x 10~
u
p
=
, =
D =
L =
oc
16
pC
p C
10
460
11.91
3.45 X 10~
\\
Cn
d) 
;)
s 1024
The high current gain obtained in this example illustrates the superiority of the NPN BJT. All
data in this example are the same as those used in the PNP J'!' of Hxample 8.2, yet . is more than
four times greater. This is mainly a result of the differences in the mobilities of the minority carriers
in the base.
;
R e c o m b i n a t i o n Current in t h e B a s e
We have indicated earlier that b o t h the c o m m o n  b a s e current gain, a, and the com
monemitter current gain, , are m a d e large by reducing the ratio of t h e width of the
base to the diffusion length of minority carriers in t h e base. Reduction of this ratio
decreases t h e recombination current in t h e base. Let us disgress briefly to investi
gate this by developing an expression for the hole density in the base of a P N P tran
sistor based on the condition for which alpha and beta were defined, namely
operation in the active m o d e .
O p e r a t i o n in the active region results w h e n V is positive and V is negative.
We assume that both have magnitudes that are m u c h greater than kT/q. Using
these conditions, Eq. (8.28) can be approximated by
KB
pMo)^
s i n h [ (
^V/f
sinh
W /L
B
CB
L p ]
46
t )
Section 8.9
245
Figure 8.18 Normalized hole density distribution in the base of a PNP transistor
as a function of normalized distance in the base for a diffusion length L = .
where p(0) is the hole density at = 0 in the base and is given by p exp (q
V /kT).
Calculations based on t h e above relation are plotted in Fig. 8.18. T h e ordinate
is the normalized hole density and t h e abscissa is normalized distance. Curves are
plotted for a diffusion length of 10 microns and various values of base widths.
It is quite evident from the plots that for (W /L
) 1 the distribution is
exponential, while for W L , the distribution approaches a straight line. This
can be concluded by setting W L and hence (W  x) L in E q . (8.46). The
first two terms of t h e Taylor series expansion of the hyperbolic sine are given as
0
EB
sinhw = u +
u /6
For a small u, sinhw can b e replaced by u so that t h e expression for the hole
density b e c o m e s
{WB
= P(0)
w
] ,Lp
= P(0)(1 
ix/Wj)
(8.47)
Cp
246
Chapter 8
Expressions f o r A l p h a a n d B e t a
To obtain expressions for a a n d in terms of the physical constants of the transis
tor, we will simplify Eqs. (8.29), (8.33), and (8.35) for operation in the active m o d e
(V
> 0 and V
< 0) as
EB
CB
qA
hp = Lr
p
uZ
/r/L))if
smh{W
B
qA
nEn
^ /
c o s h
( B/ P))
HI
8 48
( )
= f exv(qV /kT)
En
'" 4
ex
(8.49)
EB
"
V E
"
l k T )
c o s h
< 8
5 0 )
EB
= (qAD p(0)/L )
Ep
coth (W /L )
(8.51)
sinh W /L )
(8.52)
= q A D p(Q)/(L
Cp
EB
^ V^
S e c h
W P)
a n d assuming
2
ing W
IEP
j =
hp
hn
By substituting for I
a n d I from Eqs. (8.51) a n d (8.49), and rearranging
terms, the injection efficiency b e c o m e s
E
En
D L n
E
QE
tanh
DL p
p
nE
(8.54)
(W /L )
B
For (W /L)
1, tanh (W /L )
= W /L .
Using this simplification and
replacing the ratio of the equilibrium carrier densities, n /p , by t h e ratio of the
doping of the base t o the emitter, t h e injection efficiency b e c o m e s
B
QE
Section 8.9
247
(8.55)
DL N
p
nE
AE
_L
 f
nEW N
2\L j
DB
1_/WV2D W N
DL N
nE
nk
2\L )
AE
DB
( 8
5 6 )
DL N
nF
Ab
E X A M P L E 8.6
Use the data of Example 8.2 and Eq. (8.56) to calculate a value for .
Solution
The second term in the expression is calculated lo be 1/250 and the third term is 1/(250 X 2312)
so that
1
2312
250
250 X 2312
,. = 225
The value of calculated in Example 8.5, is a little higher than the average value of transis
tors used in current and voltage amplifiers operating in the commonemitter connection. The ques
tion is then, what are typical values of ^and what major parameters influence those values?
The discrete silicon transistors used in current and voltage amplifiers have a normal ^. range
from a low of approximately 75 to a high of 300. These are nominal values and the actual values
may vary over a range of plus or minus 20 percent of these. Their ability to dissipate power is quite
small and hence they are of small physical size. Their power dissipation is of the order of 100 milli
watts.
For applications in power amplifiers, where the AC output power is in watts or tens of watts,
transistors may have a beta of about 10 since their function is to transfer power rather than amplify
ing voltage. These transistors have a large size and large surface area since, in acting as agents of
power transfer from the DC source to the load, they are required to dissipate a sizable fraction of
the power delivered by the DC power.
For certain applications, such as in the input stage of an integratedcircuit, operational ampli
fier, transistors may have a beta of approximately 10,000. These are known as superbeta transistors.
For silicon transistors and as was shown by the results of Example 8.5, control over the design
of the beta is achieved through control of the ratio of the base width to the diffusion length of
minority carriers and control over the ratio of N to N .
AE
pB
248
Chapter 8
It is quite evident from the above calculation that the central control on the value of is
achieved by increasing the value of the injection efficiency given by Eq. (8.54). The advantage of a
small value of (W /L ) has been exhausted in modern day transistors by the ability to manufacture
devices whose base widths are a fraction of the diffusion length. Furthermore, these small base
widths have probably reached the limits of manufacturing technology.
8.10 E B E R S  M O L L M O D E L
T h e basic equations for the transistor currents, E q . (8.34) for I a n d E q . (8.40) for I ,
are general e n o u g h so as to apply for all four combinations of V a n d V and thus
applicable in all four regions of the characteristics. W e will n o w simplify Eqs. (8.34)
and (8.40) by assuming that W L .
The hyperbolic functions in these equations are expanded into series as
E
EB
CB
cosh u = 1 + / 2 + . . .
3
sinh u = u + u /6 + ...
For small values of u, they b e c o m e
cosh u = 1, sinh u = u, coth u l/u
Using the simplifications for the hyperbolic functions, Eqs. (8.34) and (8.40)
become
D
q
PP0
+ nE
0E
[exp (q V /kT)
 l]
EB
LnE
(a)
~
[exp (q V /kT)
 1]
CB
gA
S^
c =
[exp
(q V /kT)
 l]
EB
oc [exp (q V /kT)
q A
(b)
 l]
CB
W,
(8.57)
L nC
EF
= q A
CR
qA
pP0
nE 0E
[exp (q V /kT)
EB
DpPo
nC "PC
V /kT)
CB
1] (b)
LnC
DpPo
pPo
pPo
W
Dn
nE
(c)
LnE
[exp (q
 l ] (a)
pPo
, D nC
"OC
L nC
(d)
(8.58)
EbersMoll Model
Section 8.10
249
B=
CT
I
E ~ C
EF ~ R
I =a I
c
()
(b)
C "
p)
+ hR
"
(8.59)
R)
()
cs
ES F
= CS R
where I
=
D^jWg.
The EbersMoll equations for I
and (8.59) as
= SM
6 0
SM
ex
h = i s[ P
(q E /
V
and I
i] 
e x
c s t P (q c /
B
k T
 i]
()
(8.61)
I = a I [cx
c
ES
(q V /kT)
 1]  7 [ e x p (q VjkT)
EB
 1]
cs
(b)
We also define
=
ol /(1
 a)
(a)
= a /(l
 a)
(b)
(8.62)
By setting V
= 0 in Eqs. (8.61), the ratio of I to I b e c o m e s a and likewise
the ratio of I to I b e c o m e s We can, therefore, define
as t h e commonemitter
shortcircuit (B to C) current gain and a as t h e commonbase shortcircuit current
gain. It is to b e n o t e d that only three p a r a m e t e r s of t h e transistor are necessary and
sufficient to solve for t h e currents given the voltages. These three p a r a m e t e r s are ,
, and I . The p a r a m e t e r s ^. and are used to calculate a and a , and by using
I in E q . (8.60), I and I are available.
A circuit m o d e l for the transistor using E q u a t i o n s (8.59) and (8.61) is shown in
Fig. 8.19.
CB
SM
ES
cs
REVIEW QUESTIONS
Q810 Identify the currents and voltages in the commonbase and commonemitter connec
tions.
Q811 Why is the base width an important dimension in the quality of a transistor?
250
Chapter 8
ES
~ ES
exp
I e of diode
r
CR
 cs exp (
^Tf^
l)
Figure 8.19 The EbersMoll model for the PNP transistor. Current directions are actual.
Q812
Q813
Q814
Q815
HIGHLIGHTS
The boundary conditions in the emitter and collector are functions of the emitterbase
and collectorbase voltages respectively and assuming that the excess minority carrier
densities at the metal contacts are zero. In the base, the boundary conditions are deter
mined by the emitterbase and collectorbase voltages.
The emitter current and the collector current components of the PNP, I , I , I , and
I are determined as diffusion currents at each of the emitterbase and collectorbase
junctions. The recombination current in the base, I , is the difference between I
and I .
En
Ep
Cn
cp
REC
Cp
The emitter current is given by the sum of I and I , the collector current is the sum
of I and I , and the base current is the algebraic sum of I , I , and I
The EbersMoll relations are expressions for the emitter and collector currents as
functions of the emitterbase and collectorbase voltages and in terms of a , a , I ,
and/ .
The EbersMoll relations are applicable in all modes of operation of the device.
En
Cn
En
Cn
REC
c s
EXERCISES
E83 Consider a PNP bipolar transistor at 300K with the following parameters:
ES
Section 8.10
N
D
18
AE
= 10 crrr
nE
= 52cm /s
= 10 cm
DB
16
= 40cm /s
= 10" s
= 10 s
nE
15
EbersMoll Model
251
AC
= 10 cm
nC
= 116.5cm /s
= 10 s
W = 4
A = 0.1mm
Determine: a) a , b) , c) I , d) a , e) , f) I
B
ES
cs
13
CB
I =
0LpI
+I
cs
(1  )
+ I
(8.63)
CBO
where I
has b e e n defined as t h e collectortobase current with t h e emitter opencircuited (I = 0). E q u a t i o n (8.63) represents t h e commonbase characteristics of
t h e P N P B J T in t h e active region, as shown in Fig. 8.12(b). T h e commonemitter
characteristics in t h e active region are obtained from E q . (8.63) by replacing I by
(I + I ) and solving for I t o obtain
CBO
I = (a /l
c
 a )I
p
+ I /1
~ a=
CB0
+ I
(8.64)
CE0
where I
has b e e n defined as t h e collectortoemitter current with t h e base opencircuited.
E q u a t i o n (8.64) represents t h e commonemitter characteristics of t h e P N P
B J T in t h e active region, as shown in Fig. 8.12(a).
We will n o w determine numerical values for t h e four constants in Eqs. (8.57)
using t h e data of E x a m p l e 8.2.
CEO
EXAMPLE 8.7
Use the information given in Example 8.2 to calculate I , I , <x , and a .
f:s
Solution
The constants I
and J
ES
cs
qA[D /W
I=
qA[D /W +
cs
I=
m
cs
pPo
pPo
D n /L ]
nE
{)E
nE
D njL ]
nC
nC
= 18.6341 X 1 0 ~ A and l
ES
i5
c s
= 27.9637 X 1 0 ~ A .
and are
252
Chapter 8
The value of ^. is different from the one calculated in Example 8.2 for two reasons. First, the
hyperbolic functions used in Example 8.2 have been simplified and second, a small inaccuracy in
calculating a leads to large errors in
We observe very small values for a and . These are to be expected since the roles of the
emitter and collector have been interchanged.
We caution the reader that in all of these calculations, we have used the same value for the
areas of the base, emitter, and collector. The area of the collector is generally much larger than that
of the emitter and this difference has a significant tendency to cause a further reduction in the value
of ,
The EbersMoll relations in Eqs. (8.61) are applicable to the NPN transistor, when the rela
tions for a , , l , and I are modified in accordance with the corresponding changes made in
Eqs. (8.42) to (8.45), provided also that V replaces V and V replaces V .
Subject to the assumptions that have been made in arriving at these equations, it is important
to realize that these relations are valid in all four regions of operation, namely: active, saturation,
cutoff, and inverse active.
In the following example, we will illustrate the numerical values that result from using the
p
ES
cs
BE
EB
gc
CB
E b e r s  M o l l r e l a t i o n s in t h e l o u r r e g i o n s .
E X A M P L E 8.8
Use the results of Example 8.6 to calculate the values of the currents in the four regions of
operation.
kT
For the active region use V = 0.63 V,  V  . and V < 0
EB
c a
CB
kT
For the cutoff region use  V \ = \ V \ , V
EB
= 0.53V assuming V
CB
FB
CB
EC
< 0, and V
EB
cg
= 0.1 V
< 0
kT
, V = 0.63V, V < 0
Solution We determined the values for the constants in the EbersMoll equation as
I = 18.6341 X 10~ A / = 27.9637 X 10" A
a = 0.996
u
0.663
For the active region:
For the cutoff region:
l = 0.0941 X 10^ A
I = 0.6827mA
I = 0.680mA
j = q.404: 10" A
I = 0.0027mA
I = 9.31 X 1 0  A
=249
For the inverse active region:
For the saturation region:
l = 0.6792mA,
I = 0.6684mA
7 = 1.0245mA
I = 0.658437mA
1 = 0.009996mA
1 = 0.344mA
66
The new F (negative of old I )
0.6792mA
The new I' (negative of old I )
.0245mA
I' /I = 1.97 =
For the inverse active region use \V
CB
1S
EB
15
ES
15
15
,3
Chapter 8
Problems
253
It is worthwhile noting the very small ratio of I' to /. Furthermore, it can be shown that the
magnitude of V at saturation in the inverse active region is much smaller than the corresponding
magnitude in the active region.
We note that il is in the third quadrant of the / V . characteristics of Fig. 83)(b) that inverse
active operation is displayed. In this region, / is positive whereas the actual directions of the emit
ter and collector currents are the reverse of our actual references.
The bipolar junction transistor is operated in the inverse active mode in only certain applica
tions in digital switching circuits. As we indicated earlier, the value of is very small because of the
relative dopings of the various regions and also because of the much smaller area of the emitter
(which is the collector in the inverse active mode) when compared to that of the collector.
c
EC
t(
PROBLEMS
8.1 For an N P N silicon transistor,
a)
draw the energy level diagram at equilibrium, clearly identifying E , E , and
+
b)
repeat part (a) when the device is biased in the active region with \V
WV .
8.2 A silicon PNP transistor is operating in the active region with V = 0.668V and
V = 2V. Given that N = 4X 10 cm" , N
= 2 X 10 cm , N = 2X
10 cm" , W = 5, A = 1.8 X lO^cm , and = = , determine at = 300K,
B
BE
EB
17
CB
16
AE
14
AC
hn
e)
DB
18
16
DE
a)
15
AB
DC
Sketch the minority carrier distributions in the base for saturation and cutoff
and identify, by crosshatching, the minority carrier storage charge area when
the device is switched from cutoff to saturation. At saturation V = 0.8V
and V = 0.6V.
BE
BC
b)
Give the values of the minority carrier densities in the base at both junctions.
