Anda di halaman 1dari 497

APPENDICES

A: UNITS
FUNDAMENTAL QUANTITIES
Quantity

Unit

Symbol

Capacitance (C)

Farad

Conductance (G)

Siemens

C u r r e n t (I)

Ampere

Electric Charge (Q, q)

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)

Ohm-m

-m

Velocity (v)

Meters/second

m/s

Wavelength ()

Meters

465

466

Appendix : Physical Constants and Factors

B: PHYSICAL CONSTANTS AND FACTORS


B.1. PHYSICAL CONSTANTS
Quantity

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

k = kilo, = mega = 10 , G = giga = 10 , m = milli = 10~ ,


10"
10
pico
micron = 10 , = n a n o
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

C: THE DENSITY OF STATES N(E)


By assuming the free electron m o d e l for a metal, the energy states for electrons will
be the same as for electrons in a three-dimensional potential well where the poten
tial is V = 0 inside the well and infinite elsewhere.
Schrodinger's equation in rectangular coordinates inside the well for electrons
having constant total energy becomes
2

8 \
+ + j + = 0
dx
dy
dz
h

27

(C1)
'

This equation is solved by the m e t h o d of separation of variables by assuming


* =

g (x)g (y)g (z)


x

(C.2)

Appendix C: The Density of States N(E)

467

By substituting this solution in E q . ( C . l ) we have


2

1 dg

1 dg

g dx

8-n mE

g ^f

1 dg _
g dz

Since the right-hand 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)

where n , n and n are q u a n t u m n u m b e r s and A , B, and C are new constants.


Using the above expressions in E q . (C.3) we have
x

,
,
,
n\ + n) + n =

8mEa

(C.7)

In the three coordinate system that contains , , and n , we assume that


their m a x i m u m values are N , , and N . T h e reason for this is to d e t e r m i n e the
n u m b e r of states that have energies less t h a n some value given by

^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

Appendix C: The Density of States N(E)


n,
y

Figure C.F.I Section of sphere having n , n , and n


as coordinates.
x

By substituting Eq. (C.8) in E q . (C.9), we obtain


/8 ', 2 \ 3 / 2
V o l u m e =
2

(CIO)

E a c h point inside the section of Fig. C.F.I, corresponding to integer values of


n , n and n represents an energy state that contains a pair of electrons. Thus the
total volume inside this section represents t h e total n u m b e r of states. Since each
q u a n t u m state, because of t h e q u a n t u m spin, can a c c o m m o d a t e u p to two electrons,
at the absolute zero of t e m p e r a t u r e the volume represents one half the n u m b e r of
electrons N. We can then write
x

(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)

= volume density of states per unit energy at energy

N(E)dE

= volume density of states with energies from to + dE

N = volume density of occupied states where


s

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)

Appendix D: Einstein Relation

469

Thus E q . (C.13) represents t h e increment in the density of states in an incre


m e n t d E at energy E. This is what we defined as N(E) so that
(G14)
A t absolute zero, electrons occupy the lowest levels possible. We assume that
all states u p to = y = = a are occupied and for > a there are n o electrons at
all. Thus one can d e t e r m i n e the m a x i m u m energy that electrons can have at
absolute zero from E q . ( G i l ) . We label that energy t h e Fermi energy E . O n c e E is
fixed, the highest q u a n t u m states at absolute zero are fixed and E is defined as the
m a x i m u m energy that an electron in a certain material can have at absolute zero.
We d e t e r m i n e f r o m E q . ( G i l ) as
F

(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)

We show, in Fig. D . F l b the potential distribution in the slab, and in Fig. D . F l c


the energy b a n d diagram. The electron density in Fig. D.F1 can be expressed as

Appendix D: Einstein Relation


E

n{x) = n exp

,-(0) - ,(*)

- ,(x)
kT

, exp

kT

The electric field intensity and the potential energy of an electron


to the potential by

E (x)=

-qtfx)

(b)

Since (0) = 0 and hence (0) = 0 we can write Eq. (D.2) as


;

n(x) = n exp
i

~q<b(x)
for 0 =s d
kT

By using Eqs. (D.3) and (D.4) in E q . ( D . l ) we have

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 '

The expression in E q . (D.6) is k n o w n as t h e Einstein relation.

Appendix : Frequently Used Symbols

471

E: FREQUENTLY USED SYMBOLS


2

acceleration ( c m / s ) , metallurgical channel half width for J F E T


(cm)

area (cm )

B,E,C

base, emitter, collector of a B J T

BV

BJT b r e a k d o w n voltage collector to base with emitter o p e n

BVCEO

BJT b r e a k d o w n voltage collector to emitter with base open

speed of light ( c m / s )

C.

diode junction capacitance, depletion capacitance, transition capac


itance (F)

diode charge storage capacitance, diffusion capacitance (F)

je

B J T emitter-base junction (depletion region) capacitance (F)

B J T base storage (diffusion) capacitance

CB0

C , C
d

D, D
n

D,

gs

nE

nC

B J T collector-base capacitance

J F E T and M O S F E T gate-drain capacitance, gate source capaci


tance

diffusion constant for electrons, holes ( c m / s )

diffusion constant for electrons in the emitter, in t h e collector


(cm /s)

D,G,S

drain, gate, source of an F E T

electric field intensity ( V / c m )

energy (J, eV)

E ,E
E, E

=
=

acceptor, d o n o r energy level (J, eV)


energy level of b o t t o m of the conduction band, t o p of t h e valence
band

equilibrium Fermi level

E
E

b a n d g a p or forbidden gap (J, eV)

intrinsic Fermi level

zero v a c u u m level of energy

Force (N)

frequency (Hz)

f (E)

Fermi function for conduction b a n d

f(E)

F e r m i - D i r a c distribution function

fj

unity current-gain frequency

small-signal drain conductance

small-signal transconductance

n u m b e r of p h o t o g e n e r a t e d electron-hole pairs per second p e r unit


volume

472

Appendix : Frequently Used Symbols


G

= channel conductance of J F E T , M E S F E T in t h e absence of a deple

tion region
h

Planck's constant (J-s, eV-s)

hv

p h o t o n energy

current ( A )

/ , I~ I'

= B J T D C base, collector, emitter current

i, i, i

B J T varying c o m p o n e n t of base, collector, emitter current

B J T total collector current ( A C + D C )

= B J T varying c o m p o n e n t of collector current

C B O

C E O

I, I
cs

B J T D C collector to base current with emitter o p e n

B J T D C collector t o emitter current with base o p e n

= B J T reverse saturation current with collector, emitter shorted t o

ES

base

F E T drain current

I I

MOSFET D C drain current, saturation drain current for square-law model

I , I

MOSFET D C drain current, saturation drain current for bulk-charge model

= F E T varying c o m p o n e n t of drain current

OV

D1S

D2S

F E T total drain current ( A C + D C )

p h o t o g e n e r a t e d current

= F E T saturation drain current

SAT

= D i o d e recombination current in depletion layer

= P N junction diode reverse saturation current

J,J

current density electrons, holes

Boltzman constant

wave vector

M O S F E T device p a r a m e t e r

= dimension of lattice cube, F E T channel length

L'

= F E T effective channel length (reduced by modulation)

L ,L

L,

= diffusion length of electrons in emitter, of holes in base

nE

pB

diffusion length of electrons in P, of holes in

= carrier multiplication factor

= electron rest mass

m*^ m*

mass

= effective mass of electron, of hole

n'

= excess density of electrons over equilibrium

density of electrons (cm ),

quantum number

t h e r m a l equilibrium electron density

= t h e r m a l equilibrium electron density in collector

= t h e r m a l equilibrium electron density in region

oc

QE

Qp

t h e r m a l equilibrium electron density in emitter

Appendix : Frequently Used Symbols


n

n(0)

electron density at origin (x = 0)

intrinsic density of electrons and holes

0n

t h e r m a l equilibrium electron density in region

surface density of electrons

N,P

semiconductor d o p e d with donors, with acceptors

N ,P

473

heavily d o p e d ,

density of acceptors, d o n o r s

AC

density of acceptors in emitter, donors in base, acceptors in collec


tor

N(E)

density of states function per volume per unit energy

N, N
A

./V
N

N, N

effective density of states per unit energy at E , at E

hole density

t h e r m a l equilibrium hole density in region

p'

excess hole density over equilibrium

t h e r m a l equilibrium hole density in region

p (0)
p

hole density at = 0, in region, in region

m a g n i t u d e of electron charge

ph

total power absorbed by optical detector

total charge

Qn

0p

P(O\PM

excess minority carrier charge stored in B J T base

M O S F E T charge density in depletion layer

dm

m a x i m u m value of Q

M O S F E T electron charge density in channel

Q, Q
s

M O S F E T charge density in semiconductor, in metal

M O S F E T oxide charge density

lattice spacing

r, R

resistance

diode small-signal resistance (AC)

B J T output resistance (AC)

B J T A C resistance from collector to base

B J T A C input resistance from base to emitter

r.,r ,r

BJT ohmicresistance;base,emitter,collector

storage time

O N

time to turn device O N

OFF

time to turn device O F F

ox

thickness of insulator in M O S F E T

depress Kelvin

potential

474

Appendix : Frequently Used Symbols


V,

B J T Early voltage

applied voltage

M O S F E T b o d y voltage

built-in voltage

V.

junction voltage = V V

b r e a k d o w n voltage

B J T varying c o m p o n e n t of voltage, collector to emitter, base to


emitter

F E T varying c o m p o n e n t of voltage, drain to source, gate to source

bi

br

ce> be

as

gs

bi

EB"*

CB>

V, V
V (V ),
V (V )

B J T D C voltage; emitter to base, collector to base, emitter to col


lector
D C supply voltage to BJT; collector circuit, base circuit

F E T D C voltage; gate to source, drain to source

V,

M O S F E T D C voltage; source to bulk, channel to source

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

F E T D C threshold (turn-on) 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

depletion layer width, potential energy

SAT

lh

B J T base width

W ,W

width of region, region

F E T depth

Greek
a

= BJT: D C ratio of collector to emitter current


=

ot

BJT: D C ratio of collector to emitter current with collector shorted to


base in active region
BJT: D C ratio of collector to emitter current with emitter shorted to
base in inverse active region

=
=

BJT: D C ratio of collector to base current


BJT: D C ratio of collector t o base current with collector shorted to base
in active region

BJT: D C ratio of collector to base current with emitter shorted to base


in inverse action region

BJT: A C ratio of collector to base current with collector shorted to base

Appendix : Frequently Used Symbols

BJT: emitter injection efficiency

BJT: base transport factor

permittivity or dielectric constant

relative permittivity

= permittivity of free space

= permittivity of oxide

wavelength

M O S F E T channel length modulation factor

mobility
mobility of electron, holes

M O S F E T effective mobility of electrons, holes

frequency of light

=
=

resistivity

volume charge density

conductivity

= conductivity of region, of region

= carrier base transit time

B
T

V
,

lifetime of holes in region, of electrons in region


=

potential, surface potential

potential from E. to E in bulk of M O S F E T

(>

metal work function voltage

semiconductor work function voltage

^,

= affinity of semiconductor, E E

work function difference from metal to semiconductor,


Q

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 large-scale 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~ ohm-cm and are used for low-resistance wiring and interconnections in
electric circuits. Insulators that have resistivities greater than 10 ohm-cm 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 II-VI, IV-VI, and
III-V compounds, t h e most i m p o r t a n t of which is gallium arsenide.
Their mid-range 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

Atomic Structure and Quantum Mechanics

capacitors, and resistors. In the fabrication process, controlled a m o u n t s of particular


impurities are a d d e d to the silicon in order to form the various regions of a transis
tor. Certain sections of the wafers are oxidized to form silicon dioxide, which serves
as an insulator and as an isolator of different circuit parts. After the various devices
are formed, a good conductor, such as aluminum, is deposited on the surface to
interconnect t h e various parts of the circuits. Finally, the complete circuit, with
external leads attached, is encapsulated in a plastic or ceramic package.
O u r objective in this book is to study, first, some of the properties of pure
semiconductors and those semiconductors to which impurities are added. Having
studied these properties, we then investigate the operation and the current-voltage
characteristics of semiconductor devices, such as diodes and transistors. To do that,
we n e e d to consider the internal structure of semiconductors, t h e types of current
carriers, and the modes of transport of these carriers.
1.1

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 three-dimensional 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

a) Determine the number of atoms in each cell of the diamond lattice.


b) Determine the density of atoms in silicon, given the lattice constant L = 5.43A.
Solution
a) Eight atoms are shared by eight cells, six atoms are shared by two cells, and four atoms are
internal to the cell so that:
the number of atoms in each cell = 8/8 + 6/2 + 4 = 8.

Section 1.2

W h a t Are W e Looking For Now?

b) Volume of cell = (5.43 x 1(T ) cm .


8

Density = 8/volume = 5 x 10 atoms/cm .


22

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

Figure 1.1 Unit cell of the diamond


lattice. Source: Electronic Materials
Science by Mayer/Lau, 1990.
Reprinted by permission of PrenticeHall, Inc., Upper Saddle River,NJ.

Gallium Arsenide, a III-V 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

Atomic Structure and Quantum Mechanics

In a semiconductor, there are b a n d s of energy in which electrons can exist and


other b a n d s in which they cannot. We n e e d to establish the bases for the formation
of these bands. The existence of b a n d s results from t h e allocation of specific energy
levels to an electron in an a t o m and the consequent displacement of these energy
levels by introducing t h e effect of t h e forces that atoms exert on each other. A s a
result, each electron in t h e solid has a specific energy level; the combinations of
these levels form bands.
Before we study the formation of bands, we will present a perspective of the
theories that have evolved and have a t t e m p t e d to explain experimentally observed
properties of h e a t e d materials. Classical mechanics' concepts fail to explain these
properties as well as all other p h e n o m e n a that take place at the atomic levels. Such
explanations required the evolution of the science of Wave Mechanics, which treats
electrons as particles that have wave-like properties.

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

Bohr's Classical Model of the Atom

p
Wavelength in mm

Figure 1.2 Relative intensity of


radiation from a blackbody versus
wavelength at various temperatures.
From E. Uiga, Optoelectronics, Prentice
Hall (1995)

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
wave-like 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 planetary-like motion of elec
trons a r o u n d an a t o m and included the quantum-like 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

1.5 B O H R ' S C L A S S I C A L M O D E L O F THE A T O M


The work of R u t h e r f o r d c o m p a r e d the nucleus of an a t o m to t h e sun and t h e elec
trons that revolve a r o u n d t h e nucleus to t h e planets that revolve a r o u n d the sun.
U n d e r the influence of gravitational forces, the planets move a r o u n d the sun in a
quasi-circular orbit.
B o h r suggested that within t h e atom, gravitational forces (due to masses)
b e c o m e negligible w h e n c o m p a r e d to the electrostatic forces (due to charges) that

Chapter 1

Atomic Structure and Quantum Mechanics

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

(1-2)

where is the permittivity of free space in F a r a d s / m e t e r , r is t h e separation


b e t w e e n the electron and t h e nucleus in meters, and F i s in newtons. We assume that
the mass of the p r o t o n is much greater than that of the electron so that the p r o t o n is
fixed at the center of the system. To maintain motion in a fixed circular orbit having
radius r, the electrostatic force is balanced by an equal and opposite centripetal
force. We assume that the positive reference direction for the force is o u t w a r d so
that t h e attraction force in E q . (1.2) is negative.
Because of the presence of the positively charged nucleus, an electric field is
created a r o u n d it. Electric field intensity, %, defined as t h e force on a unit positive
charge, is directed o u t w a r d and is thus positive. F r o m E q . (1.2), we obtain the
expression for % at t h e location of the electron as
0

4 /

(1.3)

where, for the units used in E q . (1.2), % is in v o l t s / m e t e r . T h e electric field intensity


approaches zero as r approaches infinity. The electrostatic potential, V, due to t h e
electric field, is defined by
V = - \

% 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 )

The potential energy of t h e electron at a distance r, defined as t h e product of


the potential at that point and t h e charge, becomes
W(P.E.) = V{-q)

= ^

(1.6)

T h e negative sign implies that in moving an electron from infinity to r, nega


tive work is d o n e in contrast to t h e positive work that needs to be e x p e n d e d in mov
ing a positive charge towards the nucleus. This implies that an electron at point r has
to b e restrained to prevent it from moving towards the nucleus. A sketch of the
potential energy is shown in Fig. 1.3. We n o t e that the actual drawing has spherical
symmetry so that it has t h e shape of a funnel.
We are interested in t h e total energy that t h e hydrogen electron possesses,
which, in addition to potential energy, includes the kinetic energy resulting from t h e
velocity of the electron. A s it moves in t h e circular orbit, t h e electron is subjected to
a centripetal force (mv /r),
which exactly balances the attraction force of the
nucleus, causing an angular acceleration v /r.
By setting t h e m a g n i t u d e of t h e force in E q . (1.2) equal to t h e m a g n i t u d e of
the centripetal force, we have
2

(17)

After solving for mv , the expression for the kinetic energy becomes
K.E.

(1-8)

The total energy, E, of the electron is


2

= K.E. + VT(P.E.) = -q2 - +, -2


= 3 q
4
8
8 ?0

(1.9)

In accordance with t h e references established earlier, the electron has the


highest energy (least negative) as its distance from the nucleus approaches infinity.

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

Atomic Structure and Quantum Mechanics

the energy it gains and, t h e orbital radius is d e t e r m i n e d from E q . (1.8.) Hence,


according to this classical theory, the electron can take on continuous values of
energies E, depending u p o n t h e radii of the orbit. This, however, contradicts the
results of diffraction experiments. Diffraction is defined as the scattering that light
rays u n d e r g o when they pass t h r o u g h slits of a solid. T h e light rays emerge
deflected, forming fringes of parallel light, dark, and colored bands. The diffraction
experiments, using electrons, indicated that electrons can only occupy certain dis
crete energy levels as they emit radiation only at certain discrete frequencies.
The question is then: H o w does the electron emit radiation?
E l e c t r o n motion at constant speed results in a constant electric current. Since
the electron that is orbiting about a p r o t o n is subjected to centripetal acceleration,
its velocity increases with time, resulting in a time-dependent current. F r o m the laws
of electromagnetic theory, a time-varying current generates a time-varying magnetic
field. By Maxwell's equation, a time-varying magnetic field induces a time-depen
dent electric field. Electromagnetic fields and electromagnetic waves are generated,
leading to radiation that is similar to that emanating from a television or radio
transmitting a n t e n n a . Such radiation of energy may be possible over a wide b a n d of
frequencies.
If the hydrogen electron is to emit radiation, it must continuously lose energy,
revolve in smaller helical orbits with decreasing velocity, and get closer and closer to
t h e nucleus. H e n c e , if a charge is accelerated, it radiates energy and therefore expe
riences a loss of energy. This dilemma, faced by Bohr's classical model, led him to
qualify it by his quantized model.

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

Bohr's Quantized Model

= 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

Substituting t h e above expression for in E q . (1.11) and solving for r , we get


2

nhs

(1.12)

T h e expression for the velocity becomes

2s nh
0

The electronic orbits assumed in t h e B o h r theory are circular so that the


expressions derived earlier for the potential energy and total energy, Eqs. (1.6) and
(1.9), are valid.
By substituting the relation for r from E q . (1.12), as developed from Bohr's
theory, in the expression for total energy, E, of an electron, as given by E q . (1.9), we
obtain an expression for the quantized energy of the hydrogen electron as
E

- =

<

If we use rationalized M K S units, as found in A p p e n d i x A , for the constants in


Eq. (1.14), the unit of energy will be t h e joule. We now define a unit of energy
k n o w n as t h e electron-volt, eV, where one electron-volt is the energy acquired by an
electron w h e n elevated through a potential difference of one volt. O n e e V = 1.6 X
1 0 ~ joules. Replacing all t h e constants by their values, E q . (1.14) becomes
19

( 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

Atomic Structure and Quantum Mechanics

E{eV),
0
-0.56 I
-0.87 |

5
4

-1.53

-3.41

13.60

- 1

Figure 1.4 Energy level diagram for the


hydrogen atom. Six of the infinite
number of possible transitions are
indicated.

level by the absorption of radiation, again in accordance with E q . (1.10). In addition


to the energy, both the orbital radius and velocity are quantized, as expressed by
Eqs. (1.12) and (1.13) respectively. The allowed energy levels in the hydrogen atom,
as d e t e r m i n e d from E q . (1.15), are usually represented by an energy level diagram,
as shown in Fig. 1.4.
Later, Sommerfeld, following u p o n Bohr's conclusions, verified that the elec
tron orbits w e r e not circular but elliptical. This was later confirmed by others.
Planck's hypothesis provided an explanation for the discreteness of radiation
and the particle-like theory of light. B o h r confirmed the quantization of t h e energy,
the orbital radius, and the orbital velocity of the hydrogen electron. Thus, the quan
tum concept was extended to p h e n o m e n a taking place at atomic dimensions.
The B o h r theory constituted a big step forward in the explanation of the
hydrogen spectra but it did not provide complete answers. While the emission fre
quencies predicted by Bohr, for the hydrogen atom, are quite close to those that
were experimentally observed, a major weakness of Bohr's theory is that it cannot
be e x t e n d e d to t h e modeling of atoms that are m o r e complex than the hydrogen
atom. M o r e complete answers were provided by de Broglie and Schrodinger, who
provided the basis of t h e new science of Wave Mechanics.
REVIEW QUESTIONS
Q l - 1 Give examples of physical systems to which the theories of classical mechanics apply.
Ql-2 On what was Planck's hypothesis based?
Ql-3 What was Bohr's contribution to the science of Wave Mechanics and what were the
weaknesses of his postulates?

Section 1.7

W a v e Mechanics

11

Ql-4 What prompted Bohr to quantize the energy of an electron?


Ql-5 Why was the hydrogen atom selected for consideration?

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 wave-like properties but particle-like 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
El-1 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

El-2 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 wave-motion 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

Atomic Structure and Quantum Mechanics

particle and wave properties w e r e related so that de Broglie's two conclusions,


stated in equation form, are
- = h,

(a)

Kmv = h

(b)

(1.16)

where h is Planck's constant and is the de Broglie wavelength. Wavelength, K, in


meters, is the ratio of the velocity of light in m e t e r s / s e c o n d to the frequency in hertz
and 1/v is t h e spatial period of the wave.
The experimental observations of other scientists provided conclusive proof
of the relationships formulated by de Broglie.
It might be useful to emphasize t h e generality of de Broglie's hypothesis,
which h e predicted from relativity theory. According to his theory, a hypothetical
wave is associated with any particle such that Eqs. (1.16a) and (1.16b) hold. The
magnitude of the intensity of the d e Broglie wave, at any point, is a m e a s u r e of the
probability of finding t h e particle at that point. In essence, the m o t i o n of the particle
is describable in probability terms r a t h e r than in terms of the deterministic laws of
classical mechanics. This probability interpretation and the a t t e n d a n t uncertainty tie
in immediately with Heisenberg's Uncertainty Principle, which will b e discussed
shortly.
T h e "plausibility" of Eqs. (1-16) may b e established as follows:
1. The relation = hv is an expression of Planck's q u a n t u m hypothesis that
energy is absorbed or emitted in quanta, t h e size of a q u a n t u m being hv.
2. A light q u a n t u m is labeled a p h o t o n , so that the energy of a p h o t o n of light of
frequency is hv.
3. According to Einstein's theory, m a t t e r and energy are related by = mc :
Hence, a p h o t o n would have a mass of hv/c and m o m e n t u m hv/c or = h/ as
given by E q . (1.16a), w h e r e is the m o m e n t u m of the p h o t o n and c is the veloc
ity of light.
2

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 wave-like 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 particle-like and wave-like 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

Heisenberg's Uncertainty Principle

13

to the case of electron b e a m s traveling t h r o u g h a varying force field. B o t h p h e n o m


ena are mathematically analogous. It is found that when, in the optical case, the
refractive index varies over dimensions of t h e same o r d e r of m a g n i t u d e as the wave
lengths, classical theories cannot explain the p h e n o m e n a that occur. Similarly, classi
cal mechanics fails w h e n a force field in a material varies over a distance of the
order of a de Broglie wavelength. T h e r e a s o n we have to look to the wave theory for
an explanation of t h e p h e n o m e n a occurring at atomic distances is partly a result of
the H e i s e n b e r g Uncertainty Principle.

1.8 H E I S E N B E R G ' S U N C E R T A I N T Y PRINCIPLE


Heisenberg's Uncertainty Principle, which was derived from rigorous mathematical
physics, states that t h e m i n i m u m value of t h e product of the uncertainties in the val
ues of two quantities, whose p r o d u c t has t h e dimension of action,
ML /T(kg
m / s e c ) , is given by Planck's constant, h. Two such sets of quantities are m o m e n t u m
and position, and energy and time. The Principle states that t h e p r o d u c t of the
uncertainties in m o m e n t u m and position and those of energy and time t are
2

Ap Ax > h/2-

(a)

AEAt>h/2ir

(b)

(1.17)

A numerical example will clarify this result.

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~ J-s.
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

Ax = 6.62 X lCr /(2tr X 20 X 10" X 10 X 10~ ) = 5.26 X 10~ m


b) Ax = 6.62 10- /(2 X 9.1 10" x 10~ X 10~ ) = 11.57m
34

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.

In accordance with the Uncertainty Principle, it is thus meaningless to speak


of t h e motion of an electron in a circular or elliptic orbit because such a statement
implies that we can trace t h e p a t h of the electron by making at least two successive
observations of its position. In fact, for this example, if t h e position of t h e electron is
d e t e r m i n e d very accurately, so that is small, there is a lack of precision in the
uncertainty of the m o m e n t u m and Ap increases. If the limit of Ax is zero, t h e n the
uncertainty in t h e m o m e n t u m is infinite. The best that one can h o p e for is to set u p a

14

Chapter 1

Atomic Structure and Quantum Mechanics

probability p a t t e r n a r o u n d the nucleus, showing t h e region in which there is an


appreciable probability of finding t h e electron.
Thus, t h e Uncertainty Principle highlighted a n o t h e r difficulty with t h e B o h r
theory, as it is not possible to assign a certain orbit radius or m o m e n t u m to a parti
cle.
A s was shown in the example, for large-scale p h e n o m e n a where Ax for the
bullet is negligible c o m p a r e d to the dimensions of t h e bullet, the Uncertainty
Principle does not pose any serious difficulties, whereas for the electron, t h e posi
tion is indeterminate. Because t h e position, velocity, and m o m e n t u m of an electron
cannot b e d e t e r m i n e d using classical mechanics, one has to revert to wave mechan
ics to study t h e behavior of the electron.

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 wave-like 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
steady-state 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)

w h e r e m is the mass of the particle, W is t h e potential energy of the particle, and is


its total energy. This is t h e equation of a wave and w h e n applied to electron waves,
where m is the mass of t h e electron, it gives a m e a s u r e of t h e probability of finding a
given electron at a certain location. I n d e e d , t h e quantity | | is the probability
that a particle with potential energy W and total energy will b e located in the spa
tial volume dV at t h e point resulting from expressing in terms of x, y and z. Stated
in a n o t h e r m a n n e r , | | d V is the probability that a particle having potential energy
W will b e located within dV and, if it is there, that it will then have total energy E.
The wave function is allowed to b e a complex quantity and will, in general, for the
steady-state, be a function of x,y and z.
Wave Mechanics is, in effect, t h e fundamental branch of mechanics, and ordi
nary N e w t o n i a n mechanics is derived from it. Thus, w h e n a particle moves in a force
field, such that the change of potential occurs over distances that are very large
c o m p a r e d with t h e wavelength associated with the particle, the laws of N e w t o n i a n
mechanics are applicable. H o w e v e r , if the change in potential occurs over distances
comparable to t h e wavelength, as in t h e case of t h e periodic potential of a crystal,
then N e w t o n i a n mechanics does not apply and the wave n a t u r e of the particle gives
rise to completely n e w p h e n o m e n a u n a c c o u n t e d for.
We summarize the preceding discussion as:
2

1. The electron is not to be r e g a r d e d as a wave but as a particle, with an associated


hypothetical de Broglie wave.

Section 1.9

Schrodinger's Equation

15

2. Over a small region where the electron is k n o w n to exist, a probability wave


packet could be set up. The wave packet is a q u a n t u m mechanical m e a s u r e that
can b e assumed to b e localized at a given point in space. It can b e said to be
analogous to a classical particle that can be located in a given point in space.
T h e wave packet consists of the summation of constant energy wave functions,
, assembled about a center frequency so that this packet can be localized in a
given point in space. A sketch of a wave packet is shown in Fig. 1.5. Over a dis
tance that is largely c o m p a r e d with the wavelength, the motion of the electron
can be described by the m o t i o n of t h e wave packet, such motion being governed
by the laws of classical mechanics, although there is some uncertainty in the
simultaneous values of position, m o m e n t u m , velocity, etc.
3. W h e n considering the motion of the electron over distances that are small com
p a r e d with the wavelengths, such as inside the wave packet or a r o u n d the
I
nucleus, n o a t t e m p t should b e m a d e to locate the position of the electron along
its path at two successive instants of time. A n y a t t e m p t to determine the first
position will disturb the electron and cause it to trace out a different path.
U n d e r such circumstances, it is appropriate to speak only of t h e probability of
locating the electron at a certain point. This probability can be d e t e r m i n e d by a
solution of Schrodinger's equation.
Calculations in q u a n t u m mechanics use the probability function to d e t e r m i n e
the position, velocity, and m o m e n t u m of a system. A s we shall see later, finding is
not the ultimate result we are after. This quantity is a hypothetical one; a m a t h e m a t
ical tool used to obtain, mainly, the allowed system energies. In finding , two pos
tulates are of considerable importance. These are:
Postulate 1.

T h e function and its first derivative are finite, continuous, and single-valued.

Postulate 2.

The probability per unit length (per unit volume in the three-dimensional
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 ,

Figure 1.5 The composition of a wave


packet at a certain instant. Particle is
most likely to be at x but a smaller
probability is that it could be found
somewhere in the range .
Q

16

Chapter 1

Atomic Structure and Quantum Mechanics

* dxdydz=

\\ dx dy dz = 1

J space

(1.19)

* space

where t h e p r o d u c t dxdydz is t h e incremental volume. This is equivalent to


saying that it is certain that t h e particle is somewhere in space.
In t h e next section, we apply Schrodinger's equation to a simple structure in
o r d e r to illustrate t h e concepts of Wave Mechanics. W e then apply t h e solution to
the study of t h e energy levels in t h e hydrogen atom.
Application t o the Potential W e l l
We shall assume, as shown in Fig. 1.6, that an electron is located in a potential well
such that t h e potential energy W is zero b e t w e e n = 0 a n d = a, a n d infinite else
where. T h e infinitely d e e p one-dimensional well is, in a way, typical of t h e Wave
Mechanics p r o b l e m that o n e encounters in t h e study of t h e electron energies in a
hydrogen atom.
The total energy of t h e electron is kinetic, since W = 0. Therefore, t h e electron
will b o u n c e b e t w e e n t h e walls, and, if perfectly elastic, will assume a stable state. We
will d e t e r m i n e a n expression for in t h e potential well.
Applying Schrodinger's equation in t h e steady-state (t>) inside t h e o n e dimensional b o x w h e r e t h e potential energy W is zero, we have
+

(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?