8.5 A silicon PNP BJT has N = 2.5 X 10 cm , N
= 2X 10 cm , N = 1 0 c n r ,
A = 10~ cm , W = , = , T = , and L = . Determine, at
= 300K, the collector current in the active region for V = 2V and,
a)
V = 0.62V
b)
I = 2.5.
8.6 A PNP silicon BJT has N = 10 cm , N
= 10 cm , N = 10 cm" , the metal
lurgical base width = ., and A = 3mm . For V = 0.5V and V = 5V, deter
mine at = 300K, the effective width of the neutral base.
18
AE
17
16
DB
AC
N C
BC
EB
18
AE
16
DB
15
AC
EB
CB
254
Chapter 8
= 10 s,
nE
_7
= 10 s, and T
_6
N C
= 10 s.
hp
hn
C)
hn
b)
c)
d)
e)
8.9 Assume that the minority carrier distribution in the base of a PNP BJT is linear when
the BJT is operating in the active region with  v  V  and \V \
kT/q.
Derive expressions for:
sc
a)
b)
BE
8.10 The EbersMoll model of the BJT includes three device properties only, namely: I ,
a , and a . Briefly explain how you would determine these for an NPN device from
measurements carried out on the BJT.
8.11 A silicon PNP transistor is operating at 300K and has: N = 1 X 10 cm~ , N =
1 X 1 0 c n r , N = 1 X 10 cm" , L = L = L = , W = , and A =
1mm . Determine the collector current, I , when the BJT is operating in the active
mode for
SM
18
AE
17
15
DB
AC
nE
pB
nC
a)
V = 0.6V
b)
I = 2.55mA
8.12 Briefly explain why I /I ,
for an NPN BJT, operating in saturation, decreases as V
reduced. Does I decrease? Does I increase?
EB
CE
is
DE
16
= 10 cm" , W = 5, and t
AB
pE
p C
a)
AB
C S A T
I C SATl
1 
/ ,
where = I /I .
Use the EbersMoll model equations and the equality of apI
and
ol I
to derive the above expression.
8.16 For an NPN BJT having a = 0.99 and a = 0.5, plot V
versus I /I
as lJl
varies from 20 to 95. = 300K.
8.17 An NPN BJT at = 300K has a = 0.995 and a = 0.15 at I = 1mA. Determine I
for E
equal to:
5
ES
c s
C S A T
CSAT
Chapter 8
Problems
255
a)
0.2V
b)
0.1V
8.18 A symmetrical silicon PNP BJT has N
= 10 cm" , N
= 5 X 10 cm , W =
2m, = 10ns, = , A = 0.01cm , and the device is operating in the active
region. For V = 0.65V and V = 2V, determine at = 300K:
a)
I
18
)6
AE
DB
EB
CB
b)
,
8.19 An NPN BJT has N 3 x 1 0 c n r , N = 5 X 10 cm" , and N
= 5 x 10 cm" .
Also W = , A = 10" mm and the lifetime of minority carriers in all 3 regions is
.. For operation in the active region, and at = 300K, determine I for V =
0.5V.
8.20
a)
An N P N BJT is operating in the inverse active region. Sketch the minority
18
DB
16
AB
15
DC
BE
b)
< < :
explain
why
SM
CB
chapter 9
BIPOLAR TRANSISTORS
II: LIMITATIONS,
SWITCHING, AND
MODELS
9.0 I N T R O D U C T I O N
In the previous chapter, we discussed the principle of operation of the B J T and
derived currentvoltage relations subject to certain assumptions. A l t h o u g h these
relations are remarkably accurate over a b r o a d range of currents and voltages, sec
ondary effects, which are present in actual transistor characteristics, were excluded.
T h e properties of real transistors diverge from the characteristics we derived
because of the following effects:
1. Changes in the effective width of t h e base as a result of the changes in the
reverse bias applied to the collectorbase junction.
2. Multiplication of carriers in the collectorbase depletion layer at a high reverse
bias. B r e a k d o w n at the junction.
The two effects stated above can b e discerned by a comparison of the ideal
and the actual graphical characteristics. O t h e r important effects, which d o not
a p p e a r in the graphical or analytical characteristics discussed so far, are:
1. Decrease of the current gain, ^, at very low and very high injection.
2. R e s p o n s e times of the B J T related to t h e switching of the transistor b e t w e e n
the O N and O F F states.
3. The circuit modelling of t h e transistor when it is subjected to small changes in
the applied voltages. The circuit includes the capacitances of the B J T that
b e c o m e extremely important in the operation of the transistor at high frequen
cies.
256
Section 9.1
257
T h e contents of this chapter include t h r e e major topics. We will first study the
actual static characteristics of t h e B J T and explain the factors, properties, and
processes that account for t h e differences b e t w e e n t h e ideal characteristics that
were considered in t h e last chapter and t h e actual characteristics. T h e second and
third subjects are t h e switching properties of t h e B J T and its smallsignal equivalent
circuit. These two subjects represent t h e two major applications of t h e transistor, its
use as a switch, a n d as a n amplifier.
In t h e second section, we examine t h e switching characteristics of t h e B J T and
d e t e r m i n e the t u r n  O N a n d t u r n  O F F times as functions of its properties. We then
develop a smallsignal equivalent circuit of t h e BJT, which replaces t h e transistor in
t h e analysis of amplifier circuits. T h e equivalent circuit components will be deter
mined, using t h e analytical static characteristics, by allowing incremental variations
in t h e voltages applied to t h e transistor. T h e circuit will include capacitances that
are intrinsic to t h e operation of t h e transistor. These capacitances account for t h e
time and phase delays in t h e output quantities of t h e B J T and are responsible for
the variation of t h e gains of an amplifier as t h e frequency of t h e input signal is
changed.
In a short section, we will subsequently evaluate t h e performance of a B J T as
a current amplifier based on a factor k n o w n as the gainbandwidth product.
Finally, we briefly explain t h e basis for t h e m o d e l used in simulating t h e B J T
in t h e c o m p u t e r p r o g r a m S P I C E .
T h e collector current for a fixed I in t h e active region of t h e commonemitter connection increases gradually with an increase of t h e emittertocollector voltage. Since I is d e p e n d e n t mainly on V , a n increase in V
results from an increase in t h e magnitude of t h e reverse bias (V ), since
V
is equal t o V
+ V
All of t h e above currents and voltages are positive
for a P N P transistor.
B
EB
EC
BC
EC
EB
BC
CEQ
EC
in F o r w a r d A c t i v e
In our study of the diode, we observed a change of the width of the depletion layer
as the applied voltage changes. A n increase of forward voltage causes the layer to
b e c o m e n a r r o w e r and an increase of the magnitude of the reverse bias causes an
increase of the width of the layer.
A n increase of the basecollector voltage of a P N P transistor increases the
width of the depletion layer and this reduces the effective width of the base. For a
fixed value of I , as for t h e C E connection generated by a fixed V , an increase of
V
causes an equal increase of V .
The value of the minority carrier density at the base side of the emitterbase
depletion layer is fixed by V . T h e minority carrier density at the base side of the
collectorbase depletion layer is almost zero. Thus, an increase of V
caused by an
increase of V
increases the width of t h e depletion layer and thus reduces the
effective base width while maintaining practically the same values of minority car
rier densities at the extremities of the new base. This results in an increase of the
gradient of the linear hole density distribution in the base, as shown in Fig. 9.2, and
an increase in t h e collector current. This accounts for the positive slope of the con
stant I curves in the active region of the commonemitter characteristics, shown
in Fig. 9.1(b) and Fig. 9.3. It is interesting to note that t h e extrapolations of the
constant I curves to the negative V
axis all m e e t at one point in the axis. This
point is at V
=  V.
B
BE
BC
EC
EB
EC
BC
EC
EC
Section 9.1
259
Metallurgical boundaries
W (1V)
Depletion layer
due to V
Depletion layer
forV =lV
5 C
W (30Y)
EB
Depletion layer
BC
= 30V
Figure 9.2 Reduction of the effective base width causes an increase in the slope of
the hole density distribution in the base. V is assumed to be constant.
EB
EC
EB
260
Chapter 9
day discrete BJTs and in integrated circuits BJTs, consists mainly of electrons
injected from t h e base to the emitter in a P N P BJT, a constant V
therefore corre
sponds to a constant I .
In the active region of the P N P BJT, t h e hole density distribution in t h e base is
assumed to b e linear, as shown in Fig. 9.2. T h e collector current can therefore be
written as
EB
= qAD dp/dx
= qAD p{(S)/W
T h e conductance, dI /dV ,
c
exp
qVjkT
can b e written as
EC
dI /dV
c
= (qAD /W )p
= (dI /dW )
EC
(dW /dV )
EC
= qAD p{0)/W
I /W
C
dV
EC
(9.4)
EC
EC
BC
GC
EC
EC
EC
EB
EB
EB
EB
BE
EC
EB
CE
EC
EC
EC
V=
A
W dV /dW
B
EC
(9.5)
ES
= I [exp (q VjkT)
s
For an N P N transistor, V
EB
and V
EC
 1] [1 + V /V ]
EC
are replaced by V
BE
(9.6)
and V
CE
respectively.
Section 9.1
261
Carrier Multiplication a n d B r e a k d o w n
A t t h e collectorbase junction, a n u p p e r limit is set on t h e magnitude of V
by
avalanche b r e a k d o w n . This b r e a k d o w n voltage is labeled BV
on t h e commonbase characteristics of Fig. 9.1. T h e first two subscripts refer to t h e collectortobase
with t h e third subscript indicating that t h e third terminal, t h e emitter, is opencir
cuited. O n t h e commonemitter characteristics, b r e a k d o w n occurs at a smaller volt
age, labeled BV ,
referring to t h e collectortoemitter b r e a k d o w n voltage with
the o p e n base.
A t collectortobase voltages below t h e b r e a k d o w n value, t h e energy level dia
gram at that junction is very steep. A t b r e a k d o w n , and as shown in Chapter 7 for t h e
diode, t h e holes that are crossing t h e junction from base to collector and t h e elec
trons crossing in t h e opposite direction acquire sufficient energy from t h e high elec
tric field in t h e depletion layer, causing ionizing collisions with t h e lattice. These
collisions generate electronhole pairs a n d each n e w carrier causes further ionizing
collisions. This multiplicative avalanche process increases t h e n u m b e r of carriers
crossing t h e junction a n d results in an increase of t h e collector current n e a r break
down as shown in Fig. 9.4.
We d e t e r m i n e d in C h a p t e r 7 that t h e current in a reversebiased P N junction
is multiplied by t h e factor near b r e a k d o w n because of t h e avalanche process. For
the c o m m o n  b a s e B J T connection, the current / , arriving at t h e reverse biased
basecollector junction, w h e n operating near b r e a k d o w n , can be written by using
E q . (8.13) as
BC
CBO
CEO
I = M(a I
c
+I )
(9.7)
CBO
(9.8)
CBo
where BV
is t h e magnitude of t h e avalanche b r e a k d o w n voltage for t h e junc
tion. T h e subscripts, CBO, refer to t h e collectorbase voltage with t h e emitter open.
Consequently, with I = 0, E q . (9.7) becomes
CB0
I = MI
C
(9.9)
CB0
CB0
EB
0), which m a k e s the hole density at = 0 in the base equal to the equilibrium value,
p , while the hole density at = W is almost zero.
For a transistor in the commonemitter connection, the b r e a k d o w n voltage is
labeled BV ,
as the voltage between collector and emitter with the base open.
0
CEO
Section 9.1
263
I = M( I
c
ap
+ I )/(l
 Meg
CBO
Since b r e a k d o w n is m e a s u r e d at I
(9.10)
I = MI /{iMa )
c
CB0
(9.11)
B r e a k d o w n for this connection takes place, for I = 0 as well as for all other
values of I , w h e n Ma is unity, which implies that is slightly larger t h a n unity. By
assuming that V
= V , labeling t h e voltage V
in E q . (9.8) at b r e a k d o w n as
BV ,
and setting Ma equal t o unity, we have
B
EC
BC
CEO
BC
BV
ot /
f
CEO\
\BV
= 1
(9.12)
CBO
This results in
BV /BV
CEO
= ( 
CBO
so that
BV
CEO
BV /V$
CBO
(9.13)
EC
EC
CEO
CBO
CBO
CBO
CB0
rl
J _
BV,
ICBO
Figure 9.5
BV , I
CBO
CBO
compared to BV
CE0
and I CE0
BV,
264
Chapter 9
CB0
CEO
CB
BC
CE
E X A M P L E 9.1
15
A silicon PNP BJT has a collector doping 3 X 10 cm~ and is much less than the doping
N of the base. Given that $ = 100, = 4 and % = 3 x 10 V/cm, determine:
*)BV
b)BV
CBO
cr
cm
Solution
a) From Eq. (7.9), we have
m a x
V, =
& e/2qN
9
BV
b) BV
CBO
CEO
= BV ,/(lW >
CHf
br
10
L
10
L F J
I9
98
1 0
= 31.22V
P u n c h t h rough
A n interesting destructive p h e n o m e n o n , distinct from avalanche b r e a k d o w n , may
take place if the depletion layer on the base side of t h e collectorbase junction
extends, because of t h e reverse bias, so far into the base that it reaches the emitter
Section 9.2
265
EB
EB
EB
EB
E(P)
B(JV)
C(P)
E,c
E,c
Figure 9.6 PNP energy band diagrams. Solid lines show operation in the active
mode. Dashed lines refer to punchthrough.
266
Chapter 9
and current crowding u n d e r the emitter are the factors that cause a reduction in the
current gain. A t t h e collectorbase junction, the widening of the base region w h e n I
increases causes a reduction in the r a t e of increase of the collector current.
V e r y L o w Injection a n d Current G a i n
In the depletion layer of t h e forwardbiased emitterbase junction the carrier densi
ties, because of injection of carriers from b o t h sides, are greater than the thermal
equilibrium values. A s a result, t h e r a t e of recombination exceeds the generation
rate of t h e carriers. A t low values of V , of t h e o r d e r of 0.4V and less, in a P N P sili
con transistor operating in the active region, all currents are small and t h e depletion
region is quite wide. A t these values of t h e base current, the recombination compo
nent in t h e depletion layer is n o longer negligible.
We recall that we determined, for a P N P transistor, t h e electron c o m p o n e n t of
the emitter current, I , from the slope of t h e electron density on the emitter side of
the emitterbase depletion layer. This current forms the most important of the three
components of the base current. Because of recombination in the depletion layer,
the base current c o m p o n e n t that we have assumed equal to I , on t h e base side of
the depletion layer, is in fact greater than I . This requires a higher base current.
The collector current, which consists mainly of I , has not changed. A t low values of
V
and h e n c e low values of t h e base current, the recombination c o m p o n e n t of this
current in the depletion layer b e c o m e s significant. F u r t h e r m o r e , and of m u c h less
importance, the low values of V
cause a wider depletion layer that increases the
probability of recombination.
The higher base current results in a decrease of ^ at low values of t h e collec
tor current. A plot of log I and log I , as a function of V , is shown in Fig. 9.7. In
the plot, is shown to b e constant in the active region over a wide range of collec
tor currents. D e p e n d i n g on the particular transistor, the region of constant ^ may
extend over six decades of collector current.
EB
En
En
EB
EB
EB
Section 9.2
267
kT/q
and V
< 0), the expression for I may b e approximated by I in E q . (8.29). The
expression for I is set equal to I in E q . (8.33). Assuming W /L
1 so that coth
(W /L )
= L /W
and sinh W /L
W /L
EB
CB
EN
I=
(q A DD /W )
I=
^
(qAD n /L )expqV /kT
nE
0E
exp q V /kT
(a)
EB
nE
(9.14)
(b)
EB
EB
ho
e x
i^pj
P 1 E NKT
(9.15)
= (q AD n /L )
nE
0E
exp qV /kT
nE
+ (q An,W/2^
EB
exp q V /2kT
EB
To relate the expression for the total base current to the plot of I
we write
= / ; exp q V /mkT
(9.16)
in Fig. 9.7,
(9.17)
EB
w h e r e m is an empirical factor obtained by curve fitting and its value varies from
o n e to two, and T is a n o t h e r empirical factor related to the two prefactors in E q .
(9.16).
A t low values of V , the slope of the I curve is (q V /kT),
while t h e slope of
the I' curve is smaller. While the slope of T is a result of the exponential t e r m in E q .
(9.17), t h e t e r m T is greater than t h e prefactor of I in E q . (9.14(b)) because of t h e
terms in E q . (9.16), causing T to be larger t h a n I .
The increase in I at these low levels, c o m p a r e d to the values of I , results in a
smaller ^, as illustrated in Fig. 9.8.
s
EB
EB
EB
268
Chapter 9
200
100
Figure 9.8
10mA
1mA
I,c
CE
N(x) 
I /v(x)A ]
c
(9.18)
CB
CB
cn
(9.19)
where (x x ) is t h e width of t h e spacecharge layer. In general, and at the values
of V
that are used, the electric field is high enough to cause v(x) to be t h e satura
tion drift velocity, v , which is of the o r d e r of 10 cm/s.
A t low injection levels, t h e second t e r m inside the brackets of E q . (9.18),
which represents t h e charge density of the collector current, is negligible. A t a par
ticular collector current, labeled t h e critical current I, that t e r m causes the derivative in E q . (9.18) to b e c o m e zero. This critical current is given by
c
CB
I =
o
qN(x)v A
s
(9.20)
Section 9.2
+
269
where x
E q . 9.20.
c o
DB
*From Muller and Kamins, Device Electronics for Integrated Circuits, p. 327, copyright Wiley
(1986). Reprinted by permission of John Wiley & Sons, Inc.
transistor is related to t h e other two operating modes, saturation and cutoff, which
occur in the first q u a d r a n t of t h e B J T commonemitter characteristics. In Sec. 9.3,
we investigate the processes that t a k e place w h e n a B J T is switched b e t w e e n cutoff
and saturation, and consider the properties that lead to faster switching.
REVIEW QUESTIONS
Q91 Briefly define the significance of the multiplication factor M.
Q92 Why is punchthrough of concern in modern BJTs?
Q93 To increase the punchthrough voltage, the base doping can be increased. What is the
negative effect of an increase in the base doping?
Q94 Explain the negative effect of reducing the base doping on the base resistance.
Q95 What is the effect on the emitter of high base resistance at high current?
Q96 Explain why BV
is larger than
BV .
Q97 Briefly explain the shape of the curve $ versus I in Fig. 9.8.
CB0
CEO
HIGHLIGHTS
The collectorbase voltage has dramatic effects on the commonemitter output charac
teristics in two regions: at saturation (low values of V ) and at breakdown (high val
ues of V ) for a PNP transistor.
BC
GC
Section 9.3
Transistor Switching
271
The Early effect is the reduction in the effective width of the base as the collector to
emitter voltage is increased. This is accompanied by an increase in the collector cur
rent.
Punchthrough is the process whereby the reversebias at the collectorbase junction is
so large that the depletion layer in the base extends to the emitterbase junction.
As the collectorbase reverse bias is increased in the active region, the collector cur
rent increases, due both to the Early effect and the multiplication of carriers in the col
lectorbase junction.
As \ V \ is increased, the base width is reduced, a higher slope of the minority carriers
in the base results, the emitter efficiency increases, and the transport factor is higher.
Avalanche breakdown occurs when the generation of new carriers is so large that the
collector current tends to increase to destructive values.
Emitter crowding results from the IR (where I is the base current and R is the base
resistance) drop of the base current in the direction of base current flow towards the
emitter, reducing the effective forward bias in the center of the emitter relative to the
edges. Thus, the emitter electron current in an NPN transistor is concentrated at the
emitter periphery.
CB
EXERCISES
E91
15
3
CB
Ans: W = 6.25
g
9.3 T R A N S I S T O R S W I T C H I N G
A n ideal switch is a shortcircuit in t h e O N position and an opencircuit in t h e O F F
position. Switching, although not quite, but close to the ideal variety may be
achieved in a B J T commonemitter circuit. O p e r a t i o n of the transistor in saturation
occurs at a low V
and high I , simulating the O N state (low V / I ) , whereas in
cutoff, the transistor current is nearly zero, thus resembling an opencircuit. Figure
9.10 shows the circuit and t h e current drive of a transistor together with its charac
teristics in order to illustrate switching.
A t i = 0, the B J T is in cutoff and operation is at point A , while saturation cor
responds to a high value of I , where V is virtually zero. The O N state occurs p r o
vided I is equal to or greater t h a n I J^ or (V
 V )/PpR ,
which corresponds
to point in Fig. 9.10(b).
E v e n if i is switched in zero time from zero to I , the saturation state is not
reached instantaneously because t h e speed of response of t h e transistor is limited
mainly by the storage or diffusion capacitance, which accompanies the storage of
minority carriers in the base. The distributions of minority carriers in the base of a
P N P transistor are shown in Fig. 9.11 for t h e cutoff and saturation conditions where
it is assumed that the width, W , of the base is m u c h smaller t h a n t h e diffusion
length of minority carriers so that t h e distribution is linear in b o t h the active and
EC
E C
gl
gc
gi
cc
BC
B1
Chapter 9
(a)
(b)
EC
(c)
Figure 9.10 Switching of a transistor (a) circuit, (b) BJT characteristics and load line, and (c)
base current drive.
Section 9.3
Transistor Switching
273
= ~qAD
dp'/dx
= qAD p'(0)/W
p
(9.22)
The total excess minority carrier charge stored in the base, Q , is the product
of qA and the area of the distribution in Fig. 9.12.
R
= qAp'(0)W /2
(9.23)
The ratio of the charge stored to the collector current, which has unit of time,
becomes
QB/IC
= W /2D
B
= g
(9.24)
dx = v(x) dt
(9.25)
where v(x) is the velocity of the hole. The time it takes the hole to cross the base
becomes
t =
tr
Jo
dt = \
Jo
dx/v(x)
(9.26)
D dp'/dx
p
= qA D p'{Q)/W
p
(9.27)
This current can also b e written in terms of the velocity of the carriers. We also
use the linear distribution of Fig. 9.12 to write
274
Chapter 9
v(x) = qA v(x)p'(0)
(1  x/W )
(9.28)
We set the expressions in Eqs. (9.27) and (9.28) equal in order to obtain the
expression for v(x) as
'W (lx/W )
B
,
Wl
dx = r2D
(9.30)
'
We thus conclude that t h e transit time is equal to, and has t h e same signifi
cance as, T .
B
= qA\
p'{x)dx
(9.31)
Jo
where p'(x) is the excess hole density at x. We will now determine the relation
b e t w e e n the base current and the stored charge by referring to t h e continuity equa
tion formulated in C h a p t e r 4 . E q u a t i o n (4.36) for the continuity of holes in the base
is r e p e a t e d in E q . (9.32) for one dimension after introducing the factor qA and
replacing (  ) hyp'
0
l(p>(x)qA)
at
 *
dx
Jo
Jo
as
dx I
Jo
p'{x)qAdx=\
o t
(9.32)
qA dJ
(9.33)
Cp
Cn
Cn
dt
(934)
Section 9.3
Transistor Switching
275
i (t)
= Q (t)h
(9.35)
equations.
TurnON Time
A transistor is switched from cutoff to saturation by the application of a step of base
current, I as shown in Fig. 9.10(c). For i = I , E q . (9.34) becomes
BV
J
 hi
(936)
dt
G (0
B
=V ,
[1
"
('A,)]
(9.37)
ic{t)
QM
! m l E
e x p H / , ) ] , for
Q ^Q
B
SAT
(9.38)
where
is t h e value of Q at i = I . T h e reason for restricting t h e expression
for i to t h e range below G S A T *
> whereas t h e charge continues to build u p as
long as i is at I , t h e same collector current saturates at /
that corresponds to
<2
 E v e n a n increase of i b e y o n d t h e onset of saturation, as shown by Fig. 9.13,
causes a slight increase in I b u t a correspondingly large buildup of charge. Sketches
of Q (t) a n d i (t) are shown in Fig. 9.14.
5
SAT
sm a t
C S A T
S A T
S A T
c c
ECSiCP
EC
276
Chapter 9
collector current is /
, which is approximately equal to V /R
since V
CC
cc
C S A T
(a)
Section 9.3
Transistor Switching
277
C S A T
SAT
C S A T
Q N
V /R
CC
= ^
= ^
HONA,)]
[1  exp
(9.40)
C S A T
equal to
C S A T
J_
O
'N = i n
1 "
 ,
(9.41)
(V /I R )(7 /T )
cc
B1
CC
C S A T
TurnOFF Time
To turn a transistor OFF, the excess stored charge in the base must be r e m o v e d and
the collector current must be m a d e almost zero ( / ) . T h i s can b e accomplished by
setting i to zero or b e t t e r yet by making it negative in order to aid in removing the
excess stored charge faster.
By reducing the base current to zero, the excess minority carriers that have
b e e n stored in the base decay by recombination and diffusion because there is n o i
to replenish the neutralizing majority carrier charge. For zero i , E q . (9.34) becomes
C0
dQ /dt=Q h
B
(9.42)
C S A X
SAT
SAT
C S A T
Q (t)
B
= Q (0)
eH
(9.43)
where Q (0) is the total excess charge available in the base at t h e end of the base
current pulse. W h e n Q {t) is equal to <2 t at r = t , t h e transistor is at t h e edge of
the active region so that Q (t ) = C? at ^
transistor is in t h e active
region, so that i is found from Eqs. (9.35) and (9.43) as
B
SA
o r
1 >
i (t)
c
= ( 0 / = ( C 2 / T ) e""> = /
sat
C S A T
e~*P
(9.44)
278
Chapter 9
QB(0
Figure 9.15 Sketches of Q , i , andp'(x) during the turnOFF period as i is reduced to zero
instantaneously.
B
t = ^ n(Q (Q)/Q )
s
(9.46)
SAT
C S A T
C S A T
E X A M P L E 9.2
The base current pulse into a switching transistor is shown in Fig. 9.El. Given that for Fig. 9.10.
V
= 5V, R = lKfl, = 0.258, W = 2, and D = 5cm /s, sketch the waveshapes of the col
lector current and of the charge Q and identify the values at the critical points.
2
cc
Solution
I,VM
For Q = Q
B
S A T
,,/R,
5mA
= Am
Section 9.3
279
Transistor Switching
(/1)
'is
200
200
t(ns)
Figure 9.E1
S A T
as
T
GAT W = P
S
We now determine the turnON time at which Q = <2 ,from Eq. (9.41), using
B
= 0.25 X 10~ s, V
c c
= 5.0V, I
SAT
8 J
= 127ns
The collector current remains at 5mA and, as long as i is greater than zero, the charge Q con
tinues to increase until the end of the i pulse at t 200ns, at which time Q is determined from Eq.
(9.37) as
B
Q (200na)
B
= 27.5pC
This is the Q (0) to be used in Eq. (9.46) to determine the storage time.
After i becomes zero, the transistor discharges from 27.5pC towards zero. Once it reaches
< 2 , i begins to decay towards zero as shown in the sketches below. How long after ; becomes
zero does it take Q to decrease to Q
= 20pC?
We use Eq. (9.46) to calculate the storage time as
B
SAX
S A T
Sketches of the time variation of stored charge and collector current are shown in Fig. 9.E2.
It is worth mentioning here, and as we will further study in Chapter 11, that the BJT switching
time can be considerably reduced by connecting a metalsemiconductor diode (Schottky diode)
from collector to base. The diode clamps the base to the collector and prevents the BJT from going
into deep saturation.
280
Chapter 9
Q (pc)
B
50
27.5
20
i (mA)
B
12.5
127
Figure 9.E2
279.6
t ( n s )
9.4 S M A L L  S I G N A L E Q U I V A L E N T CIRCUIT
In the previous section, we investigated the switching properties of a transistor and
our interest was restricted to the extremities of the transistor characteristics,
namely, cutoff and saturation. In this section, we will b e concerned with the active
region of o p e r a t i o n where the transistor exhibits its amplifying properties. Since
amplification is o n e of the basic functions of the transistor, our objective is to
develop a m o d e l for t h e transistor that we can use to study its amplifying properties.
The analytical and graphical static characteristics of the transistor are highly
nonlinear. In amplification, we are not only interested in magnifying a voltage or a
current, but this must b e d o n e while preserving the original shape of t h e signal. For
this to b e possible, we must restrict t h e operation, on the nonlinear characteristics,
to linear segments. This can b e m a d e possible only by selecting a certain point on
the characteristics and allowing t h e input variations to occur on the slope of the
characteristics at that point. Only then can we insure faithful reproduction of the
signal we wish to amplify. This is t h e case for amplification of small signals of voltage
and current.
For what we label as power amplifiers, large excursions of t h e signals over non
linear portions of t h e characteristics are required to obtain large a m o u n t s of output
power. Here, the nonlinearities introduced are removed by some filtering means.
In o r d e r to study the dynamic A C response of t h e transistor to a sinusoidal
excitation, for example, it b e c o m e s necessary to m o d e l the transistor by what we
label as an A C equivalent circuit, which describes the response of the device to
incrementally small variations, small c o m p a r e d to the D C values at the point where
Section 9.4
281
/
EC
,EC
AV
EB
L =
C
M
(")
v,EB
Figure 9.16
cc
cc
EC
EB
CB
CB
is negative and
b) T h e width of the base is much smaller than the diffusion length of minority car
riers in the base so that a linear distribution of hole density results. This assump
tion is quite valid in m o d e r n discrete and integrated circuit transistors. The
assumption does not exclude recombination in the base.
c) Lowlevel injection of holes from the emitter into the base.
282
Chapter 9
EFFECTS O F C H A N G E S IN V
Carrier Processes
The equivalent circuit we aim to find consists of l u m p e d passive elements and
d e p e n d e n t current generators that will replace the box shown in the commonemitter connection of Fig. 9.17. T h e lower case symbols, shown below, refer to incre
m e n t a l quantities that replace the changes in the P N P currents and voltages as
A
h = " V'c
= ~
a be =
AV
 EB>
EB
EB
1. The hole density in the base at the edge of the emitterbase depletion layer
increases. To maintain neutrality, the electron density in t h e base increases. We
assume a fixed V
at the reversebiased CB junction so that the excess hole
density at = W is zero. The hole density distribution is shown in Fig. 9.18.