The solution t o E q . (1.20) is


(,1 = Asm kx + Bcos kx
U7=c

(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

The electron cannot p e n e t r a t e outside t h e potential well, so that the wavefunction is


zero at = 0 and at = a. Thus,
= 0 at = 0

and

at = a

from which we obtain,


= 0, ka = m r

and

= Asm kx

(1.22)

w h e r e = 1 , 2 , 3 , . . . . Then, setting the expression for k, which follows E q .


(1.20), equal to /, we have
2

a 8<n mE
, ,
^2
= " ^
2

(1-23)

where is an integer. The energy corresponding to the integer in E q . (1.23)


becomes

T h e probability of locating t h e electron at any is


sin k x ) or (A sin /) . If a
particle is located at a certain point, or t h e instant it is there, E q . (1.24) expresses its
energy for that particular integer or q u a n t u m n u m b e r , n. The variations of and
|| in the well are shown in Fig. 1.7. Energy is therefore quantized (i.e., the particle
can take on only certain discrete values of energy). The particle spends most of its
time at t h e point at which | | is a maximum, as illustrated in the figure.
2

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

Atomic Structure and Quantum Mechanics

F r o m E q . (1.22) and Fig. L 7 , w e observe that, at the points at which = 0, for


the first cycle of , kx = ir. Since k is /, we write

= .

(1.25)

Thus, at these points, = a/n and since the distance b e t w e e n zeros is a half-wave
length (one wavelength corresponds to the spatial distance over one cycle), then
= /2

a/n

and
a = /2.

(1.26)

F r o m t h e above, we conclude that to satisfy the b o u n d a r y conditions for a sta


tionary state, t h e wavelength is related to the dimensions of the box, as shown by
E q . (1.26). In other words, the wavefunction represents a standing wave similar to
that which is observed on a vibrating string. The equation for the energy is also writ
ten by using the equality following E q . (1.20), as
2

hk
E

- - s

We can relate the energy to t h e m o m e n t u m by using the expression in Eq.


(1.16b) and replacing the wavelength by its equivalence from E q . (1.26) so that the
m o m e n t u m , p, b e c o m e s
p = h/X = ah/2a

(1.28)

Since k is /, t h e expression for the energy can b e expressed as


2

= /2m

(1.29)

By using E q . (1.29) in Eq. (1.27), the m o m e n t u m is expressed in terms of k as


hk/2ir.
A sketch of energy versus k and is shown in Fig. 1.8, w h e r e the dots illustrate,
pictorially only, the various allowed energy levels.
=

.*'

...**
k

>p

Figure 1.8 Energy versus k and for


k = / and = hk/2-.

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

, w h e r e they b o t h refer to t h e same state.


J

-rtl

For a particle in a three-dimensional box, assuming the same constraints in


t h r e e dimensions as for a one-dimensional case, we expect a standing wave of the
form
. , . .
. 2 . 2 . 2
= sin
sin \ sin = sin
sin
sin
a
b
c

X
\

where x, y, and have separate quantizations with q u a n t u m n u m b e r s , , and n


corresponding to the dimensions a, b, and c, respectively. W h e n this wavefunction is
substituted in Schrodinger's equation, an expression for is obtained.

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 one-dimensional 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

Figure 1.9 An electron in Region 1;


incident on Region 2.

20

Atomic Structure and Quantum Mechanics

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

</ , 1'.. +, 8TrTO


-T
( -

W) , = 0 for < W

or
- k:> = 0 where k = * ( - /.)
:

xkl

The expression, = y l e ^ + Be o, is a solution to the differential equation in Regiot


where d = 1 / k .
2

b) Applying Postulate 1 on page 15, at = 0, gives


fW,

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

The results of this example are;

A
ad

sin ax + A cos ex
0

= Ae""*^
2

The solutions are shown sketched in Fig. 1.10.

Section 1.9

Schrodinger's Equation

21

Reflected wave
Incident wave
Incident wave of
electron having
energy

Potential wall

W>E

Figure 1.10 Incident and reflected


Region 1

x= 0

Region 2

waves together with penetrating wave.

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
Ql-6
Ql-7
Ql-8
Ql-9

What is the significance of Heisenberg's principle?


What was de Broglie's contribution to Wave Mechanics?
Briefly state the significance of in Schrodinger's equation.
Interpret the solution of Schrodinger's equation to the deep potential well and how it
relates to the science of Wave Mechanics.

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 quantum-mechanical concepts in their duality applied not only
to electromagnetic waves but to electrons as well.

22

Chapter 1

Atomic Structure and Quantum Mechanics

Another link between particle-like properties and wave-like properties of matter was pro
vided by the de Broglie relationship, which expresses wavelength in terms of momentum.
The most far-reaching 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
El-3 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

El-4 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

El-5 Determine the energy in eV of an electron for = 1 in the infinitely deep well when
a = 50A.
3

Ans: = 14.92 " eV

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 cathode-ray 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

b) Calculate the number of quanta in 10 J.


a) Derive an expression for the wavelength of spectral lines emitted by the transi
tions from the excited states to the ground state using the Bohr relation for the
electron energy in the hydrogen atom,
b) Calculate the wavelength of the first four spectral lines.
The velocity of a certain free particle is 5 X 10 m/sec. The mass of the particle is
10~ kg. Determine:
a) the particle energy.
b) the de Broglie wavelength.
In accordance with physics statistics, the average energy of an electron in a medium of
free electrons at thermal equilibrium is 3kT/2, where k is Boltzmann's constant and
is in degrees kelvin. Determine, for the electron,
a) its velocity,
b) its momentum,
c) the de Broglie wavelength at = 300K.
An electron is moving with a velocity of 10 m/s. Determine, for the electron:
a) its momentum,
b) its de Broglie wavelength in m and A,
c) its energy in J and eV.
An electron has a de Broglie wavelength of 100A. Determine:
a) electron momentum,
b) electron velocity.
For an infrared radiation of 1 , determine:
a) the frequency of the radiation,
b) the energy of the photon in eV.
The uncertainty in the position of a particle having mass 10~ Kg is 10A. Determine
the uncertainty in:
a) the momentum of the particle,
b) kinetic energy of the particle.
For the electron in Problem 1.11, determine the wave vector k.
For an infinitely deep potential well having a = 100A, determine for an electron the
energy levels for = 1,2 and 3. Calculate the energy in eV and Joules.
An electron is located in a one-dimensional potential energy well having width of 3 A.
Determine
a) the kinetic energy of the electron in the ground state.
b) the frequency of the spectral radiation of an electron that drops from the next
higher state to the ground state.
For a particle that has a mass of 2 grams and energy 1.5kT, determine the de Broglie
wavelength at = 300K.
6

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 many-electron 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.

2.1 A P P L I C A T I O N O F S C H R O D I N G E R ' S E Q U A T I O N T O THE H Y D R O G E N


ATOM
The B o h r theory of the hydrogen a t o m is n o longer satisfactory for two major rea
sons. First, it assumes that the electron rotates a r o u n d t h e nucleus in a given orbit.

24

Section 2.1

Application of Schrodinger's Equation to the Hydrogen Atom

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 quantum-mechanical 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 many-electron 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)

Figure 2.1 (a) Illustration of potential


well of hydrogen atom, (b) Spherical
polar coordinates for Schrodinger's
equation.

Chapter 2

Energy Bands and Current Carriers in Semiconductors

profile of the hydrogen atom potential distribution. F u r t h e r m o r e , the potential well


is a one-dimensional p r o b l e m , whereas we wish to solve for the hydrogen a t o m in
three dimensions.
We will p r o c e e d with an outline of the general steps in the solution of
Schrodinger's e q u a t i o n for the hydrogen a t o m and then present the results of this
solution. We begin by substituting the expression for the potential energy in
Schrodinger's equation, E q . (1.18), in t h r e e dimensions, so that we have,
(2.1)
We shall not solve this equation but we will outline the steps involved in the
solution. First, the system, or the hydrogen atom, consists of a p r o t o n of charge q
and an electron of charge q and mass m . T h e position of the p r o t o n is fixed at the
origin, as shown in Fig. 2.1(b), and t h e electron is at a point whose spherical coordi
nates are (r, , and ) . Second, E q . (1.18), written in three dimensions, is trans
formed into the new coordinate system. Third, the m e t h o d of separation of variables
is used to solve the Schrodinger equation. By this method, we assume a solution of
the form
(2.2)

(r, , ) = /?(/)/()/()

W h e n t h e assumed solution is substituted in the spherical-polar form of


Schrodinger's equation, three separate differential equations result; one in r only,
one in only, and o n e in only. In the solution of the three equations, and to obtain
physically acceptable results, three q u a n t u m n u m b e r s emerge; n, i and m , k n o w n as
the principal, azimuthal, and magnetic quantum n u m b e r s respectively. This is analo
gous to the single q u a n t u m n u m b e r n, obtained for the electron in the potential well
example discussed in the previous chapter.
Let us digress briefly to explain how a solution is obtained so as to identify the
sources of these q u a n t u m n u m b e r s and their relation to the physical properties of
the electron and to t h e electronic orbits.
We will now consider a simplified Schrodinger equation and examine the
results. This equation is i n d e p e n d e n t of and but is a function of r, h e n c e labeled
the radial equation. T h e solutions to this equation have spherical symmetry. Before
proceeding any further, we note that just as in the example of the potential well dis
cussed in C h a p t e r 1, our eventual interest is not in the expressions for t h e wavefunc
tions. For this work, we are interested in the conditions for which the wavefunctions
satisfy the Schrodinger equation and the b o u n d a r y conditions.
The radial form of the differential equation in has a series of solutions. The
simplest and most basic solution, , expresses as a function of r and a factor
identified as r . By substituting this solution in Schrodinger's radial equation, we
obtain expressions for and r (see prob. 2.6). W h e n higher order solutions for ,
such as and , are included, we obtain energy as a function of t h e quantum
number, n. This n u m b e r is a m e a s u r e of the total energy of the electron in a particu
lar state as well as being a m e a s u r e of t h e diameter of t h e major axis of the elliptical
orbit, which is twice the radius of the circular orbit d e t e r m i n e d by Bohr.
e

Section 2.1

Application of Schrodinger's Equation to the Hydrogen Atom

27

T h e q u a n t u m n u m b e r , n, is an integer that takes on t h e value 1, 2, and 3.


Because of the association of with t h e energy of the electron, the energy levels in
are also designated by the electronic shells K, L, and M, corresponding to = 1,2,
and 3.
A quantum-mechanical interpretation of the orbital radius of the hydrogen
a t o m is that it is n o longer a fixed n u m b e r (and t h e electron having a fixed orbit) but
that there is a high probability of locating the electron at a distance identified by a
quantum number.
Two o t h e r q u a n t u m n u m b e r s result from the complete solution of t h e equa
tions in each of and of the Schrodinger equations.
The azimuthal number, i, results from the solution of t h e equation in such
that its value determines b o t h the quantized angular m o m e n t u m as //2 and the
d i a m e t e r of the minor axis of the quantized elliptic orbit of the electron as
minor axis diameter
n

major axis diameter

The values of t h e set of azimuthal q u a n t u m n u m b e r s cover t h e range


= 0to = n -

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 many-electron 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 .
(

T h e Pauli Exclusion Principle


In 1925, Pauli formulated the principle that t h e r e can be only two electrons of o p p o
site spin in a q u a n t u m state. H e concluded this from a study of atomic spectra at
about the same time that the extra q u a n t u m n u m b e r , m , was identified. H e , in
effect, concluded that n o m o r e t h a n two electrons can have the same distribution in
s

Chapter 2

Energy Bands and Current Carriers in Semiconductors

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 one-electron 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 many-electron 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

Quantum Numbers and Number of Electrons.

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

It is easily verified from Table 2.1 that the m a x i m u m n u m b e r of electrons in a


shell is 2n . It is also to be expected that the higher shells are farther from the
nucleus and have t h e higher energies.
Thus, the first level contains two electrons, the second contains eight, the
third 18, and so on. Since the levels were originally assigned by spectral studies, the
levels are d e n o t e d by s, p, d, f, g, etc., w h e r e = 0 corresponds to the s level, i = 1 to
t h e level, and so on. (The letter s refers to sharp, the to principal, t h e d to diffuse,
the / to fundamental, and so on.) C a r b o n , for example, with an atomic n u m b e r of 6
(six electrons), has Is 2s 2p , with the first n u m b e r referring to the shell and the
superscript n u m b e r referring to the n u m b e r of electrons in that particular level.
Silicon has is 2s 2p 3s 3p for an atomic n u m b e r of 14. The 3p level can accom
m o d a t e a m a x i m u m of six electrons but only two are there, leaving four empty
spaces in that level. A schematic representation of an isolated silicon a t o m is shown
in Fig. 2.2, illustrating t h e location of the electrons in the shells.
2

We n o t e that t h e lower levels are filled first, with t h e u p p e r m o s t energy levels


only partially filled with electrons.

Section 2.1

Figure 2.2

Application of Schrodinger's Equation to the Hydrogen Atom

29

Representation of a silicon atom.

T h e M a n y - E l e c t r o n Solid
The transition from the one-electron p r o b l e m to the many-electron 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 one-electron 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 one-electron 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 one-electron 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 hydrogen-like. 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

Energy Bands and Current Carriers in Semiconductors


r

Figure 2.3

Locations of energy levels for a single atom.

4. Theoretically, each electron in a solid represents a q u a n t u m state defined by


four q u a n t u m n u m b e r s and has energy levels different from every other elec
tron.
A m o r e complete m o d e l will b e included later.

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

Energy Band Formation

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 electron-volt 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

Energy Bands and Current Carriers in Semiconductors

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

Energy Band Formation

33

= number of atoms in crystal


4N electrons in outer shell
4N states-no electrons
6N states, 2N electrons
6N electrons maximum
6N states

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

Relative spacing of atoms

The formation of energy bands in silicon as atomic spacing is reduced (T = 0).

The process of energy b a n d formation in silicon is interesting and we will


study the splitting in m o r e detail by considering the representation shown in Fig.
2.6.
Varying t h e spacing of t h e atoms, as shown in Fig. 2.6, is obviously a theoretical
exercise. H o w e v e r , we assume that silicon a t o m s are brought n e a r each other. A t
large atomic spacing, the 3p levels in the shell and all lower energy levels are
shown as straight lines, as the atoms have not begun to influence each other. Of
course, the straight line 3p includes six q u a n t u m states and the 3s lines include
two states. T h e energies of the states are very close to each other.
A s t h e spacing is reduced, t h e 3p levels split first followed by the 3s levels. A t
still smaller spacing, the two bands formed by the 3p and 3s levels merge into a sin
gle b a n d containing eight states. Following this, the new b a n d splits into two dis
tinct b a n d s and, at the normal atomic spacing, two bands (conduction and valence)
are formed separated by the gap. T h e r e are four levels in each of the bands.
A t very low t e m p e r a t u r e s and since electrons t e n d to occupy the lower levels
first, t h e r e will b e four electrons that completely fill the valence band. T h e con
duction b a n d contains n o electrons. A t higher temperatures, electrons will b e
excited from the valence b a n d to t h e conduction band. F r o m studies of energy level
splitting, we present the following results:

A forbidden band is formed b e t w e e n t h e top b a n d and the next lower band.


Ideally, states that a c c o m m o d a t e electrons of silicon cannot exist in the for
bidden band.

T h e t o p b a n d is the conduction band and the next lower b a n d is t h e valence


band. T h e r e are m a n y other b a n d s below the valence band.

34

Chapter 2

Energy Bands and Current Carriers in Semiconductors

T h e width of t h e forbidden band, the b a n d gap energy , d e p e n d s on the lat


tice constant and t h e energy levels a r o u n d this band. The lattice constant rep
resents the n o r m a l atomic separation.

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.

A t the t e m p e r a t u r e of absolute zero, electrons occupy the lowest possible


energy levels.

The width of the b a n d gap, , decreases with t e m p e r a t u r e in accordance with


t h e e q u a t i o n in Table 2.2, in which we have shown the values of E and the
d e p e n d e n c e of E on t e m p e r a t u r e for G a A s , Si and Ge.
g

TABLE 2.2 Band Gap Variation with Temperature


(r) =
g

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

Source: C. D. Thurmond, "The Standard Thermodynamic Function of the


Formation of Electrons and Holes in Ge, Si, GaAs and GaP," Journal of
Electrochem. Society 122,1133 (1975)
It should be n o t e d that b a n d structures are not exact models of electronic energies.
In order to obtain them, m a n y approximations and simplifying assumptions have to
be made. Some experimental proof of the m o d e l is obtainable from studies of x-ray
spectra. Properties such as conductivity, magnetic effects, and optical effects
d e p e n d , to a certain extent, on the presence of foreign atoms, lattice defects, or
imperfections, whereas t h e b a n d theory is based on a hypothetically perfect lattice.
A b a n d m o d e l that results from Fig. 2.6 is shown in Fig. 2.7.

REVIEW QUESTIONS
Q2-1 Where do the quantum numbers come from?
Q2-2 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.
Q2-3 What phenomena cause energy levels to split?
Q2-4 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?
Q2-5 What particular aspects completely specify the electrical conducting properties of a
solid?
Q2-6 If is the atomic number of a solid, what is the charge of the nucleus?

Section 2.3

After splitting

iV-isolated
atoms

Energy Band Formation

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

The solutions of the time-independent Schrodinger's equation in r, , and provide


the three quantum number n, , and m respectively, where defines the total energy
of an electron in a particular quantum state.
The Pauli exclusion principle requires that in an atom that contains many electrons, no
more than one electron can exist in any one quantum state. This resulted in a fourth
quantum number m for the spin of two electrons. Thus, each electron is specified by
four quantum numbers.
As atoms of the same solid are, theoretically, brought together, the energies of the out
ermost electrons are modified. Each level, such as the 3p level, which accommodates
six electrons, is split into closely spaced levels.
Where interacting atoms of a solid are brought to the normal atomic spacing, R , the
old levels are split into very closely spaced levels. These levels are so close to each
other that bands are effectively formed. The top two bands of semiconductors in which
electrons can exist are the conduction band and the valence band. These two bands are
separated by a band in which states are available but no electrons can exist. This band
is the forbidden band.
(

EXERCISES

E2-1 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.
E2-2 How many of the electrons in an atom of silicon, at thermal equilibrium, form each of
the conduction band and the valence band?
E2-3 Determine the electron configuration, in ascending order of energy, for the following
elements: Ge, Ga, As, and P.

36

Chapter 2

Energy Bands and Current Carriers in Semiconductors

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 one-dimensional 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

Mathematical Model of Band Formation

37

n a t u r e of the potential energy distribution for an array of atoms is very complex so


that the solution of Schrodinger's e q u a t i o n for such a distribution is very difficult.
Kronig-Penney Model
In an effort to obtain a mathematical solution that will confirm energy b a n d forma
tion, including the existence of forbidden bands, Kronig and P e n n e y used a onedimensional m o d e l of the potential energy distribution in a solid. The m o d e l shown
in Fig. 2.9 exhibits s o m e similarity to the one-electron potential distribution of Fig.
2.8(c). It consists of a regular array of square-well potentials. In spite of its crude
similarity to Fig. 2.8(c), t h e solution of Schrodinger's equation using this m o d e l
describes, fairly accurately, t h e effects that are observed in real solids.
It is i m p o r t a n t to b e aware of the assumptions that are m a d e in t h e m o d e l and
its solution. These are:
1. E l e c t r o n interaction with t h e core is purely coulombic in nature.
2. E l e c t r o n to electron interaction is precluded.
3. Non-ideal effects, such as collisions with the lattice and the presence of impuri
ties, are neglected.
4. A t o m s are fixed in position, although, in fact, the atoms may be vibrating.
In the K r o n i g - P e n n e y model, t h e zero level of potential energy is assumed to
be in the vicinity of the core and t h e p e a k value W occurs midway b e t w e e n the
cores.
A n electron having a certain mass, m, and energy, E, is assumed to exist in this
potential energy distribution, which is established by the a t o m cores. The p r o b l e m is
to d e t e r m i n e the distribution of the energy levels as a function of the wave n u m b e r
k. A s e p a r a t e Schrodinger equation is set u p for each of the regions = 0 to = a,
where W = 0, and = - to = L, where W = W . The distance (a + b) is labeled
the lattice constant L.
Subject to the b o u n d a r y conditions, two solutions for t h e wavefunctions are
obtained, one for each of the regions 0 < < W and > W . It is necessary to
r e p e a t h e r e that our information is not obtained from the expressions for the wavefunctions but in the expressions that result from applying the b o u n d a r y conditions
0

w
, r

b
X

L
Figure 2.9 One dimensional Kronig-Penney model of the potential energy
distribution.

Chapter 2

Energy Bands and Current Carriers in Semiconductors

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 left-hand side of each of the equations, which include and W ,
and one term on t h e right-hand 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 left-hand
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 K r o n i g - P e n n e y m o d e l plot displays discontinuities and perturbations


w h e n c o m p a r e d to the parabolic shape of t h e free particle solution. These
modifications are greatest at t h e lower values of energy. A t the higher values
of energy, the solution for this m o d e l is similar to that of t h e free particle. The
implication is that the greater t h e electron energy, the less is the importance
of the periodic potential in the crystal.

The K - P m o d e l plot has discontinuities at kL = BTT. O n e can conclude that


at the values of energies corresponding to these discontinuities, electron
waves cannot p r o p a g a t e in the solid. These values of k m a r k the boundaries
of the energy zones that m a k e u p the forbidden bands.

A t higher values of E, t h e width of the permitted bands increases while the


width of t h e forbidden bands is reduced.

T h e left-hand 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 jc-axis.The values

Figure 2.10 (a) Energy versus wave number k using the Kronig-Penney model. The dashed curve that
is superimposed is the free particle solution, (b) Energy versus wave number for the potential well.

Section 2.4

Mathematical Model of Band Formation

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 one-dimensional 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 three-dimensional crystal, the E-k 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 E-k 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 E-k 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

Figure 2.11 Energy distribution from


the Kronig-Penney model for the lower
three allowed bands.

40

Chapter 2

Energy Bands and Current Carriers in Semiconductors

Gallium arsennide
(a)

Silicon

Figure 2.12 E-k sketches for (a) direct


band gap and (b) indirect band gap
semiconductors.

H e n c e , G a A s is known as a direct band gap semiconductor, and Si and G e are


known as indirect band gap semiconductors.
In a direct b a n d gap semiconductor, a p h o t o n of light energy, hv, can excite an
electron from the t o p of the valence b a n d to the b o t t o m of the conduction band.
Similarly, an electron in the conduction b a n d of a direct semiconductor can fall
directly into an empty state in t h e valence b a n d and emit p h o t o n s of light having
energy . In indirect semiconductors, an electron in the conduction b a n d cannot
fall directly into the valence b a n d because it must u n d e r g o a change in energy and a
change in m o m e n t u m . A p h o t o n by itself cannot excite an electron from the t o p of
the valence b a n d of an indirect semiconductor to the b o t t o m of t h e conduction
b a n d because t h e p h o t o n has sufficient energy to cause the transition but does not
possess the necessary m o m e n t u m for this transition.
A n electron moving b e t w e e n the valence b a n d and the conduction b a n d of an
indirect semiconductor can occur through a defect in the semiconductor or by the
action of phonons, which can provide sufficient m o m e n t u m to assist indirect transi
tions.
The application to which the direct semiconductors b e c o m e i m p o r t a n t is the
optical device. In this case, G a A s is a principal semiconductor used in semiconduc
tor lasers and light-emitting diodes.

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 two-dimensional 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 outermost-orbit
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

Covalent Bond Model

Section 2.5

\/ \/ \/
'

"

'

\/
%

"

41

XXX-X
//

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 III-V 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

Energy Bands and Current Carriers in Semiconductors

as we shall see later, the conductivity of semiconductors, under certain conditions,


may increase with increasing t e m p e r a t u r e .
In metals, the conduction (free) electrons are distant from their nucleus, move
freely a m o n g atoms, and are b o n d e d to different atoms at different times. A metal is
thus said to consist of an array of positive ions s u r r o u n d e d by closed-shell electronic
orbits immersed in a gas of free electrons. T h e r e is n o forbidden gap that the elec
trons have to overcome. While there are approximately 1 0 free electrons per c m
in silicon at r o o m t e m p e r a t u r e , a metal will have approximately 1 0 electrons per
cm .
10

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 two-level 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 car-length 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

Current CarriersElectrons and Holes

Section 2.6

43

Liquid motion

(a)

(b)

Figure 2.14

Analogy to bands in silicon.

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

(2-3)

w h e r e ( q) is the charge of an electron, v is its drift velocity, and t h e negative sign


indicates that t h e direction of current is opposite to t h e direction of motion (v ) of
the electrons.
Mathematically, one can also state that the current density in the valence b a n d
is m a d e u p of two c o m p o n e n t s : t h e s u m m a t i o n of t h e motion of all the electrons in a
completely filled b a n d (no vacancies) minus that current associated with t h e miss
ing electrons as,
d

^~qv
filled band
d

0 0

_^~qv
empty states
d

Since t h e current in a filled b a n d is zero, the current resulting from t h e avail


ability of empty states is the current in t h e valence b a n d given by

Chapter 2

Energy Bands and Current Carriers in Semiconductors

J =
VB

(2-5)

Because t h e direction of motion of t h e carriers is in response to an electric


field, we conclude that t h e hole has a positive charge and the free electron has a
negative charge since the electric field causes the holes and the electrons to m o v e in
opposite directions.
Electron a n d Hole Energies
The generation of an electron-hole pair in the bands is shown in Fig. 2.15(a) and in
t h e covalent picture is shown in Fig. 2.15(b). In the b a n d picture, the electron has
m o v e d u p in energy, and in the covalent b o n d illustration, t h e electron is free to
r o a m with its place becoming available for occupancy.
In the energy b a n d diagram, the word " e n e r g y " refers to the energy of the
electrons. W h e n t h e energy of t h e electron is increased, that electron occupies a
higher energy level in t h e band. This applies equally well to conduction b a n d and
valence b a n d electrons.
A n increase in the energy of a hole is caused by the raising of the energy level
of the electron. A s electrons move u p within the valence band, holes b e c o m e avail
able and m o v e d o w n t h e b a n d occupying lower levels of t h e electron energies. Thus,
an increase in the energy of a hole is associated with downward motion in the valence
band.
A n electron located at the level of the b o t t o m of the conduction b a n d at rest
has potential energy E , and zero kinetic energy. Similarly, a hole located at the top
of the valence b a n d E at rest, has potential energy E , and zero kinetic energy.
c

\/ \/ 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)

where F is t h e externally applied force, as by the electric field, m is t h e mass of an


electron in a space free of lattice forces, and m* is t h e mass of the electron that
includes t h e effect of t h e forces in the crystal. Such a mass, m*, is labeled the effec
tive mass of an electron and m* is t h e effective mass of a hole. We will now deter
mine expressions and definitions for the effective masses.
In C h a p t e r 1, we d e t e r m i n e d that for b o t h a free particle and for t h e particle in
the well, the energy is given by
Q

E = f~

2m

(2.7)
0

where is the particle m o m e n t u m and m is the free-electron mass. The motion of


an electron in t h e conduction b a n d of a semiconductor is analogous to that of a free
particle, except that t h e conduction b a n d electron is subjected to lattice forces. In
this case, we label this mass as t h e effective mass of an electron and we define the
energy as
0

where m* is the effective mass of an electron and is labeled t h e crystal m o m e n t u m


in reference to t h e forces of the crystal that act on t h e electron. In an o p e r a t i o n sim
ilar to Eqs. (1.27) and (1.29), m o m e n t u m is also written as

Chapter 2

Energy Bands and Current Carriers in Semiconductors


=

hk/2ir

(2.9)

w h e r e h is Planck's constant and k is t h e wave n u m b e r .


The expression for t h e energy of an electron in the lattice is obtained by using
E q . (2.9) in E q . (2.8),
=

h k /8TT m*

(2.10)
2

The important conclusion is that the energy is directly proportional to k and


inversely p r o p o r t i o n a l t o t h e effective mass of an electron. It is obvious from Fig.
2.11 that t h e effective mass is not necessarily constant, it is a constant only if the
relation of to k is purely parabolic.
Based on E q . (2.10), the effective mass of an electron is defined as
(2.11)
O n close study of the E-k plot of Fig. 2.11 for an electron in a varying field, it is
a p p a r e n t that the second derivative of with respect to k is negative w h e n the E-k
curve is concave d o w n w a r d s and it is positive w h e n t h e curve is concave upwards.
T h e curves n e a r the m i n i m u m of the conduction b a n d and n e a r t h e m a x i m u m of the
valence b a n d are nearly parabolic so one can safely assume that the mass of the
electron is constant in those two regions. H o w e v e r , n e a r t h e m i n i m u m of the con
duction band, t h e E-k curve is concave upwards so that t h e mass of an electron is
positive, whereas near the m a x i m u m of the valence band, the curve is concave
downwards so that the mass of an electron is negative.
T h e direction that a particle will b e accelerated w h e n an electric field is
applied is d e t e r m i n e d by t h e sign of t h e mass and the charge as
a =

q%/m

(2.12)

A positive mass electron (one n e a r the b o t t o m of t h e conduction b a n d ) having


a negative charge will b e accelerated in a direction opposite to that of the field, in
accordance with E q . (2.12). The motion of an electron opposite to the direction of
the electric field constitutes a current in the direction of the field since current is
defined as being opposite to t h e direction of t h e motion of an electron.
A n electron (negative charge) n e a r the t o p of the valence band, which has a
negative mass, m a y b e regarded, as far as its motion is concerned and in accordance
with E q . (2.12), as a positive mass-positive charge particle as t h e two negative signs
in E q . (2.12) are replaced by a positive sign. Therefore, the electron n e a r t h e t o p of
the valence b a n d is accelerated and causes current in the direction of the electric
field.
In conclusion, an electric field applied to a semiconductor accelerates elec
trons in t h e conduction b a n d in an opposite direction to t h e electric field and accel
erates positive particles in the valence b a n d in the same direction as the electric
field. The result is a current in t h e direction of the electric field consisting of elec
trons (positive mass, negative charge) moving in t h e conduction b a n d and the parti
cles (positive mass, positive charge) moving in the valence band. These particles are
the holes.

Section 2.8

Conductors, Semiconductors, and Insulators

47

A p p r o x i m a t e values of the effective masses of electrons m* and holes m* rel


ative to t h e masses in a vacuum, are listed in Table 2.3. We n o t e from the table that
the effective masses of electrons and holes in gallium arsenide are smaller than
those in silicon. We refer to the E-k diagrams of Fig. 2.12 and n o t e that t h e curvature
of the vs. k diagram, in particular near the b o t t o m of the conduction b a n d in
G a A s , is considerably greater t h a n that in Si. Consequently, and in accordance with
Eq. (2.11), electrons in G a A s have a m u c h smaller effective mass than in Si.
TABLE 2.3 Density of states effective masses of
electrons and holes in Si, Ge, and GaAs and 300K

m*/m
m*/m

Si

Ge

GaAs

1.18
0.81

0.55
0.37

0.065
0.52

These masses a r e labeled as density of states effective masses. A different def


inition for masses with different values is used for calculations of t h e mobilities
of electrons a n d holes. They are labeled conductivity
effective masses. We will
define mobility of a particle as t h e ratio of t h e velocity of the particle to t h e elec
tric field. In this b o o k , all references for effective mass will b e to that of t h e density
of states.