CB
Bo+
Equivalent circuit
o C
be
Section 9.4
283
= (M
+ MJ
En
(9.47)
S m a l l  S i g n a l Currents a n d Circuit E l e m e n t s
Because of t h e linear distribution of hole density in t h e base and neglecting b o t h t h e
collectorbase leakage current and t h e recombination current of holes w h e n com
p a r e d to t h e total collector current, and with reference to Fig. 9.18, we write
I
= I
= qAD p[(0)/W
+ M
Cp
(a)
(b)
(9.48)
qAD Ap(0)/W
p
(9.49)
p'M Po
P
e x
p[(0) =
P (<?
v
E B
k T
q(V
exp 
e x
AV
kT
EB
) [ P {q / )
EB
AV )
EB
kT
qV
exp kT
E
~!]
 l]
(9.50)
EB
to b e much
284
Chapter 9
(q/kT)
Pl
(9.51)
AV,EB
eb
eb
exp (qvJkT)
EB
= 1 + qvJkT
+ \
(^$)
2\kT
(9.52)
eb
En
(9.53)
 1]
EB
qAD n p[(0)
nE
*En
OE
( 9
'
5 4 )
^nEFO
<lAD n
nE
Ah
Ap(0)
0E
(9.55)
LePo
= qA\
[{
)/ ]
0
dx
(9.56)
Jo
where Q is t h e excess hole density charge in t h e base.
By replacingp' by p ' ( 0 ) ( l  x/W) a n d integrating E q . (9.56), we have
B
gAp W
0
2T
qV
kT
El
exp
 1
(9.57)
EB
Al
mc
= (qAW /2T )
B
(0)
(9.58)
The change in t h e base current is t h e sum of Eqs. (9.55) and (9.58). By using
E q . (9.51), t o replace Ap(0), we have
AI = (0)^AV
B
Pl
n E
EB
L Po
nE
Q E
2x
(9.59)
Section 9.4
285
We replace M a n d AV
by i and v respectively and we also replace
p[(0)qA by its d e p e n d e n c e on I from E q . (9.48(a)). T h e result is
B
EB
be
D n W
nE
kT
0E
, W%
(9.60)
lc
D
pP0 nE
^p p
c = fflcv
(961)
be
= K K e= y y
(962)
h = S v J^
m
= iJ^
(9.63)
EB
CB
D n W
nE
OE
DpP L
0
(9.64)
2L' _
nE
This expression for agrees very well with E q . (8.56) for ^,.
E q u a t i o n (9.63) m a y b e rewritten as
0
h = SA/^o
= J
(965)
where r^ = v Ji
is t h e ratio of to g a n d is k n o w n as t h e C E input resistance.
We draw n o w a smallsignal low frequency equivalent circuit for t h e B J T in
the commonemitter connection as shown in Fig. 9.19. This is a lowfrequency circuit
because capacitive effects a r e n o t included.
It is i m p o r t a n t t o n o t e that t h e equivalent circuit is valid for b o t h t h e P N P and
N P N transistors.
It is i m p o r t a n t t o point out that t h e values of b o t h g a n d r are d e p e n d e n t on
t h e magnitude of t h e collector current at t h e quiescent point. H e n c e , a change in t h e
location of t h e quiescent point on t h e static characteristics changes t h e magnitudes
of r and g and hence . We n o w n e e d to address t h e significance of t h e smallsig
nal current gain, .
We have defined earlier t h e commonemitter current gain, ^, t o b e I /I . T h e
symbol is t h e ratio of t h e collector smallsignal current t o t h e base smallsignal
current. A l t h o u g h t h e values are often assumed to b e t h e same, but because of the
d e p e n d e n c e of on I , as shown in Fig. 9.8, it is instructive t o d e t e r m i n e a m o r e
b
286
Chapter 9
Bo+
<
Figure 9.19 Lowfrequency commonemitter equivalent circuit of the
transistor.
iji
= M /M
0
iji
dl,
(9.66)
Capacitance Effects
We still have to determine t h e capacitive effects associated with a change in AV .
T h e r e is a storage capacitance or diffusion capacitance as defined for the diode in
Sec. 7.3. This is a consequence of the associated change in stored charge in the base.
We n o t e from Fig. 9.18 that
EB
AQ
qAW Ap(0)/2
B
BS
= AQ /At
B
= qAW Ap(0)/At/2
= (qAW p[(0)
(q/kT)
AV /2At
EB
(9.67)
'/,
A/
r B
e s
= / (Wg/2D ) ^^
c
= (Wl/2D )
p
' "^
(9.68)
Section 9.4
Figure 9.20
287
EB
only.
<*>
where C is the base storage or diffusion capacitance and using E q . (9.30) is given by
b
C  g
b
(W /2D )
= gr
m
(9.70)
EB
dv,
i C ~f
bt
(9.71)
je
where C. is the emitterbase junction capacitance and has b e e n defined for the diode
in Sec. 7.3.
Based on the variations caused by a change in V , we can draw t h e equivalent
circuit shown in Fig. 9.20. While t h e elements shown in the figure may seem to
include the major c o m p o n e n t s in the equivalent circuit, there are two additional
p a r a m e t e r s that are important, one of which is extremely influential in determining
the high frequency response of the transistor.
EB
EFFECTS O F C H A N G E S IN T H E M A G N I T U D E O F V
CB
Carrier Processes
O u r emphasis on the active region performance of the transistor may have under
mined t h e influence of the collectortobase voltage. This voltage has dramatic
effects on the saturation region response and on activities at and n e a r breakdown.
E v e n in the active region, as we have seen earlier, an increase of reversebias at the
288
Chapter 9
EB
CB
CB
EB
EB
CB
EB
CB
AI
AV
die AW
qApKO)
dW AV
BC
AW
BC
AV
BC
"
A ^ . z k ^
AV
F r o m E q . (9.5) and for V
BC
BC
AV
(9 V3)
BC
(9.74)
BC
BC
Section 9.4
U
289
= W I /V
BC
(9.75)
cl
AI
ii = \ V ^ = V J / ^
(976)
R e c o m b i n a t i o n Current C h a n g e
A n increase of V causes a reduction in t h e total minority carrier charge stored in
t h e base, which results in a reduction of t h e rate of recombination in t h e base and
thus a reduction in t h e base current. Since t h e recombination current is defined as
Qh,
we can write
BC
/
T h e effect of change in V
and letting =
= AQ h
B
(9.77)
BC
e c
AI
AV
rec
BC
AQ
i AV
p
_  (7cTB)
AV
BC
BC
AI
0
(9.78)
BC
IEC
BC
BC
bc
AV
PO BC
PO A
C O M P L E T E E Q U I V A L E N T CIRCUIT
The change in t h e stored charge accompanying a change in t h e collector base volt
age causes a storage capacitance. By using E q . (9.75), we obtain t h e expression for
the capacitance caused by a change in V , which is
BC
AQ
C
_ A(I T )
A V
(9 80)
AV
BC
290
Chapter 9
C. = ^
(9.81)
We replace the expression for the transit time by its equivalence from E q .
(9.70) and use the definition of g from E q . (9.62) so that
m
= ?TZ
T7?
(982)
C.
S be
m
C
g V v /& V
m
bc
bu
change in
EB
EB
CB
CB
c =c v,/v
u
0
+
II
be
EB
Bo
K<
= Je
<
^gmVbe
Section 9.4
g V v /V
m
bc
291
CB
be
ec
cV
hi = g (V /V )v
m
= g (V,/V )v
bc
+ g
be
^ v
'
(9.83)
ec
The first t e r m is in parallel with, and much smaller than, the current generator
gv
and may therefore b e neglected. T h e second t e r m may b e replaced by a resis
tance connected from collector to emitter given by
m
be
=f
l
7V
( 9
cl
8 4 )
5m t
bV
r = f
u
(9.85)
( /
EC
Name
Relevant Equation
transconductance
CE current gain
CE input resistance
CE output resistance
storage capacitance ( V )
EB junction capacitance
storage capacitance ( V )
CB junction capactance
base ohmic resistance
emitter ohmic resistance
collector ohmic resistance
collector base resistance
EB
CB
(9.64)
ty/ (9.84)
(9.70)
(7.16)
(9.80)
(7.16)
c
(9.875)
Typical Value
Unit
38.6 X 10"
100
2600
50.000
15
2
.008
2
100
2
20
5 10
none
ohms
ohms
pF
F
pF
pF
ohms
ohms
ohms
ohms
292
Chapter 9
B'o
' <^
be
Cb + Cje 
n be
<C o
6E'
Figure 9.22 Complete highfrequency smallsignal equivalent circuit of the BJT.
9.5 F I G U R E O F M E R I T
A m e a s u r e of t h e quality of a high frequency transistor is its figure of m e r i t / A s we
shall see,f is a m e a s u r e of t h e ratio of g to t h e total capacitance of the transistor.
By neglecting r , r , a n d r , we will determine an expression for t h e shortcir
cuit current gain. (/), of the transistor. It is defined as the ratio of the current I in
a shortcircuit placed at the output, as shown in Fig. 9.23, to an input driving current
We use R M S quantities for I , I , and V .
Because of the shortcircuit, C is in parallel with C^. By neglecting the current
in C c o m p a r e d to g V , I is g V
a n d V is
+
rent gain is given by
r
()
be
be
IJL
be
be
= (/) = / [ 1 + / r . ( C + C )]
0
(9.86)
where g = / ^ .
The symbol, f , is defined as the frequency at which the magnitude of the
shortcircuit current gain is unity, so that at / = f , (/) = 1 and
m
Section 9.5
Figure of Merit
293
Bo
I y
gm h
e
'
Eo
fr
= ^
c?m
F r o m E q . (9.70), C / g
6
is , where
<m
2
= W /2D
B
+ !L1k
c?m
so that
C + C
(9.90)
REVIEW QUESTIONS
Q98
To switch a transistor (PNP) from cutoff to saturation, what should happen to the
294
Chapter 9
HIGHLIGHTS
There are two major fields of application of the BJT: switching circuits and amplifier
circuits. While operation is in the active region in amplifiers, switching takes place
between cutoff and saturation.
The switching process is studied by using the two charge control equations. A base cur
rent pulse initiates the switching from cutoff to saturation, causing a buildup of minor
ity carrier charge in the base. Rapid turnON time is enhanced by a large amplitude of
the base current, a smaller saturation collector current, and a small lifetime.
The turnOFF time is smaller if the lifetime is smaller, resulting in a fast removal of
the minority carrier charge buildup in the base.
In the use of a BJT in amplifiers and because of the nonlinearity of the BJT character
istics, changes in the input and output variables must be restricted to very small values.
To analyze a transistor circuit, an equivalent smallsignal circuit is determined, which
represents the transistor for incremental changes about an operating point on the out
put characteristics.
EXERCISES
E92
Ans: r = 1.3.
E93
Ans: C. = 10.28pF
9.6
NPN TRANSISTORS
In the early part of the previous chapter, we indicated that, in the interest of consis
tency with t h e P N junction diode, we would use t h e P N P transistor in the analysis of
the operation, the equations, and the models of the device.
H o w e v e r , it is i m p o r t a n t to m e n t i o n that because of circuit yield and economy,
integrated circuit fabrication is geared towards the N P N transistor. A l t h o u g h it is
quite possible to p r o d u c e P N P transistors on the same chip as N P N devices, they
will not possess the high quality properties of the NPN.
The major advantage of N P N transistors rests in the attractive properties of
the minority carrier electrons in the base. Electrons have a higher mobility than
holes, h e n c e resulting in faster and higher gain devices. The higher $ of the N P N in
comparison to that of the P N P was confirmed in the examples of Chapter 8.
F
Section 9.7
295
9.7 T H E G U M M E L  P O O N M O D E L
In the early sections of this chapter, we c o m p a r e d actual characteristics of t h e B J T
to t h e simple relations that we derived in C h a p t e r 8 for the EbersMoll model. We
also identified the physical processes that are the causes of t h e discrepancies
b e t w e e n actual and derived characteristics.
The EbersMoll equations, while representing a simple and elegant model, do
not provide very precise results w h e n applied to m o d e r n  d a y minimal geometry
transistors. A modified m o d e l is n e e d e d to represent the second order effects. The
results of t h e modifications a r e included in the G u m m e l  P o o n model. This m o d e l is
the basis for the c o m p u t e r simulation p r o g r a m S P I C E . S P I C E is the acronym for
Simulation Program with Integrated Circuit
Emphasis.
The G u m m e l  P o o n m o d e l incorporates three important second o r d e r effects.
They are:
EB
I =
s
qADfo/W,II
qAD nj
qADn
N W
qAN W
DB
DB
(9.91)
296
Chapter 9
where
= qA
n{x)dx
(9.93)
Jo
T h e symbol Q
GQ
N dx
(9.94)
DB
Jo
Because of the effects of high injection and basewidth modulation,
becomes
Q
= Q
G0
+ Q +Q +Q +Q
E
(9.95)
where
Q
Q
G0
and Q
Q and Q
p
The highinjection effects are m o d e l e d using the charge control relations and
the chargecontrol time constants and , while basewidth modulation effects
are m o d e l e d by a forward Early voltage, V , and a n equivalent, V , for reverse oper
ation together with t h e depletion layers capacitances, C and C .
The process in steps two and t h r e e requires the addition of four new p a r a m e
ters. These plus the four required for depletion layer recombination and I , and
m a k e u p the eleven p a r a m e t e r s n e e d e d to describe the transistor static charac
teristics in the G u m m e l  P o o n model. To m a k e the m o d e l m o r e complete, the values
of r., r and r are added.
jc
O'
16
DB
AC
BE
CE
18
DE
16
CB
DC
nB
BE
CB
Chapter 9
Problems
297
16
AB
DC
16
16
15
DB
pB
AC
CB
o c
oc
BC
pB
BC
1 + T/T>
S
where is the transit time of minority carriers in the base and is the lifetime of
minority carriers in the base. Explain the physical significance.
9.9 A base current pulse of 250 with a duration of 300ns is used to turn on a silicon
PNP BJT in the circuit of Fig. 9.10(a). Given V = 5.2V, R = 1 = ,
W = 5, and D = 10cm /s, determine, at = 300K,
cc
9.10
pB
(a)
(b)
(a)
BE
(b)
9.11 The transistor in the circuit of Fig. 9.24 is a silicon PNP device operating at = 300K,
and at I = 1mA, V = 0.7V, and V = 5V. The device has = 200, N
=
10 cm , N = 10 cm , W = 0.8m, V = 100V, and D = 10cm /s.
Determine:
c
17
EB
16
AC
BC
DB
pB
0)
9.12 For the device of Problem 9.11, determine, for N
C = 0.2pF,
AE
(a)
(b)
C
C
je
18
3
_3
= 1 0 c m , A = 10 cm , and
298
Chapter 9
(d)
20
16
_3
DB
pB
chapter 10
JUNCTION FIELDEFFECT
TRANSISTORS
10.0 I N T R O D U C T I O N
FieldEffect transistors ( F E T s ) are labeled as such because the primary action is the
effect of a transverse electric field o n t h e longitudinal m o t i o n of t h e carriers. Both
their construction and o p e r a t i o n are considerably different from bipolar transistors.
W h e r e a s the currents in a B J T include b o t h holes and electrons, the current in an
F E T involves one carrier only. In a BJT, the study of t h e currents is based mainly on
the study of the diffusion of the minority carriers. In an F E T , the current is a result
of carrier drift under the influence of a longitudinal electric field. Finally, F E T s are
labeled as unipolar devices to highlight t h e fact that t h e current carriers are either
holes or electrons.