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

Energy Bands and Current Carriers in Semiconductors


Metals

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

Figure 2.16 Conditions of energy bands in metals, semiconductors, and insulators


at (a) very low temperatures and (b) at = 300K.
A t low temperatures, the conduction b a n d of a semiconductor is practically
empty. Applying an electric field to the solid accelerates the electrons in the valence
b a n d and increases their kinetic energy. This increase in energy is n o w h e r e sufficient
to m o v e the electron across the gap into t h e conduction band. Therefore, n o current
flows. By raising the t e m p e r a t u r e of t h e semiconductor, some electrons at the t o p of
the valence b a n d gain enough t h e r m a l energy to m o v e into t h e conduction band.
B o t h t h e electrons in t h e conduction b a n d and t h e holes in the valence can now b e
accelerated, gain kinetic energy, and therefore carry a current w h e n an electric field

Section 2.8

Conductors, Semiconductors, and Insulators

49

is applied. The conductivity at 300K is small, c o m p a r e d to metals, hence the n a m e


semiconductor.
The b a n d gap energy of an insulator, such as d i a m o n d or silicon dioxide, is sev
eral times greater than that of semiconductors, such as silicon or gallium arsenide.
E v e n at very high t e m p e r a t u r e s (high enough to a p p r o a c h the melting point), very
few electrons acquire enough t h e r m a l energy to be raised to the conduction band.
Therefore, the resistivity of insulators is very high and their conductivity is very low.
In metals, the top b a n d is only partially filled with electrons. Bands b e c o m e
partially filled if either the n u m b e r of electrons is not sufficient to fill the b a n d or,
m o r e commonly, a completely filled valence b a n d overlaps an empty conduction
band. A n electric field applied to the m e t a l at a t e m p e r a t u r e of 300K easily moves
electrons to higher empty levels and causes a current to flow. Thus, the conductivity
of metals is very high.

REVIEW QUESTIONS
Q2-7
Q2-8
Q2-9
Q2-10
Q2-11
Q2-12
Q2-13

Distinguish between a direct and an indirect semiconductor. Give an example of


each.
Suggest an application for indirect semiconductors in relation to their absorption and
emission of photons.
Trace on Fig. 2.12(b) the motion of a hole in a semiconductor to which an electric field
is applied.
Why is the mass of an electron negative for an electron located at the top of the
valence band?
Why are the effective masses of holes and electrons smaller in GaAs than in Si?
Why is the ability to conduct electricity, or conductivity, higher in metals than in semi
conductors at room temperature?
Why is the conductivity of insulators, when compared to that of a semiconductor, neg
ligible?

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 negative-mass, negative-charge electron that is subjected to an electric field is accel
erated in the direction of the electric field, opposite to that of a negative-charge, posi
tive-mass electron. The negative-mass, negative-charge electron is interpreted to be a
hole, which has positive charge and positive mass.

EXERCISES
E2-4

The energy gap of GaAs is 1.42V. Determine:


a)
the minimum frequency of light that will cause the transition of an electron
from the valence band to the conduction band.

50

Chapter 2

Energy Bands and Current Carriers in Semiconductors


b)

The wavelength of this light.


14

Ans: a) / = 3.43 X 10 Hz

b) = 8.74 X 10" m
10

E2-S 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 one-dimensional 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 E-k 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 k-direction. Determine:
a)

The signs of the effective masses of the wavepackets made Up of groups of


states near A, B, C and D.

b)
c)

The direction of the velocity of each of the wavepackets.


The direction of acceleration of each wavepacket.

Chapter 2

Problems

51

2.5 The E-k 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 time-independent 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

Reduce the equation to one in r only, independent of and .


For the equation determined in (a), first, assume a solution of the form *Pj =
C exp (r/r ) and, second, use Postulate 2 in Chapter 1 for the volume of a
spherical shell of radius r and thickness dr in order to determine an expression
forC.
Determine the expression for r by substituting the solution for in the dif
ferential equation found in (a).
Determine the expression for the total energy of the electron in the lowest
energy state. Given:
0

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 current-voltage 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)

where N(E) is t h e density of states function for free electrons in a metal m e a s u r e d


per unit volume, per unit energy, located about the energy level E. Its significance is
clearer if we state the following: To obtain the total n u m b e r of energy states per unit
volume in a given energy range, dE, about E, we multiply N(E) by dE. In fact, it
b e c o m e s an integration process.
We will assume that the function given by E q . (3.1) is valid for semiconduc
tors, provided we use the effective mass for t h e electron, since t h e electron moves in
the periodic potential of the crystal lattice and it may b e assumed to b e a free elec
tron. The densities of electrons and holes will be d e t e r m i n e d by using this function
t o g e t h e r with a probability of occupancy of a state function.
The energy corresponds to a set of q u a n t u m n u m b e r s and can therefore
take on only certain discrete values. A sketch of the distribution of states is shown in
Fig. 3.1. This curve is not continuous but is m a d e u p of a set of discrete points with
the adjacent states so close to each other that, for all practical purposes, o n e can
safely consider it continuous.
A n analogy should help clarify the significance of the variation of N(E) and E.
Let us consider an oval-shaped stadium, with seats at all elevations a r o u n d t h e field,
and label its volume as a unit of volume. A t the higher elevations, o n e can count
m o r e seats a r o u n d the stadium for a given level t h a n at t h e lower elevations. Thus,
we can say that the n u m b e r of seats per unit volume (the whole stadium) per m e t e r
of elevation corresponds to N(E). The n u m b e r of seats per m e t e r of elevation at the
higher elevations is larger than t h e n u m b e r of seats per m e t e r of elevation at the
lower elevations. Similarly, the density of states N(E) is larger at higher energy levels E.
Classical mechanics states that at absolute zero all electrons have zero energy,
and that w h e n a material is h e a t e d from absolute zero, each particle absorbs an
a m o u n t of energy equal to kT, w h e r e k is Boltzmann's constant and is the
absolute t e m p e r a t u r e . H o w e v e r , Wave Mechanics theories obviously do not agree
with those of classical mechanics. First, only very few electrons have zero energy at
zero t e m p e r a t u r e . Second, w h e n the material is heated, only those electrons close to
a certain energy level, k n o w n as the Fermi level, can be excited to higher unoccupied
levels. T h e significance and relevance of the Fermi energy level will be clarified in

N(E)
0

Figure 3.1 Distribution of energy states;


versus N(E).

54

Chapter 3

Intrinsic and Extrinsic Semiconductors

the following section. Therefore, w h e n a material is h e a t e d from absolute zero, not


every electron gains energy kT, as classically happens. According to Wave
Mechanics, only those electrons within an energy range of the o r d e r of kT of the
Fermi level can b e thermally excited. F u r t h e r m o r e , in t h e study of energy b a n d for
mation, we found that t h e solid has states of varying energy, even at absolute zero,
and that is because of t h e Exclusion
Principle.

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 Fermi-Dirac 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 lower-level, less-numerous 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 lowest-level 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

Fermi-Dirac Distribution Function

55

f(E)

Figure 3.2 Fermi-Dirac distribution function for three temperatures. Temperature


is measured in K.

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 Fermi-Dirac distribution function for
three temperatures. N o t e t h e rounding-off 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]

F r o m E q . (3.4), we can then calculate t h e incremental density of free electrons


corresponding to a specific interval of energy dE, for a fixed Fermi level, and as a
function of t h e t e m p e r a t u r e T.
A word of caution is in o r d e r here concerning the location of E in semicon
ductors. In a later chapter, we shall n o t e that t h e level E lies in t h e forbidden band.
You may t h e n question our statements h e r e concerning occupancy of E , even
p

56

Chapter 3

Intrinsic and Extrinsic Semiconductors

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

(b) The probability of vacancy is [1 f(E)] so that at 0.4V it becomes


^ ^
1 + exp0.0259

1-/( ) = 1

(c) The probability of vacancy at 0.01 V becomes


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)

Figure 3.3 Distribution of the density of


states in the conduction and valence
bands.

Chapter 3

Intrinsic and Extrinsic Semiconductors

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)

is t h e probability of vacancy of a state in the valence b a n d and is

The Boltzmann Approximation


To simplify t h e integration in Eqs. (3.7) and (3.8), we m a k e an important assump
tion that is related to t h e location of t h e Fermi level. This assumption, k n o w n as the
Boltzmann approximation,
consists in dropping t h e unity t e r m in t h e d e n o m i n a t o r
of the expression for the Fermi function given by E q . (3.2). Based on this approxi
mation, E q . (3.2), for the probability of t h e occupancy of a state by an electron,
becomes
f (E)
c

= 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

'

A t values of (E E) = 3kT, t h e exponential term in t h e d e n o m i n a t o r of E q .


(3.10) becomes very small so that the validity of this equation is restricted to values
of (E - E) > 3kT. E q u a t i o n (3.10) b e c o m e s
F

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

It is possible that in a highly d o p e d semiconductor, as we shall see later, t h e


Fermi level moves to within less than 3kT from the b a n d edges and even into the
bands. Such semiconductors are said to b e degenerate.
We conclude that for a n o n - d e g e n e r a t e semiconductor, t h e Fermi-Dirac equa
tion r e p r e s e n t e d by f for electrons and f, for holes respectively becomes
f (E)

= exp - ( E -

f (E)

= exp-(E -E)/kT)

Ep/kT)

(a)
(b)

(3.12)

Expressions f o r Electron a n d Hole Densities


Using t h e expressions in Eqs. (3.5) and (3.6) together with Eqs. (3.11) and (3.12) in
Eqs. (3.7) and (3.8), we have for electrons

For holes, t h e expression is

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

Intrinsic and Extrinsic Semiconductors

N (E)xf (E)

N(E)

N {E)

N (E)x[l-f (E

Ey

N (E)xf (E)
n

Ei
E
F

N(E)x[l-f (E
c

N (E)

~l-fc(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

A similar expression is obtained for as


l2m*kWn
P = 2(j -j

(E

Ey\

(3-17)

T h e expressions for and in Eqs. (3.16) and (3.17) can be written as


= Njxp-^j?*)

(3-18)

(3.19)
where TV and TV are given by
(3.20)

(3.21)
3

For m in kg., k in J / K and in J-s, TV is in m~ .


The terms TV. and TV^ are referred to as the effective density of states of t h e
conduction and valence bands respectively. The values for TV., and N are shown in
Table 3.1 where the difference b e t w e e n TV and TV result from t h e different values of
v

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

(3-22)

Intrinsic Carrier Density


In an intrinsic semiconductor, o n e to which n o impurities have b e e n a d d e d , at >
OK, the density of electrons in the conduction b a n d must equal the density of holes
in the valence b a n d because for every electron that is excited to the conduction

62

Chapter 3

Intrinsic and Extrinsic Semiconductors

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)

where is a constant i n d e p e n d e n t of t e m p e r a t u r e and , equal to (E E ), is


the b a n d gap energy of the semiconductor. High t e m p e r a t u r e s and smaller b a n d
gaps favor large values of intrinsic carrier density. It is i m p o r t a n t to n o t e that the
in t h e e x p o n e n t has a m u c h greater effect t h a n the factor . We will illustrate
these effects in E x a m p l e 3.2.
By substituting for the constants in E q . (3.24) and at = 300K, we find that
t h e intrinsic carrier density in silicon is approximately 1 X 1 0 c m , signifying that
this is the density of electrons in t h e conduction b a n d and the density of holes in the
valence band. Mainly because of t h e larger b a n d gap, the intrinsic carrier density in
gallium arsenide at = 300K is 2.49 X 1 0 c m .
T h e r e are 5 1 0 atoms per cubic centimeter in silicon and each a t o m has
four electrons in t h e covalent b o n d representing t h e 3s and 3p levels. These same
levels form t h e conduction and valence b a n d s in silicon. This results in a density of
2 1 0 electrons p e r c m available for conduction, all residing in t h e valence
b a n d at = OK. A t = 300K the density of electrons (in the conduction b a n d ) is
1 X 1 0 c m ~ . Approximately, therefore, o n e in 1 0 of t h e electrons available at
= OK at the t o p of the valence b a n d is elevated at r o o m t e m p e r a t u r e from the
t o p of the valence b a n d to t h e conduction band. This represents a very small propor
tion of t h e available electrons.
By substituting for t h e constant K in Eq. (3.24) from t h e expressions in Eqs.
(3.20) and (3.21), we have

1 0

- 3

- 3

22

23

10

13

. = f f f OWT expg)

(3.25)

We observe that t h e intrinsic carrier density is i n d e p e n d e n t of the location of


the Fermi level and d e p e n d s strongly o n the band gap energy, , and o n the tem
perature. T h e strong d e p e n d e n c e o n t e m p e r a t u r e is a result of trie rapidly varying
t e r m in t h e exponent since the n u m b e r of carriers that acquire t h e r m a l energy is
increased as m o r e of t h e electrons that are d e e p e r in t h e valence b a n d are able to
move u p to t h e conduction band. This d e p e n d e n c e is illustrated in Fig. 3.6. The
intrinsic carrier density decreases exponentially with an increase of , accounting
for the very low conductivity of insulators.
Calculations of t h e intrinsic carrier density are illustrated by E x a m p l e 3.2.
E X A M P L E 3.2
Calculate the intrinsic carrier density of silicon at:
a ) = 300K
b ) = 600K
Include effect of variation of with temperature.

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 2-2, to be 1.032eV. Using Eq. (3.24)
and (300K) = 10'cm " , we have
g

r(600K) _ / 6 0 0 \ e x p ( - 1 . 0 3 2 / ( 8 . 6 2 x 1 0 " X 600)


5

(300K)

UoO/

exp ( - 1 . 1 2 / ( 8 . 6 2 10"" X 300) .


5

n,.(600K) = 3.337 x 10 cm
15

= 1.137 10

63

64

Chapter 3

Intrinsic and Extrinsic Semiconductors

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

Location o f Fermi Level


O n e major assumption that we m a d e in deriving t h e expressions for t h e electron
and hole densities for t h e intrinsic semiconductor is that t h e Fermi level is at least 3
kT e V distant from t h e edges of t h e conduction and valence bands. We investigate
t h e strength of t h e validity of this assumption for an intrinsic semiconductor by
equating t h e expressions for t h e intrinsic electron and hole densities in Eqs. (3.18)
and (3.19) and solving for E , so that
F

We use t h e expressions for N a n d N from E q . (3.20) a n d (3.21) t o obtain


c

TV,
The expression for E
E

\m*l

(3.27)

\m*

becomes

, _ ^2

WS_SLS-W!S)
4

m*

< 3

2 8 )

\m*j

w h e r e (E + E )/2 represents t h e middle of t h e b a n d gap. T h e Fermi level for an


intrinsic semiconductor is labeled E so that
c

E = E =E
i

f- -kT<n(^j

(3.29)

We observe that E is t e m p e r a t u r e - d e p e n d e n t so long as t h e effective masses of


electrons a n d holes are unequal. T h e second t e r m in t h e right-hand side of Eqs.
(3.26) a n d (3.28) is of t h e o r d e r O.OleV for silicon at r o o m t e m p e r a t u r e . For all
practical purposes, o n e can state that t h e Fermi level in an intrinsic semiconductor,
E is approximately halfway b e t w e e n E and E , since E in E q . (3.28) is given by
E + E so that (E + E )/2 b e c o m e s (E + [E / 2 ] ) .
In t h e following example, we obtain o r d e r of m a g n i t u d e values for t h e loca
tion of E .
i

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

Bv using the values of and N. from Table 3.1, we calculate


2

For intrinsic silicon,


r/
h = (E

+ 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

For intrinsic gallium arsenide,


E = (E
t

+ 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
Q3-1
Q3-2
Q3-3
Q3-4

Determine the number of allowed states in a crystal that has JV atoms.


Why is the Boltzmann approximation used?
For what values of (E - E ) is Eq. (3.9) valid?
In the equations derived so far, the effective mass has been assumed constant. Is this
true? Why?
c

Q3-5 In Eq. (3.15), the upper limit of integration was set at infinity. Why?
Q3-6 Briefly explain why the narrower the band gap, the higher is the intrinsic carrier den
sity in a semiconductor.
Q3-7 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 Fermi-Dirac 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 one-half.
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

Intrinsic and Extrinsic Semiconductors

The intrinsic carrier density represents the density of each of the electrons in the con
duction band and holes in the valence band.

EXERCISES
E3-1 Determine the number of electronic states in the conduction band of silicon located
16

between energies of 1.0 and l.leV. Assume a volume of 10~ m .


E3-2 Calculate the intrinsic carrier density of Ge.
Ans:

13

n = 2.17 X 10 cm~

E3-3 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 N-type or P-type, where the electron density is increased
in the N-type and the hole density is increased in the P-type.
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.
N-type 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 jV-type
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 5-electron donor phosphorus atom to the covalent
bond structure generates a free electron since only four are required to complete
the bond.

E a c h impurity a t o m occupies the space of one silicon a t o m among, say, a mil


lion atoms of t h e semiconductor. For covalent bonding with the four nearest semi
conductor atoms, four electrons are n e e d e d and the fifth electron of t h e d o n o r a t o m
is not required in the crystal binding. This electron becomes a free electron influ
enced only by the crystal structure and it is said to be "loosely" b o u n d . A sketch of
t h e resulting structure is shown in Fig. 3.7.
In the b a n d picture, the d o n o r atoms occupy energy levels in the forbidden
b a n d that are slightly below the b o t t o m of the conduction band.
The free electrons that are generated intrinsically escape from a semiconduc
tor covalent b o n d w h e n that b o n d absorbs t h e r m a l energy. Using the b a n d picture,
t h e intrinsically generated semiconductor electron gains enough energy to j u m p
across the b a n d gap and becomes free. T h e b a n d gap energy is about l . l e V in silicon
at = 300K. The question is then, how m u c h energy is required to release one elec
tron from a d o n o r a t o m ?
We will use the binding of energy of t h e hydrogen electron to obtain an esti
m a t e of t h e binding energy of the d o n o r electron. The hydrogen electron in its
ground state, 1, has energy given by B o h r ' s relationship, E q . (1.14), to be

A n energy of 13.6eV is required to free t h e hydrogen electron.


To obtain a similar quantity of energy for our d o n o r electron, we have to cor
rect the expression for t h e energy of t h e hydrogen electron by replacing the relative
value of the mass of the d o n o r electron and t h e relative permittivity of the semicon
ductor.
Because t h e hydrogen electron is in t h e field of t h e hydrogen nucleus, whereas
the d o n o r electron is u n d e r t h e influence of t h e potential field of t h e semiconductor

Chapter 3

Intrinsic and Extrinsic Semiconductors

lattice, we expect the masses to b e different. The d o n o r electron is moving in a semi


conductor lattice that has permittivity where , the relative permittivity, has a
value of 11.8 for silicon.
Using the above information, we obtain the approximate expression for the
binding energy of a d o n o r electron, E , in silicon to be,

~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 charge-neutral.

P-Type S e m i c o n d u c t o r
A semiconductor is said to be P-type 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

Intrinsic and Extrinsic Semiconductors

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.

r o o m t e m p e r a t u r e , all d o n o r atoms are positively ionized, each a t o m having lost an


electron to the conduction band. Similarly, each acceptor a t o m is negatively ionized
at r o o m t e m p e r a t u r e . Listed in Table 3.2 are the results of m o r e accurate calcula
tions of t h e ionization energies of donors and acceptors in silicon. The
ionization
energy is defined as the energy required to ionize a donor or an acceptor atom.
T A B L E 3.2
Silicon

Ionization Energies of Impurities in

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

Source: Reprinted from S. M. Sze and J. C. Irvin,


"Resistivity, Mobility and Impurity Levels in GaAs,
Ge and Si at 300K," Solid State Electronics 11,599
(1968), with kind permission from Elsevier Science
Ltd.,The Boulevard, Langford Lane, Kidlington
0X5#1GB,UK.

In addition to t h e controlled addition of donor and acceptor impurities, other


impurities may exist at energy levels m o r e distant from the b a n d edges than the
controlled impurities. These impurities are randomly available and usually act as
centers of recombination generation or traps. Because they exist a r o u n d t h e middle
of the energy gap, they can act as energy levels at which electrons and holes m e e t to
recombine. M o r e will be said later about recombinations.

Section 3.6
3.5

Densities of Carriers in Extrinsic Semiconductors

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 N-type semicon
ductors, t h e m o r e n u m e r o u s carriers, as the majority carriers. Holes in N-type semi
conductors are minority carriers. In P-type 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)

In order to determine the values of n and p for a d o p e d semiconductor, we


n e e d a n o t h e r relationship that relates them.
0

Chapter 3

Intrinsic and Extrinsic Semiconductors

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

ND-NA

(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 N-type semiconductor, either N
N or N is zero. F r o m E q . (3.32),
we find n for an TV-type semiconductor to be,
D

1/2

=
2

N\
D

(3.33)

The density of the holes is t h e n calculated from E q . (3.30) to be,


Po = Z-

(3-34)

In solving Eqs. (3.30) and (3.31) simultaneously and w h e n numerical values


are given for N and N , it is always advisable to obtain a quadratic equation for the
larger expected carrier density, the expected majority carrier density. The r e a d e r
should try solving for the minority carrier density to appreciate this caution.
If the d o n o r density N is m u c h greater than , which is quite c o m m o n in
most extrinsic semiconductors, E q . (3.33) can b e simplified so that
D

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

in silicon at = 300K, where n. = 1 X

1 6

n = 10 cm
Q

- 3

and p = 1 X 1 0 c m " .
Q

Section 3.6

Densities of Carriers in Extrinsic Semiconductors

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 .

We note, therefore, that the density of majority carriers is m u c h greater than


the density of minority carriers.
A n interesting supplement to t h e results of this section is obtained by plotting
the ratio of the majority carrier density to the d o n o r density as a function of temper
ature, as shown in Fig. 3.11.
We assume that a d o n o r impurity having density N of 1 0 c m ~ is a d d e d to a
silicon sample. A t t e m p e r a t u r e s n e a r zero K, the electron density in the conduction
b a n d is zero since t h e thermal energy at that t e m p e r a t u r e is not sufficient to ionize
any d o n o r atoms and certainly not enough to excite electrons from t h e valence b a n d
into the conduction band. A s the t e m p e r a t u r e is increased, some d o n o r atoms
b e c o m e ionized, losing their electrons to the conduction band, while t h e r e is still not
sufficient energy to m o v e electrons out of t h e valence band.
We notice that u p to about 150K, the electron density is smaller t h a n the
d o n o r density because not all d o n o r atoms are ionized and very few electrons have
b e e n excited from t h e valence b a n d to the conduction band. The region 0-150K is
k n o w n as the freeze-out region. A t t e m p e r a t u r e s b e y o n d 150K, all d o n o r atoms are
assumed to b e c o m e ionized so that = TV^.This concentration of electrons remains
constant at the value of N u p to r o o m t e m p e r a t u r e , at which N
n The range
from = Np = N u p to where n begins to increase b e y o n d its r o o m t e m p e r a t u r e
value is k n o w n as the extrinsic region. This range extends from about 150 to 400K.
A b o v e 400K the intrinsic carrier density begins to increase rapidly and eventually
reaches a value greater than N . The material begins to b e c o m e intrinsic again and
the electron density b e c o m e s m u c h larger than N because of t h e large n u m b e r of
electrons that are thermally excited from the valence b a n d to the conduction band.
For t e m p e r a t u r e s exceeding 400K, t h e region is labeled the intrinsic region. In the
intrinsic region of Fig. 3.11, the densities of b o t h holes and electrons increase, how
ever, their thermal equilibrium values are still d e t e r m i n e d by Eqs. (3.30) and (3.31).
16

74

Chapter 3

Intrinsic and Extrinsic Semiconductors

Calculations of electron and hole densities in extrinsic silicon are shown in


Examples 3.4 and 3.5.

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)

Neutrality required that


+N
+9

= + N

x 10 = + 10"
16

(2)

Solving equations 1 and 2 simultaneously, we have


16

= 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)

Solving equations 3 and 4 simultaneously, we have


16

= 1.1 x 1 0 c n r
M

= _;-= 1.014 1 0 ' W

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

Densities of Carriers in Extrinsic Semiconductors

a) = 300K

75

b) = 600K

Solution
a) At = 300K, we find that
10

n- = 1 X 10 crrr

np

1 x HPcm *

(1)

Neutrality requires that


~

+10"

(2)

Solving equations 1 and 2 simultaneously, we have


17

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)

Solving equations 3 and 4 simultaneously,


,7

n = 1.0011 X 10 cm-

The hole density is calculated to be


P=jp=

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

We have d e t e r m i n e d in Section 3.3 that, for an intrinsic semiconductor, the Fermi


level E lies approximately halfway b e t w e e n the b o t t o m of the conduction b a n d
and the top of t h e valence b a n d at (E + E )/2. We have d e n o t e d this intrinsic
Fermi level by t h e symbol E . Since n is useful in calculating carrier densities in
extrinsic semiconductors, we will find E useful as a reference level when dealing
with extrinsic semiconductors.
E q u a t i o n (3.18) can therefore b e modified by replacing E by E. and n by , so
that
F

n, =

^^^

(3.36)

Similarly, and since we are referring to an intrinsic semiconductor. Eq. (3.19)


becomes

76

Chapter 3

Intrinsic and Extrinsic Semiconductors

N zx (^-B

(3.37)

/V exp^ = n exp^

(3.38)

7V exp^ = n exp^;

(3.39)

n =

E q u a t i o n s (3.36) and (3.37) give


c

and
t )

Replacing t h e terms on t h e left-hand side of Eqs. (3.38) and (3.39) by their


equivalence from Eqs. (3.18) and (3.19), and solving for and p, we have
3

n = n e x p p T ^ ) =

(^ =

40

( )

(3.41)

For an extrinsic s e m i c o n d u c t o r in t h e r m a l equilibrium, t h e and in Eqs.


(3.40) and (3.41) h a v e b e c o m e n and p . I n each of these two equations, t h e vari
ables a r e t h e t h e r m a l equilibrium carrier densities a n d t h e energy distance
b e t w e e n E a n d E . Thus, for an intrinsic semiconductor, E is at E., w h e r e a s for
an extrinsic s e m i c o n d u c t o r a n d as n and p change, t h e distance b e t w e e n E
a n d E changes. T h e location of E. with respect to t h e c o n d u c t i o n and valence
b a n d edges has n o t changed. Therefore, E provides a useful reference for extrinsic
semiconductors.
0

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 N-type semiconductor,
where = n = N , as
Q

- Ei = kTn^
n

= kT in
i

(3.42)

For a P-type s e m i c o n d u c t o r , = p N , and from E q . (3.41) we have


Q

Ei - E =
F

kTln^

= kT In ^

(3.43)

We observe from Eqs. (3.42) and (3.43) that for an TV-type semiconductor the Fermi
level is higher t h a n E and moves u p closer to the conduction band. For a P-type
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

Fermi Level In Extrinsic Semiconductors

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

(c) Repeat part (a) if = 600K.


18

(d) Repeat part (a) if 10 atoms replace the 10 * atoms of phosphorus.


Solution
From Example 3.3, we have E at 300K to be 0.0073eV below the middle of the band gap.
(a) By using Eq. (3.42) for phosphorus donor atoms, where n = N . we have
i

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.
(

(b) We use Eq. (3.43) for boron acceptor atoms, w h e r e = N , as

- E = 0.02583 in 7 - ^ 0 = - 98 V
F

and from Example 3.3 we have


E + E
, =

v1

+ 0.255eV and /,.. ~ /,,. - 0.865cK

E **E
F

,
- 0.0073, also E = E + 1.12, we find

15

(c) At = 600K, . is determined in Hxample 3.2 as 3.337 x 10 cm \ E, = 1.032cV and


1 2 X 10'*
n = 1.2 10 cm"" , so that E = 0.05166 3
i5- F
/ 0.065eK
16

3 3 7

Assuming that E. is approximately at (E + )/2,


of the band gap.
c

18

1 Q

E is located approximately in the middle

10

(d) E - E, = 0.02583 Cn (10 /10 ) = 0.475eF, E = . + 0.475.


E + Ey
, ..
.
Using
= ^ \ we have E = E + 0.56 + 0.475 = E + 1.035 =
F

- 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

Intrinsic and Extrinsic Semiconductors

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 semi-conductor.
r

19

REVIEW QUESTIONS
Q3-8
Q3-9

In an intrinsic semiconductor, explain why E is not located exactly midway between


P

the valence and conduction bands.


As the temperature of a sample of semiconductor that is doped with donor impurities
is increased, in what direction does E move?
Briefly define the phrase "thermal equilibrium"?
What is a degenerate semiconductor?
The configuration of electrons in a certain element is 4s 4p . Is it a donor or an accep
tor element?
By referring to the periodic table, determine which element is represented by the
configuration 4s 4p .
The density of atoms in silicon is 5 X 10 cm~ . Since each atom has four valence elec
trons, the density of states becomes 2 10 . Why are N and N so much smaller than
this number?
P

Q3-10
Q3-11
Q3-12
Q3-13

Q3-14

22

23

HIGHLIGHTS

By adding controlled amounts of impurities to semiconductors, the densities of holes


and electrons can be markedly changed. These impurities may be elements either from
Column V of the periodic table, known as donors, or they may be from Column III,
known as acceptors. Phosphorus and arsenic are donors; boron and gallium are accep
tors.
When donor impurities are added to an intrinsic semiconductor, the more numerous
carriers, electrons, are called majority carriers and the holes are minority carriers.
When both donors and acceptor impurities are added, the one with the larger density
determines which carrier becomes the majority.
A unique property of semiconductors is the large changes that are made in the densi
ties of the carriers by the addition of impurities, also known as dopants. The changes in
the carrier densities are translated into changes in conductivity.
A semiconductor that has been doped is known as an extrinsic semiconductor.
Donor and acceptor atoms occupy energy states in the forbidden band, donor atoms
are close to E , and acceptor atoms are close to E . In these locations it takes very lit
tle energy to excite a donor electron to go into the conduction band and also to excite
an electron from the valence band to an acceptor level, thus generating a hole. A
donor atom that loses an electron and an acceptor atom that gains an electron are said
to be ionized.
In intrinsic semiconductors, the Fermi level E is located close to the middle of the
band gap. The Fermi level moves up towards the conduction band when donors are
added and moves closer to the valence band when acceptors are added.
c

Which Semiconductor?

Section 3.8

79

EXERCISES
E3-4 A sample of semiconductor is doped with N
is 10 cm~ . Determine and p.

13

13

13

= 10 cm~ . The intrinsic carrier density

Ans: = 1.61 X 10 cm~


0 5

E3-5 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

E3-6 Repeat Ex. 3-6 for silicon at 300K


Ans: 4 "
3

10

E3-7 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: = 1.08 X 10 cm16

E3-8 A sample of silicon is doped with N = 5 X 1 0 c n r at room temperature. Determine


the location of the Fermi level relative to the conduction band edge.
D

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

Intrinsic and Extrinsic Semiconductors

results because a higher t e m p e r a t u r e increases t h e density of electrons that are


excited to the conduction band, thus increasing the current that results in higher
heat dissipation, which increases t h e t e m p e r a t u r e . The m o r e serious p r o b l e m is the
difficulty of introducing controlled a m o u n t s of impurities into small selected areas.
Because of this, one has to work with large areas. This results in carriers having to
take longer times in covering distances within t h e device. This longer travel time
m e a n s slower switching speed.
Silicon has important advantages that are u n m a t c h e d by germanium. It is
a b u n d a n t in n a t u r e in t h e form of sand and quartz. Thus, t h e cost of t h e starting
material is negligible. Because silicon has a wider energy gap (forbidden b a n d ) than
germanium, it can b e used at higher temperatures, thus greatly reducing a cause of
instability. Silicon devices can b e o p e r a t e d safely at t e m p e r a t u r e s of about 200C,
whereas g e r m a n i u m devices are limited to a safe operating t e m p e r a t u r e of 80C.
Silicon has a major processing advantage in that it forms a stable oxide, silicon
dioxide. The silicon dioxide offers a t o p quality insulator and w h e n used in device
processing, it provides a good barrier for the diffusion of impurities in certain
selected areas. F u r t h e r m o r e , it b e c o m e s possible to operate with very small dimen
sions of the o r d e r of one micron or less. Faster switching and higher frequency limits
are the benefits.
Thus, silicon is a t o p quality semiconductor while providing an insulator with
excellent properties.
Gallium arsenide has t h e major advantage of a mobility that is approximately
five times that of silicon, since a higher mobility is translated into a higher velocity
of t h e carriers for a fixed electric field. Of course, this property offers possibilities of
faster switching. It is, however, m o r e difficult to process and is m o r e expensive than
silicon. O n e of its major applications, as we shall see later, is in optical devices.
We have listed in the accompanying table some properties of t h e t h r e e major
semiconductors.
TABLE 3.3 Properties of Silicon, Germanium and Gallium Arsenide at = 300K
Property

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

3.3 Repeat Problem 3.2 if the phosphorus doping is at a level of 10 cm~ .


16

3.4 Repeat Problem 3.2 for a boron doping of 10 cm~ .


3.5 The Fermi level in a silicon sample at equilibrium is located QAOeV below the middle
of the band gap. At = 300K,
a)
Determine the probability of occupancy of a state located at the middle of the
band gap.
b)
Determine the probability of occupancy of the acceptor states if the acceptor
states are located at 0.04eV~ above the top of the valence band.
c)

Check if the assumption of complete ionization in part (b) is valid.


19

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)

For Problem 3.5, determine the thermal equilibrium electron density.


Determine the thermal equilibrium hole density.
Calculate the p n product.
0

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

3.8 A silicon sample at 300K has an acceptor density of 10 cm~ . Determine:


a)
The hole density.
b)
The electron density.
10

3.9 A certain silicon sample at 300K has p = 4 X 1 0 c n r . Determine:


a)
The electron density.
b)
The acceptor density if the donor density is 1.25 10 c m .
0

10

16

3.10 A certain silicon sample at 300K has 7V = 1 X 10 cm- and N


Determine:
a)
The majority carrier density.
b)
The minority carrier density.
D

-3

16

= 0.8 X 10 cm- .

Chapter 3

Intrinsic and Extrinsic Semiconductors

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 TV-type 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
cross-sectional area of the bar and t h e velocity with which the carriers move. For an
/V-type 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
}

The question we n e e d to answer is: H o w d o we d e t e r m i n e the m e a n velocity of


the carriers? Obviously, t h e velocity d e p e n d s on the acceleration, which in turn
d e p e n d s on t h e electric field intensity. For a constant electric field, w h a t is t h e limi
tation to the velocity that t h e carriers can acquire?
In our discussions in earlier chapters, we assumed that the solid possesses a
perfectly periodic lattice and disregarded any irregularities in it. In a metal at >
OK, electrons are continuously in motion due to the t h e r m a l energy they acquire. If
a constant force is applied to the electrons, as by an electric field, t h e electrons in the
perfectly periodic crystal would b e accelerated, gain energy, move to empty states at
higher energy levels within the same band, and continue to gain in velocity and
energy.
Electric current is proportional to the n u m b e r and velocity of the electrons
and, therefore, the current in a material to which an electric field is applied, in the
absence of o t h e r p h e n o m e n a that might limit its value, will tend to increase without
limit as t h e velocity increases. However, we k n o w from O h m ' s law that the current
reaches a fixed value indicating a constant velocity. The velocity we are referring to
is a mean velocity, averaged over t h e velocities of all t h e free electrons.
T h e m e a n velocity is constant because of a resistive force that prevents further
acceleration. This resistive force is a direct result of t h e collision of t h e electrons
with atoms or with the regions in t h e lattice of t h e material that disturb t h e motion
of the electrons. T h e collison could also b e with foreign atoms.
Before further discussion of t h e n a t u r e or effect of t h e collisions, let us present
the condition of the electrons in a semiconductor w h e n they are not u n d e r t h e influ
ence of an electric field.

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

Carrier Processes: Drift, Diffusion and Generation Recombination

A s a m e a s u r e of t h e energy of the electron, we establish t h e level of the energy


of an electron at rest to b e at the b o t t o m of t h e conduction b a n d E . The kinetic
energy of an electron is m e a s u r e d by the energy separation above E , as E
It has b e e n established that, at thermal equilibrium, the m e a n - s q u a r e thermal
velocity of the electron is related to t e m p e r a t u r e by t h e relation*
c

(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 low-intensity 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 field-free 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 quantum-mechanical 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 non-stationary
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

Collisions and Scattering

87

trons and t h e atomic core. It is caused m o r e by t h e variation in the potential distri


butions in the solid w h e n the crystal structure is not perfectly periodic. Such a solid
is labeled aperiodic. Aperiodicity is caused by t h r e e factors, one of which is a tem
p e r a t u r e of the solid above zero. Aperiodicity may also result from imperfections in
t h e solid and from the presence of impurities. T h e r e are therefore t h r e e p h e n o m e n a
that individually or jointly cause departures from t h e periodic potential.
A t n o r m a l operating temperatures, t h e lattice atoms vibrate about their equi
librium positions. These vibrations disturb t h e periodicity and change the periodic
potential, which generates an electric field that scatters t h e carriers. A n electron
travelling through t h e lattice undergoes scattering, which causes changes in the
magnitude and direction of its velocity. T h e scattering results b o t h from collisions
with the vibrating atoms themselves and with t h e varying potential field that results
from the displacement of the vibrating atoms. The higher t h e t e m p e r a t u r e , the
larger is t h e amplitude of the atomic vibrations, the larger is the scattering cross sec
tion, the larger is the disturbance of t h e potential field, and therefore t h e higher the
probability of the scattering taking place.
A second mechanism that causes scattering results from the presence of ion
ized impurity atoms. These atoms p r o d u c e an electric field that changes t h e poten
tial distribution in the crystal. This type of scattering is m o r e important at lower
t e m p e r a t u r e s w h e n the t h e r m a l scattering is weaker. The higher the impurity con
centration, t h e higher is the probability of this type of scattering.
A third cause of scattering, although less important t h a n thermal or impurity
scattering, results from imperfections such as vacancies in the crystal structure and
crystallographic defects.
Crystal imperfections cause a change in t h e periodic potential and thus create
an electric field. A n electron passing near an imperfection interacts with the field
resulting in a change of the direction of motion.
We conclude that at lower temperatures, impurity scattering is m o r e domi
nant, while at the higher temperatures, t h e r m a l (lattice) scatterings is m o r e
effective.

88

Chapter 4

Carrier Processes: Drift, Diffusion and Generation Recombination

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 x-direction, obeys t h e equation,
d
m

dx
-

(4-2)

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 steady-state, 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

w h e r e T is the m e a n time b e t w e e n collisions, k n o w n as the m e a n scattering time,


and the drift velocity is given by
c

(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

O n comparing Eqs. (4.3) and (4.4), we find that . = m*/w


becomes
1 exp

v[t) =

(4.4)
so that E q . (4.4)

(4.5)

Thus, after a certain time t > , the particle velocity will have a steady-state
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 steady-state 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 field-directed 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 N-type 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

Carrier Processes: Drift, Diffusion and Generation Recombination

at A is zero. Since t h e b o t t o m of t h e conduction b a n d at E is t h e level of the poten


tial energy of an electron at rest, t h e n the energy b a n d diagram for an electron in
this illustration shows a decreasing potential energy and hence a decreasing level of
E as we p r o c e e d from to A in Fig. 4.2(b). This explains t h e slope of t h e energy
b a n d diagram.
A n e l e c t r o n r e l e a s e d from rest at p o i n t in t h e bar is accelerated by t h e
electric field in Fig. 4.2(a), acquires kinetic energy, a n d just before p o i n t C, its
total e n e r g y is u n c h a n g e d as it replaces t h e p o t e n t i a l energy it lost by kinetic
energy. W h e n it collides with t h e lattice structure at p o i n t C, it loses all t h e kinetic
energy it acquired, its total energy is p o t e n t i a l , a n d t h e electron d r o p s to t h e level
E c o r r e s p o n d i n g to p o i n t C. It is accelerated again, collides with t h e lattice, and
d r o p s b a c k to t h e lower E level at p o i n t D. This process continues until t h e elec
t r o n r e a c h e s p o i n t A , w h e r e it has lost all t h e qV p o t e n t i a l energy that it h a d at
p o i n t B. This p o t e n t i a l energy has b e e n c o n v e r t e d , along t h e way, into h e a t in t h e
lattice.
c

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

where v is the drift velocity of an electron in c m / s , is its mobility in c m / v o l t - s e c


and % is t h e electric field intensity in V / c m . The negative sign denotes that, for an
electron, the direction of the velocity is opposite to the direction of the electric field
intensity. By using E q . (4.3), in E q . (4.7), the mobility of t h e electron becomes
d

< 4

8 )

The analogous expressions for hole m o t i o n are


= ^

(4-9)

and the mobility of the hole is

- %

410

< >

Because the time interval b e t w e e n collisions depends mainly on t e m p e r a t u r e and


impurity concentration, the mobility depends on those two p a r a m e t e r s as well as on
the material through the effective mass. With reference to Fig. 2.12, which showed a
m u c h greater curvature near the b o t t o m of the conduction b a n d of G a A s t h a n that
of Si, the effective mass of an electron in G a A s is m u c h smaller than in Si. A s a
result, t h e electron mobility in G a A s is four to five times greater than in silicon.

Section 4.5

Mobility

91

GaAs electrons

I0

1(P

10

LO

io

Electric field (V/cm)


Figure 4.3 Variation of drift velocity with electric field intensity for gallium
arsenide and silicon. Source: Reprinted with permission from Ruch and Kino,
"Measurement of the Velocity-Field Characteristics of Gallium Arsenide," Applied
Physics Letters 10,40 (1967) copyright American Institute of Physics; and Caughey
and Thomas "Carrier Mobilities in Silicon Empirically Repated to Doping and
Field," Proc. IEEE 55,2192 (1967) IEEE.

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

The d e p e n d e n c e of the mobility on the impurity concentration, through the m e a n


scattering time T , results from the scattering of the carriers by the coulombic effect
of the ionized impurities as t h e carriers travel in the solid. High doping density
c

92

Carrier Processes: Drift, Diffusion and Generation Recombination

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.

m e a n s m o r e scattering, lower drift velocities, and mobilities. The variation of mobil


ity with scattering through its d e p e n d e n c e on impurity concentration and t e m p e r a
ture is shown in Fig. 4.4 and highlighted in the inset.
The question t h e n is: Why is it that, in the inset of Fig. 4.4, the mobility is seen
to increase with an increase of t e m p e r a t u r e in a certain t e m p e r a t u r e range, yet
decreases with an increase of t e m p e r a t u r e in a n o t h e r t e m p e r a t u r e range?
A t low t e m p e r a t u r e s , lattice vibrations are so small so that as t h e t e m p e r a t u r e
increases, t h e carriers receive t h e r m a l energy that increases their velocity, thus
reducing the time that the electron spends in the vicinity of the lattice and impurity
atoms. This reduces t h e scattering cross section, which causes the mobility to
increase with t e m p e r a t u r e .
T h e increase of mobility with t e m p e r a t u r e continues until, at a certain temper
ature, t h e lattice vibrations d o m i n a t e and the scattering cross section increases, thus
increasing the possibility of collision with t h e lattice and decreasing the mobility
with an increase of t e m p e r a t u r e .

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

W h e r e . and refer to the collision times due to impurities and t e m p e r a t u r e


respectively. The mobility can be expressed in terms of the mobilities due to impu
rities and d u e to t e m p e r a t u r e and , respectively as
(

(4.12)

It has b e e n shown that the mobility due to lattice scattering ( t e m p e r a t u r e ) , ,


is proportional to 1/T .
The mobility caused by impurity scattering, , has
b e e n d e t e r m i n e d to be proportional to T /N,
where is the total impurity density
(

312

3I2

(N

N ).
A

Empirical expressions for the mobilities of electrons and holes in silicon in


c m / V - s , as a function of doping density and t e m p e r a t u r e have b e e n derived,* and
shown below, w h e r e is in , = /300, and is in c m .
2

- 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

A t = 300 these expressions reduce to:


1252

"

(4.15)
17

1 + 0.698 X 10- 7V
54.3 +

407
1 + 0.374

17

10~

These relations agree fairly well with experimental results.


*From Arora et at, "Electron and Hole Mobilities in Silicon as a Function of Concentration and Tem
perature," IEEE Transactions on Electron Devices, Vol. 29, No. 2, p. 2192, February 1982. 1982, IEEE.

(4.16)

Chapter 4

Carrier Processes: Drift, Diffusion and Generation Recombination

Impurity concentration (cm )


Figure 4.5 Mobilities of electrons and holes for Si, Ge and GaAs at = 300K, as
a function of total impurity density (N + N ). Source: Reprinted from S. M. Sze
and J. C. Irvin, "Resistivity, Mobility and Impurity Levels in GaAs, Ge and Si at
300K," Solid State Electronics, 11, 599 (1968) with kind permission from Elsevier
Science Ltd., The Boulevard, Langford Lane, Kidlington 0X5 1GB, UK.
D

Plots of m e a s u r e m e n t s of the mobility versus total impurity concentration


(N + N ), at = 300K, are shown in Fig. 4.5. We observe that at low concentration
levels t h e mobility is i n d e p e n d e n t of the concentration.
We conclude this section by summarizing the factors that influence the mobility:
D

1. T h e mobility d e p e n d s very strongly on the t e m p e r a t u r e and on irregularities in


the crystal, which may b e d u e to the growing of the crystal or t h e impurities.
2. So long as the drift velocity is less than t h e thermal velocity, the mobility varies
directly with, and is proportional to, t h e time b e t w e e n collisions.
3. If t h e drift velocity approaches t h e saturation velocity through an increase of
the electric field intensity, the mobility decreases with an increase of t h e electric
field.

Drift

Section 4.6

95

Current and Conductivity

4. The mobility, therefore, is affected to varying degrees by: irregularities in the


lattice structure, t e m p e r a t u r e , concentration of added impurities, and the elec
tric field intensity.

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)

w h e r e A is the cross sectional area in c m , n o r m a l to the direction of current flow, q


is t h e charge in coulombs, is t h e electron density in c m " , v is the electron drift
velocity in c m / s e c , and I is in amperes. T h e direction of t h e current is opposite to
that of the motion of t h e electrons. By replacing for the velocity from E q . (4.7) into
E q . (4.17), we have
3

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

The factor multiplying % is k n o w n as the conductivity

so that

J = %

(4.19)

w h e r e the unit of the conductivity is ( o h m - c m ) " . Analogously the current density


for holes becomes
J =
p

p%

= %

qVlp

(4.20)

The total drift current density, / , is


(

J = J +J
t

=%+%

(4.21)

'

The total conductivity is

= <? +

(4-22)

The addition of small traces of impurities has dramatic effects o n t h e conduc


tivity of semiconductors. We illustrate this by an example.

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

Carrier Processes: Drift, Diffusion and Generation Recombination

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^ohm-cmF

;)

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 /V-s and =
350cm /V-s. The extrinsic conductivity becomes
2

s 1.6 X 1 0 " (825 X 10' + 350 10 ) = miohm-cm)23

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 ohm-cm. In general,
t h e semiconductor is either /V-type or P-type. For a semiconductor in t h e r m a l equi
librium, t h e resistivity becomes

= Vi^ 0 +
n

Po)

23

(- )

where n and p a r e t h e electron a n d hole densities, in thermal equilibrium, respec


tively in c m . For an /V-type semiconductor n
p and n = N , E q . (4.23)
becomes
Q

- 3

as \/{qv N )
n

(4.24a)

For a F-type 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
Q4-1 Identify the processes that cause scattering.
Q4-2
Q4-3
Q4-4
Q4-5

Upon what physical factors does mobility depend?


Why is the electron mobility in GaAs about three times that in silicon?
Why and how does the mobility depend on doping? Explain.
An electric field is applied to a sample of semiconductor in thermal equilibrium. If the
field is in the positive x-direction, in what directions are the motions of holes and elec
trons?
Q4-6 In the above question, in what directions are the electron and hole currents?

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

Resistivity and Resistance

2 3 456 8

18

1 n

l9

2 3 456 8
1 Q

97

2 3 456 8

20

1 n

2l

-3

Impurity density (cm )


Figure 4.6 Resistivity versus impurity density at 300K for Si, Ge, and GaAs.
Source: Reprinted from S. M. Sze and J. C. Irvin, "Resistivity, Mobility and Impurity
Levels in GaAs, Ge and Si at 300K," Solid State Electronics, 11, 599 (1968) with
kind permission from Elsevier Science Ltd.,The Boulevard, Langford Lane,
Kidlington 0X5 1GB, UK.
HIGHLIGHTS

Electrons in a semiconductor at > 0 are continuously moving at a very high velocity


in all directions. This velocity is known as the mean thermal velocity. At thermal equi
librium, the net velocity of all electrons in any one direction is zero, therefore the cur
rent is zero.
When an electric field is applied to a semiconductor, electrons acquire energy and a
velocity that is much smaller than the mean thermal velocity. Now there is a net mean
velocity opposite to the direction of the electric field. This velocity is known as the
drift velocity.
The reason for the limited drift velocity is the electrical interaction of the electron
with the imperfect solid.
The factor that relates the velocity to the electric field intensity is labeled the mobility.
At low values of electric field the velocity increases linearly with an increase of the
field resulting in a relatively constant mobility. Just as electrons in the conduction
band acquire drift velocity and mobility, so do the electrons in the valence band.

98

Chapter 4

Carrier Processes: Drift, Diffusion and Generation Recombination

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
E4-1

15

A sample of silicon is doped with 10 c m


boron. Determine:
a)

the density of holes and electrons

b)

the mobility of holes and electrons


Ans:

15

- 3

15

of phosphorus and 5 X 1 0 c n r of

a) = 4 X 10 ~ , = 2.5 X 10 cnr
2

b) = 350cm /V-s, = 1500cm /V-s

E4-2

15

b)

the conductivity of the sample


Ans:

a) v (n) = 5 X 10 cm/s, v (p) = 2 X 10 cm/s


d

b) = 0.16 (ohm-cm)
E4-3

-1

For the sample in E4-2 calculate a) the resistivity b) the drift current density /
Ans:

E4-4

An electric field is applied to a sample of Ge that is doped with 1 0 c n r of boron.


The field has intensity of 100 V/cm. Calculate:
a)
the velocity of electrons and holes

a) = 3.12 ohm-cm,

b) / = 32 A / c m

Determine the resistance of the sample in E4-2 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 ohm-cm. Assume that arsenic is the only dopant.
Determine the impurity concentration.
Solution

Arsenic is a donor impurity; hence the resistivity is given by,


1
=

<7/V*

Both and are unknown.


One method to determine is to use Fig. 4.6 and read N to be approximately 9 x 10 cm .
To confirm this value, we use Fig. 4.5 to read the mobility for N - 9 10 and find as
1400 cm /s from which we calculate the resistivity to be 4.96 ohm-cm.
One can, therefore, assume a starting value for N determine the mobility from Fig. 4.5, calcu
late the resistivity, and then refer to Fig. 4.6 to obtain a new value for N .
The resultant impurity concentration is 9 X 10 cm" .
14

14

[r

14

Section 4.7

Resistivity and Resistance

99

E X A M P L E 4.3
2

A sample of silicon at 300K of length 2.5cm and a cross-sectional 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,

We need to determine and /; from the following relations


+N

+ N

and pn = nj

The results were determined in Example 3.4 to be


= 10"'cm

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 /V-s

and

, = 292cm /V-s so that,

r - U X 1CT X 626 10 = 1.001 (ohm-cm)"


19

16

= 1.6 x 10~ x 292 X 1 10 = 4.67 X 10


19

13

(ohm-cm)"

b) The total conductivity becomes


1

= ^ + a s 1 (ohm-cm)" and the resistivity is = 1/


f

= /='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

Carrier Processes: Drift, Diffusion and Generation Recombination

resistivity has two c o m p o n e n t s : one i n d e p e n d e n t of t e m p e r a t u r e and t h e other tem


perature-dependent.
Since resistance can b e defined as the ratio of applied voltage to current, an
increase of current with t e m p e r a t u r e signifies decreased resistance. T h e current is
proportional to the n u m b e r of carriers and to their net m e a n velocity in the direc
tion of t h e applied electric field. In a metal, w h e n the t e m p e r a t u r e is increased, the
density of free electrons hardly changes because of the already large density, but the
average drift velocity decreases because of the increased vibration of the a t o m s and
molecules. Thus, with an increase of t e m p e r a t u r e , the resistance of a metal increases.
Semiconductors, on the other hand, have a negative t e m p e r a t u r e coefficient of
resistance. Their resistance decreases with t e m p e r a t u r e . In some ways, this property
is an advantage of semiconductors; on t h e o t h e r hand, it is a source of instability.
W h e n t h e t e m p e r a t u r e increases, the incremental velocity of the carriers d u e to the
electric field decreases, but the density of the carriers (free electrons and holes)
increases appreciably. The increase in the density of carriers m o r e than offsets the
decrease of their velocity. Electrons from t h e valence b a n d acquire t h e r m a l energy
and m o v e into the conduction b a n d to b e c o m e free electrons, leaving vacancies in
t h e valence b a n d that account for the holes. Thus, the current increases with an
increase of t e m p e r a t u r e , which decreases t h e resistance, thus increasing t h e current,
which then increases the t e m p e r a t u r e . Instability results and a runaway situation
arises, which causes the destruction of the semiconductor.
Thus, we observe that irregularities in a material cause it to resist the flow of
current, and that, w h e n the t e m p e r a t u r e is increased, the resistances of metals and
semiconductors vary in an opposite m a n n e r .

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. Electron-hole 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

Particle Diffusion and Diffusion Current

"

= -

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 P-semiconductor; (c) hole diffusion and hole diffusion
current; and (d) electron diffusion and electron diffusion current.

will cross in the direction of decreasing density. T h e particle flow across x


positive direction, is given by

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 x-direction. In t h r e e dimensions, E q . (4.25)
becomes
1

F = -DVP

(4.26)

In the process of particle diffusion in semiconductor devices, the flux consists


of the m o t i o n of electrons or holes across a surface. We will translate t h e r a t e of
motion of carriers into an electric current since the motion involves charged elec
trons or holes. We consider in Fig. 4.7(b) a P-semiconductor region into which elec
trons are either injected at = 0 from an N-semiconductor or are generated by a

102

Chapter 4

Carrier Processes: Drift, Diffusion and Generation Recombination

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

w h e r e t h e gradient dn/dx has unit of electrons per c m per cm of length, D is in


c m s so that the unit of F is electrons c m ~ s . Because the gradient is negative,
F represents a positive flux of electrons moving from left t o right (increasing x)
at Xy
n

_ 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 )

For a negative gradient, indicating a decrease of with x, the current density is


negative. This is to be expected since a motion of electrons in the positive direction
results in a current in the negative direction. If holes are injected into an region,
E q . (4.27) applies as well and is multiplied by the charge +q to obtain an expression
for the hole current density. T h e direction of hole motion is in t h e direction of posi
tive current flow causing J to be positive since dp/dx is negative.
In general, the current densities due to the diffusion of electrons and holes are
given by
(4.29)

W h e n J and / are in a m p s / c m , q is in coulombs, the diffusion constants are


in c m s and the derivatives represent t h e gradients in one dimension. Diffusion of
electrons and holes and the directions of the diffusion currents are shown in Figs.
4.7(c) and 4.7(d).
The diffusion constant for either holes or electrons is related to the mobility
by Einstein's relation in E q . (4.30) (see A p p e n d i x D ) . A t a fixed t e m p e r a t u r e , D is
directly proportional to . This is to be expected since b o t h mobility and diffusion
constants are factors in the expressions for drift and diffusion currents respectively.
Therefore,
n

_ 1

D - (f),

(4-30)

Section 4.9

Carrier Currents

103

A t 300K. kT/q = 0.0259 V and for in c m / V - s e c , D is in c m / s . Thus, D has


the same d e p e n d e n c e that has on doping density. However, because of kT/q, D is
proportional to , where lattice scattering is dominant, and to T~ when impu
rity scattering is the major scattering p h e n o m e n o n .
Application of the expression for the diffusion current is shown in Example
4.4.
0 5

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

p = 10 cm~ and = 1 X 10 cm"


0

Using the density of impurity atoms, we determine the mobility of electrons in the region
from Eq. (4.15) to be:

= 1331 cmV(V-s)

The electron diffusion current density is given by:


H

"
"dx
The diffusion constant is obtained from Einstein's relationship at = 300K,
D

"

~q "

The slope of the electron density


v

259

is -

dx

'

1 3 3 1

3 4

'

4 8

m2//S

'
14

10 x 10~

] [ = - 10 cm"
/

s o t h a t 7 = 1.6 X 10" x 34.48 ( - 1 0 ) = ~0.55mA/cnr


The motion of electrons is in the positive x-direction, causing a current in the negative
x-direction.
19

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

Carrier Processes: Drift, Diffusion and Generation Recombination

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 steady-state, 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 R-G
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 electron-hole 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

Recombination and Generation

105

Of course, the tendency for systems u n d e r nonequilibrium conditions is to r e t u r n to


equilibrium. For example, w h e n excess minority carriers are injected into one end of
a semiconductor bar, electron and hole densities will tend to obtain equilibrium val
ues by causing t h e r a t e of recombination to exceed the rate of generation. Thus, the
rates of generation and recombination are not balanced w h e n the semiconductor is
p e r t u r b e d , causing t h e concentration of one type of carrier to exceed its equilibrium
value. T h e forces of n a t u r e tend to restore conditions back to where they stood prior
to the p e r t u r b a t i o n .
A n N-type semiconductor has equilibrium majority and minority carrier den
sities of n and p respectively. If this semiconductor is exposed to a source of energy
that increases the density of the minority carriers, then the process of generationrecombination acts to reduce the concentration of these carriers by causing the
recombination rate to exceed t h e generation rate. If, on the other hand, due to
extraction of carriers, the carrier density were to decrease, then the generation rate
will exceed the r a t e of recombination until equilibrium is established.
Q

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 electron-hole 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 generation-recombination processes are
very unlikely in silicon and germanium, as discussed in Section 2.4, because of the
shape of t h e E-k diagram of Fig. 2.12. Direct recombination generation is the com
m o n process in III-V 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 recombination-generation are shown in Fig. 4.8a.
Direct generation of electron-hole 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 electron-hole 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 recombination-generation ( 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

Carrier Processes: Drift, Diffusion and Generation Recombination

generation

recombination

generation

(a) direct
Figure 4.8

recombination
(b) indirect

Direct and indirect processes of R-G.

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

Recombination and Generation

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 R-G center is an acceptor type.

T h e localized states are also k n o w n as recombination-generation centers or


traps.
The obvious question becomes: W h e r e d o these centers or traps c o m e from?
These traps occur because of lattice imperfections
(such as a missing a t o m in
the covalent b a n d structure), which result from dislocations or defects in t h e lattice
or result from the presence of impurities, which may b e deliberately introduced.
The lattice imperfections normally result from the p r e p a r a t i o n of the semicon
ductor sample and hence from the fabrication process. Such defects may b e greatly
r e d u c e d by very high-quality growth or by the production of very p u r e silicon.
The impurities that cause traps may b e inherent to the solid but are commonly
introduced so as to control t h e lifetimes of minority carriers. For t h e impurities in
silicon to introduce efficient recombination-generation centers, the energy levels of
these impurities must lie in the immediate vicinity of the middle of the forbidden
band, in contrast to C o l u m n III (boron) and C o l u m n V (phosphorus) impurities,
where energy levels are close to t h e valence and conduction b a n d s respectively.
Gold and iron in silicon and copper in germanium introduce such recombina
tion centers, some above (acceptors) and s o m e below (donors) the middle of the
forbidden band, but in b o t h cases, n o m o r e t h a n O.leV from the middle of t h e band.

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 /V-type 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 low-level injection. This condition of
low-level 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

Carrier Processes: Drift, Diffusion and Generation Recombination

minority carriers are injected. Because t h e change in n , An, is so small c o m p a r e d to


n , we assume for all practical purposes that n is unchanged.
To illustrate t h e concept of low-level injection, assume n = 1 0 c m ~ in sili
con, so that p = n /n
= 10 cm~ . A n increase in p , Ap , of 1 0 c m ~ will cause an
increase in n , An , of 1 0 c m ~ , which is negligible c o m p a r e d to n .
G e n e r a t i o n - r e c o m b i n a t i o n is a continuous process, but since t h e hole density
has b e e n increased to a value greater than its thermal equilibrium value, t h e r a t e of
recombination m u s t exceed t h e generation rate. We are assuming a t r a p density N ,
located at a t r a p energy level E n e a r t h e middle of t h e forbidden band. Since the
material is N-type, t h e Fermi level E is above t h e middle of t h e energy gap and E
is greater t h a n E Because E < E , t h e Fermi-Dirac relationship predicts that the
E level is practically full of electrons. The recombination process consists of
the c a p t u r e of holes from t h e valence band. O n c e the holes are captured, some of
the m u c h m o r e n u m e r o u s electrons are also captured.
Q

15

10

10

A n a l y t i c a l Relations
Expressions are n e e d e d for the time-variation 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

n,p = electron, hole density, c m


n ,p = t h e r m a l equilibrium electron, hole density cm~
0

An = n = deviation of electron density from t h e t h e r m a l equilibrium


value
Q

deviation of hole density from the t h e r m a l equilibrium value.

An and Ap can b e b o t h positive or negative. A positive deviation implies


an excess of carriers while a negative deviation corresponds to a deficit of car
riers.
N

= n u m b e r of recombination-generation centers (traps) per cm .