In general, the F E T has several advantages over t h e bipolar junction transis
tor. First, it has a m u c h higher input resistance, thus causing negligible loading of a
voltage source connected at the input. Second, the F E T is relatively insensitive to
t e m p e r a t u r e and i m m u n e to radiation. Third, and in particular for siliconbased
devices, it is less noisy. The main disadvantage of t h e F E T is its lower transconductance, g , and hence lower gain. The F E T is m o r e economical to p r o d u c e than the
B J T as it requires fewer fabrication steps and occupies m u c h less chip area.
Because of its low noise and relative insensitivity to cosmic radiation, the
J F E T amplifier is primarily used in satellite communications operating in t h e giga
hertz frequency range. It is also used as an amplifier such as in operational ampli
fiers and in comparators.
T h e r e are two basic classes of F E T s : t h e junction F E T ( J F E T ) and the metaloxidesemiconductor F E T ( M O S F E T ) . The junction in t h e J F E T may b e at a semi
conductorsemiconductor surface or at a metalsemiconductor surface, and we
m
299
300
Chapter 10
10.1 C O N S T R U C T I O N A N D O P E R A T I O N
Construction a n d t h e Basic Functions of t h e Terminals
We will use t h e perspective sketch of Fig. 10.1(a) and the cross section in Fig. 10.1(b)
to explain the construction of the J F E T .
The transistor has t h r e e terminals: the source, the drain, and the gate. The
source is t h e semiconductor terminal from which carriers are emitted that travel
through a semiconductor channel to b e collected by the drain. In the JFET, t h e
refers to the type of semiconductor of the channel and the carriers are electrons. A
positive voltage applied from drain to source is used to accelerate t h e electrons
through the channel. The third terminal, labeled the gate, is used to control the flow
of electrons. In t h e single gate device, the gate is situated above the channel. In the
twogate model, an additional gate is shown below t h e channel with the two gates
operating at the same voltage. The gate, a P semiconductor in the N J F E T , is of
opposite conductivity to t h e channel. A negative voltage is applied to the gate with
the source grounded. With a positive voltage applied to t h e drain, with respect to
the source, the gatechannel junction is reversebiased and the resulting electric
field is transverse to the direction of motion of t h e electrons. A s the gate voltage is
changed, or as t h e drainsource voltage is changed, t h e reverse bias across the gatechannel junction changes. The reverse bias changes the width of the depletion layer,
thus changing the vertical dimension and the cross sectional area of t h e channel that
is n o r m a l to t h e direction of m o t i o n of the electrons travelling from source to drain.
+
Operation
Effect of gatetosource voltage
T h e channel of an N J F E T is a bar of sil
icon, whose resistance or conductance d e p e n d s u p o n its length, L, its width, 2a,
its depth, Z , and its conductivity. For a fixed drainsource voltage, the width of
the depletion layer, formed b e t w e e n t h e gate and t h e channel, is controlled by
a voltage applied b e t w e e n gate and source. Because of t h e high doping of the
P
gate and the reversebias across the gatechannel junction, t h e depletion
layer extends almost entirely in t h e channel. T h e gate becomes t h e control terminal
since by varying t h e reverse bias, b e t w e e n gate and source, and hence b e t w e e n
gate and channel for a fixed drainsource voltage, t h e width of the channel and
therefore the effective crosssectional area of the channel n o r m a l to the carrier flow
is changed.
The change in the crosssectional area of the channel causes a change in its
conductance. The process of controlling t h e conductance of the channel is k n o w n as
+
Section 10.1
301
Figure 10.1 (a) Perspective view of an NJFET; (b) simplified physical crosssection; (c) symbol for the and JFET; and (d) basic structure of a onegate
model.
302
Chapter 10
IV
R = resistance of channel
R
<D1
R = , > \
2
oV
ID2 <
oV
D2
IDI
= AV
R  R > R
3
fD3
(c)
Figure 10.2 Making the gate voltage more negative (a to c) at constant V results
in decrease of the width of the channel. Note the decrease of width at the source
end.
D
well as its cross sectional area decreases, and its resistance increases, resulting in a
decreasing drain current. This effect is illustrated in Fig. 10.2.
Effect of the drain voltage
The previous discussion may have left t h e
impression that t h e width of t h e channel is d e t e r m i n e d only by t h e gatesource volt
age. In fact, t h e drainsource voltage is an integral part of the process of channelwidth modulation.
A n increase of t h e drain voltage, with respect to a g r o u n d e d source for an
N J F E T at a fixed negative gate voltage, has obviously n o effect on t h e width of the
depletion layer at the source end of the channel. However, as we move from the
source towards t h e drain end of t h e channel, the reverse bias b e t w e e n t h e negative
gate and the channel, which is positive (with respect to the source because of the
positive drain voltage), is increased. This causes an increase in t h e width of the
depletion layer, hence a decrease in the effective width of the channel, and conse
quently a decrease in t h e crosssectional area of t h e channel. The channel, therefore,
has the smallest area at t h e drain end. This effect is exacerbated as the drain voltage
is increased, as shown in Fig. 10.3
Section 10.1
O
303
Vr. = W
V =V
n
Depletion layer
i n
= v
D2
'02
V >
D2
D1
v =rv
n
v =v
D
D3
oV
()
'D3
D7,>V
D2
(b)
Figure 10.3 Making the drain voltage more positive (top to bottom) results in a (a) decrease
of the area of the channel and (b) smaller increase of the drain current with increasing drain
voltage, as indicated by the decreasing slope with increasing voltage.
A s we indicated earlier, the channel has the smallest area at t h e drain end. A
further increase of the drain voltage causes t h e t o p and b o t t o m ends of t h e deple
tion layers, in a two gate model, t o touch at the drain end. This condition is k n o w n as
pinchoff.
Drain Current U p to Pinchoff
With the gate voltage fixed, an increase of
the drain voltage from zero, a n d also at small drain voltages, causes small changes in
the channel resistance so that the drain current increases fairly linearly with the
drain voltage.
Two conflicting p h e n o m e n a , which influence the current in opposite ways,
accompany further increase of the drain voltage for a fixed gate voltage. T h e resis
tance of the channel increases because of t h e decrease of t h e width of the channel
and simultaneously the drain voltage, which caused the increase of resistance, has
increased. We will later analytically d e m o n s t r a t e that the combined p h e n o m e n a
tend t o increase the drain current u p to the point where the channel is pinchedoff
at the drain end.
304
Chapter 10
SAT
D3
Figure 10.4 (a) Increase of drain voltage at V = 0 results in pinchoff. (b) Drain current
increases with V and saturates at 7, for V s= V. .
Q
Additional increase of the drain voltage, and as shown in Fig. 10.4, causes the
drain current to increase at a smaller rate and, eventually, increase of the drain volt
age, causes t h e channel to be pinched at the drain end. The drain voltage at the
onset of pinchoff is k n o w n as the saturation drain voltage, V . We illustrate the
onset of pinchoff by t h e sketches of Fig. 10.4. It is important to point out that pin
choff results from a particular combination of drain and gate voltages.
SXT
Beyond Pinchoff
The question becomes: O n c e pinchoff has occurred,
what h a p p e n s to the channel resistance and to the drain current as t h e drain voltage
is increased, so that V > V 7
For voltages greater t h a n the saturation drain voltage, further pinching of the
channel takes place and the pinchedoff region spreads out in t h e channel towards
the source, effectively isolating the channel from t h e drain, as shown in Fig. 10.4. The
D
SAT
Section 10.1
SAT
SAT
SAT
SAT
SAT
SAT
CG
SAT
SAT
V = V
S A T
V
(10.1)
We n o t e that the builtin voltage at the junction has not b e e n included. We will
consider that in the next section.
For every V t h e r e corresponds a V
at which pinchoff occurs. The differ
ence at pinchoff b e t w e e n V
( > 0 for an N J F E T ) and t h e corresponding V ( < 0
for N J F E T ) is always V . T h e pinchoff voltage, V , is a property of the particular
device.
We have also shown in Fig. 10.5, the locus of all the pinchoff points, labeled the
pinchoff line, where each point represents t h e difference b e t w e e n V
and V . This
difference is equal to V .
We concluded earlier that, by assuming that the length of t h e isolation region
is m u c h smaller t h a n the channel length L, the drain current remains constant at the
value where V = V . The region on the characteristic curves where V > V
is
G
SAT
S A T
SAR
SAT
SAT
306
Chapter 10
k n o w n as the saturation region. The region to the left of the pinchoff line in Fig. 10.5
and where V < V
is labeled the linear region, although only for small V , is the
variation of I with V being linear. Cutoff represents the condition corresponding
to V
= 0 and hence zero I .
W h e n the transistor is t o b e used as a n amplifier, the operating point is
selected to b e in the saturation region. T h e principal use of t h e J F E T in the linear
region is as a variable resistor and, in particular, in the region very n e a r the origin of
Fig. 10.5.
Since the gatechannel junction is operating as a reversebiased diode, minor
ity carrier electrons from t h e gate cross into t h e channel. However, because the dop
ing of the P gate is so high, the density of electrons in the gate is very small and the
n u m b e r of electrons injected into the channel is negligibly small compared to the
n u m b e r of electrons available in t h e channel.
It is to be n o t e d that the gate in the N J F E T has b e e n assumed to o p e r a t e at
V ss 0. It is possible to o p e r a t e t h e gate at a very small positive voltage, equal to or
less than the builtin voltage of the gatechannel. However, operating the gate at a
higher forward bias defeats the purpose of having the gate act as a control element
since a positive gate voltage may form a forwardbiased diode b e t w e e n the gate and
D
SAT
SAT
Section 10.2
the channel. This diode t h e n exhibits a low input resistance and causes an u n w a n t e d
gate current.
10.2 C U R R E N T  V O L T A G E C H A R A C T E R I S T I C E Q U A T I O N
Preliminary Conditions
We will now derive the relation b e t w e e n t h e drain current and the drain and gate
voltages. Before we do that, we establish certain basic equations related t o the
derivation. E q u a t i o n (5.30) gives, for an abrupt P N junction, the relation b e t w e e n
the depletion layer width and the voltages across the layer. For the dimensions
shown in Fig. 10.6 and at pinchoff, the depletion layer width, at the drain end from
channel to either of the gates, b e c o m e s t h e distance a and the resulting voltage drop
from drain to gate, which is t h e voltage across the depletion layer and in accordance
with E q . (5.30), b e c o m e s
+
DC
= V
+ V V
SAT
bi
= qN ay(2s)
(10.2)
V = qN aV(2s)
p
= V
SAT
+ V  V
bi
(10.3)
308
Chapter 10
( ) = p =
A
q[x N 2[a
n
 W(x)]Z
(10.4)
v
D e r i v a t i o n of CurrentVoltage Relationship
For the device structure shown in Fig. 10.6, the drift current of electrons that
constitutes t h e drain current I shown in Fig. 10.6 is given by
D
= qN v(x)A
= 2qN v(x)Z[a
 W(x)]
(10.5)
where is t h e channel depth, 2[a  W(x)] is the channel width at x, and v(x) is the
drift velocity of the electrons at a point in the channel and is given by
dV
= ;
() = 
(6)
W(x) = [2s(V
bi
 VJ/qNj^
(10.7)
where V is the builtin voltage across t h e channelgate depletion layer and V is the
voltage d r o p from gate to channel at any x. The voltage drop V is given by
bi
= V V
c
(10.8)
Section 10.2
Wipe) = [2e(V
V )/qN ] 1/2
+ V(x) 
(10.9)
By substituting, in E q . (10.5), for v(x) from E q . (10.6) and for W(x) from E q . (10.9),
we obtain
2 \o.5
dV
qN
(10.10)
We now separate the variables and V , and integrate along t h e channel from
= 0, V = 0 to = L, and V = V as
x
Vnr
I dx
2q^ N Z
n
qN
(v
+ v
bi
r dV
(10.11)
By using the fact that I is constant for all x, we perform the integration to
obtain
D
0.5
IL
= 2q^N z\aV
 (^
+ V  Vf
D
 (V  V Y ]
bi
(10.12)
bi
2q^ N Za/L\V
n
2 ( 2
3
\O.S
l(v
\qN a'
+ v
bi
y )
1 5
 (y
b l
v y ]
G
(10.13)
We introduce V from E q . (10.3), so E q . (10.13) can b e written as
2V (V
y
1
bi
+ v
y 2
2 (\'u
v
V,, 2
V
(10.14)
SAT
S A T
hi
SAT
bi
310
Chapter 10
V >0
U'(.v)
P~
X
xO
xL
(a)
(b)
Figure 10.7 Singlegate model showing (a) depletion layer and (b) pinchoff.
E X A M P L E 10.1
18
16
SAT
Solution
a) From Eq. (10.3). V =
p
14
The dielectric constant e for silicon is 11.8 X 8.85 X 10" F/cm, so that
19
"
16
1.6 X 10 X 10 X (10 )
2 X 11.8 X 8.854 X 1 0
 1 4
~ "
GO 
19
10 S
kT
c) From Eq. (5.17), V
= l"
bi
G = 4.16
(N N /nf)
A
1q34
= .026 ^7^20 = 0.838V
10
153 /0.838\3/2
3 \ 7.65
6.73mA
Section 10.2
e) From Eq. (10.2), for V = 2V, V =V  V + V = 7.65  0.838  2 = 4.812V, and for
V = 0, V
= 7.65  0.838 = 6.812V
G
SAT
bi
S A T
Additional Remarks
We remind the r e a d e r that for an channel JFET, V , V , and V are positive quan
tities, whereas V is zero or negative. E q u a t i o n (10.14) predicts that at V = 0, the
current is zero and all characteristic curves pass through t h e origin of the I  V
axes. The drain current is also zero for all V w h e n (V = V  V ) . By assuming
that V is m u c h smaller than V and for V = V , the current is zero, V
is zero,
and the locus of the current curve is the V axis. Based on E q . (10.2), the above con
clusion indicates that setting V = V causes the channel t o b e completely
pinchedoff from source to drain at all V , its thickness and cross sectional area are
both zero, its resistance b e c o m e s infinite, and t h e current is zero.
We define the threshold voltage or turnOFF voltage, V , as that value of V
that m a k e s V
= 0 and I = 0. This value of V turns the transistor OFF. The V
axis represents cutoff for V =s V , as shown in Fig. 10.5, so that the expression for
V i s d e t e r m i n e d from E q . (10.3) as
bi
bj
bi
S A T
SAT
V =V V
T
bi
(10.15)
S A T
S A T
^SAT
bi
(10.16)
^ 0
3 V
V
P
bi
REVIEW QUESTIONS
Q101 For a fixed V , explain how the drain current in an Nchannel JFET is reduced as the
gate voltage is made more negative.
Q102 For a fixed gate voltage, explain how the drain current is increased as the drain volt
age is increased.
Q103 Briefly define and explain in an equation the significance of the pinchoff voltage.
Q104 Explain the reason for using the gradual channel approximation.
Q105 What is the difference between the pinchoff voltage and the threshold voltage?
Q106 Explain how there is drain current after the channel is pinched off.
D
312
Chapter 10
HIGHLIGHTS
The JFET is basically a 3terminal device, which contains, in its simple form, a terminal
from which carriers originate, a channel through which they drift, a second terminal
that collects the carriers, and a third terminal that controls the flow of the carriers. The
sender is the source, the receiver is the drain, and the control is exerted by the gate.