W h e n the source of excitation that caused the p e r t u r b a t i o n is removed, it will


take some time for t h e density of holes to r e t u r n to its equilibrium value via the
indirect recombination-generation process. The question becomes: O n what factors
does t h e time r a t e of change of hole density d e p e n d ?
First, to discard an excess hole, that hole must move from its location in the
valence b a n d to E , where the recombination traps are. These traps, of density N ,
are filled with electrons because E > E Therefore, the r a t e of decay of hole den
sity, dp/dt, is proportional to N since t h e larger the density of traps, t h e faster the
recombination will occur. F u r t h e r m o r e , t h e n u m b e r of holes scrambling to rejoin
electrons increases with t h e n u m b e r of excess holes since t h e greater t h e n u m b e r of
holes available for recombination, t h e m o r e of t h e m will recombine per unit time
and t h e faster the recombination rate. H e n c e , the rate of hole density decay is p r o
portional to Ap. B o t h factors can b e included in an equation that states
T

Section 4.10

Recombination and Generation

dp/dt = -K^Ap

109
(4.32)

where is a constant of proportionality (positive) and t h e negative sign indicates a


decrease in t h e hole density because of recombination. Similarly, t h e analogous
expression for minority carrier electrons in a p e r t u r b e d semiconductor is given by

dn/dt = -KJST^n
1

The units of K^Nj. and K N


constants as
2

(4.33)

are ( s e c ) " so that o n e can define t h e two time

= 1/

(a)

= 1/K N
2

(b)

(4.34)

By using the relations of E q . (4.34) in E q . (4.33), we write


dp/dt = -/

(a)

dn/dt = ~/

(b)

(4.35)

It is to b e emphasized h e r e that t h e relations in Eqs. (4.35) are for indirect


t h e r m a l generation-recombination process and are:

for minority carriers

for low-level injection.

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 P-type 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 high-frequency
response of devices.

110

Chapter 4

Carrier Processes: Drift, Diffusion and Generation Recombination

REVIEW QUESTIONS
Q4-7

Explain, in your own words, what is meant by diffusion in general and, as relating to
semiconductors, in particular.
Q4-8 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?
Q4-9 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.
Q4-10 What is the rate of recombination and what does it depend on?
Q4-11 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 cross-sectional
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

E4-5 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
E4-6 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

The Continuity Equation

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 N-type 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
low-level 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

Because of recombination within the slice, the n u m b e r of holes entering the


slice at is larger than the n u m b e r of holes leaving at + Ax as shown in Fig. 4.11.
Since the n u m b e r of holes within the slice (pAAx) is greater than the equilib
rium value (PQAAX),
t h e n u m b e r recombining, per second, within the slice is
/
and t h e n u m b e r generated is ^/^
Thus, the rate of change of the
n u m b e r of holes due to recombination and generation is

-(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

A = area of section normal


to the plane of the paper

J (x)
P

] (x + Ax)
P

+ Ax
Figure 4.10
equation.

Section of N-type semiconductor for the formulation of the continuity

Chapter 4

Carrier Processes: Drift, Diffusion and Generation Recombination

light pulses
C
metallic contact
D

/ metallic contact

- L-

-jf

(a)

hole density

(b)
Figure 4.11 (a) Haynes-Shockley experimental setup; (b) the effects of drift,
diffusion, and recombination on the minority carrier hole pulse.

Because of the assumed one-dimensional n a t u r e of the flow of minority carri


ers, holes enter and leave t h e slice only across t h e faces and ( + ). The total
n u m b e r of minority carriers within the slice is pAAx. Thus, we can write
-(pAAx)
dt

= = ^ ^ A A x

[ J p i x )

- "

{X

In effect, the statement of t h e equation is: R a t e of change of holes in =


rate of (generation-recombination) + the r a t e at which holes enter t h e slice less
the rate at which they leave. The partial derivative has b e e n used to indicate a varia
tion of with time and distance. By dividing t h e previous equation by AAx, we
obtain the continuity equation in three dimensions for holes in N-type semicon
ductors

^
= - ^ - qW
dt

(4.36)

The analogous expression for electrons in P-type semiconductor is given as


dn

dt

div/

(4.37)

In the chapter on diodes, we will use t h e continuity equation, after replacing


the current density by its diffusion c o m p o n e n t equivalence from E q . (4.29), to

Section 4.13

Haynes-Shockley Experiment

113

d e t e r m i n e the steady-state distributions of minority carriers in the regions subject


to the relevant b o u n d a r y conditions. We will assume that the drift currents of minor
ity carriers are negligibly small c o m p a r e d to the diffusion currents, and determine
expressions for t h e currents as a function of distance. The applied voltage will
a p p e a r in the b o u n d a r y conditions.
T h e results referred to above include three properties of the minority carriers.
They are: mobility, diffusion constant, and lifetime. In the following section, we
briefly describe an experimental p r o c e d u r e that was first used to calculate these
properties. O u r interest in this experiment is m o r e in the illustrations it provides
of t h e carrier processes rather than in t h e detailed determination of the three p r o p
erties.

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 current-voltage 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 electron-hole pairs generated at C, because of low-injection, 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 X-time scale of
t h e scope is synchronized with t h e timing of the light pulse generator. In brief,
pulses of holes are light-generated 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

Carrier Processes: Drift, Diffusion and Generation Recombination

b e c o m e s wider because of diffusion of holes b o t h to the right and to t h e left of the


center of t h e pulse. The area of t h e pulse is r e d u c e d because of
recombination.
The electric field intensity in the bar is d e t e r m i n e d from the voltage source
and t h e distance d. The time it takes t h e pulse to travel from C to D is m e a s u r e d on
the scope so that t h e average drift velocity is calculated and hence t h e mobility is
determined. O t h e r results of m e a s u r e m e n t s enabled H a y n e s and Shockley to calcu
late t h e diffusion constant and the lifetime.
A l t h o u g h nowadays t h e r e are m o r e precise m e t h o d s of measuring minority
carrier properties, the experiment explicitly illustrates the effects of the three car
rier processes that minority carriers are subjected to.

PROBLEMS
5

4.1 The intrinsic resistivity of a semiconductor at 300K is 3 10 ohm-cm. Given


= 1700 and = 350, cm /V-s, determine the intrinsic carrier density.
4.2 An intrinsic semiconductor has resistivity of 2 X 10 ohm-cm at = 300K and has
= 4000 and = 1000cm /V-s. Calculate the resistivity for
a)
An acceptor doping of 10 cm~ .
b)
A donor doping of 10 cm~ .
Assume that the mobilities are constant with doping.
4.3 A semiconductor (TV-type) bar is injected at one end by minority carrier holes and an
electric field of lOOV/cm is uniformly applied along the length of the bar that moves
the holes a distance of 2cm in . Determine
a)
The drift velocity of holes.
b)
The diffusion constant of holes at = 300K.
4.4 An N-type semiconductor bar is 2cm long, has a cross-sectional area of 0.1cm , an
electron density of 5 X 10 cm~ and a resistivity of 10 ohm-cm. A 10V battery is con
nected across the ends of the bar. Determine
a)
The time it takes an electron to travel the length of the bar.
2

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 /V-s. 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 ohm-cm. Donor and acceptor
atoms are added with densities of 10 cm~ and 5 X 1 0 c n r respectively. Given =
1600 and = 600cm /V-s, 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 /V-s, 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 /V-s.
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 /V-s, 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 /V-s, 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 P-silicon, 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 N-type Ge bar has a resistivity of 5 ohm-cm. 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 N-semiconductor having = 3900 cm /V-s
b)
The velocity of an electron at a distance of lcm if the space between the con
tacts is a vacuum.
A bar of N-type 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

Carrier Processes: Drift, Diffusion and Generation Recombination


b)

The total drift current density.


4

4.20 A hole current of 10" A/cm is injected into the side (x = 0) of a long N-silicon 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 steady-state excess hole density at = 0.
b)
Repeat part (a) at = .
Given = 400cm /V-s, = 1600cm /V-s, 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 electron-hole pairs at = .
b)
The rate of recombination of electron-hole 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 non-equilibrium 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 current-voltage 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 field-effect transistors.
a

117

118

Chapter 5

The PN Junction Diode

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) Current-voltage characteristic of diode; (b) Current-voltage
characteristic in the positive I and V regions; (c) Circuit from which (b) is obtained
showing diode symbol.
5.1

SPACE-CHARGE 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

(a) Two-dimensional diode section and symbol, (b) Typical construction.

Section 5.1

Space-Charge Region

119

place within an extremely small distance. T h e r e is thus a sudden change in doping in


going from to N.
In general, diodes are fabricated with one region m o r e highly d o p e d than the
other, the m o r e highly d o p e d region identified by a superscript plus, such as P N
and N P . T h e majority carrier density in the m o r e highly d o p e d region may b e about
three orders of magnitude greater t h a n t h e majority carrier density of the other
region.
In our illustration, we will use a P N diode, where in Fig. 5.3(a) we have two
sketches, one to illustrate the hole densities in b o t h regions and one to illustrate the
electron densities in the two regions. The equilibrium hole density in the P-region is
p and in the region is p , whereas n and n represent the equilibrium electron
densities in the and regions respectively.
Because of t h e large gradients of charges (electrons and holes) that exist near
the surface of contact, holes diffuse from to and electrons diffuse from to P.
The diffusions are shown in Fig. 5.3(b) for each type of carrier.
By assuming that N and N , in t h e and regions respectively, are each
m u c h greater t h a n t h e intrinsic carrier density, then p ~ N and n ~ N .
For every electron that leaves the region, because of diffusion across the
surface, an u n c o m p e n s a t e d positively ionized d o n o r is left behind, and for every
hole that leaves the region, an u n c o m p e n s a t e d negatively ionized acceptor is cre
ated. These carriers leave from points n e a r the surface of contact. A s a result, t h e
and regions are separated by what is k n o w n as a depletion region, a region that is
depleted of holes and electrons but contains positively ionized d o n o r atoms on one
side and negatively ionized acceptors atoms on t h e other side. Because this region
consists mainly of ionized charged impurities, it is also k n o w n as the space-charge
region. The P N junction diode in equilibrium consists of charge-neutral and
regions separated by a space-charge region. Because of the high electrostatic field
p r o d u c e d in the depletion region, the concentrations of mobile carriers, holes and
+

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

The PN Junction Diode

qN

p0

-qN -

Figure 5.4 Space-charge region,


direction of electric field, and built-in
voltage at thermal equilibrium.

metallurgical
junction

hi

electrons, are very small in comparison to t h e impurity concentrations over most of


the space-charge region. Conditions after contact are shown in Fig. 5.4.
We n o t e in Fig. 5.4 that the space-charge region has b e e n assumed to contain
ionized impurity atoms only and we also assume that outside of t h e depletion region
the materials are neutral. The side of the depletion region is positively charged
and t h e side is negatively charged, and since the electric field intensity is defined
to b e the force o n a unit positive charge, it is shown in Fig. 5.4 to b e directed from
right to left.
We observe in the figure that the N-side of t h e depletion layer extends from
= 0 to = x , whereas t h e side extends from = 0 to = C
J . T h e line sepa
rating the two regions is at = 0, which is t h e metallurgical junction of t h e and
regions. The subscripts 0 refer to t h e r m a l equilibrium.
In a later section of this chapter, analytical relations will b e developed and
expressions will b e derived for V and for the a m o u n t of bending of the bands.
nQ

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 built-in voltage, V , and shown in Fig. 5.4.
The built-in 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

Space-Charge Region

121

distance

Bending of energy bands.

Figure 5.5

Since in Fig. 5.4, t h e voltage at x is higher t h a n the voltage at - i . by V , an


electron at > x has less potential energy t h a n an electron at < _ . We n o t e in
Fig. 5.5 that the respective energy levels in the neutral region are lower than those
in the region by qV . H e n c e , the energy b a n d s in t h e region are said to b e n d
down relative to those in t h e region.
We recall that electrons reside in the conduction b a n d and holes reside in the
valence band. F u r t h e r m o r e , electron energy is m e a s u r e d upwards, whereas hole
energy is m e a s u r e d downwards.
The presence of charges on either side of the junction create an electric field
directed from to P. The positive and negative charges (impurities ions), shown in
Fig. 5.4, are generated because the electrons and holes that are associated with t h e m
left by diffusing to the other side. We label the built-in voltage as the contact potential
and also as the barrier voltage.
n0

n0

bi

bi

Drift a n d Diffusion Currents


Holes from to and electrons from to continue to diffuse because of t h e gra
dients of carrier densities resulting in a total diffusion current from to N.
Simultaneously, t h e electric field at t h e junction forces the holes to drift from to
and t h e electrons to drift from to resulting in a total drift current from to P.
We have two currents, drift and diffusion, for both electrons and holes in opposite
directions.
The total current, t h e sum of the two drift and two diffusion currents, must be
zero. A non-zero current causes energy dissipation, which is not possible since t h e r e
is n o source in the diode to provide this energy.
The student may w o n d e r if the total electron current and the total hole cur
rent are each zero. The current of each carrier has to b e zero because of the follow
ing reasoning: Since the diode is at thermal equilibrium, the product np is equal
to nj. If we assume that t h e electron current is not zero, then t h e hole current is
equal in m a g n i t u d e and opposite in direction to the electron current. T h e r e is thus a

122

Chapter 5

The PN Junction Diode

net transfer of electrons and holes in t h e same direction causing an increase in


t h e densities of b o t h in o n e of the regions and violating the condition of thermal
equilibrium.

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 built-in 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)

where is the potential energy of the electron. A n electron in the conduction b a n d


has potential energy of E , the b o t t o m of the band. The gradient of could b e
replaced by the gradient E . Since the gradients of E , E and E are the same, we
choose to use E. in E q . (5.1) for reasons that will b e c o m e a p p a r e n t shortly. Using E
E q . (5.1) in o n e dimension b e c o m e s
c

q dx

(5.2)

The electrostatic potential, , is defined in one dimension by

dx

(5.3)

U p o n comparing Eqs. (5.2) and (5.3), we write


=

(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

Analytical Relations at Equilibrium

123

d*<j>
dx

~P/

5 5

()

where is t h e volume charge density, ( = ) is the permittivity, e is the relative


dielectric constant, and is the permittivity of free space.
By using Eqs. (5.3) and (5.4) in E q . (5.5) we have

(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)

By using the expressions in Eqs. (5.9) in E q . (5.8), replacing D by


by its equivalence from E q . (5.2), and setting J = 0, we have

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

Built-in V o l t a g e in Terms of Fermi P o t e n t i a l


In our earlier discussion, we concluded that in a diode in equilibrium, an electric
field and a potential are established across t h e junction, and that t h e energy bands
are b e n t in b o t h t h e and regions. We also just established that, whereas E , E
c

124

Chapter 5

The PN Junction Diode

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 space-charge region, we define
Fermi potentials in each of the charge neutral and regions, cb and ^ , as
v

jp

EF Ejp

(PFp ~

EF

<PFn

--"

P-J-lj

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 built-in 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

Built-in V o l t a g e in Terms of D o p i n g Densities


Since the electron and hole currents across the junction are each zero, we use Eq.
(4.31(b)) for the hole current in one dimension as
(5.12)

Section 5.2

Analytical Relations at Equilibrium

125

By using the Einstein relation t o replace D , we have


(kT

\dp

or
%

*1<%

(5 13)

By integrating across the depletion layer from x ,


, t o x , w h e r e the hole density is p , we have
p0

n0

where t h e hole density is

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 right-hand side of E q . (5.14) is the negative of the built-in
voltage, V , across the depletion layer so that after integrating we have
u

b i

= in -^
Q

(5.15)

where p is the equilibrium hole density in t h e region and p is t h e equilibrium


hole density in the region. Since N in and N in are each m u c h greater than
n2 / ,
n a n d p s N and n = N , so that p == nf/N
we have
Qp

0n

Qn

0n

5 16

( )

Pap-NAtnAPan-JfUsing t h e relations of Eqs. (5.16) in E q . (5.15), we have

(5.17)

. ' ^

The built-in voltage V , also k n o w n as t h e barrier voltage, d e p e n d s on the


doping of the a n d regions a n d on t e m p e r a t u r e and b a n d gap, through nf. T h e
barrier voltage for an abrupt silicon junction, which has energy b a n d gap of 1.2eV,
varies b e t w e e n 0.7 and 1.0 volt while that for an abrupt gallium arsenide junction,
whose b a n d gap energy is 1.4eV, varies b e t w e e n 1.0 and 1.4 volts.
The a p p e a r a n c e of a voltage w h e n two materials are in contact raises the pos
sibility of associating this voltage with voltage sources. This contact voltage is n o t a
source of voltage, hence it is n o t a source of energy. A n y a t t e m p t to m e a s u r e this
voltage by connecting a voltmeter across the external and contacts will provide
an indication of zero volts. So what h a p p e n e d t o the barrier voltage V ? T h e metal
lic contacts b e t w e e n the voltmeter leads and the and regions provide a contact
potential equal to and opposite to t h e diode barrier voltage. F u r t h e r m o r e , if the
diode voltage w e r e a source, connecting a resistance across the diode terminals
would cause a current a n d energy will b e dissipated. This is absurd since t h e r e is n o
source of energy in the circuit.

bi

bi

126

Chapter 5

The PN Junction Diode

Electric Field a n d Potential in t h e S p a c e C h a r g e R e g i o n


In this section, we will derive expressions for t h e electric field in the space charge
region, also labeled the depletion region, and for the potentials of t h e region. We
start with Poisson's equation as defined by E q . (5.7), written in one dimension for
the general case

f-J-fo+

(*>

We identify t h e limits of t h e depletion region as extending from = ~x


to
= x , as shown in Fig. 5.7. F r o m = 0 to = - , the portion of the depletion
region extracted from t h e region is included, and from = 0 to = x t h e bal
ance of t h e depletion region is in the region.
To obtain expressions in closed form, some simplifying assumptions, classified
u n d e r the depletion approximation,
are m a d e .
The assumptions of the depletion approximation are:
1. In the space charge region from = -x
to = x , only ionized impurities
exist. The acceptor density covers t h e distance = ~x to = 0, and from =
0 to = x only d o n o r impurities are available. Throughout the depletion
region, t h e densities of electrons and holes are m u c h smaller t h a n N and N
and are neglected.
2. Outside the space charge region, two neutral regions, one on each side of the
depletion region, exist. They include electrons and holes. The sum of all the
charges is zero so that these regions are labeled as neutral and t h e electric field
is zero.
p0

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 space-charge 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

= 0, since we are at t h e b o u n d a r y of the neutral region. So,


C

= - ^ x

and the expression for the electric field b e c o m e s

Section 5.2

Analytical Relations at Equilibrium

%(x) =&&(
For -x

pQ

=s

for 0

127

(5.20)

nQ

0, w h e r e = qN , we derive an analogous expression as


A

%{x) = - ^ - f (x + Xpo) for -

=s = 0

(5.21)

A t = 0, the electric field must be continuous since there is no layer of charge at


that point. We obtain the expression for the largest value of %, ^ , from Eqs. (5.20)
and (5.21). Since both of the two expressions have a m a x i m u m at = 0 we have
m a x

= - *

= -,
'

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

The PN Junction Diode

%{x) = % ( \

+ -f]

for - x ^ ^ O

We d e t e r m i n e an expression for t h e potential


tion layer by starting with E q . (5.3) in integral form.
()

(5.23)

across t h e side of the deple

= J %{x)dx for -x

(b)

= =s 0

pQ

By using E q . (5.23b), integrating and arbitrarily assuming the zero of potential


to b e at = ~x in order to determine the constant of integration, we have
p0

()

= -% [x

max

+ ^ )

for -x

^O^x

p0

(5.24a)

Similarly, is d e t e r m i n e d by using E q . (5.23(a))

After integrating, and since t h e potential is continuous at = 0, we set =


at = 0 so that

() = -% (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 space-charge region is given by set
ting = x in the expression for so that

p0

nQ

bl

= ~^(x

+ xj

w h e r e V is the total voltage from to and W is the total width of t h e depletion


layer. Since ' 8
is a negative quantity, V is a positive quantity. The distribution of
total potential energy is d e t e r m i n e d from g and q$ .
bi

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 built-in 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)

(5-26)

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

Analytical Relations at Equilibrium

/
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)

We will use an example to illustrate t h e application of the expressions and we


will calculate typical orders of magnitudes for the various quantities.

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 built-in 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

from Eq. (5.27) and then x


}

- (
{

1 1

8 5 4

= 0.715V

1 Q 2 0

from Eq. (5.19)

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

The PN Junction Diode


15

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 reverse-biased 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 space-charge 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

Conditions in the Diode with Voltage Applied

131

and the space-charge 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 built-in voltage, V , when V is connected,
to m a k e positive with respect to and V aids the built-in 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 built-in 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)

It is i m p o r t a n t to remind the reader that the derivations in this chapter are


based on the assumption that t h e and regions form a step junction, or an abrupt
junction, at the metallurgical boundary.

REVIEW QUESTIONS
Q5-1 For a PN junction diode in thermal equilibrium, explain why each of the electron and
hole currents are zero.
Q5-2 Briefly explain why the section separating the and regions is labeled (a) a space
charge region (b) a depletion region.
Q5-3 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?
Q5-4 Since a voltage appears across the depletion region, which region has the higher volt
age, the or the region?
Q5-5 The depletion region extends deeper into the lightly doped (N) region of the PN diode.
Explain why.

132

Chapter 5

The PN Junction Diode


Equilibrium

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

Conditions in the Diode with Voltage Applied

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

The PN Junction Diode

Q5-6 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

When a semiconductor and an semiconductor are placed in intimate contact,


holes diffuse (rush) from to and electrons rush from to P.
As a result of the transfer of carriers, positively charged donor ions are left behind on
the side of the junctions and negatively charged acceptor ions accumulate on the
side. The whole region that consists of ions is known as the space charge region.
The existence of positive charges on one side and negative charges on the other side
causes an electric field directed from to P.
The electric field acts to force electrons to drift from to and holes from to so
that a balance is established at which each of the hole and electron currents is zero.
The existence of an electric field between x and x requires a voltage difference
between the and regions. This voltage is known as the built-in voltage.
Because a voltage difference is formed between the and regions, an energy differ
ence exists between them. Thus, an energy hill or barrier is established across the
depletion region. Consequently, the energy bands are bent and the Fermi levels of the
two regions are aligned.
p0

nQ

EXERCISES
ES-1

16

E5-2

15

A P N silicon junction diode is doped with N = 10 cm~ and N = 10 cm~ .


a)
Determine the location of the Fermi levels, with respect to the bottom of the
conduction band, in each of and N.
b)
Show that the separation between the Fermi levels is equal to the built-in
voltage calculated from the doping.
Ans: a) f o r N , E - E = 0.21eV
for P, E - E = 0.86eV
For the diode of the above exercise, determine:
a)
b)
c)

The width of the portion of the depletion region in N.


The total width of the depletion layer.
The electric charge in each of the two segments of the depletion region given
the area is 0.2 X 0.3cm.
Ans: a) 0.088
b) 0.968

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 current-voltage 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

1. Electrons in the conduction b a n d have potential energy E and kinetic energy


(E E ), where is the energy level located above E of the electron. Energy
of electrons is m e a s u r e d upwards in the energy diagram.
c

2. Holes in t h e valence b a n d have potential energy E and kinetic energy


(E - E), where is the energy level of the hole. Energy of the hole is mea
sured d o w n w a r d s in the energy diagram.
y

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

Illustration of carrier motion at equilibrium.

d r i f t

distance

136

Chapter 5

The PN Junction Diode

Figure 5.12 Energy band diagram with forward bias. Note the negative slopes of
the energy levels in the space-charge 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 pyramid-like 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 Fermi-Dirac 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 N-region 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

The PN Junction Diode

n u m b e r of electrons crossing from to than those crossing from to and, simi


larly, a larger n u m b e r of holes crossing from to than from to N, all c o m p a r e d
to equilibrium.
Thus, the current is limited mainly by the availability of thermally generated
minority carriers and is, in fact, i n d e p e n d e n t of t h e applied bias for an applied
reverse bias of a few tenths of a volt or larger.
A s s u m p t i o n s f o r Ideal Diode E q u a t i o n
The expression we will derive for the current-voltage characteristic of the diode is
labeled ideal because of the following assumptions that are m a d e :
1. T h e space-charge region boundaries represent a step-junction from the bulk
and regions.
2. N o carriers exist in the space-charge region, they just traverse it. This region
consists of ionized impurities only.
3. In the bulk of t h e diode, outside the depletion layer, the semiconductor is neu
tral. This is valid due to the limited conductivity that causes the ohmic voltage
drop to b e negligibly small.
4. O p e r a t i o n is at a t e m p e r a t u r e such that all impurity atoms are ionized.
5. Perfect ohmic contacts are m a d e to the ends of the and regions. A perfect
ohmic contact is one that is assumed to have zero resistance and allows easy
current flow in b o t h directions so that the voltage across the contact is zero. A s
we shall see later, t h e minority carrier density at the ohmic contact is the value
at equilibrium.
6. W h e n the diode is forward-biased, low
electrons from and holes from are
density of t h e carriers at the b o u n d a r y
minority carriers, is m u c h smaller than
carriers in that region.

injection is assumed. This m e a n s that as


injected into the opposite region, the
of the new region, where they b e c o m e
the equilibrium density of the majority

7. N o recombination or generation takes place in t h e depletion region so that b o t h


electron and hole current are constant across this region.
We will derive t h e diode equation by determining expressions for t h e minority
carrier currents in each of the neutral regions. The sum of the minority carrier cur
rents at the junction constitutes the total current.
The reason for using the minority carriers in each region is because of assump
tion t h r e e (the electric field, and hence the voltage across the neutral regions, is
zero). This is not quite true, as we will determine. T h e r e is an electric field that
causes, at low-injection levels, negligible minority carrier drift currents but large
majority carrier drift currents.
Hence, the minority carrier currents in t h e neutral region result from diffusion
only. The p r o c e d u r e we will follow consists of determining expressions for the distri
bution of minority carriers, subject to the b o u n d a r y conditions, from the continuity
equations. Expressions for t h e minority carrier currents will b e obtained as a func
tion of distance. The two minority carrier currents at t h e edges of the depletion layer
will b e added to obtain the total current.

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

where is t h e hole density at x,p


is t h e equilibrium hole density in N, and is t h e
lifetime of holes in the region.
We let p' represent the excess hole density, (p p ), and E q . (5.31) becomes
0n

Qn

SO'
-2

dx

p'
=

Dr
p

5 32

( )

>

In o r d e r to m a k e the expressions less cumbersome, we will label t h e point x


as our new point of reference, x' = 0, so that x' = x , where x and x are the
limits of t h e depletion layer with bias applied in contrast to x and ~x , which are
for t h e r m a l equilibrium. The general solution to E q . (5.32) becomes
n

nQ

p' = B e x p ( ^ ) + B
1

p0

exp^-J

(5.33)

where B and B are constants of integration and L , k n o w n as t h e diffusion length


for holes in the region, is given by
}

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 I-1

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

*See subsection titled "Boundary Condition at Junction."

140

Chapter 5

The PN Junction Diode


(2W,

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

Replacing ' by ( p ) and x' by (x x ), the excess hole density in t h e


region for each of the two cases becomes*
Qn

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

- >;

By differentiating Eqs.(5.37) and replacing t h e derivatives in Eq.(5.39), we


obtain
qDpPOn

qVa
kT

exp

exp

forW

(a)
(5.40)

p\

'

qVa
-1
exp
kT

forVy

(b)

A t x' = 0 or = x , t h e edge of the transition layer, Eq.(5.40(a)), becomes


n

JJx

qTJ p
p

0n

exp

*Details shown in the last section of this chapter.

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

The PN Junction Diode


qD n
" "
n

J (-x )
n

0p

exp

-1

kT

for W

(5.42)

where D , n , and L refer to electrons in the region, and W is t h e length of the


neutral region. Since we have assumed that t h e minority carrier currents are
continuous across the transition layer, then t h e total current density across t h e junc
tion, which is also the total current density t h r o u g h o u t t h e diode, b e c o m e s
n

Qp

[J (x) +
J(-* )l
T h e diode current b e c o m e s
P

I =

I = A

qDpPon

A[J (x )+J (-x )]


p

qD nop

JlVa
expkT

for W

and W

(5.43)
n

For a reverse-biased diode, with V negative, t h e expression for the current is


/ = I , w h e r e I is the reverse saturation current given by
a

for

W.

and

L
(5.44)

W h e n W L , the hole current, given by Eq.(5.40(b)), is constant through


out the region. If it is also assumed that the width of t h e region is m u c h smaller
than L , then the diode current expression given by Eq.(5.43) appliesexcept that
L and L are replaced by the widths, W and W , of t h e and regions respec
tively.
For a P N diode, N N so that n p , the second t e r m within t h e first
set of brackets in E q . (5.43) and E q . (5.44), b e c o m e s negligibly small c o m p a r e d to
the first term.
Sketches of the distributions of current components are shown in Fig. 5.16. We
have assumed that t h e r e is n o recombination in the depletion layer, h e n c e the
minority carrier current densities at = for / and at = for / have the
n

Qp

0n

same values as those at = x and = x respectively. In the region where


the current is carried by majority carriers, we obtain t h e majority current density as
the difference b e t w e e n the total current density and the minority carrier current
density. T h e total current density t h r o u g h o u t the diode, / , is d e t e r m i n e d by t h e sum
mation of J and / in the depletion region.
In t h e following example, we illustrate t h e application of Eq.(5.43).
p

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

Distributions of current densities for forward bias.

Lifetime of holes in region =


Lifetime of electrons in region = 50|JLS
D of electrons in region = 34 cm /s.
D of holes in region = 13 cm /s.
For V = 0.66, volt calculate: J (-x ),J (x )
2

and the total diode current. Operation is at 300K.

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

Using Eq.(5.34), we calculate


L = (D T )

1/2

(10 X 10~ X 13)

1/2

= 11.4 X 10~ cm

Similarly, L = ( 0 ) = 4.1 X 10~ cm.


By using Eqs.(5.41) and (5.42), we calculate J and J , keeping in mind that x is at x' = 0 and
-x is at A" = 0.
n

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

The PN Junction Diode

The total diode current becomes


1

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 forward-biased 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 metal-semiconductor 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
metal-N 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

question is: W h a t is t h e mechanism of conduction by which these currents occur?


We refer to the earlier section on the ideal diode equation, in which we
assumed that the drift current of minority carriers in the neutral regions was negligi
ble because of the very low electric field intensity in those regions w h e n low injec
tion is assumed. In spite of t h e very small magnitude of t h e electric field intensity, it
does cause large drift currents, J in and / in P, of t h e majority carriers. The
majority carrier densities are m u c h larger than t h e minority carrier densities so that
the p r o d u c t of t h e majority carrier density and t h e electric field intensity results in
large drift currents.
For a reverse bias, the carriers that cross the depletion layer, and as was shown
in Fig. 5.13, are electrons from to and holes from to P. This is in contrast to the
forward-biased diode in which electrons diffuse from to and holes from to N.
n

Section 5.4
electrons drawn by V from contact

electrons brought from

electrons move from to


contact to replenish those drawn by V

>
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

electrons drawn from contact


to recombine and drift

o>

Currents in Diode

-*
depletion
region

electrons drift in
<

N-region

metallic contact

metallic contact

electrons moved by V

|.

electrons moved by V

electron
hole
( Q t ) electron-hole
recombination
Figure 5.17

Current loop in forward-biased P N junction diode.

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 electron-hole 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 electron-hole 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 semiconductor-contact to enter the external
circuit and m o v e towards the metal-P semiconductor contact.

146

Chapter 5

The PN Junction Diode

S a t u r a t i o n Current
The current-voltage 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 current-voltage
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/)

an increase of the intrinsic carrier densities.

a decrease of t h e b a n d gaps in accordance with Table 2.2.