The drain, in an Nchannel device, is biased positively with respect to the source. The
gate, which in the Nchannel JFET is a region, is biased negatively with respect to
the source and sits on top of the channel. By controlling the gatetodrain voltage,
hence the gate to the channel voltage, a depletion region is formed in the channel that
constricts the depth of the channel throughout its length. This process changes the
crosssectional area of the channel, hence the resistance of the channel, and conse
quently the current through the channel.
It is to be recalled that for all combinations of gate and drain voltages, including V =
0, V 0, a depletion region exists and account must be taken of the builtin voltage.
For very low values of V and a variety of values of V , the relation between the drain
current and drain voltage is linear and the device may be used as a variable resistor.
The value of this resistor is controlled by V .
An increase of V beyond the linear region for a fixed V results in further constric
tion of the channel and an increase of current accompanies the increase of V at a
slower rate as V increases. This process continues until the channel is pinchedoff at
the drain end and the current reaches its saturation value 7 , corresponding to a
drain voltage V .
G
SAT
SAT
EXERCISES
E101 An Nchannel doublegate silicon JFET, operating at 300K, has N = 10 cm~ ,
N. = 5 X 10 cm~ , a = 1.5, L = , and Z/L = 5. Determine: a) the builtin
voltage and b) the pinchoff voltage.
15
18
Ans: a) V = 0.816V
b) V = 1.72V.
hi
19
Ans: a) I = 0.21mA,
b) V = 1.02V.
10.3 C H A N N E L C O N D U C T A N C E A N D J F E T T R A N S C O N D U C T A N C E
The channel conductance,
V ,is
 V
relations at a certain
^ d I
/ d V
(10.17)
on V
S = G {1  [(V
d
bi
+ VD
V )/V r]
G
(10.18)
Section 10.3
( V V ), E q . (10.18) becomes
bi
(;
8 G [l[(V V )/V r}
d
bi
313
(10.19)
S A T
dip
dV Vr
 (^V/
constant
10 2
(  )
E X A M P L E 10.2
For the JFET of Example 10.1, determine:
a) The channel resistance for V 0 at V = 0.
b) The saturation drain current for V = 2V.
c) The transconductance at V = ~2V.
Q
Solution
a) From Example 10.1, we have
3
= 4.16
10"
= 2.78
10" S
4 A T = W X 10"
S A T
I 7.65
= 3.58mA
= 4.16
10
1 
/2.838\o.5
= 1.62
7.65
X 10*$
314
Chapter 10
10.4 S E C O N D A R Y EFFECTS
ChannelLength Modulation
In the previous section, we concluded that at t h e onset of pinchoff, the drain current
becomes 7
and remains a t that value as V increases b e y o n d V . However,
m e a s u r e d characteristics shown in Fig. 10.8 indicate that a gradual and slow increase
of drain current accompanies an increase of V b e y o n d pinchoff.
The n o n z e r o slope of t h e I  V characteristic, which is exhibited in the satu
ration region, is a result of the decrease of the effective channel length as V is
increased b e y o n d V . The decrease in effective L , resulting from t h e extension of
the depleted isolation region towards t h e source, reduces the channel resistance and
causes the current to increase. This increase is confirmed by E q . (10.16), in which G
is inversely proportional to L. This effect is in a way analogous to basewidth m o d u
lation of t h e B J T and is k n o w n as channellength
modulation.
S A T
SAT
SAT
Breakdown
Avalanche b r e a k d o w n occurs in a J F E T w h e n the reverse bias on t h e gatechannel
junction, at the drain end of t h e channel, equals the b r e a k d o w n voltage of t h e junc
tion, so that
V
= V  V  VBI
(10.21)
D
G
l
where V is the magnitude of t h e b r e a k d o w n voltage d e t e r m i n e d by the physical
properties of the junction and given by E q . (7.11). B r e a k d o w n results in a very sharp
K
BR
bi
10
Breakdown fixed
<
10
15
20
25
30
V (V)
D
Section 10.5
315
increase of the current and, furthermore, as shown in Fig. 10.8, at m o r e negative val
ues of V b r e a k d o w n occurs at lower values of V , as evidenced by E q . (10.21).
G
Variation in M o b i l i t y
O n e of t h e assumptions that we m a d e in deriving t h e expression for t h e currentvoltage characteristic is that t h e mobility of t h e electrons in t h e channel is constant
and therefore not d e p e n d e n t on t h e voltages applied to the channel.
The assumption is quite valid for long channels, those identified as having
L a. H o w e v e r , at smaller values of L and for high drainsource voltages in satu
ration, the electric field intensity in the ^direction of Fig. 10.6 is high and the mobil
ity decreases with increasing field intensity.
This effect was illustrated in Fig. 4.3, which showed the variation of drift veloc
ity with electric field intensity for silicon. We observed that at values of field inten
sity up to a b o u t 1 0 V / c m , the drift velocity in silicon increased in a linear m a n n e r
with an increase in field intensity, so that t h e mobility was assumed to b e constant.
A t higher values of field intensity, t h e slope of t h e curve decreased, indicating a
decrease of mobility. A t a value of field intensity of approximately 8 X 1 0 V / c m , the
velocity saturates at the t h e r m a l value and the relationship of = % is n o longer
valid.
Therefore, in a J F E T that has a short channel, and for a fixed drain voltage, the
higher electric field intensity decreases the mobility. This leads to lower values of
drain current through a decrease of G in E q . (10.14).
4
Temperature Effects
The major influence of t e m p e r a t u r e on the currentvoltage characteristic is exhib
ited t h r o u g h t h e decrease of mobility that accompanies an increase of temperature.
Mobility is d e t e r m i n e d by carrier scattering and t h e scattering at higher t e m p e r a
tures causes a decrease in the mobility, which results in a small decrease of current
at high temperatures. This is in contrast to the effect of t e m p e r a t u r e o n t h e collector
current of a BJT, which increases with t e m p e r a t u r e and in some cases m a y lead to
t h e r m a l runaway of t h e device.
10.5 S M A L L  S I G N A L E Q U I V A L E N T CIRCUIT
The smallsignal lowfrequency equivalent circuit represents the operation of the
transistor as changes in the gate and drain voltages are m a d e about an operating
point, on t h e characteristic, which is d e t e r m i n e d by I , V , and V . These changes
are initiated by a change in V , , which causes the changes in I and thereby V . In
general, we write
D
= I (V ,V )
D
(10.22)
316
Chapter 10
D + h = (D + v , V
ds
+v)
(10.23)
gs
where I is the operating point D C current, i is the incremental change in the cur
rent, v and v are the incremental changes in the drain and gate voltage respec
tively, and i is the total instantaneous current. By using Eqs. (10.22) and (10.23), the
current change b e c o m e s
D
ds
gs
h = h (V
+ v V
+v)  I
gs
(V , V )
D
(10.24)
^^
1!
. ...
III.
v
VD
ds
VG
(J.U.Z3)
= 8v
m
gs
1 0
2 6
( )
where g and g are the transconductance and the channel conductance respec
tively, which have b e e n defined in Section 10.3. Using E q . (10.26), we can draw the
smallsignal lowfrequency equivalent circuit shown in Fig. 10.9. T h e current and
voltages in E q . (10.26) represent instantaneous values, usually, of sinusoidal waves.
The equivalent circuit has b e e n labeled as linear so that a sinusoidal input gate volt
age will generate sinusoidal drain current and drain voltage. It is lowfrequency
since capacitances have b e e n neglected.
m
10.6 F I G U R E O F M E R I T O F THE J F E T
The Figure of Merit of an active device, such as a transistor, is a m e a s u r e of b o t h the
gain and the high frequency response. The highfrequency response of a device is
d e t e r m i n e d by its internal capacitances. We show in Fig. 10.10 t h e highfrequency
equivalent circuit of t h e J F E T . This circuit is m a d e u p of t h e lowfrequency circuit
plus the capacitances.
 \
Section 10.6
317
5
Figure 10.10 Highfrequency equivalent circuit of the JFET.
C = 2sZL/a
(10.27)
Za/2
q\i, ZaN
n
= RC
(10.29)
b e a p p r o x i m a t e d by
1
= 2 L /^a /V
 2 ^ r
_
c
qy a N
n
~^
(30)
318
Chapter 10
T h e expression in E q . (10.30) can also be obtained by determining t h e shortcircuit current gain and then finding the frequency at which that gain is unity. By
placing a shortcircuit across the output of Fig. 10.10 and applying a sinusoidal gate
current having an R M S value of I., we d e t e r m i n e the equation for the shortcircuit
current gain, / //., where / is the R M S value of the current through a shortcircuit
placed from drain t o source. By neglecting the current through C , as c o m p a r e d to
g V , where V is t h e R M S value of , t h e current gain becomes
gd
I /Ii
U M C
(10.31)
+ C)
gd
1 0
3 2
The expression for f , given by E q . (10.32), can b e reduced to that given by Eq.
(10.30) by making the following substitutions: replace g by the m a x i m u m value it
can have and that occurs at V = 0, replace 2a by a for the onejunction model,
neglect the small value of V , and replace the sum of the two capacitances by the
expression for C in E q . (10.27). The expression f o r / becomes
T
bi
 * & 
(10.33,
where V is the drainsource voltage and v is the m e a n drift velocity. Of course, the
derivation of E q . (10.34) is based on the assumption of constant mobility. In the
region where the velocity has saturated at v , the transit time becomes L/v . The
transit time is usually small c o m p a r e d to the R C time constant.
D
10.7 H I G H  F R E Q U E N C Y LIMITATIONS
The highfrequency limit of operation of a J F E T is d e p e n d e n t on the dimensions
and physical constants of the transistor. To improve the highfrequency response, let
us examine E q . (10.29) and n o t e how t h e p a r a m e t e r s have to vary.
1 . Decreasing the channel length, L, decreases the capacitance and increases g ;
hence, t h e r e is an improved gainbandwidth product. F r o m the point of view of
transit time, decreasing L has two desirable effects: (a) A t a fixed V , a lower L
increases % and hence increases the drift velocity, (b) The transit time
m
Section 10.7
HighFrequency Limitations
319
REVIEW QUESTIONS
Q107 At values of V greater than V
and for shortchannel JFETs, two phenomena
work in opposite directions, one to increase the current and another to decrease it.
Explain.
Q108 What is the major source of capacitance in the equivalent circuit of the JFET?
Q109 Why is the gatetosource connection opencircuited in the smallsignal equivalent
circuit?
D
S A T
HIGHLIGHTS
Increase of V beyond pinchoff causes a gradual increase of the current, based on the
assumptions listed, because of channellength modulation.
Breakdown at the reversebiased diode between gate and channel takes place beyond
saturation when the draintogate voltage causes a sufficiently large electric field that
is normal to the direction of current flow.
The gate voltage controls the drain current just as the emitterbase voltage controls
the collector current in a BJT. A BJT has a base current whereas the JFET has a negli
gible gate current, it being the current of a reversebiased diode.
The equivalent circuit of the JFET at low frequencies includes a transconductance g
and an output conductance g . The transconductance is the rate of change of drain cur
rent with gate voltage at the DC operating point in the saturation region of the output
characteristics. The output conductance is a measure of the change of the drain current
as the drain voltage is changed at an operating point. The equivalent circuit is relevant
when the device is used as an amplifier where operation is in the saturation region.
D
EXERCISES
E103 A doublegate Nchannel silicon JFET operating at 300K has N = 1 0 c m ,
N = I0 cm" , and a = 1.5. Determine: a) the builtin voltage, b) the effective
channel width for V = 0 and V = 0, c) the effective channel width for V =  I V
a n d V = 3V.
19
16
3
320
Chapter 10
Ans: a) V = 0.894V
b) w' = 2.32
c) W = 2
bi
E104
Ans: a) g = G
b) For FET gjl
m
= 3/V
= 38.6
E X A M P L E 10.3
'
. / "
Solution
^
2e.ZL
2 X 11.8 X 8.854 10~ X 250 x 10 x 25 X 10
a) From Eq. (10.27) C =
=
~
14
4veL
19
16
f = 253MHz
T
PROBLEMS
Unless otherwise indicated, all devices are doublegate silicon at = 300k.
10.1 An Nchannel silicon JFET has a gate doping of 5 X 10 cm~ , a channel doping of
10 cm~ , and width a = 0.2. Determine:
a)
the pinchoff voltage.
b)
the gate bias required to make the width of the undepleted channel equal to
0.15 at V = 0.
10.2 An Nchannel silicon JFET has N = 5x 10 cm" , N = 10 cm" , and a = 1.2.
Determine:
a)
the builtin voltage.
b)
the pinchoff voltage.
10.3 A Pchannel GaAs JFET has N = 5x 10 cm , = 10 cm" , and a = 0.2.
Determine:
18
17
15
17
18
17
Chapter 10
Problems
321
a)
b)
c)
16
fl
18
v/4
V =
V /9
0
C = 2V
10.6 An Nchannel silicon JFET has N = 10 cm , N = I0 cm , a = lm, L = 25,
and = 1mm. Determine:
a)
the builtin voltage.
b)
the pinchoff voltage.
c)
the drain current at V = 7V and V = 3V.
d)
I atV =2V.
10.7 Experimentally determined NJFET characteristics in saturation can be roughly
approximated by the expression
p
19
16
SAT
'SAT = I
(1 + V /V r
DSS
where I
is the saturation drain current at V = 0. Use the data of Prob. 10.6 to
determine I
from Eq. (10.14) and repeat part (d) of Prob. 10.6. Comment on the
accuracy of this approximation.
10.8 A Pchannel silicon JFET has a = 2.5, L = 8, = 300, N = 10 cm , and
N = 2 X 10 cm~ . Use = 4 0 0 c m / l  ^ in the channel. Determine:
DSS
DSS
ls
18
a)
The pinchoff voltage.
b)
The saturation drain current at V = IV.
c)
V at the current of part (b).
10.9 An Nchannel silicon JFET has = 24, L = 4, a = 1.2, N = 10 cm" ,
N = 5 X 10 cm . Use = 1200cm /Vs.
G
DS
19
15
At V = 3V, determine:
a)
the drain current at saturation.
b)
the drain voltage at saturation.
10.10 An Nchannel silicon JFET is to be used as a voltagevariable resistor at very low
values of V . The channel has resistivity of 5 ohmcm, = 1200cm /F5,
N = 1 0 c n r , a = 4, = , and L = 15. Determine:
a)
the minimum resistance achievable.
b)
the value of (V /V) that is required to double this resistance.
G
17
322
Chapter 10
17
S A T
bi
DSS
= lmA(V
= 0).
chapter 11
METALSEMICONDUCTOR
JUNCTIONS AND
DEVICES
11.0 I N T R O D U C T I O N
Metalsemiconductor junctions have long b e e n used as linearly conductive (ohmic)
metallic connections to devices and in integrated circuits. D e p e n d i n g on the metal
and on the type of semiconductor, such junctions may pass current easily into and
out of a junction (labeled ohmic contacts) or they may be rectifying in allowing easy
current flow in only o n e direction. The rectifying property has m a d e possible b o t h
the rectifier diode, k n o w n as the Schottky barrier diode, and t h e M E S F E T (metalsemiconductor F E T ) .
We are interested in studying the current voltage characteristics and proper
ties of such a junction. We begin, just as we did with the P N diode, by examining the
properties of the contact, such as the potential barrier, the depletion layer, and the
mechanisms of carrier transfer. These basic properties are uniquely d e t e r m i n e d by
the energy barriers that are formed w h e n a metal and a semiconductor are brought
into intimate contact.