Ge

GaAs

Si

10 -

6
"3

2 -

0.2

0.4

0.6

V (V)
a

0.8

1.0

Figure 5.18 Current-voltage


characteristics of silicon, germanium, and
gallium arsenide diodes.

Section 5.4

Currents in Diode

147

a decrease of the mobilities.

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 built-in voltage (barrier voltage), V , given by Eq.(5.15)
relates the equilibrium hole densities on b o t h sides of the space-charge 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 space-charge region and the values of the carrier densi
ties are those at the edges of the region.

REVIEW QUESTIONS
Q5-7
Q5-8
Q5-9
Q5-10
Q5-11
Q5-12
Q5-13

Explain the movement of the energy bands when a diode is forward-biased.


Some holes diffuse throughout the region. Explain what happens to them when
they arrive at the metal contact of that region.
What is meant by low-level injection?
The diode equation (5.43) is sometimes labeled the diffusion equation. Explain.
Define the term "diffusion length" of a minority carrier.
If the diffusion length of holes in N, L , is much greater than W , what is the shape of
the excess hole distribution in as determined from Eq. (5.37)?
Briefly explain the difference between the I-V relationships of a silicon and a gallium
arsenide diode.
p

148

Chapter 5

The PN Junction Diode

HIGHLIGHTS

In a PN diode at thermal equilibrium, a built-in voltage is generated at the junction


that makes positive with respect to P. Consequently, the drift components for each
of the electron and hole currents are balanced by equal and opposite diffusion cur
rents so that each of the electron and hole currents are zero.

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 forward-biased.
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 I-V 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
E5-3

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

E5-4

18

16

Determine the diode currents at (a) V = 0.6V and (b) V = 0.66V


a

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 space-charge 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
' = 17-26 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.

It is quite evident that t h e diffusion currents at equilibrium are m a n y orders of


m a g n i t u d e greater than those at forward bias for low injection. It follows that the
drift currents at equilibrium have values that are m u c h larger t h a n the currents at
forward bias. Thus, the values of the diffusion currents calculated at forward bias rep
resent the difference between large diffusion and drift
components.
Using the results and conclusions of this section, we will verify the b o u n d a r y
condition that we used in Eq.(5.35), namely that

(>'>)=
where p'(0) = (p(0) - p ) is the excess hole density in at the edge of the spacecharge region.
The assumption of low-level 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

E q u a t i o n (5.15) was derived assuming, in E q . (5.12), that at equilibrium the


drift and diffusion currents are equal. Since that assumption is also valid at forward
and reverse bias, we replaced V" (at equilibrium) in E q . (5.15) by (V* V ) so that
E q . (5.15) may be rewritten as
bi

bi

so that
p(-x )
p

= M*)(exp ^ )

p ( ^ )

4 8

(5- )

Since from Eq.(5.15), exp-y^f = , we have


kT
p
Qn

rf-^-jWg-P^)

(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

The PN Junction Diode

T h e excess hole density at = x ,p'(0)

is p(0) p ,

so we can write

0n

exp-

(5.51)

kT

A similar relation can b e obtained for electrons in the region.


In t h e next chapter, we will c o m p a r e t h e actual diode characteristics with
those described by the relations derived here. We will then discuss t h e causes for the
deviations b e t w e e n the two.
G e n e r a l E q u a t i o n for Hole Distribution in t h e R e g i o n of t h e
PN J u n c t i o n D i o d e
The continuity equation, (Eq. (5.31), can b e written as
2

d p'

p'

(5.52)

dx

where p' is the excess hole distribution in and L


\

= (D )
V

'

The solution to this equation is


B

where

B.

exp

P ' ( 0 ) = P exp
0

(5.53)

exp

(5.54)

Using the b o u n d a r y conditions in E q . (5.53) we kT


have
p'(0)

= B, + B

(5.55)

= 0

p'(W)

(5.56)

We substitute the relation of E q . (5.56) in E q . (5.53) to get


B-

n\

(5.57)

Substituting E q . (5.56) into E q . (5.57) and solving for B and B , we have


x

-p'(0)

p'(0)
1

2W,
exp

exp (2W

(5.58)

exp

Substituting these values for and B into E q . (5.53), we have

p'(0)

exp

exp

[) ~ l2W -x'Y\
n

(2W \
n

(5.59)

Problems

151

This equation can be simplified for two cases.


Case I: For W

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') > 3L , E q . (5.61) reduces to


p

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%

We have neglected higher order terms t h a n the first, since x'


E q . (5.63) as

p'(0)

X'

- +

2p'(0) ' -

L . We rewrite

Ep_

2W
L

2W,

or
' = ' ( -

(5.64)

wherep'(0) is given by E q . (5.54).


PROBLEMS
For all diodes assume: W L and W L , unless otherwise indicated. Also in and
regions p = N and n = N respectively. = 300K
n

0p

0n

152

Chapter 5

The PN Junction Diode

5.1 An abrupt junction PN germanium diode has N = 10 cm and N = 10 cm .


Assuming all impurities are ionized and at = 300K determine, at equilibrium:
a)
The built-in voltage V .
b)
The widths of the depletion region, x and x .
c)
The electric field at = 0.
5.2 Assume abrupt junction PN diodes having N = I 0 c m r and N = 1 0 c n r .
Calculate the built-in voltage at = 300K for:
a)
A silicon diode.
b)
A germanium diode.
5.3 At = 300K, the region of an abrupt junction diode has a resistivity of 0.1 ohm-cm
and the region has a resistivity of 1 ohm-cm. Assume n. = 2.5 X 10 cm~ , =
3600cm /V-s and = 1700cm /V-s. Determine the built-in voltage V .
5.4 The forward current across the depletion layer of a PN junction diode consists of elec
trons from to and holes injected from to N. The ratio of the hole current cross
ing the junction to the total current is known as the injection efficiency. Determine an
expression for the injection efficiency as a function of:
17

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

'

5.5 For a forward-biased abrupt junction diode, W L , L = , and at = x ,


(x' = 0) the ratio of the hole current to the electron current, / //, is 100 in the steady
state. Determine I /I at x' = .
5.6 A germanium P N diode operating at = 300K has the following properties: =
= , N = 10 cm- , N = 1 0 c n r , D = 39cm /s, D = 26cm /s and the area A
= 1.25 X lO^cm . Determine:
a)
The diode's current when 0.2V is applied in the forward direction.
b)
Repeat part (a) but in the reverse direction.
c)
Repeat part (a) for a silicon diode having the same properties as the Ge
diode.
d)
Repeat part (b) for the silicon diode.
5.7 A P N silicon diode has 7V = 10 cm~ and N = 10 cm~ , = i = and A =
1.2 X 10" cm . Determine at = 300K:
a)
The reverse saturation current.
b)
The current when the applied voltage in the forward direction is 0.7V.
c)
The current when the applied voltage in the reverse direction is 0.7V.
5.8 An N P diode has the following properties: N = 10 cm~ , N = 10 cm~ , A =
1 0 - W , W = 2, W = 200, = = 0.28, D = 20cm /s, and D =
10cm /sec. Plot at = 300K the individual electron and hole currents in P, as a func
tion of distance, for V = 0.5V, given n] = 10 cm .
5.9 A P N junction diode has the following properties: N = 10 cm" , N = 1 0 c m , A
= 10- cm , = 2000cm /V-s, = 4000cm /V-s, L = 2X lO^cm and L = 3 X
10 cm. Given n = 1 0 c m and the relative permittivity = 16, determine:
a)
The conductivities of the and regions.
b)
The built-in voltage.
c)
The reverse saturation current of the diode.
d)
The diode current for a forward bias of 0.25V.
+

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
ohm-cm and that of the region is 0.2 ohm-cm. 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 ohm-cm
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 majority-carrier current equals the minor
ity-carrier 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 /V-s and = 0.1 , and for holes in , = 480cm /V-s and =
. Use = 10~ cm :
A

a)

Determine at = 300K the diode current for V = -10V, V = -0.1V and


V = 0.4V.
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 low-level 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 low-level injection is reached.
A

17

15

18

5.16 A P N germanium diode has N

-3

= 1 0 c m and N

15

= 10 cm" . Determine:

a)

The excess hole concentration at the edge of the depletion layer in N, at =


x ,for V = 80mV.

b)

At what value of V is the limit of low injection reached.

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

The PN Junction Diode

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)

An expression for the depletion layer width, W.


abrupt junction silicon diode has N = 10 cm~ and N = 5 X 10 cm~ .
in = 0.1 5, in = 0.028, and A = 10~ cm , determine:
The reverse saturation current.
For V = V /2, the excess minority carrier concentration at x' = 0 (x = x )
and at x' = into N.
c)
For V = 0.75V , the injected minority carrier currents on both sides of the
depletion layer.
5.19 For a PN silicon junction diode having N = 1 0 c n r and N = 1 0 c n r , determine
V for:
a)
= 300K.
b)
= 450K.
+

17

15

bi

bi

16

13

bi

5.20 One of the shortcomings of semiconductor devices is their sensitivity to temperature.


Semiconductors with narrow band gaps have a higher intrinsic carry density. In this
problem and in Prob. 5.21, you are to establish the effect of temperature on the
reverse saturation current.
(a)
Assuming that the effects on I of changes in band gap energy and mobility
are negligible, show that the reverse saturation current of a diode is given by
s

I=

CPexp(-E /kT)

where C is a constant that depends on diode properties.


(b)

5.21

a)

Calculate the factor by which I increases as the temperature increases from


27C to 100C for
i) A Ge diode
ii) An Si diode
Show that the fractional change in the reverse saturation current of a diode
per unit change in temperature is given by
s

1 dE
L dT

__
3

E
kT

Chapter 5
b)
5.22

a)

Problems

155

Determine the fractional change in the reverse saturation current at


= 300K for a germanium diode and a silicon diode.
Show that the fractional change in the forward current of a diode,
operated at a given voltage V , is given approximately by
1 dl _ E _ qV
I dT
kT
g

b)

5.23

a)

Calculate the fractional change in the forward current of a diode


operated at a bias of 0.2V at = 300K for a germanium diode and a
silicon diode.
Show that the expression for the electric field in the region of a sili
con PN junction diode at equilibrium is given by
kTldp
q dx

b)

18

If the impurity distribution in is given by = 10 exp~ '/QA, where


x is in microns, calculate the magnitude of % at x' = 0 in the region.
The point x' = 0 occurs at the edge of the depletion layer.
Compare the electric field in the region to that in the depletion layer
by calculating the maximum value of the electric field in the depletion
region of an abrupt junction diode at equilibrium if N = 1 0 c n r
and N = 10 cm" .
1

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

Illustration of wafer and chip.

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 cylindrical-shaped
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 metal-semiconduc
tor field-effect transistor, and a metal-oxide-semiconductor field-effect 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

The fundamental processes in t h e fabrication of discrete devices and inte


grated circuits are the same. In this chapter, we will explain and illustrate the various
processes that are required to form the devices and the integrated circuits.
The fabrication of semiconductor devices has b e e n based on t h e use of silicon
as the p r e m i e r semiconductor. Two other semiconductors, germanium and gallium
arsenide, present special p r o b l e m s while silicon has certain specific advantages not
available with the others.
A t 300K, silicon has a b a n d gap of 1.12eV, while germanium's b a n d gap is
0.66eV. Because of this small b a n d gap, t h e intrinsic carrier density of germanium at
= 300K is about 2.5 X 10 cm . A t t e m p e r a t u r e s of about 400K, this density
b e c o m e s 1 0 c m ~ , which is c o m p a r a b l e to t h e lower range of doping densities used.
This p r o p e r t y limits its use to low t e m p e r a t u r e applications at less than 350K.
The o t h e r semiconductor of major interest is gallium arsenide. In spite of its
attractive electrical properties, gallium arsenide crystals have a high density of crys
tal defects, which limit t h e performance of devices m a d e from it. F u r t h e r m o r e , com
p o u n d semiconductors, such as G a A s (in contrast to elemental semiconductors such
as Si and G e ) are m u c h m o r e difficult to grow in single crystal form. B o t h silicon
and g e r m a n i u m do not suffer, in t h e processing steps, from possible decomposition
that may occur in c o m p o u n d semiconductors such as G a A s .
O n the plus side, G a A s has a low-field electron velocity that is larger than sili
con so that electron devices using G a A s are faster t h a n those using Si. Also, G a A s
has a lower saturation electric field than Si so that G a A s devices have a smaller
power-delay product. Devices m a d e from substrates of G a A s tend to have smaller
parasitic capacitances, which contribute to their speed advantage over silicon
devices. A n o t h e r advantage of G a A s results from its direct b a n d gap, which m a k e s it
possible to provide certain functions not possible in Si, such as coherent and inco
h e r e n t light emission.
A major advantage of silicon, in addition to its a b u n d a n t availability in the
form of sand, is that it is possible to form a superior stable oxide, S i 0 , which has
superb insulating properties and, as we shall see later, provides an essential and
excellent ingredient in the fabrication and protection of devices.
Lastly, at the present time, silicon remains the major semiconductor in the
industry.
13

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 cost-effective 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

The Purity of Silicon

159

fabrication, silicon must b e in crystalline formvery pure, free of defects, and


uncontaminated. T h e r e is n o other industry that places d e m a n d s as severe as those
required by the semiconductor industry on the purity of silicon.
In the following discussion, we will d e t e r m i n e an estimate of t h e expected
purity of silicon. We start with the expected results.
Normally, and for the greatest n u m b e r of applications, the d o p a n t densities in
devices vary from 1 0 t o 1 0 c m . T o be effective, these densities must be very pre
cisely controlled. Impurities that are inherently present in the silicon or that m a y b e
accidentally introduced, w h e t h e r in the bulk (volume) form or o n the surface of the
material, interfere to a serious extent with the operation of t h e fabricated devices. In
o r d e r for the impurities to b e at a sufficiently low and acceptable level, their density
should b e n o greater than two or t h r e e orders of magnitude lower than 1 0 c m ~ . A t
this level, t h e density of u n w a n t e d impurities should not exceed 1 0 c m ~ . We deter
mined in C h a p t e r 1 that there are approximately 1 0 atoms per cubic centimeter in
silicon. We are requiring, therefore, that there b e n o m o r e than one u n w a n t e d impu
rity in 1 0 silicon atoms. This is a purity of one in a billion.
15

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 90-99 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

Silicon tetrachloride is a liquid that can b e distilled and b e m a d e of very high


purity. T h e SiC is subjected to hydrogen reduction to p r o d u c e silicon. M o r e com
monly, t h e metallurgical grade silicon is purified by combining it with hydrochloric
acid to form trichlorosilane, S i H C . The trichlorosilane is reduced by hydrogen to
form very p u r e silicon, k n o w n as semiconductor
grade silicon, which is vapor
deposited on high purity silicon rods.
A t this stage, the solid is polycrystalline composed of m a n y small crystals (of
submicron dimensions) having r a n d o m orientation and containing m a n y defects.
The silicon contains one u n w a n t e d impurity a t o m in about 10 atoms of silicon. The
4

160

Chapter 6

Fabrication Technology

purity is excellent. For silicon to be used in the fabrication of devices, however, it


must be nearly perfect and crystalline in nature. We, therefore, now n e e d to produce
single crystals of silicon. This is d o n e by a m e t h o d k n o w n as crystal growth. Crystal
growth consists of converting a r a n d o m oriented or polycrystalline material into
one that is orderly and crystalline.
T h e easiest a p p r o a c h is to melt silicon and let it freeze. O n e m e t h o d to carry
this out is in a growth process k n o w n as the Czochralski process. In this process,
crystalline silicon, to which a d o p a n t is added, is grown.

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 quartz-lined
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

Solid-liquid interface

Solid single crystal

RF heating coil

Rotate and lift


Quartz
crucible
Figure 6.2 The Czochralski method for crystal growth and purification.

Section 6.3

The Czochralski Growing Process

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

T h e r e is also a limit t o the a m o u n t of impurity that can be added as the con


centration in t h e melt must not exceed about 2 percent, otherwise single crystal
growth is h a m p e r e d . This together with t h e segregation coefficient limit t h e a m o u n t
of doping in the crystal.
The following example illustrates the mass of a dopant required for a certain
doping density.
E X A M P L E 6.1
16

A silicon ingot that should contain 10 phosphorus atoms/cm is to be grown by the


Czochralski method.
a)
Determine the concentration of phosphorus atoms in the melt to give the required con
centration in the ingot.
b)
The crucible initially contains 50Kg of molten silicon. Determine how many grams of
phosphorus should be added:
Given: for phosphorus k = 0.35, density of silicon = 2.53g/cm and atomic weight of
phosphorus = 30.975g/mole, Avogadro's number = 6.023 x 10 atoms/mole.
3

23

Solution
a)
b)

k = C /C C = 10 crrr , C = 2.85 10 atoms/cm


volume of silicon = 50 X 1072.53 = 1.976 X 1 0 W
Number of phosphorus atoms = 2.85 10 X 1.976 X 10" = 5.63 10
Amount of phosphorus =
" ' ; ^ W - * = 28.95mg.
l6

)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.

It is important t o note that for a given impurity there is a m a x i m u m density of


the impurity, at a given t e m p e r a t u r e , that is allowed in or can be absorbed by a crys
tal. This m a x i m u m concentration is known as solid solubility. Because the solubility
decreases with t e m p e r a t u r e , if an impurity is introduced at its m a x i m u m concentra
tion and the t e m p e r a t u r e is reduced, t h e crystal precipitates the excess impurity to
achieve equilibrium.
S e e d Crystal
After having set u p the melt, a seed crystal (a small highly perfect crystal), attached
to a holder and possessing the desired crystal orientation, is dipped into t h e melt
and a small portion is allowed to melt. Very slowly, t h e seed is rotated and pulled up

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
mirror-like 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

Metallizations and interconnections.

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 resistance-heated 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

controlled by microprocessors, which m o n i t o r b o t h the gas flow sequence and the


furnace t e m p e r a t u r e . A t h e r m a l oxidation system, together with an example of S i 0
growth, are shown in Fig. 6.3.
A silicon surface oxidizes very rapidly so that even at r o o m t e m p e r a t u r e a
layer of silicon dioxide about 2nm thick is p r o d u c e d on the wafer. In the thermal
oxidation process, the thickness of the oxide varies from 0.1 for VLSI gate oxide to
one micron. F r o m a consideration of densities and relative molecular weights of Si
and S i 0 , an oxide film of thickness x consumes 0.44JC of the silicon. This relation
ship is d e t e r m i n e d in the example that follows.
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

the volume of 1 mole of SiO, =


2

the volume of 1 mole of Si =

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

volume of 1 mole of silicon


volume of 1 mole of S i 0
2

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

Thus, to grow 200 Angstroms of S i 0 , a layer of 88 Angstroms of silicon is used up.


2

O n c e the layer of S i O has b e e n formed on the surface of t h e wafer, it is selec


tively r e m o v e d (etched) from those surfaces where impurities are to b e introduced
and k e p t as a shield, for the underlying silicon surface, where n o dopants are to b e
allowed.
Oxide layers are relatively free from defects and provide stable and reliable
electrical properties.
z

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 radio-frequency 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 solid-state 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 drive-in 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

depth into substrate


Figure 6.4 N-Impurity distributions into P-substrate, for predeposition and drivein diffusion.
of their concentration gradient and into the locations that the silicon atoms vacated.
Predeposition tends to produce, near the silicon surface, a shallow but heavily
d o p e d layer. The second step of diffusion, drive-in, is used to drive t h e impurity
atoms d e e p e r into the surface, without adding any m o r e impurities, thus reducing
the surface concentration of the dopant. A sketch of the resulting impurity distribu
tions is shown in Fig. 6.4.
C o m m o n dopants are boron for P-type layers and phosphorus, antimony, and
arsenic for N-type layers. Diffusion, however, is rarely performed using the p u r e ele
m e n t s themselves. R a t h e r , c o m p o u n d s of the elements are used and impurities may
be introduced from either solid, liquid, or gaseous sources.
B o r o n and p h o s p h o r u s have two desirable properties: (1) they have a high dif
fusion rate (diffuse easily and quickly) into silicon and low diffusion rate in S i 0
and (2) they are b o t h highly soluble in silicon. We refer to S i O since, prior to carry
ing out the process of diffusion, windows are o p e n e d , in a previously deposited layer
of S i 0 , for the impurities to diffuse in. The o p e n windows are in silicon but the
adjacent regions are shielded with silicon dioxide. The process of window opening is
covered later.
A typical a r r a n g e m e n t of t h e process of diffusion is shown in Fig. 6.5. The
wafers are placed in a quartz furnace t u b e that is heated by resistance heaters sur
rounding it. So that t h e wafers may b e inserted a n d r e m o v e d easily from t h e fur
nace, they are placed in a slotted quartz carrier k n o w n as a boat. The wafers are
m o u n t e d on their side, as illustrated in t h e figure.
To introduce a p h o s p h o r u s dopant, as an example, p h o s p h o r u s oxychloride
( P O C l ) is placed in a container either inside the quartz tube, in a region of rela
tively low t e m p e r a t u r e , or in a container outside the furnace at a t e m p e r a t u r e that
helps maintain its liquid form. For a dopant, b o r o n is used. The p r o p e r vapor pres
sure is maintained by a control of t h e t e m p e r a t u r e .
2

Section 6.4

Fabrication Processes

167

resistance heaters
quartz furnace tube
quartz tube

Vent
end caps
Liquid impurity source

Figure 6.5

silicon wafers

Physical layout of equipment used in diffusion.

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 drive-in 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
drive-in 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

Diffusion of dopant atoms in silicon.

168

Chapter 6

Fabrication Technology

Expressions f o r t h e Diffusion of D o p a n t Concentration


The process of diffusion of dopants is similar to that of the diffusion of holes and
electrons that was discussed in C h a p t e r 4. We will relate expressions for the rate of
d o p a n t diffusions. First, we define relevant terms:
C = d o p a n t concentration in c m " at surface (x = 0).
3

= diffusion depth into t h e substrate (cm).


t = duration of diffusion (s).
2

klT

D = diffusion coefficient ( c m / s ) exp~ ,

k is a constant.

Vz)i.

L = diffusion length =
C(x,t)

= d o p a n t concentration at depth and time t.


Assuming that the diffusion coefficient is i n d e p e n d e n t of doping concentra
tion, the diffusion equation is given by
dC/dt

= >0

(6.1)

By using the appropriate b o u n d a r y conditions, it can b e shown that the doping


concentration C after the predeposition
step at a d e p t h and after time t is given
by
v

QOcii) = C, erfc

(x/V4D t )

(6.2)

i l

w h e r e the error function a n d the error function c o m p l e m e n t are defined as


erf(x) = 2/VTT f e^dy
Jo

(6.3)

and
erfc(x) = 1 - e r f ( )

(6.4)

These functions are tabulated in mathematical handbooks. We list below a few


of their properties:
erf(0) = 0
erf(OO) = 0

dx

j eric(x)dx
Jo

= 1/VTT

The expression for C after the drive-in diffusion becomes


C (x,t )
2

= 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 drive-in 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

Impurity profiles for predeposition and drive-in diffusion.

relevant diffusion coefficients at the given t e m p e r a t u r e s are = 2.5 X 10


and D = 2.5 X 1 0 ~ c m / s and the concentration at = 0 is C = 1 0 c m ~ .

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 high-energy 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:

D o p i n g levels can b e precisely controlled since the incident ion b e a m can b e


accurately m e a s u r e d as an electric current.

T h e d e p t h of the d o p a n t can b e easily regulated by control of t h e incident ion


velocity. It is capable of very shallow penetrations.

E x t r e m e purity of the d o p a n t is guaranteed.

The doping uniformity across t h e surface can b e accurately controlled.

Because t h e ions enter the solid as a directed b e a m , t h e r e is very little spread


of the b e a m , thus t h e doping area can b e clearly defined.
Since this is a low-temperature process, t h e m o v e m e n t of impurities is mini
mized.
This process has o n e major shortcoming in that it may create considerable
crystal d a m a g e because of t h e collisions of the high energy ions with t h e silicon.
T h e r e are two types of collisions: First, electronic collisions take place b e t w e e n the
incident electrons and t h e target atoms; second, nuclear collisions occur involving
the nuclei of t h e incident and target atoms. Such d a m a g e results in inferior perfor
mance of devices m a d e by this process. The d a m a g e may b e so extensive that it
transforms the silicon from a crystalline to an a m o r p h o u s structure (Sec. 1.1).
If t h e d a m a g e is not extensive, the lattice structure is restored by the process
of annealing. In one process, k n o w n as furnace annealing, the wafer is h e a t e d in an
inert a t m o s p h e r e at a t e m p e r a t u r e of 600 to 1000C for about 30 minutes. This
restores t h e crystalline a r r a n g e m e n t of t h e silicon.
In a n o t h e r process, k n o w n as rapid t h e r m a l annealing ( R T A ) , optical radiant
energy is generated and delivered to t h e surface of the wafer. The energy is gener
ated by a tungsten-halogen l a m p at a wavelength of 0.3 to 4 micrometers in a quartz
enclosure. Because of the n a t u r e of the process, the quartz walls do not acquire the
light energy that is directed at the wafer and thus t h e wafer is not in t h e r m a l equilib
rium with the walls of the system. A s a result, this process serves to considerably
reduce t h e annealing time c o m p a r e d to that required in the furnace.
A n o t h e r shortcoming of ion implantation is t h e investment in t h e e q u i p m e n t
mainly because of its complexity.

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

function is to expose t h e selected regions and protect t h e others. T h e r e m a y b e hun


dreds of identical dice (or ICs) on a wafer with each circuit containing h u n d r e d s of
thousands, or millions, of devices. Identical steps are carried out simultaneously for
each process, such as t h e diffusion of the regions of a diode on one or several cir
cuits, to b e r e p e a t e d over the whole wafer. For each process, a separate mask is
needed.
T h e mask production starts with a drawing using a computer-assisted graphics
system with all the information about t h e drawing stored in digital form. C o m m a n d s
from t h e c o m p u t e r are p r e p a r e d that drive a p a t t e r n generator, which uses an elec
tron b e a m to write (photoengrave) the particular pattern, for one or several dice, on
a glass plate covered with a thin c h r o m i u m film. W h e n t h e glass plate is p r e p a r e d for
one or several dice on the wafer, it is k n o w n as a reticle. A mask usually refers to a
glass plate that contains a p a t t e r n for t h e whole wafer. The reticle p a t t e r n is p r o
jected o n t o the wafer and a wafer stepper is used that reduces the reticle circuit onto
the photoresist-covered (see next section) wafer that steps across over t h e surface
until the entire array of circuits is built up.
The use of a single mask for all t h e circuits on a wafer is not feasible for print
ing very small ( < 1 ) features because of alignment p r o b l e m s resulting from t h e
heat that t h e wafer is exposed to, which causes slight distortion of the surface. Such
systems are still in use, however, for fabricating simple logic circuits and analog
devices such as L E D s .

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 light-sensitive 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

1. The wafer is b a k e d at 100C to solidify the resist on the wafer.


2. The reticle is placed o n t h e wafer and aligned by c o m p u t e r control.
3. The reticle is exposed to ultraviolet light with t h e transparent parts of the reti
cle passing the light onto the wafer. T h e photoresist u n d e r t h e o p a q u e regions
of the reticle is unaffected.
4. The exposed photoresist is chemically r e m o v e d by dissolving it in an organic
solvent and exposing the silicon dioxide u n d e r n e a t h . This is a process very simi
lar to that used in developing photographic film.
5. The exposed silicon dioxide is t h e n etched away using hydrofluoric acid, which
dissolves silicon dioxide and not silicon. T h e regions u n d e r t h e o p a q u e part of
the reticle are still covered by the silicon dioxide and t h e photoresist.

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

mask placed above


photoresist and UV
light directed at it

'////////////////////,
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.

6. The photoresist u n d e r the o p a q u e regions of the reticle is stripped using a


p r o p e r solvent and the silicon dioxide is exposed.
All surfaces are protected except those covered by silicon only in which diffu
sion or ion implantation is to t a k e place. The surfaces covered by silicon dioxide do
not permit any entry of dopants.
T h e photoresist used in this discussion is labeled a positive resist, whereby
windows are o p e n e d wherever the ultraviolet light passes through the transparent
parts of t h e mask. Negative resist is also used and remains on t h e surface, which is
exposed to U V light, and windows are o p e n e d u n d e r t h e o p a q u e parts of the mask.
T h e r e is a practical limit, at linewidths of about 1-2 micrometers, to using
ultraviolet light. The printing of smaller features can b e accomplished by using very
short wavelength radiation such as electron-beam or x-ray lithography.

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 10-20 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 boat-shaped 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

gas valve and


flow gauge
vent

so

cl

- temperature bath

Figure 6.9

SiCI

System for growing an epitaxial layer.

174

Chapter 6

Fabrication Technology

outlets for the gases. The t u b e is h e a t e d by induction using t h e heating coils w o u n d


a r o u n d the tube.
All the chemicals that are introduced and that take part in the reactions are in
the form of gases, hence the process is k n o w n as Chemical Vapor Deposition
( C V D ) . The epitaxial layer is grown from t h e vapor phase onto the silicon, which is
in the solid state. T h e thickness of the layer varies from 3 to 30 microns and the
thickness of t h e layer and its doping content are controlled to an accuracy of less
than 2 percent. The reactions are carried out at a t e m p e r a t u r e of approximately
1200C. T h e high t e m p e r a t u r e is necessary so that t h e d o p a n t atoms can acquire a
sufficient a m o u n t of energy to allow t h e m to m o v e into the crystal to form covalent
b o n d s and b e c o m e an extension of the single crystal lattice. Because the layer is
grown o n the substrate, epitaxy is a growth technique w h e r e the crystal is formed
without reaching t h e melting point of silicon.
We list below, and with reference to Fig. 6.7, the sequence of operations
involved in t h e process:
1. H e a t wafer to 1200C.
2. Turn o n H to reduce t h e S i O on t h e wafer surface.
2

3. Turn on anhydrous H C to vapor-etch t h e surface of t h e wafer. This removes a


small a m o u n t of silicon and other contaminants.
4. Turn off H C C
5. D r o p t e m p e r a t u r e to 1100C.
6. Turn on silicon tetrachloride (SiC ).
4

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 N-type 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 P-layer 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 silicon-based 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

An aluminum interconnection between two diffused silicon regions.

176

Chapter 6

Fabrication Technology

subjected to a high intensity electron b e a m formed by an electron gun, which vapor


izes the a l u m i n u m that travels to t h e wafer. By t h e use of a mask and photolithogra
phy, the a l u m i n u m is deposited on the identified regions on the wafer surface.
In t h e sputtering process, t h e material to b e deposited is placed in a container
maintained at low pressure in t h e vicinity of the substrate. T h e material to be
deposited is labeled the cathode or target, while the a n o d e is the substrate. A D C or
a radio-frequency high voltage is applied b e t w e e n a n o d e and cathode. This high
voltage ionizes the inert gas in the chamber. The ions are accelerated to the cathode
(in this o p e r a t i o n the a n o d e is usually g r o u n d e d ) where, by impact with the alu
m i n u m target, atoms of aluminum are vaporized. A gas of aluminum atoms is gener
ated and deposited on t h e surface of the wafer.
Following the deposition of aluminum, the silicon wafers are placed in a fur
nace to solidify the connections so that low resistance metallic contacts are made.
The interconnections b e t w e e n elements of an integrated circuit are m a d e by
aluminum lines having a thickness of about 0.5. These are laid on t o p of the sili
con dioxide layer, which covers the surface of the wafer. By using photolithography,
openings are m a d e in the silicon dioxide so that the aluminum layer is connected to
the silicon or to t h e ohmic contact on the silicon. In very complicated integrated cir
cuits, it is necessary to have two or three vertically stacked layers of interconnec
tions separated by silicon dioxide layers. The interconnecting lines terminate at
aluminum pads from which connections to the outside are made.
The connections from t h e chips to the outside world are p r o d u c e d by the met
allization p a t t e r n a r o u n d t h e periphery of the chip. These are k n o w n as bonding
pads and they are connected by wires to the package, as shown in Fig. 6.11. The
bonding pads are about 100>m square and the bonding wires are m a d e of gold with
a diameter of about 25. Following these connections, the individual dice are
encapsulated and hermetically sealed in a variety of packages.

leads

Figure 6.11 Typical die package. Source:


J. Mayer and S. Lau, Electronic Materials
Science: For Integrated Circuits in Si and
GaAs, Macmillan (1990).

Section 6.5

Planar PN Junction Diode Fabrication

177

It is i m p o r t a n t to maintain t h e resistance of the interconnections to small val


ues. A large value for this resistance combined with parasitic capacitances can result
in a time constant R C , which may end up, in a superfast device, to b e the limiting
factor in the overall speed of a circuit.
O h m i c Contacts
It cannot be assumed that depositing aluminum on a d o p e d semiconductor forms a
very good ohmic contact. A n ideal ohmic contact exhibits a perfectly linear relation
ship b e t w e e n current and voltage with t h e plotted current-voltage characteristic
passing through the origin. In certain cases, and as we shall discuss in later chapters,
a contact b e t w e e n aluminum and a semiconductor may result in a rectifying contact
having I-V characteristics of a diode, whereby current conduction is only in one
direction.
O n P-type silicon having a concentration of 1 0 c m , or m o r e , aluminum
forms a good ohmic contact. However, aluminum on lightly d o p e d N-type silicon
forms a rectifying contact. In o r d e r to prevent the formation of such a contact, an
N diffusion is placed over t h e N-type silicon.
16

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 short-circuit between two of the three terminals of the transistor (collec
tor to base).
Figure

Process Description
+

(a)

A n N substrate grown by the Czochralski process is t h e starting metal


~ 150 thick.

(b)

A layer of N-type silicon (1-5 ) is grown on the substrate by epitaxy.

(c)

Silicon dioxide layer deposited by oxidation.

(d)

Surface is coated with photoresist (positive).

(e)

Mask is placed on surface of silicon, aligned, and exposed to U V light.

(f)

Mask is removed, resist is removed, and S i O u n d e r the exposed resist is


etched.

(g)

B o r o n is diffused to form region. B o r o n diffuses easily in silicon but not


in S i 0 .

(h)

Thin aluminum film is deposited over surface.

(i)

Metallized area is covered with resist and a n o t h e r mask is used to identify


areas w h e r e metal is to be preserved. Wafer is etched to r e m o v e u n w a n t e d
metal. Resist is then dissolved.

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)

Figure 6.12 Typical steps in the fabrication of a planar PN junction diode.

(j)

Contact metal is deposited on t h e back surface and ohmic contacts are


m a d e by heat treatment.

In t h e following chapters, we will b e discussing o t h e r devices. W h e r e v e r appro


priate and necessary, reference to the particular fabrication techniques will be illus
trated.

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 N-epitaxial 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) cross-section, (b) rectangular
dimensions, and (c) meander pattern.

R =

pL/Wd

where is the average resistivity of t h e layer in ohm-cm, d is its thickness, L is its


length and W is its width.
Resistances in monolithic circuits are defined by a term k n o w n as the sheet
resistance. The sheet resistance has units of o h m s p e r square, it being the resistance
of a square having W = L.
Assuming 100 to 200 o h m s p e r square, practical resistors may have values
ranging from 100 to several kilohms. H i g h e r resistances are obtained by using a
m e a n d e r pattern, as shown in Fig. 6.11(c). T h e major p r o b l e m with resistors of high
values is that they tend to occupy a large area on t h e chip. A resistor of 50 Kohms
uses u p an area of the semiconductor that may b e occupied by h u n d r e d s of transis
tors. Ion implantation can b e used, however, to generate lightly d o p e d regions suit
able for precision high-value resistors.
Capacitors

O n e type of capacitor in monolithic circuits is m a d e by using the capacitance


formed b e t w e e n t h e and regions of a reverse-biased diode. Such a capacitance
is shown in Fig. 6.14.
These capacitors, just like resistors, are formed in the same diffusion processes
that are used to form devices. A s we will see in a later chapter, bipolar transistors
are m a d e of t h r e e regions in which either of t h e two P N junctions may b e used as
capacitors, w h e r e b y the b r e a k d o w n voltage of t h e capacitor m a y vary considerably
from o n e to the other.

180

Chapter 6

Fabrication Technology

V/////////.

'V/////////////A

Figure 6.14

A capacitor in a monolithic circuit.

The disadvantage of junction capacitors is the d e p e n d e n c e of the capacitance


on t h e voltage applied t o t h e junction. Capacitors that are voltage-independent can
be formed from metal insulator N semiconductor layers, as used in M O S struc
tures. M e t a l oxide semiconductor junctions and devices are studied in Chapters 12
and 13.
+

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

where C(x,t) = C erfc [*/(2/))]


Use the equation for the predeposition concentrations to determine an expression for
Q in terms of D, t, and C .
6.5 A certain process requires a boron predeposition diffusion step into an N-type wafer.
The wafer has been uniformly doped, prior to this step, by 1 10 phosphorus
atoms/cm . The predeposition step is carried out at 1000C for 30 minutes. The diffu
sion coefficient, D, of boron in silicon at 1000C is 1.78 X 10~ cm /s and the solid sol
ubility of both boron and phosphorus in silicon is assumed to be 3 X 10 /cm at the
relevant temperature. Determine:
a)
The diffusion length for this step.
s

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 drive-in 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)

The surface concentration after the drive-in step.


Q after the drive-in step.
The junction depth after the drive-in step, recalling that the phosphorus con
centration is 1 X 10 /cm .
Note that since no new dopant is introduced during the drive-in step, the Q at the ini
tiation of this step is the same as that at the completion of the predeposition step.
6.7 List the principal advantages and disadvantages of ion implantation.
6.8 Why are two diffusion steps normally used when one would be less costly?
6.9 What is the main reason for the use of epitaxy in the fabrication of microcircuits?
15

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 current-voltage exponential relationship, we observe deviations
b e t w e e n t h e two. These deviations occur mainly at the low-current and t h e high-cur
rent ends for forward bias, and in the reverse-bias 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 current-voltage 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

Deviations in the Forward Region of the Characteristic

183

1. N o recombination of carriers takes place in the depletion region w h e n a for


ward bias is applied to the diode.
2. N o generation of carriers occurs in the depletion region during reverse bias.
3. L o w injection. This refers to a range of diode currents that causes t h e highest
value of minority carrier density, which occurs at the edge of the depletion
region at forward bias, to b e m u c h smaller than the equilibrium majority carrier
density.
4. Coupled with assumption 3 is t h e assumption that the voltage applied to a
diode appears totally across the depletion region. Because there is n o voltage
change in t h e regions outside the depletion region, n o electric field exists there
and h e n c e n o charges. These regions are therefore labeled t h e neutral regions.
This justifies t h e process of calculating the currents in these regions by using t h e
diffusion of minority carriers (refer to the earlier section, Electric Field in and
Regions).
In the absence of any recombination in the depletion region, we conclude that
t h e r e is n o change in the m a g n i t u d e of the minority carrier currents as they traverse
the depletion region. We d e t e r m i n e d the ideal diode equation by finding the current
d u e to holes at the edge of t h e depletion region in and t h e current due to elec
trons at the edge of the depletion region in P. Since these currents were assumed to
be constant t h r o u g h o u t the depletion region, their sum d e t e r m i n e d t h e total diode
current, which of course is constant t h r o u g h o u t the diode.
R e c o m b i n a t i o n Currents
By relaxing t h e first assumption, requiring that n o recombination takes place in the
depletion region of the diode w h e n a forward bias is applied, we will show that the
actual I-V characteristic of t h e diode at low voltages deviates from the ideal rela
tions derived in C h a p t e r 5. R e c o m b i n a t i o n takes place in t h e depletion region
because the carrier densities are greater than their equilibrium values.
In fact, we will show that expressions for t h e ideal and the actual total diode
current m a y b e considered and derived as being m a d e u p of electron or hole recom
bination currents, b o t h in t h e neutral regions and in the depletion region.
We identify the three regions of t h e forward-biased diode, illustrated in Fig.
7.1, as the neutral regions 1 and 2, and the depletion region 3.
Figure 7.1 is similar to Fig. 5.17, except that we are including carrier recombi
nations in all three regions. We will trace t h e path of t h e total flux of electrons, I/q,
consisting of a n u m b e r F of electrons per unit time drifting from t h e metal contact
in the external circuit through the voltage supply to the metal contact. To com
plete the loop, for a constant current / at a certain voltage through the diode, these F
electrons will disappear completely by t h e time they reach the end of t h e metal con
tact that is in touch with the region. A new batch of F electrons starts out again.
Obviously, these electrons must disappear by recombining with holes. Some of
these electrons r e c o m b i n e with holes in N, that diffuse into the region, within one
hole diffusion length from the end of t h e depletion layer = . Some of the
remaining electrons r e c o m b i n e with holes, which are sent from t h e region, inside

184

Chapter 7

Limitations to Ideal Diode Theory


- metallurigcal junction

(a)

N
2

W(V )
a

depletion layer (3)


metallic contact

metallic contact

electron
hole
(o) recombination

(b)

Figure 7.1 (a) A forward-biased diode and (b) diode currents made up of
recombination currents.

the depletion region. T h e rest of t h e F electrons recombine with holes in t h e


region within o n e electron diffusion length from = x . While we have used elec
trons in t h e above explanation, o n e can arrive at t h e same result by considering t h e
motion a n d recombination of holes within t h e diode.
Based on t h e above discussion, it is just as accurate t o derive t h e ideal I-V
relations for t h e forward-biased diode by considering only t h e two c o m p o n e n t s of
the electron or hole recombination currents that occur in t h e neutral regions. We
will n o t carry that out b u t we will use t h e results in E q . (5.43) for t h e ideal I-V diode
relations in which recombination in t h e depletion region was excluded. T h e total
current was obtained as t h e sum of t h e electron diffusion current at = x , which
in t h e absence of recombination is t h e same at = a n d t h e hole diffusion current
p

at = . T h e ideal diode current, I is given by E q . (5.43) as


p

exp

(7.1)

kT

where I is determined from E q . (5.44) by replacing p


s

Aqn]

l(L N )
p

by n /N

0jj

D
{L N )\
n

and n

Qp

by n /N
i

as

Deviations in the Forward Region of the Characteristic

Section 7.1

185

A n expression for t h e recombination current in the depletion layer, I , has been


obtained* as
R

(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 reverse-biased. For that,
RO

^GEN ~~

IRO-

T h e expression for t h e total diode current becomes

/ = /, I = I [exp ( f H + 4 o [exp (gj)-l


+

(7.3)

In a reverse-biased diode, the total current becomes I = (I + I


S

R 0

).

Comparison of Real a n d Ideal D i o d e Characteristics


A comparison of t h e I-V characteristics, predicted by the ideal diode equation, E q .
(7.1), and the actual diode equation, E q . (7.3) for a forward-biased silicon diode, is
shown on semilog scales in Fig. 7.2.
Low-Injection
We observe in the figure that at low values of forward bias, the current predicted by
the ideal diode equation is smaller than t h e actual current, whereas at high values of
voltages, t h e current given by t h e ideal equation is larger than t h e actual current.
Let us first consider the low-voltage region.
The difference b e t w e e n the two currents at low voltages is the recombination
current in the depletion region given by E q . (7.2). A s a result of this recombination,
m o r e electrons enter the depletion region from than leave and similarly for holes
from P. Thus, the total hole or electron current entering t h e depletion region is
larger than the hole or electron current leaving it (the difference in either case being
the recombination current).
The complete expression for the diode current includes, in addition to the
ideal diode equation, (which represents only the sum of the recombination currents
in t h e two neutral regions) a t e r m that accounts for the recombination current
inside t h e depletion region. We label t h e two currents that m a k e u p the ideal equa
tion as diffusion currents, whereas I in E q . (7.2) is labeled t h e recombination cur
rent.
We observe, in Fig. 7.2, that for a silicon diode at r o o m t e m p e r a t u r e , the
recombination current dominates at small forward voltages and, at voltages larger
R

*From Tyagi, Introduction to Semiconductor Materials and Devices, p. 198, copyright Wiley,
(1991). Reprinted by permission of John Wiley & Sons, Inc.

186

Chapter 7

Limitations to Ideal Diode Theory

Figure 7.2 Real and ideal diode forward


characteristics for a silicon diode at
= 300K.
than about 0.4V, the diffusion current is dominant. This is indicated by the change in
the slope from 1/2 to 1, of t h e curves, as the voltage increases. The reason for the
dominance of the recombination current at low values of voltage is that I
in E q .
(7.3) is greater than I by a factor of approximately three orders of magnitude for a
silicon diode at = 300K. The exact voltage at which the transition b e t w e e n the
two slopes occurs d e p e n d s on the t e m p e r a t u r e as exhibited through the d e p e n d e n c e
of the two currents on the intrinsic carrier density n It is also i m p o r t a n t to point out
that the width W of the depletion layer, which appears in I , is itself a function of
the applied voltageit being larger at lower values of voltage, thus contributing to
making the recombination current larger.
A n empirical relation of the forward I-V characteristic for a forward-biased
diode based on E q . (7.3) and using t h e ideality factor has b e e n d e t e r m i n e d * as
R0

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)

Deviations in the Forward Region of the Characteristic

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 I-V 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 : high-level injection and bulk resistance.
High-level 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 low-level injection and high-level 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

For 0.4V bias:

P {) ~ 1 0 c m ~ , at t h e edge of the depletion layer in N.

For 0.7V bias:

p (Q) = 1.1 X 1 0 c m - , and Ap (0)

16

n" = 1 0 c m - , and An (0)


)n

so that,

16

16

= 1.1 X 1 0 c m "

(0)

2-1

16

Ap (0)
n

1 0 c m ~ , at the edge of the depletion layer in N.

The increase of the minority carrier concentration, at high injection levels,


results from the increased forward bias and, to maintain neutrality, causes an increase
in the majority carrier concentration. The increase of both majority and minority car
rier concentration leads to higher currents and thus increased voltage drops (IR) in
the neutral regions so that the total applied voltage is d r o p p e d across both the deple
tion and the neutral regions. Furthermore, both the mobility and lifetime are functions
of the majority carrier concentrations in the region. We observe, at voltages exceed
ing 0.7V in silicon, a slower increase in the current with increased voltage.

p,n

p,n

Po

"On

POn
"Op

"Op

low injection

Figure 7.3

Carrier densities at low and high injection.

high injection

188

Chapter 7

Limitations to Ideal Diode Theory

A n approximate expression for t h e current at high injection level has b e e n


obtained as*
(7.5)
where V = V - IR and R is the total series resistance of the neutral regions. O n e can
conclude that the slope of the I-V curve becomes 1/2 instead of 1 assuming that the
plot is m a d e as a function of V . E q u a t i o n (7.5) is valid only for a P N step junction.
It is worthwhile noting that the increase in the densities of both majority and
minority carriers t h r o u g h o u t the and regions, resulting from t h e increased bias,
causes a small increase in the conductivities of these regions or a decrease in the
resistivities. However, the decrease in the resistances of the bulk regions is much
smaller t h a n the increase in t h e current, hence the result is a relatively larger volt
age drop.
d

Electric Field in a n d Regions


In our derivation of t h e ideal diode equation, we excluded any calculation of major
ity carrier currents. T h e reason is, we have n o analytical m e t h o d of determining an
expression for the electric field in t h e so-called neutral regions, although the exis
tence of an electric field is basic to the operation of the diode. This existence is veri
fied by the following explanation: Starting at the side of the depletion layer,
electrons diffuse towards the metallic contact of t h e region. In diffusing, they
r e c o m b i n e with holes until, at t h e contact, t h e density of electrons has decreased to
t h e equilibrium value. Because of recombination, the electron current in has its
largest value at the edge of the depletion layer and decreases to zero at the metal
lic contact. The total diode current in has, however, a n o t h e r c o m p o n e n t , the hole
current within the region. The hole current in t h e region is caused mainly by the
drift of holes away from the metal contact. The motion of holes away from the con
tact is a result of the acceleration acquired from the electric field.
A similar explanation holds for the motion of majority electrons in the region.
The minority-carrier drift current is assumed to be negligible because the drift
current of holes in the region, given by (q %), is m u c h smaller t h a n the drift
current of electrons in N, (q A %), due to the difference in carrier concentra
tions.

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.

Real Diode Characteristics in the Reverse Direction

Section 7.2
ideal diode
equation

-Vbr

189

Figure 7.4 Actual reverse


characteristics compared to that
predicted by the ideal diode equation.

of the o r d e r of 4kT/q, as is evident from E q . (7.1), the current saturates at a value of


-I , whereas the actual characteristic indicates an increase of reverse current with
an increase in the reverse bias. Second, and beyond that, at certain voltage labeled
(V ), the current increases rapidly unless limited by an external resistance. This
voltage is k n o w n as the breakdown
voltage.
A n increase of the reverse bias applied to a diode has the following effects:
First, t h e minority carrier densities at t h e edges of the depletion layer b e c o m e zero
for all practical purposes (p exp q VjkT, V < 0) since the barrier at the junction
has b e c o m e so steep that minority carriers are immediately swept across the junc
tion to the other region. Away from the vicinity of the depletion layer, the minority
carrier density approaches its equilibrium value. Second, the depletion layer
b e c o m e s wider as the reverse bias increases. Third, t h e potential barrier separating
t h e two regions b e c o m e s steeper (V - V , V < 0) and the m a x i m u m value of the
electric field intensity, which occurs at the metallurgical junction, increases.
The high energy barrier at the junction makes it impossible for majority carri
ers t o climb over, while minority carriers in or n e a r the vicinity of t h e junction easily
fall d o w n this barrier. This causes t h e minority carrier density in the depletion
region to be smaller than the equilibrium value. The result is an increase of the gen
eration rate of carriers, a process that is opposite to t h e increased recombination
rate in the forward-bias case. This process causes an increased reverse current and is
e n h a n c e d by a wider depletion layer as the reverse voltage increases.
The ideal diode e q u a t i o n predicts an almost constant current, I , for a
reverse bias applied to the diode. The actual characteristic indicates a non-zero
slope of the reverse characteristic for voltages less than b r e a k d o w n . To analytically
d e t e r m i n e the reverse current, we assume a long-base diode and an applied reverse
voltage of about 4kT/q. The total reverse current is obtained from E q . (7.3) as
s

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

Limitations to Ideal Diode Theory

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 electron-hole 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 electron-hole pairs by hole
ionization is not shown.

Section 7.2

Real Diode Characteristics in the Reverse Direction

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 electron-hole 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

(7-7)

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

for 2 < < 6

(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

Limitations to Ideal Diode Theory


N N ,
A

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

For a constant critical field, we n o t e that the magnitude of the b r e a k d o w n


voltage is inversely proportional to the doping of the w e a k e r region of the junction
diode.
We will illustrate the calculation of the b r e a k d o w n voltage by example 7 . 1 ,
while example 7.2 is used to illustrate the d e p e n d e n c e of the multiplication factor
on the reverse voltage.

E X A M P L E 7.1
3

The doping densities of an abrupt-junction 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

where V. = V - V , = , e for silicon is 11.8 and = 8.85 x 10" F/cm.


We substitute the values so that,
bi

3 10

X 1.6 '" 10

1.04 10

12

V,

10
17

15

+ 8 10 )

Solving for V., we obtain


V \V ,~V \
r

= 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

Real Diode Characteristics in the Reverse Direction

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

From Eq. (7.8), is given as

=
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

The characteristic of avalanche and Zener breakdown compared.

194

Chapter 7

Limitations to Ideal Diode Theory

Figure 7.7

(a) Tunnelling through a barrier and (b) tunnelling across a junction.

The b r e a k d o w n in Z e n e r diodes is a result of the p h e n o m e n o n of tunnelling.


The illustration of tunnelling through a barrier and tunnelling in a diode are shown
in Fig. 7.7. (Refer to E x a m p l e 1.4.)
Tunnelling may take place w h e n electrons in the valence b a n d of the region
of a diode are opposite empty conduction b a n d energy levels in TV. The probability
of tunnelling increases as t h e width of t h e depletion layer is reduced. We observe
from E q . (5.28) that the width of the depletion layer decreases substantially as the
doping of the two regions is m a d e very large of the order of 1 0 c m ~ . With reverse
bias, t h e decreased width of t h e depletion layer results in a high electric field inten
sity.
It is possible to have the relevant energy levels aligned for any P N junction
diode by the application of the p r o p e r reverse bias, but tunnelling and Z e n e r break
down will not occur unless the depletion layer is r e d u c e d by high doping so as to
cause an electric field intensity in silicon of about 1 0 V / c m . It is worthwhile noting
that the actual separation, d, b e t w e e n the relevant energy levels in and is
smaller than the depletion layer width W. This effect is shown in Fig. 7.7(b).
Commercially available diodes have b r e a k d o w n voltages that range from o n e
volt to h u n d r e d s of volts. These diodes are fabricated to possess certain b r e a k d o w n
voltages and are very useful in voltage regulator applications. While they are all
k n o w n collectively as Zener diodes, silicon diodes with b r e a k d o w n voltages of 5
volts or less are normally of the Z e n e r type, the others are of the avalanche type.
18

Section 7.2

Real Diode Characteristics in the Reverse Direction

195

REVIEW QUESTIONS
Q7-1 Briefly explain the mechanism by which the recombination current, in a certain voltage
Q7-2
Q7-3
Q7-4
Q7-5

range, exceeds the current predicted by the ideal diode equation.


Briefly explain the effect of high-level injection on the diode current.
How does an increase of diode reverse bias increase the diode reverse current?
Explain the role of the avalanche process in breakdown.
Clearly explain the main requirements for Zener breakdown to occur.

HIGHLIGHTS

Actual diode I-V 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 high-level 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
E7-1

a) Determine the value of the recombination current in a PN silicon diode at


= 300K operating at a voltage of 0.3V given N = 10 cm- , N = 10 cm" ,
A = 10~ cm , W = 0.8 , and r = T = .
b) Calculate the diode current using the ideal equation
17

15

10

Ans: a) / = 2 X 10~ A, b) I = 6.52 10"


E7-2

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

E7-3

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

Limitations to Ideal Diode Theory

CAPACITANCES OF THE DIODE


A P N junction diode possesses two intrinsic capacitances. O n e of these, the transi
tion capacitance, or junction capacitance, has physical characteristics similar to those
of a parallel-plate capacitor. The expression that defines the transition capacitance
in terms of physical p a r a m e t e r s is identical to that of the parallel-plate capacitor.
The second capacitance, k n o w n as the storage capacitance, or diffusion capaci
tance, also possesses the charge-storing p r o p e r t y of a capacitor but has n o other
physical characteristics similar to t h e parallel-plate capacitor.
A simple parallel-plate capacitor has capacitance that depends only on the
physical dimensions and on the dielectric constant of the insulator separating the
plates. It is k n o w n as a linear capacitance because a plot of the charge Q on each
plate versus voltage is a straight line passing through the origin. The diode capaci
tances, on the other hand, are nonlinear, indicating that the capacitance is a function
of the applied voltage and it is given by

C-

(7.X2)

The transition capacitance exists w h e n a diode is reverse-biased 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

<

>

where V. = V - V and dV. = dV since V is a constant. The quantity of charge


stored in each side of the depletion layer is given by Q .
The magnitude of the charge in the depletion layer is obtained from E q . (5.19)
modified for a biased condition as
bi

bi

Q = AqN x
s

bi

= AqN x
A

(7.14)

Using x from E q . (5.27) in E q . (7.14) and replacing V


V), t h e expression for Q is written as
p

(V

bj

by V., which becomes

Section 7.3

Capacitances of the Diode

197

depletion layer (V ) a

- depletion layer
\(V + AV )
a

x=0

Q
+Q

electrons flow into


this layer to reduce
positive charge by AQ

-Qs

holes flow into


this layer to reduce
negative charge by AQ

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

The expression for the transition capacitance becomes


C,

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,

(7-18)

- MR

respectively. Voltage V. becomes

(7 19)

Using E q . (7.19) in E q . (7.17), we have


C,

(7.20)

H e n c e , the transition capacitance is a voltage-dependent capacitance (through


its d e p e n d e n c e on x ). Such a capacitor finds application in the electronic tuning cir
cuit of a television receiver.
n

198

Chapter 7

Limitations to Ideal Diode Theory

1 v {V)

Figure 7.9

Plot of l / C vs. V .
a

It is interesting to n o t e that, in contrast to a parallel-plate capacitor in which


the charges reside on the plates, the charges of transition capacitances arc located in
t h e depletion layer.
A plot of 1/C versus V using E q . (7.17) and shown in Fig. 7.9 yields a straight
as in test line having an intercept of V on the V' axis. The slope can be used to
d e t e r m i n e t h e d o p a n t density N .
E x a m p l e 7.3 illustrates the d e p e n d e n c e of the transition capacitance on the
applied voltage. We n o t e t h e larger capacitance at forward bias.
2

hi

E X A M P L E 7.3
17

The doping densities of an abrupt-junction silicon PN diode are N. = 10 atoms/cm and


N = 8 X 10 atoms/cm .The area of the junction is 2 X 10"'cm .
Calculate the junction capacitances at: a) zero bias, b) reverse bias of 6V, and c) forward bias
of 0.7 V.
,5

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

b) At V = -6V, V = 0.77 + 6 = 6.77V


j

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

Capacitances of the Diode

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 forward-biased.
;

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 space-charge 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(exp-l)^'

The excess hole charge Q


L , is given by

"

(7.21)

stored in t h e region d e t e r m i n e d for a large

= qA\
Jo

p'dx'

= ^ L ^ e x p ^ - l )

(7.22)

The storage capacitance C , also labeled the diffusion capacitance b e c o m e s


s

= 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 forward-biased diode has a larger value t h a n that of
a reverse-biased diode since the depletion layer is narrower. However, for a for
ward-biased 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

Limitations to Ideal Diode Theory

E X A M P L E 7.4
17

The doping densities of an abrupt-junction silicon diode are = 10 atoms/cm and N =


10 atoms/cirr .The cross sectional area of the diode is 2 x 10 cm and the lifetime of holes in the
region is 0.18. Given the diffusion constant for holes in is 16cm /s, calculate at room tempera
ture the storage capacitance at a) V' = 0.6V and b) V = 0.65 V.
A

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

The equilibrium density of holes in the region,p , is given by


Qn

10

20

The diffusion length for holes is


0 ,

f ,

L = (D , ) ' = (16 X 0.1 X 10 )


p

()

a 5

= 1-26 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)

The incremental resistance is found as

"

(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

Small-Signal Equivalent Circuit

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

and for V = 0.65V,


r=
d

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 small-signal 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 reverse-biased diode includes the

WW

(a)

Figure 7.10

(b)

Equivalent circuit of diode (a) reverse-biased and (b) forward-biased.

202

Limitations to Ideal Diode Theory

Chapter 7

junction capacitance C. and an incremental resistance whose expression is deter


mined from E q (7.26) as
(7

"= / " V

28)

Because the diode operates at relatively high voltages in t h e reverse direction,


the incremental resistance is very high and may often b e assumed to b e infinite.
In addition to t h e parallel combination of the capacitance and t h e incremental
resistance, a resistance is a d d e d in series with the combination. This is the bulk resis
tance, R of t h e neutral regions. It is calculated as t h e product of the resistivity and
the length of each region divided by the cross sectional area. It is evident that in
both the forward and reverse directions t h e resistivity changes as the current
changes. Thus, complicated relations are required to calculate this resistance.
The equivalent circuit in t h e forward direction includes b o t h the storage and
transition capacitances and the incremental resistance. This incremental resistance
is a very small quantity. O n the o t h e r hand, the effect of the series resistance, R ,
becomes quite i m p o r t a n t in the forward direction and mainly at high values of cur
rent in t h e high injection m o d e .
s

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 short-base 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 .
+

Referring to E q . (5.37(b), which was obtained by a series expansion of the


exponential and shifting t h e origin to x' = 0 (x = x ), the expression for t h e excess
hole density in b e c o m e s
(7.29)
p'(x)
exp

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 short-base 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

U p o n comparing E q . (7.30) for the short-base diode with that of t h e corre


sponding long-base diode (Eq. 7.25), we n o t e that the current in the short-base
diode exceeds that in t h e long-base diode by t h e factor
L /W .
p

Section 7.5

Properties of the Short-Base Diode

203

=0
Figure 7.11 Excess hole density distribution p'(x') in the base of a short-base
diode (W L ).