Fabrication of t h e Schottky diode and t h e M E S F E T is discussed in Sec. 11.6.
11.1 E N E R G Y  B A N D D I A G R A M S OF M E T A L A N D N  S E M I C O N D U C T O R
B e f o r e Contact
The b a n d diagrams for a metal, such as gold, and an Ntype semiconductor (N sili
con) that are not in contact are shown in Fig. 11.1. In order to compare the relevant
energy levels within and b e t w e e n the two solids, we seek an energy level that is
c o m m o n to both solids and that is fixed with respect to both. That energy level is the
vacuum level E . T h e vacuum level, E , is defined as t h e energy that an electron is
Q
323
324
Chapter 11
Semiconductor
N(E)
1
(0
Figure 11.1 (a) Energy levels of metal (gold) and semiconductor (silicon) (not in contact),
(b) Fermi function, (c) Carrier density distribution function. Above the E level of silicon and
relative to E , there is a larger density of electrons in silicon than in gold. Crossed areas are
filled with electrons..
c
assumed to have if it were at rest outside and just free of the solid. We remind the
r e a d e r that E is t h e zero level of energy and all other energy levels shown in the fig
ures represent negative energies.
T h e Fermi energy represents the average energy of a n electron in t h e system.
For a metal, the Fermi energy is the average energy of the most energetic electrons.
The energy difference b e t w e e n the vacuum level and the Fermi level is labeled the
work function of the solid. The work function is therefore defined as the energy
required to move an electron from the Fermi level to E , w h e r e it is at rest and free
of the influence of the solid.
In an extrinsic semiconductor, the location of the Fermi level is d e t e r m i n e d by
the degree of doping of the semiconductor. Therefore, the Fermi level is not located
at a fixed level with respect to the conduction and valence bands. In a given metal,
however, the Fermi level is located slightly above the b o t t o m of the conduction
band and at a fixed energy separation from the vacuum level E . In a semiconduc
tor, the b o t t o m of t h e conduction band, E , is located at a fixed energy separation
from E . This energy difference b e t w e e n E and E in t h e semiconductor is labeled
the electron affinity, which is d e n o t e d by q\ , where the symbol \ is t h e G r e e k letter
chi. The affinity is therefore the energy n e e d e d to move an electron from t h e bot
t o m of the conduction b a n d and place it at rest outside t h e solid.
F r o m Fig. 11.1, we n o t e that the work function for the metal is (E ? ) and
for the semiconductor is (E  q<t> ), where q<i> is the energy location of the Fermi
level below E . We observe that if <i> is less than <t> while t h e solids are separated,
Q
{)
Section 11.1
325
t h e n the electrons in the metal have, on the average, a total energy that is higher
than that of t h e electrons in the semiconductor. O n t h e other hand, if is greater
than , t h e electrons in the semiconductor have an average total energy that is
greater than t h e electrons in the metal. In Fig. 11.1, we have used a metal and a
semiconductor such that <> > , so that the work function of Ntype silicon is
smaller than that of gold (<> = 4.75eV) and the affinity of silicon is 4.15eV.
Also in Fig. 11.1, we have shown the Fermi distribution function variation ver
sus energy, centered at E which shows the probability of occupancy, f(E), of a state,
at energy E, by an electron. The density of states distributions functions, N(E) (refer
to Sec. 3.1), shown for electrons, have their zero energy level located at the level of
the b o t t o m of t h e conduction b a n d s of the metal and of the semiconductor. The dis
tributions of the densities of electrons are formed from the product N(E)f(E)
and
are shown within t h e solid lines of the curves of N(E). In order to c o m p a r e at t h e
same energy level the densities of electrons in the metal to those of t h e semiconduc
tor, t h e d a r k e n e d areas are shown m e a s u r e d above the b o t t o m of the conduction
b a n d of the semiconductor. Since t h e work function of the semiconductor is smaller
t h a n that of the metal and t h e average energies of electrons in the semiconductor
are higher t h a n those of t h e metal, it stands to reason that above the energy level,
E of t h e semiconductor, the density of electrons in the semiconductor is greater
than that of electrons in t h e metal.
A t the instant of contact and since the Fermi level of the silicon is higher than
that of t h e metal, electrons will transfer from the semiconductor into the metal until
equilibrium is r e a c h e d and the Fermi levels are aligned. A s a result, and at t h e r m a l
equilibrium, t h e semiconductor is charged positively with respect to the metal. The
effects of intimate contact on t h e energy levels and energy b a n d s is shown in Fig.
11.2. After contact and at t h e r m a l equilibrium, t h e r e will b e a continuous flow of
electrons at t h e same rate in b o t h directions with t h e result that the net current
across the junction is zero.
326
Chapter 11
Metal
q%
Us
N(E)
EF
Ec
N(E)
(a)
dEj
dx
Pc
C/cw?
qN
Q~ = q
=q
Nw
D
Qi C/cm2
w
)
(c)
Figure 11.2 (a) Conditions at thermal equilibrium after contact, (b) Charge density
distribution, (c) Electric field. Dark areas in the distribution functions represent the electrons
that have enough energy to cross the junction.
Section 11.1
327
semiconductor. This is shown by the equal areas above E that represent the density
of electrons that are able to cross t h e junction.
After contact and simultaneous to t h e bending of the bands, an energy barrier
qV is formed in the semiconductor at t h e surface of contact with the metal. In con
trast to t h e depletion layer, having width x = W , that is formed in t h e semiconduc
tor, a surface sheet of electrons is formed in the metal. T h e r e can b e n o depletion
layer and n o voltage d r o p in t h e metal. This is because the metal is assumed to b e a
perfect conductor with zero resistance, which does not permit t h e formation of an
electric field and the sustaining of a voltage drop.
A t t h e r m a l equilibrium there will b e a flow of electrons in b o t h directions and
the net current across the junction is zero. Just as with t h e P N junction diode, there
are potential barriers, which the electrons must overcome, on b o t h sides of t h e junc
tion. The barrier for electrons in t h e metal shown in Fig. 11.2 is k n o w n as a Schottky
barrier.
O n e of the i m p o r t a n t characteristics of a Schottky barrier is the barrier height
q& , which, as we observe from Fig. 11.2, is t h e energy difference b e t w e e n the
aligned Fermi levels and the semiconductor b a n d edge at t h e surface with t h e metal.
The Schottky barrier, q , represents the energy barrier that electrons in the metal
must o v e r c o m e to move into the semiconductor. This barrier, shown in Fig. 11.2 at
the surface of contact, is written for an ideal contact as
Q
bi
= ?(*  x )
(iii)
The barrier for t h e electrons that are in t h e bulk of the semiconductor, which
prevents t h e m from moving into t h e metal and which they must overcome, is qV
and is given by
bi
= (,,)
(H.2)
The builtin voltage, V , can b e expressed in terms of the barrier, , and the poten
tial difference, b e t w e e n the Fermi level and E , defined by E q . (3.18) as (kT/q)
in NJN , where n = N so that
bi
(H.3)
TTiis barrier is smaller than t h e barrier , which the electrons in t h e metal face.
However, at thermal equilibrium and because of the larger density of electrons in
the metal, as shown by the tail of t h e distribution curve, there are as m a n y electrons
in the metal as t h e r e are in the semiconductor that have energies greater than the
barrier they face. These electrons are continuously in motion, so that t h e r e is an
equal current of electrons from t h e metal to t h e semiconductor as there is from the
semiconductor to the metal.
We italicized t h e phrase ideal contact earlier, prior to E q . (11.1), to indicate
that calculated values for , using E q . (11.1), are in general considerably smaller
than actual values. This is because the surface states that are p r o d u c e d as a result of
the disruption of the semiconductor crystal lattice w h e n t h e contact is formed tend
to have an important effect on the barrier height.
M e a s u r e d values for the barrier height, , of various metals on silicon and
N  G a A s are shown in Table 11.1.
328
Chapter 11
Si(N)
Si(P)
GaAs(N)
Au
PtSi
0.72eV
0.58eV
0.80eV
0.80eV
0.34eV
0.90eV
0.67eV
0.45eV
0.80eV
0.85eV
11.2 S C H O T T K Y B A R R I E R D I O D E
Rectifying M e t a l  N S e m i c o n d u c t o r Contact
Metal
E
Semiconductor
E
0 x,
Metal
1
(a)
Semiconductor
TOT
(b)
Figure 11.3 (a) Schottky junction with forward bias applied, (b) Circuit to indicate voltage
reference. (Dark areas refer to electron density.)
Section 11.2
Metal
Figure 11.4
reference.
329
Semiconductor
(a) Schottky junction with reverse bias applied, (b) Circuit to indicate voltage
A n applied voltage causes a change in the bending of the bands in the semi
conductor and a corresponding change in t h e electric field in the semiconductor at
the junction with the metal. T h e barrier in t h e semiconductor is reduced w h e n a for
ward bias is applied which m a k e s t h e metal positive with respect to the semiconduc
tor. This bias reduces the electric field and the degree of bending of t h e bands, as
shown in Fig. 11.3(a). A s the barrier is reduced, m o r e electrons, as c o m p a r e d to ther
mal equilibrium, cross from the semiconductor to t h e metal. O n t h e other hand, t h e
n u m b e r of electrons that cross from the metal to the semiconductor is unchanged
from the n u m b e r that crosses at thermal equilibrium because the barrier height,
qQg, they have to s u r m o u n t is unaffected by the applied voltage and unchanged
from its t h e r m a l equilibrium value of
330
Chapter 11
Figure 11.5
W = V(2e/qN )(V
D
_ V ) for N N
bi
(11.4)
(11.5)
where the applied voltage, V , is positive for forward bias and negative for reverse
bias. T h e symbol W replaces x in E q . (5.20) for based conditions.
The charge density in t h e depletion layer, assuming it consists uniformly of
d o n o r ions, and t h e corresponding capacitance per unit area are obtained as
n
Q
C
J =
= qN W
= V2qeN (V
dV = V ( e / V ) / 2 ( y
?
 V)
(11.6)
 V ) = /W
(11.7)
hl
b i
= 2(V
hi
 V )/qeN
a
(11.8)
E X A M P L E 11.1
ls
The silicon in an aluminumsilicon Schottky barrier diode has N = l() cm \ Determine the
builtin voltage, V .
D
Solution We determine > from the density of states in the conduction band .. and the doping
density N as
;
Section 11.2
= (kT/q)
(n
331
(NJN )
D
= .0259 in
3 22 X lO ''
'
= 0.269V
5
The builtin voltage is determined from Eq. (11.3) by using Table 1.1 to obtain
V.. <l>,  = 0.72
0.269 = 0.451V
Rectifying M e t a l  P S e m i c o n d u c t o r J u n c t i o n
The discussion so far has b e e n limited to conditions at the contact b e t w e e n a metal
and an semiconductor. A t that contact, the work function of the metal is greater
t h a n that of t h e semiconductor. We concluded that a barrier exists at this contact
and is rectifying, allowing passage of current in only one direction. In a later section,
we will show that w h e n the work function of the metal is smaller than that of the
semiconductor, t h e contact is nonrectifying and is labeled ohmic. A n ohmic contact
provides n o barrier to the flow of current and allows easy current flow in both direc
tionsfrom metal to semiconductor and from semiconductor to metal.
We will now consider the contact between a metal and a semiconductor,
wherein < and the work function of the metal is smaller than that of the
m
332
Chapter 11
Metal
Vacuum level
Semiconductor
lis
(a)
(b)
qN
(c)
W
(d)
Section 11.3
333
Ptype semiconductor
(c)
Figure 11.7 Energy band diagrams for metal Nsemiconductor (<I> > ) and
metal semiconductor ( < ) at (a) equilibrium, (b) forward bias, and
(c) reverse bias.
m
11.3 C U R R E N T  V O L T A G E C H A R A C T E R I S T I C S O F M E T A L
NSEMICONDUCTOR SCHOTTKY DIODE
Simplifying A s s u m p t i o n s
Before proceeding with analytical relations for t h e currentvoltage characteristics
and in o r d e r to simplify the analysis, we m a k e t h e following assumptions:
1 . All electrons in the conduction b a n d move with the same average thermal
velocity and t h e motions are randomly distributed in direction.
2. All electrons in the semiconductor, which are incident on the barrier, cross into
the metal.
334
Chapter 11
(h
T h e r m a l Equilibrium Currents
The energies of t h e electrons at the surface of the conduction b a n d of the semicon
ductor are separated from the Fermi level by q so that by using E q . (3.18) the
density of surface electrons, n , at t h e r m a l equilibrium is given by
B
n = 7V exp(^)
s
(11.9)
where is the density of conduction b a n d states that are assumed to b e all located
at t h e b o t t o m of t h e conduction band.
A t t h e r m a l equilibrium and from Fig. 11.2, the relation for q is found to be
B
9 * B = 4 bi + ( c
 F)
1 0
= N exp [(E
c
 E )/kT]
(11.11)
n = N [exp (qVJ/kT]
s
[exp {E
 E )/kT]
(11.12)
n = N exp(qV /kT)
s
(11.13)
ms
I^qAnpJA
(11.14)
*From D. Pulfrey and N. Tarr, Introduction to Microelectronic Devices, 1993, p. 271, Prentice
Hall. Reprinted by permission of Prentice Hall, Upper Saddle River, New Jersey.
Section 11.3
335
A t thermal equilibrium, t h e diode current is zero, so that we can set the cur
rent from the semiconductor to the metal equal and opposite to t h e current from
the metal to the semiconductor as
4 , = ~Ls
= ~<lAn vJ4
(11.15)
The expression for the currents at thermal equilibrium can also be written in
terms of the doping density N by using E q . (11.13).
D
= " 4 s = ~qA(vJ4)
exp {qVjkT)
(11.16)
Currents W i t h Bias A p p l i e d
We have indicated earlier that w h e n a voltage is applied to t h e junction, the voltage
appears totally in t h e semiconductor and the barrier to electron flow from t h e semi
conductor to the metal is changed, whereas the barrier to electron flow from the
metal to the semiconductor is unchanged at q& Applying a forward bias to the diode by connecting the metal to the positive
terminal of the voltage source increases the density of t h e electrons that can cross
from the semiconductor, while keeping the density of electrons that can cross from
the metal to the semiconductor at the thermal equilibrium value. Conversely, apply
ing a reverse bias to the junction increases the height of the barrier that electrons at
the semiconductor surface must s u r m o u n t in order to cross into the metal. Again,
this does not change t h e height of t h e barrier that electrons in the metal face.
Therefore, the current with reverse bias, which m a k e s the metal negative with
respect to the semiconductor, is the current of electrons that cross from the metal to
the semiconductor at thermal equilibrium.
We will consider the positive reference direction of current to be from t h e
metal to the semiconductor across t h e junction. With reverse bias applied, t h e diode
current is from semiconductor to metal and is d e t e r m i n e d by using E q . (11.9) in Eq.
(11.14) as
B
4 s = ~qA(vJ4)N
(11.17)
n = N
s
exp [q(V
bi
= [N
exp {qVjkT)]
VJ/kT]
[exp qVjkT]
(11.18)
w h e r e V is t h e applied voltage.
By replacing the first factor in the right h a n d side of E q . (11.18) by its equiva
lence from Eqs. (11.9) and (11.13), we have
a
[exp qVjkT]
(11.19)
336
Chapter 11
(11.19), so that we have the expression for the current from metal to semiconductor
at forward bias given as
4 s = qAivjyj/4
exp
(11.20)
I = 4
+ ' s m = qA(vj4)N
exp
(11.21)
(q<S> /kT)
B
112
and
/ = ART [exp
where R is qv N
(<& /kJ)]
B
[exp(qVjkT)
 1]
(11.22)
/4.
th
(11.23)
/ = ART
s
(11.24)
11.4 C O M P A R I S O N O F S C H O T T K Y D I O D E W I T H P+N D I O D E
A t this point, we highlight t h e advantages and disadvantages of t h e Schottky diode
by comparing it with the P N diode. A s we indicated earlier, most Schottky barrier
diodes used in integrated circuits use platinum silicide on silicon so that the bar
rier height q<i> is about 0.9eV.
We will use an example to illustrate t h e o r d e r of magnitudes of the currents
and voltages for a Schottky barrier diode and a P~N diode.
+
E X A M P L E 11.2
A Schottky barrier is formed from platinum silicide on NSi, which has a doping density of
10 cm~ and an area of 10~ cm . A PN diode has the same area and N = 10 cm ',',, = 10 cm~ ,
and i = = lxs.
16
19
16
Section 11.4
337
l5
0n
:;
::':;::
:;
' ::
+
Because the Schottky diode has a much higher current density than the P N
diode, its t u r n  O N voltage is m u c h lower. The current voltage characteristics are
c o m p a r e d in Fig. 11.8. It is also important to n o t e that PtSi forms a high barrier
whose advantage is a smaller reverse saturation (leakage) current than that caused
by aluminum or tungsten.
O n e of the important applications, as d e m o n s t r a t e d from the above calcula
tion, is the use of the Schottky diode in lowvoltage, highcurrent rectifiers.
The lower cutin voltage also m a k e s t h e Schottky diode a valuable device used
to clamp t h e collector to the base of a transistor at about 0.4V ( P N P ) , making the
the emitter to the collector voltage approximately 0.3V, and thus preventing the
transistor from going into d e e p saturation. The use of the Schottky diode connected
from collector to base considerably increases the switching speed of t h e B J T by up
I(mA)
from metal
Figure 11.8 Currentvoltage characteristics of PN and Schottkybarrier diodes
and symbol for SBD.
338
Chapter 11
PN
Figure 11.9 Reverse recovery transient for (a) P N junction diode and a
(b) Schottky diode. Variations with time of currents and voltages.
Section 11.4
339
REVIEW QUESTIONS
Q l l  1 How is the work function of a metal determined?
Q l l  2 Show, in equation form, how the work function of a semiconductor varies with the
doping.
Q l l  3 The Fermi level of a metal does not change when electrons are added to the metal.
Why?
Q l l  4 Explain why a depletion layer is not formed in the metal at a metalsemiconductor
contact.
Q l l  5 Why is there no diffusion (storage) capacitance in a Schottkybarrier diode?
Q l l  6 Why is the Schottkybarrier diode much faster, in switching, than the PN diode?
Q l l  7 What requirement must be met to make tunnelling possible?
Q l l  8 Identify the three uses to which metals are put in electronic circuits.
HIGHLIGHTS
For the same applied voltage, the Schottkybarrier diode (SBD) conducts considerably
larger forward and reverse currents than a PN junction diode.
340
Chapter 11
The PN junction diode is a minority carrier device, whereas the SBD is a majority car
rier device.
EXERCISES
15
3
0.
Ans:
a) V = 0.5V
bi
0.57
E l l  2 ( a ) Calculate the values of the two capacitances of the PN diode of Example 11.2 at 7 =
5mA. (b) Repeat (a) for the SBD.
Ans:
a) C = 68pF, C = 0.193
j
b) C. = 72pF, C = 0
s
11.5 N O N  R E C T I F Y I N G O H M I C CONTACTS
In Section 11.2, we showed that t h e Schottky barrier diode represents a rectifying
contact just like the P N junction diodes. A rectifying metalsemiconductor junction
permits easy current flow w h e n t h e junction is forwardbiased. A Schottky diode is
formed w h e n the metal work function is greater than that of the semiconductor
or w h e n t h e metal work function is smaller t h a n that of t h e semiconductor. Such
diodes, as indicated earlier, have special applications.
A n o t h e r major application of metalsemiconductor junctions is in the forma
tion of nonrectifying ohmic contacts.
In the fabrication of discrete devices and in the interconnections of integrated
circuits, it is necessary to establish ohmic metallic contacts to t h e devices, to the con
nections b e t w e e n t h e semiconductor region and its external terminal, and to inter
connect elements in an integrated circuit. Such contacts should not in any way
interfere with t h e operation of the device or a circuit and therefore should exhibit
negligible resistance to the flow of current into and out of the device. This resistance
should b e negligible c o m p a r e d to the resistance of t h e bulk regions of t h e device.
T h e contact should consequently support a negligible voltage d r o p c o m p a r e d to the
drop across t h e active region of the device. Such contacts are k n o w n as ohmic con
tacts.
+
M e t a l  S e m i c o n d u c t o r O h m i c Contacts
A n essential condition that is m e t by an ohmic contact is to permit easy and u n o p
posed transfer of majority carriers b e t w e e n the metal and semiconductor.
A n ohmic contact is formed w h e n the metal work function is smaller t h a n t h e
semiconductor work function with < . E n e r g y b a n d diagrams for t h e metal
Section 11.5
341
Figure 11.10 Energy band diagrams for metal Nsemiconductor contact with
<E> < (a) before contact, (b) after contact, (c) negative bias on the
semiconductor, and (d) positive bias on the semiconductor.
m
342
Chapter 11
Section 11.5
Metal
343
Semiconductor
Electron tunnels
c
(a)
Electron tunnels
(b)
Metal
/V region
type
semiconductor
(a)
(b)
+
Figure 11.12 (a) Band diagram for a metalN N~ ohmic contact and (b) currentvoltage characteristics of a Schottky barrier diode and an ohmic contact.
344
Chapter 11
The energy b a n d diagram, showing tunnelling in b o t h directions, for a metalN  N ~ contact is shown in Fig. 11.12(a).
+
11.6 T H E M E S F E T
The basic structure and operation of the M E S F E T are identical to the J F E T . While
the main disadvantage of the J F E T is its low gainbandwidth product, two major
differences in t h e fabrication of the M E S F E T m a k e it m o r e attractive. In the first
place, the semiconductor gate of t h e J F E T is replaced by a metal to form a Schottky
barrier. This advantage is related to the simplicity of the Schottky barrier and since
diffusion is not used, fabrication to close geometrical tolerances and smaller chan
nel lengths is m a d e possible. The second advantage is related to t h e use of N  G a A s
for the channel in which t h e electron mobility is five times that of silicon (only at
low fields of a few k V / c m ) . B o t h of these advantages serve to decrease the transit
time of electrons from source to drain and provide a significant increase in the gainbandwidth p r o d u c t of t h e M E S F E T s , thus leading to their use in monolithic
microwave integrated circuits and in high speed digital circuits. For t h e same dimen
sions, G a A s M E S F E T s have a speed advantage over Si M E S F E T s of about a factor
of 3. In addition, t h e use of semiinsulating G a A s provides excellent isolation
b e t w e e n adjacent devices.
Fabrication of t h e M E S F E T
The starting material of the substrate is a G a A s wafer cut from an ingot p r o d u c e d
by the Czochralski m e t h o d and chromium doped, which places the Fermi level n e a r
the center of the bandgap, resulting in a material having a high resistivity of t h e
order of 10 ohmcm. It is thus k n o w n as semiinsulating gallium arsenide. Devices
and circuitconnections m a d e on substrates of semiinsulating G a A s have lower
capacitances, which lead to high speeds in integrated circuits.
Cross sections of a depletion M E S F E T and an adjacent Schottky barrier,
formed on t h e same wafer, are shown in Fig. 11.13.
By using photolithography to define the regions, ion implantation is used to
form the drain, source, and channel of the M E S F E T and the base of the Schottky
8
Au/Ge/Ni
Ohmic contact
Al Gate
Au/Ge/Ni
ohmic contact
Al Schottky
barrier
Au/Ge/Ni
ohmic contact
Schottky diode
Section 11.6
The MESFET
345
15
Photoresist
Silicon wafer
Metal film
Metal
Figure 11.14 Illustrating the liftoff
technique.
346
Chapter 11
M o d e s of O p e r a t i o n
M E S F E T s are m a d e in two types, k n o w n as t h e enhancement (normallyoff) type
and depletion (normallyon) type. The two m o d e s are not possible in the same
device. Thus, we are referring to two different types of devices whereby the type is
characterized by the width of the channel that is formed u n d e r the gate. The width
of the channel of t h e depletiontype device is about double that of t h e e n h a n c e m e n t
type.
In t h e M E S F E T , just as in the JFET, the gate voltage modulates the width of
the depletion region in the semiconductor of t h e Schottky barrier diode (formed by
the gate and the channel) and h e n c e the resistance and therefore the resulting elec
tron flow from source to drain in an Nchannel device. In the e n h a n c e m e n t device,
the builtin voltage causes the channel to be completely closed at zero gatesource
bias (assuming V = 0) and therefore o p e r a t i o n of the device requires a positive
V . In t h e depletion device, with V = 0 (and V = 0), the depletion region extends
only partly across the channel, as shown in Fig. 11.15, and it b e c o m e s possible to
m o d u l a t e t h e width of the channel and decrease it by applying a negative voltage to
the gate and by applying a positive voltage to t h e drain.
D
Threshold V o l t a g e
We will identify two reference voltages for t h e M E S F E T by reproducing E q . (11.4)
for the width of the depletion layer at a Schottky barrier given as
(11.4)
In a M E S F E T , V represents the voltage that is applied from metal to semi
conductor bulk and is a function of b o t h V a n d V . For V = 0, V = V , while V
is t h e builtin voltage across t h e depletion layer. It is a positive quantity for both
types of devices.
We d e t e r m i n e d from E q . (10.3) that for a channel width a, the voltage that
must exist across the depletion region to completely pinch the channel is defined as
the pinchoff voltage and is given by
a
SchottkyBarrier gate
Source
1
,\'+
11
/Vchannel
a
Semi insulting GaAs substrate
hi
Section 11.6
V
The MESFET
(11.25)
347
a qN /2e
D
(11.26)
However, w h e n t h e builtin voltage is m o r e t h a n sufficient to close the chan
nel, a positive voltage must b e applied to t h e gate in order to reduce the width of the
depletion layer so that t h e channel is just closed. For this condition, V is greater
t h a n V and the threshold voltage is positive, given also by E q . (11.26). This is what
h a p p e n s in an e n h a n c e m e n t type device. To avoid confusion, a clear understanding
of the relations derived in this section is obtained by assuming that V = 0. The
relations in Eqs. (11.25)(11.26) are valid, however, regardless of the value of V .
Thus, t h e threshold voltage, V , is positive for an e n h a n c e m e n t device and neg
ative for a depletion device, and the relations in E q . (11.26) apply to b o t h types. The
conditions discussed above are illustrated by Fig. 11.16.
bi
D e p l e t i o n Device Characteristics
We observe from Fig. 11.15 that the depletion region b e c o m e s wider as we m o v e
from source to drain. This is due to t h e effect of the drain voltage, which increases
the reverse bias across t h e Schottky junction and has its highest value at the drain
end, causing the channel to have its smallest width at t h e drain.
A t low values of V , and with an increase of V , the narrowing of t h e channel,
at constant V , does not cause a sufficient increase of t h e resistance of t h e channel
for that change of resistance to have an important effect on the drain current. Since
it is assumed that for a constant resistance the current is directly proportional to V ,
the current increases linearly with an increase in V . A t higher values of V , the nar
rowing of t h e channel due to an increase of V becomes important, and although
the current continues to increase with an increase of V , the increase is slower and
therefore less linear. These changes are illustrated in the characteristics shown in
Fig. 11.17(a). A less negative value of V for a constant V decreases the width of
the depletion layer and increases the width of t h e channel, causing an increase of its
cross sectional area and a consequent decrease of resistance. Thus, a higher drain
current results.
D
E n h a n c e m e n t Device Characteristics
In this device, the channel is pinched off (completely closed) by t h e builtin voltage
at t h e r m a l equilibrium. To o p e n u p t h e channel and to reduce t h e depletion layer
width, a positive voltage has to b e applied to the metal gate at zero drainsource
voltage. Because the applied voltage may cause the Schottky junction to b e forwardbiased, it becomes i m p o r t a n t to limit the forward current of that junction. With a
348
Chapter 11
V =0
V <0
V =0
V <0
V =V
V .<0
q(V V )
bi
= qV
q<y V )
bi
(a)
V
V >0
T
V =V
n
V >V
qv
bi
q(v ~v )
bi
= v
q
q(v v )
bi
(/>)
Figure 11.16 Relations between V , V , and V for (a) a depletion device and (b)
an enhancement device.
T
bi
positive voltage at the drain V > 0, the largest forward bias and therefore the
largest gate current occur at t h e gatesource end of the channel. A forward current
at a forward voltage has the effect of reducing t h e input resistance of the device
since a high input resistance is one of t h e advantages of these devices.
The input gate current can be limited by keeping t h e forward bias of the junc
tion below the t u r n  O N voltage, identified in Fig. 11.8. The t u r n  O N voltage of
G a A s Schottky diodes is of t h e o r d e r of 0.7V, c o m p a r e d to the 0.4V t u r n  O N volt
age of silicon Schottky diodes.
The currentvoltage characteristics of an e n h a n c e m e n t type device are shown
in Fig. 11.17(b).
D
Section 11.6
(a)
The MESFET
349
(b)
Figure 11.17 Currentvoltage characteristics and symbol for (a) depletion device
and (b) enhancement device.
Relations B e t w e e n t h e V o l t a g e s
With respect to t h e currentvoltage characteristics shown in Fig. 11.17, we will intro
duce the effect of the drainsource voltage into the relations for V , V , V , and V
It is to b e n o t e d that both t h e gate voltage and the drain voltage, herein referred to
as V and V , are m e a s u r e d with respect to the source, which is assumed to be
g r o u n d e d and connected to the substrate.
B o t h V and V and therefore V a r e constants for a given device. T h e two
other voltages that control the width of the depletion layer and hence the extent of
constriction of t h e channel are V and V .
We first consider t h e relations for the depletiontype device where V is nega
tive, so that the relation b e t w e e n the voltages that will cause pinchoff at the drain
end of t h e channel is
bi
bi
(11.27)
K : " Vr
= VSAT
V
y
 V
v
(11.28)
bi
where V
refers to t h e drain voltage at the point of saturation of the drain current.
For a negative V , where \V \ < \V \, pinchoff occurs at a smaller value of V and
hence a smaller current, as shown by Fig. 11.16.
For V = 0, the channel is pinchedoff t h r o u g h o u t its length when V = V
V so that V is a particular value of V and V is negative and a constant for a
SAT
bi
350
Chapter 11
bi
V =V V
T
(11.29)
2Vp(V ,
V \V2
Vr.
VP
'SAT
^ 0
v..
+ v</
Lebih dari sekadar dokumen.
Temukan segala yang ditawarkan Scribd, termasuk buku dan buku audio dari penerbitpenerbit terkemuka.
Batalkan kapan saja.