We will n o w d e t e r m i n e expressions for t h e storage capacitance and t h e incre


m e n t a l resistance of a short-base diode. T h e expression for t h e transition capaci
tance is t h e same as that given for t h e long-base diode.
T h e excess minority carrier charge stored in t h e region is given by t h e prod
uct of q and t h e area of t h e triangle, shown in Fig. 7.11, as

Qs = 9A(^J

XW ^

qApJ^

( e x p ^- l )

(7.31)

The storage capacitance becomes


dQ

_ q A W

n P o n

qV

The incremental resistance of t h e short-base diode is given by


( dl\_,
r

The p r o d u c t r C
d

<

W.

w ~

(7.33)

(^W^expf

becomes
rC
d

(7.34)

O n comparing t h e equations in t h e forward direction for t h e currents a n d for


the storage capacitances of t h e short a n d long-base diodes having t h e same p , A,
V , and N N , we have
0n

/(short) = ^ / ( l o n g )

C (short) = (Wj2L )
p

(a)

C (long)

(b)

(7.35)

204

Chapter 7

Limitations to Ideal Diode Theory

Therefore, the short-base diode has a higher current and a smaller storage
capacitance t h a n the long-base diode.
U p o n comparing E q . (7.23) for the storage capacitance of the long-base diode,
E q . (7.32) for t h e capacitance of the short-base diode, and the conditions of Eqs.
(7.35), we observe that t h e capacitance of the short-base diode is much smaller than
that of the long-base 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 short-base 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 short-base diode because the width of the base is
much smaller t h a n a diffusion length. We recall that t h e long-base 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 reverse-biased 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 low-current 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
steady-state voltage to t h e steady-state 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.
+

Turn-ON T i m e
To d e t e r m i n e the O N time, we consider a long-base diode, W L , in t h e circuit
of Fig. 7.12(a) that is in t h e reverse-biased 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

Diode Switching Characteristics

|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 steady-state value and
V is the instantaneous value of the diode voltage.
a

Since the diode current is changed instantaneously from I to 1 , the forward


voltage V becomes zero in a time of approximately 1 0 " s since it takes very little
time for electrons in the region and holes in the region to m o v e to neutralize
the d o n o r and acceptor ions. We will therefore begin counting time, t = 0, from
V = 0.
Because the forward current is constant I = (qAD )(dp/dx'),
the slope of the
hole density distribution at the side of the depletion layer is constant while the
hole density builds up, with time, inside the region as holes diffuse from to
and as shown in Fig. 7.12(b). The current supplies holes for b o t h the buildup of
excess holes and to feed the recombination with electrons. The buildup is termi
nated when the hole density at the side of the depletion layer reaches the value
that is d e t e r m i n e d by the voltage across the diode, which corresponds to the current
I as d e t e r m i n e d from the static I-V characteristic of the diode. A t that time, the
hole density has reached the value shown at t = t in Fig. 7.12(b).
Since the forward current delivers a certain n u m b e r of hole-coulombs per unit
time, and knowing the final accumulated n u m b e r of hole-coulombs, neglecting
recombination, then the t u r n - O N time is determined by the relation b e t w e e n the
current and the total of excess hole charge accumulated.
The equation, for the hole current density, E q . (5.41), at the edge of the deple
tion layer, labeled / = J and the equation for the excess stored hole charge,
E q . (7.22), are r e p e a t e d h e r e as
s

10

Qn

J = I /A
F

= qD

pjL

[exp (q VjkT)

- 1]

(a)

206

Chapter 7

Limitations to Ideal Diode Theory


Q

= qAL

[exp (q VjkT)

Qn

- 1]

(b)

(7.36)

By solving for Q in terms of I , we have


B

= (L /D )l =7 I

(7.37)

p F

T h e result in E q . (7.37) demonstrates that the smaller t h e a m o u n t of final


stored charge, t h e faster t h e device turns O N since I is fixed by t h e circuit. Based
on our earlier definition and in accordance with E q . (7.37), the t u r n - O N time is
therefore equal to the lifetime of holes in a long-base diode.
The relation b e t w e e n total stored charge and the current, in a short-base
diode is d e t e r m i n e d from Eqs. (7.30) and (7.31) to b e
F

= I W J2D
F

(7.38)

The factor W /2D


can b e shown to r e p r e s e n t t h e transit time of a hole in t h e
region. It is the time it takes a hole to travel from the edge of the depletion layer
in to the metal contact.
Therefore, the t u r n - O N time for a long-base P N diode is t h e lifetime of holes
in t h e region and t h e t u r n - O n time for a short-base P N diode is t h e transit time
of a hole in N.
p

Turn-OFF 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

A s long as t h e r e is a large n u m b e r of holes n e a r the edge of the depletion layer


in N, the reverse current will b e large and holes will diffuse in the direction from
to , with the hole density in at t h e edge of the depletion layer having a positive
slope (since I = I = qD A dp/dx). This process continues until t h e excess
hole density at t h e junction is r e d u c e d to zero, so that t h e hole density at that point
b e c o m e s equal to t h e t h e r m a l equilibrium value, causing V to b e zero. T h e time it
takes t h e excess hole density to b e c o m e zero (p = p ) is k n o w n as t h e storage time,
i^, as shown in Fig. 17.13. This is t h e time at t h e end of t h e constant current phase of
the reverse current transient. A s m o r e holes are transported back to the region by
diffusion, t h e hole density at = x drops below the equilibrium value, V becomes
negative and opposes V in o r d e r to reduce the magnitude of t h e current to b e less
t h a n I , as illustrated in the figures.
A s m o r e holes are m o v e d to N, t h e hole density at t h e junction b e c o m e s m u c h
smaller t h a n t h e equilibrium value, V b e c o m e s large and negative, and t h e current
D

0n

Section 7.6

Diode Switching Characteristics

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.

approaches -I T h e time for the current to fall from -I to 0 . 1 / is k n o w n as the


recovery time, t . T h e current 0 . 1 / is used as a reference in place of I . T h e total
turn-off time, t = t + t, is called the diode reverse recovery time.
A n approximate expression for the storage time, t , has b e e n obtained* for a
long-base diode as
y

t = n
s

1+

(7.40)
A

The recovery time, t , is, in effect, the time t aRk-I


e n to charge the junction capaci
tance to the reverse voltage, V . A n estimate of t is given as the time constant
23R Cj, w h e r e C. is the m e a n junction capacitance between zero reverse bias and
reverse bias at ~V .
The total turn-off time becomes
r

(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

Limitations to Ideal Diode Theory

We conclude that t h e storage time in a long-base diode is r e d u c e d by a large


reverse current and by a small carrier lifetime. For a short-base diode, t h e storage
time is r e d u c e d by a large reverse current and by a r e d u c e d region width. A large
reverse current sweeps t h e holes out faster and a short lifetime causes a fast reduc
tion of the carriers by rapid recombination. A small forward current causes a small
stored charge, thus causing a small-storage charge removal time. A small width, W,
reduces t h e storage time in a short-base diode by reducing t h e total hole charge
storage.
The lifetime of minority carriers can b e reduced by doping t h e semiconductor
with an e l e m e n t that introduces additional recombination centers. This is accom
plished by doping silicon with gold. A disadvantage of the reduction of minority car
rier lifetime in a long-base diode or the reduction of region width in a short-base
diode is the increased reverse saturation current of the diode.
We can n o w relate t h e switching times of a diode to t h e capacitances. To turn a
diode OFF, the excess minority carriers stored must be carried away, which in effect
causes a discharge of t h e storage capacitance and a charging of the transition capac
itance. The time t a k e n to turn a diode O N is essentially t h e time required to dis
charge the junction capacitance while the voltage across t h e diode reverts from a
large negative value to a small forward bias. This is accompanied by t h e charging of
the storage capacitance.

REVIEW QUESTIONS
Q7-6
Q7-7

Explain why a small capacitance is preferred when a diode is used as a switch.


Which diode capacitance is significant in the forward directions and which one is sig
nificant in the reverse direction?
Q7-8 Why and how can the diode transition capacitance be used in the tuning of a televi
sion receiver?
Q7-9 Of what significance is the equivalent circuit of the diode?
Q7-10 In accordance with Eq (7.40), a large I increases the storage time and a large I
decreases it. Explain the effect of each current.
P

HIGHLIGHTS

The capacitance of a semiconductor device is nonlinear in that the capacitor charge is


not directly proportional to the voltage, as is the case in a linear capacitor. A nonlinear
capacitance is defined as the ratio of the change in the charge to the change in the
voltage that produced it (dQ/dV).
Positive and negative charges are found in the space-charge layer of a diode. They con
stitute a transition (junction) capacitance, as the change in the width of the layer does
not vary directly with the applied voltage.
The storage capacitance, on the other hand, refers to the rate of change of charge stor
age of minority carriers as the voltage is changed.
A small-signal equivalent circuit is made up of circuit elements that represent the
diode when incremental changes are made in the applied voltage. The diode circuit is

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 turn-ON time is favored by a small lifetime
for a long-base diode and by a small transit time for a short-base diode.

EXERCISES
E7-4

A silicon abrupt-junction diode has a junction capacitance of lOpF at a reverse bias


of 8.8V and 20pF at a reverse bias od 1.6V. Determine the built-in voltage.
Ans: V,. = 0.8V
bl

E7-5

A P N diode has the following properties:


= 2 . 5 8 , W = 0.5am, and D = 10 cm /sec. Determine:
2

a)
b)

'

the transit time of the holes in the region.


the turn-ON time for this diode and the approximate turn-ON time if W =
.
Ans: a) 1.25 X 10- s
n

10

10

b)1.25 x 10" s,2.5 x 10" s

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 low-voltage 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

Limitations to Ideal Diode Theory

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

1/C = 7.48 X 10 (13 - 20V )


a

Determine, at = 300K,
(a)
the built-in 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

= 10 cm~ and A = 2X 10~ cm , determine:

the built-in voltage.

(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

the depletion layer thickness at breakdown.

7.8 An abrupt junction silicon PN diode, operating at = 300K, has N = 5 X 10 cm~ ,


7V = 1 0 c m , and A = 10" cm . Determine for a critical field of 3 X 10 V/cm
(a)
the applied voltage at which the low level injection assumption is violated.
(b)
the breakdown voltage.
(c)
the junction capacitance at breakdown.
7.9 Repeat problem 7.8 for N = 1 0 c n r and YV = 1 0 c n r .
7.10 For a P N junction diode, determine the effect of an increase of N on the following:
15

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

If the series resistance, R , of the neutral regions and metal contacts of a


diode is 300 ohms, determine the current at which the applied voltage devi
ates from that predicted by the ideal diode equation by 10 percent (given
I = 10~ A).
(b)
Repeat part (a) if R = 5 ohms.
7.12 An ideal silicon N P long-base diode has N = 10 cm~ , N = 1 0 c m , = =
10~ s, and A = 10~ cm . Determine, at = 300K:
(a)
the small-signal incremental resistance at a forward bias of 0.1,0.5, and 0.7V.
(b)
repeat part (a) for a reverse bias of 0.1,5, and 20V.
7.13 Use the parameters of the diode in problem 7.12 to calculate each of the junction
capacitances, C, and the diffusion capacitance C for:
(a)
forward bias of 0.1,0.5, and 0.7V.
(b)
reverse bias of 0.5 and 20V.
7.14 In a silicon PN junction, the doping density is linearly graded and symmetrical from
-W/2 to + W/2, where N N = Bx. Derive the expression for the junction capaci
tance showing that
s

14

18

c =\
A

16

13/

-3

chapter 8
BIPOLAR TRANSISTORS I:
CHARACTERISTICS AND
FIRST-ORDER 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

Structure and Basic Operation

base

collector

)
B

213

Figure 8.1 (a) Two-dimensional representation of the structure of a PNP BJT;


(b) a cross section of a discrete PNP BJT with approximate hole paths and current
direction indicated by dotted lines; and (c) a cross section of an IC NPN BJT.

214

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

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 emitter-base
junction is forward-biased and the collector-base junction is reverse-biased. 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 reverse-biased 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 transfer-resistor, 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 forward-biased
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

Fabrication of the Bipolar Integrated Circuit Transistor

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 3-10 ohm-cm 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 sub-collector 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 high-temperature
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 N-epitaxial
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 ohm-cm, with a thickness of 0.5 to 5 for high-speed digital circuit
applications and 10-20 for linear analog circuits. A layer of silicon dioxide about
5000 to , thick is grown thermally on t h e surface of the epi-layer.
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 P-isolation 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 reverse-biased 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 N-epitaxial 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

Steps in the fabrication of an NPN BJT.

Section 8.2

Fabrication of the Bipolar Integrated Circuit Transistor

217

///////////W//M //1kY//////////////////////////

TV

Figure 8.3

Transistor tubs.

Mask number three is used to o p e n a window for the P-type base of the transis
tor. P-type 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 metal-semiconductor 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 computer-controlled
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

Bipolar Transistors I: Characteristics and First-Order Model

(b)
Figure 8.4 Path taken by electrons from emitter to collector (a) without a buried
layer through N-epilayer 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 sub-collector. 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

Terminology, Symbols, and Regions of Operation

(a)

(b)

(a) PNP symbol and (b) NPN symbol.

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 emitter-base 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 forward-biased junctions, b o t h reverse-biased
junctions, and o n e forward-biased with a n o t h e r reverse-biased 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 emitter-base junction is forward-biased

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

Bipolar Transistors I: Characteristics and First-Order Model

and the collector-base junction is reverse-biased. For a P N P transistor in the active


region, V is positive and V is negative, while for a N P N transistor V is positive
and V
is negative. Bipolar transistors used in amplifier circuits are m a d e to oper
ate in this region, where t h e largest voltage and current amplification are obtained.
In the saturation mode, which corresponds to t h e first q u a d r a n t in Fig. 8.3,
both junctions are forward-biased. W h e n the transistor is to be used as a switching
device, this m o d e corresponds t o t h e O N condition of t h e switch because t h e ratio
of the voltage to the current is small, hence simulating a low resistance.
W h e n both junctions are reverse-biased, the transistor currents are so small
that this condition is referred to as cutoff and corresponds to the third q u a d r a n t of
Fig. 8.3. Since the voltages causing these very small currents are non-zero, the result
ing resistances are very large, indicating an open-circuit. A transistor operating as a
switch in t h e O F F position is in cutoff.
The fourth m o d e , known as the inverse-active mode, is the one least used and
corresponds to the second q u a d r a n t of Fig. 8.3. In this m o d e , the collector takes the
role of t h e emitter and the emitter becomes the collector. Certain digital circuits use
transistors that are m a d e to o p e r a t e in this mode.
The doping of t h e emitter is m u c h greater t h a n the doping of t h e base, which
in turn may b e greater than the doping of the collector. These differences in doping
are dictated by the requirements of the amplification properties of the transistor,
which take place w h e n the device is operating in the active m o d e . In a later section,
we will discuss the effect of the doping on the currents. Because of the differences in
both the cross sectional areas and doping, the collector and emitter are not inter
changeable. This interchange h a p p e n s only when the device is operating in the
inverse-active m o d e , where the device practically loses all its amplifying properties.
EB

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 common-base,
common-emitter,
and common-collector
con
nections. T h e most commonly used connection is the common-emitter type, with t h e
common-collector 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 two-dimensional 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

Transistor Currents in the Active Region

221

+
V

EB

"Si

EB

CB

(a)

(c)

(*>)

Figure 8.7 (a) Common-emitter, (b) common-base, and (c) common-collector 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)

For a B J T operating in the active m o d e , and, as shown in Fig. 8.3, V


is posi
tive and V
is negative, resulting in a forward-biased emitter-base junction and a
reverse-biased collector-base junction and V
= V ,
V
= ~V ,V
= V .
EB

CB

EB

BE

CB

BC

CE

EC

Emitter Current
For a forward-biased emitter-base 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

(8-2)

+ 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

Bipolar Transistors I: Characteristics and First-Order Model

Figure 8.8 (a) Two-dimensional 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 forward-biased emitter-base
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 reverse-biased 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

Transistor Currents in the Active Region

Emitter

Base

223

Collector

(b)

Figure 8.9 Energy-level 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 reverse-bias the electric field in the collector-base 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 collector-base 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

Bipolar Transistors I: Characteristics and First-Order Model

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 collector-base 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

(8-3)

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

where / , = I and it is directed into the base.

Figure 8.10

Current components in the active region.

(8-4)

Section 8.6

The BJT as a Current Amplifier

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 modern-day silicon BJTs, we will
d e m o n s t r a t e that a B J T in the common-emitter 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 reverse-biased collector-base 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

Base Current as t h e Control Current


A s we shall see later, t h e base current defined in this section as consisting of I only
is of the o r d e r of 1 percent of the collector current. In spite of its small size, its
i m p o r t a n c e lies in the fact that it is t h e control current supplied to the input of the
common-emitter BJT.
m

226

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

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 short-base 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)

= (qA D /W )cx {qV /kT)


E

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 emitter-junction, W is the effective width
of the base, and D is the diffusion constant of holes in the base.
E

The base current, I , calculated at t h e emitter side of the junction, is a diffu


sion current in an emitter whose width is assumed to b e m a n y times the diffusion
length of electrons, L . This current is given by
h = hi = (qA D )dn/dx'
= qAD n (0)/L
=
m

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"

(8-8)

Because I is directly related to I , i n d e p e n d e n t of the voltage V , a linear


relationship exists b e t w e e n I a n d I , and this m a k e s t h e B J T in t h e C E connection
a linear current amplifier.
Fixing l or V 7
c

EB

BE

F r o m our discussion, it is a p p a r e n t that either I or V could b e fixed to obtain the


current amplification. By fixing V , problems m a y arise, as illustrated in Fig.
8.11(a). Because of the exponential d e p e n d e n c e of I on V , attempting to fix I
through t h e fixing of V
by a voltage source represents a fragile situation, since a
small change in V
causes a large change in I . The voltage supply, V , is not the
problem. It is t h e shifting of the characteristic due to a change in t e m p e r a t u r e that is
of concern.
A superior arrangement, used to fix I , is to place a battery in a series with a
resistor across t h e input of the C E arrangement. By adjusting either the battery
voltage, V , or the resistor, R , a stable current, I , can b e fixed, as shown in Fig.
B

EB

EB

EB

EB

EB

EB

BB

Section 8.7

Transistor Parameters

Figure 8.11 (a) Illustrating the effect of shift in the characteristic I - V on /


and (b) a battery in a series with the resistor used to fix I . Note the effect of the
shift is minimized.
B

BE

227

8.11(b). The corresponding value of V


is d e t e r m i n e d at the intersection of the cir
cuit equation, I = (V
- V )/R ,
and the characteristic curve of t h e input of the
device.
EB

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

(8-9)
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 collector-to-base 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

Bipolar Transistors I: Characteristics and First-Order Model

We recall that I and /


En

are directed out of t h e base terminals and

is
CBO

REC

directed into t h e base a n d out of t h e collector terminal.


We define t h e emitter injection efficiency, , as t h e ratio of t h e hole current
injected from t h e emitter t o t h e base to t h e total emitter current defined by E q .
(8.9(a)).
J^f

(8-10)

We also define t h e base transport factor, , to be t h e ratio of t h e hole current


entering t h e collector t o t h e hole current entering t h e base from t h e emitter as
8 = *Ep
and this product is k n o w n as t h e DC

T h e product ~y8 equals I /I


base current gain, a, a n d given as
CP

= = 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

(8-13)

To introduce I in E q . (8.13), we replace I by (I


B

+ I ) and we solve for I , so


C

that
a

I CBO
= z
I + f ^
1 a
1 a

We define two n e w symbols, and l


I =VI
C

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 collector-to-emitter 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 reverse-biased C-B 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 forward-biased
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 common-emitter
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

Graphical Characteristics and Modes of Operation

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

b) The injection efficiency is calculated as


y = hplh.

= hJihj-

+Q

= /(

01

+ - ) = 0.995

c) a = 7 = 0.985
= /(1 - ) = 65.67
D

>

"

En

= 2.01mA,/ = 1.981mA,4 = 0-029mA

CBO

'CEO

c~

115

CBOIQ ~ " ) = 76.67

e) The base transport factor becomes = 1.99/2 = 0.995


a = y = 0.995 x 0.99502 - 0.990
= /(1 - ) = 99.49

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 common-base 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)

If either the C E or CB relations were available as complete expressions, then


by using Kirchhoff s current and voltage laws, one can determine expressions for
each of the three transistor currents as a function of any two of t h e voltages so that
either of the above relations completely define the I-V characteristics of the transis
tor. It is i m p o r t a n t to point out that, n o matter which connection of t h e B J T is used,

230

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

Cutoff

CB0

Figure 8.12 (a) Common-emitter circuit and output characteristics of the PNP transistor
(b) Common-base circuit and output characteristics of the PNP transistor.

and which variables are t h e d e p e n d e n t and i n d e p e n d e n t quantities, the I-V charac


teristics of the transistor are uniquely determined. Analogously, either connection
of the B J T represents t h e I-V characteristics completely.
M o d e s of O p e r a t i o n
B o t h t h e C E and CB characteristics exhibit the t h r e e major m o d e s of operation,
namely: active, saturation, and cutoff. In certain switching circuits, the B J T in the C E
connection is in the inverse active m o d e shown in the third q u a d r a n t of those char
acteristics of Fig. 8.12(a).

Section 8.8

Graphical Characteristics and Modes of Operation


V

231

CB

Inverse active

Saturation

Cutoff

Active

Figure 8.13 Distributions of minority carriers in the four operation modes of the
PNP BJT.

In analyzing the diode currents in C h a p t e r 6, we used the distribution of


minority carriers in t h e and regions. In Fig. 8.13, we have displayed t h e distribu
tions of minority carriers in each of t h e four o p e r a t i o n m o d e s of the BJT. These will
b e used in discussing t h e characteristics of t h e transistor. In sketching t h e minority
carriers distributions, we have assumed that t h e width of t h e base is m u c h smaller
t h a n the diffusion length of holes in t h e base and that t h e widths of t h e emitter and
collector regions are m u c h greater than t h e diffusion length of electrons. Just as with
the diode, the values of t h e minority carriers densities at t h e junctions are given by
In the base at = 0,p(0) = p exp (qV /kT)
0

EB

(a)
(8.19)

In the base at = W ,p(W )

= p exp (qV /kT)

In the emitter at x' = 0, n (0)

= n

CB

0 E

(b)

exp (qV /kT)

(a)

exp (qV /kT)

(b)

EB

(8.20)
In the collector at x" = 0, n ( 0 ) = n
c

QC

CB

Based o n t h e above, we can d e t e r m i n e expressions for the various currents by


assuming that the B J T currents may b e obtained from t h e diffusion of the minority
carriers.
CE A c t i v e M o d e
With I fixed, V
fixed, p(0) fixed, and assuming that the distribution of hole den
sity in the base is linear, as shown in Fig. 8.13, t h e hole current in the base is equal to
the hole current in t h e collector (since t h e r e is n o recombination in t h e base) and
B

EB

232

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

this is equal to the collector current, l . A s long as t h e E - B junction is forwardbiased and t h e C-B 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)

The active m o d e is terminated when V


= V a n d hence V
= 0. With V
< 0, t h e C-B junction is forward-biased and t h e device is in t h e saturation mode.
Thus, V
= 0 is t h e dividing line b e t w e e n active a n d saturation modes.
EC

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 C-B junction is forward-biased 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 C-B 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 reverse-biased, 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 C-B junction is reverse-biased, 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 emitter-base junction is, therefore,
slightly forward-biased, 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 emitter-base junction is reverse-biased
and t h e collector-base junction is forward-biased. 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

Graphical Characteristics and Modes of Operation

233

shape of t h e characteristic in this m o d e , shown in the third q u a d r a n t of Fig. 8.12(a),


is identical to that in t h e first quadrant, except that t h e collector current for the
same base current (as in the active m o d e ) is m u c h smaller than that of the first
quadrant. T h e main r e a s o n for the smaller collector current is that the doping of the
new emitter is considerably smaller than the former emitter. This is evident from
Eq. (8.8), r e p e a t e d h e r e by replacing the p a r a m e t e r s of the emitter by those of the
collector, as
I /I
C

(inverse active) =

AED poL
AD n W
p

nC

0C

(8.22)

nC

U p o n comparing E q . (8.22) to E q . (8.8), we n o t e that n


n
because of
the doping of t h e emitter N
N . F u r t h e r m o r e , A A . This m a k e s I /I
quite small and slightly greater than unity.
QC

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 forward-biased 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 C-B 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 C-B 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 open-circuited and V
> 0, I becomes the reverse
saturation current of the C-B 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

Bipolar Transistors I: Characteristics and First-Order Model

junction is zero. This reduces the hole gradient at = W , which b e c o m e s smaller


than w h e n the E-B junction is short-circuited with V
= 0.
B

EB

REVIEW QUESTIONS
Q8-1 Identify the junction voltages and their signs in each of the four modes of operation for
an NPN transistor.
Q8-2 Identify the components of each of the three terminal currents for an NPN transistor.
Q8-3 What is meant by emitter efficiency?
Q8-4 What is meant by base transport factor?
Q8-S State, in equation form, the relationship between the three voltages for a PNP transis
tor.
Q8-6 Why is the collector current in a PNP transistor in the inverse active region much
smaller than in the active region?
Q8-7 Why does the collector current decrease when the transistor is operated in the satura
tion region?
Q8-8 What is the function of the buried layer?
Q8-9 Why are the devices placed in tubs?

HIGHLIGHTS

The transistor is labeled a transresistance because in the normal or active mode of


operation, the emitter-base junction is forward-biased at a low E-B voltage, resulting
in a small input resistance. The collector-base junction is reverse-biased at a high C-B
voltage, resulting in a large resistance.

Transistor output characteristics are displayed in the common-emitter connection with


I vs. \ V \ at various values of I , or in common-base connection with 7 vs. \ V \ at
various values of I . The same information is contained in both.
When used in an amplifier circuit, the BJT is operated in the active mode. When used
in a switching circuit, the transistor is switched between cutoff and saturation.
An important property of a transistor is that in normal operation, small changes in the
base-emitter voltage cause large changes in the collector current.
The alpha of a transistor in the forward active region is the ratio of the collector to the
emitter current, while beta is the ratio of the collector to the base current.
c

CE

CB

EXERCISES
E8-1 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

Ans: a) 0.996, b) 0.99985, c) 0.99585, d) 240


E8-2 Sketch the energy band diagram for the electrons in an NPN transistor for (a) thermal
equilibrium and (b) saturation.

Section 8.9

Analytical Relations for the Currents

235

8.9 A N A L Y T I C A L R E L A T I O N S FOR THE C U R R E N T S


A s s u m p t i o n s a n d Procedure
In t h e preceding sections, we have described t h e operation of t h e bipolar junction
transistor and d e t e r m i n e d certain relations basic to its operation in the active mode.
In this section, we will derive complete expressions for t h e currents subject to the
following assumptions:

Low-level injection.

T h e electric field intensity in t h e bulk regions outside t h e depletion regions is


so small that the drift current of minority carriers is neglected.

N o recombination and generation in the depletion regions.

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.

E a c h of the three bulk regions is assumed to b e uniformly d o p e d and b o t h


junctions are considered to be step junctions so that the change in impurity
density, from one region to another, is abrupt.

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
ter-base depletion layer.
We start with the continuity equation for holes in the steady-state 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

Bipolar Transistors I: Characteristics and First-Order Model


Depletion regions

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

a n y w h e r e in t h e base. Subject to t h e assumptions we m a d e , t h e hole current in t h e


base is a diffusion c u r r e n t so t h a t we can rewrite E q . (8.23) as

We define t h e excess hole density p' as (p - p ) a n d use L


E q . (8.24) b e c o m e s
Q

d p'
2 dx
2

2
L

= / ,

so t h a t

(8-25)

T h e general solution t o E q . (8.25) can be written as


xjL

p> = B e
1

x/L

+ B e2

(8.26)

T h e low-level 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 )

Analytical Relations for t h e Currents

[exp (qVjkT)

P o

- 1]

(b)

237
(8.27)

Applying the b o u n d a r y conditions to E q . (8.26),


the expression
8.26), we determine
deti
for the hole density = p' + p
Q

=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

w h e r e W is the base width.


The portion of the emitter current due to holes, 1
B

is found from

(~qD

Ep

A)(dp/dx)

at = 0

so that

l

Ep

[ ^ (qV /kT)

- l][coth

EB

- [exp (qV /kT)

(W /L)]
B

(8.29)

- l]/[sinh ( W / L ) ] j

CB

f l

Now we will determine an expression for the electron current c o m p o n e n t of


the emitter current. The distribution of excess electron density in the emitter region
n' is determined from an equation similar to E q . (8.25), written as
E

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

The solution to E q . (8.30) is of the form

n' = ()

e x

(8.32)

(-x'/L )
nE

The electron current c o m p o n e n t of the emitter current is


q

( a t %

'

0 )

[exp (qV /kT)

- l]

(8.33)

[exp (qV /kT)

- 1]

EB

^nE

The total emitter current is


=
.

+
hp

h
Ln

= [(q A 7 ) / L ) ( c o t h (W /L ))
A

+ (qA D njL )}
nE

[q A D ^ / ^ sinh (W /L ))][ xp(qV /kT)


B

CB

nE

- 1]

EB

(8.34)

238

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

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 C-B
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)

To d e t e r m i n e I , we will use t h e distribution of excess electron density in t h e


collector region, n' , in an equation similar t o E q . (8.30), as
CN

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' = n' (0")


c

= n

[exp(q V /kf)

- l ] at x" = 0

CB

(a)

n^ = 0 a t x " = OO

(8.37)

(b)

The solution t o E q . (8.36) is


n' = < ( 0 ) e x p {-x"/L )
c

(8.38)

nC

The electron c o m p o n e n t of t h e collector current is


dn'
= qAD ^( tx"

Cn

nC

= 0) =

oc

L nC

[exp (q V /kf)
CB

- 1] (8.39)

The s u m of t h e current in Eqs. (8.35) and (8.39) forms t h e collector current


qA
L

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)

The base current, / , is


I

'c

(8-41)

By substituting Eqs. (8.34) a n d (8.40) in E q . (8.41), we obtain t h e expression


for t h e base current. This expression includes three terms: t h e electron current
across t h e emitter junction, t h e electron current across t h e collector junction, and a

Section 8.9

Figure 8.15

Analytical Relations for the Currents

239

Transistor currents and their components. Arrows indicate actual directions.

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

Bipolar Transistors I: Characteristics and First-Order Model

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 /V-sec,
= 21.4cm /sec,
= 4.62 x 10~ cm,
2

, = 447cm /V-sec,
D - 11.6cm /sec,
L = 3.4 X 10^ cm,
2

|.t. = 1330cm /V-sec


7> = 34.5cm /sec
L = 5.87 X 10~ cm
(

nC

Section 8.9

241

Analytical Relations for t h e Currents

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 ;

< i ^ < " n r

| ( )

, ,

= 0.0294117:
sinh

(Wg/Lp = 0.0294159, coth (W /L )


B

= 34.0099

10

For F = 0.63 V, exp (qV /kT)


= 3.66 10 . Active mode operation requires V
ative .mil its magnitude lo be much greater lli.m A / q.
i j

EB

cg

a) Using Eqs. (8.29) and (8.33) for l

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

h n = 'r, - 'cp = 0-293 X 10 "A


I = l = 9.4 10~
P

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 collector-base 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 /V-s, D

tlE

= 26.3cm /s

242

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

nE

= 5.13 X l O ^ c m a n d

16

= 1.6405 X 10" A

By repeating the calculations of Example 8.2 we have


f = 0.684mA
r

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

Solution The change in W causes changes in the hyperbolic functions of W /L


W /L is now 0.0588235.
B

The value of

We evaluate the hyperbolic functions as


sinh (W /L )
B

= 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

Analytical Relations for the Currents

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^f-1
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

Bipolar Transistors I: Characteristics and First-Order Model

E X A M P L E 8.5
,7

16

An NPN silicon transistor has the following properties:


= 10 cm- , N = 10 cm~ ,
N
10 cmT ,yl = 10~ cm , and W = . Assume that the minority carriers' lifetime in each
of the three regions is = KT . Use nf = 10 .
For active mode operation use V = 0.63V and \ V \ V , calculate:

15

AS

DC

20

BE

b) / . , /
C

Cp)

and/

HC

BE

d)
20

17

Solution For minority carriers: p


= 10 /10 , n = ^/
are calculated from Eqs. (4.15) and (4.16) D = 1/? and L =
OE

Emitter
3

cmcm /V-s
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~

I.,. - I 57. sinh (W,, I.J - (1.(117544759. colli f " '-) = 57

\\

a) I = 1908.47,/ = 1.76438 x 10 A . / = 910.2343


b) I = 1908.358./ = 5.523 ^ , ^ = 1908.35
c) l = 1.8639 10~
En

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
mon-emitter 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

Analytical Relations for the Currents

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)

A perfectly linear distribution of hole density in the base m a k e s I = I , hence


zero recombination current. T h e constant slope m a k e s the second derivative of the
excess hole density distribution zero, which from E q . (8.25) implies that either p' is
zero (p = p ), which is obviously wrong, or that L is infinite, which would result
from an infinite lifetime, which is also impossible.
In fact, by making W L , the hole density distribution approaches, but
never becomes, a straight line, for there will always be some recombination.
E

Cp

246

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

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 )

We further simplify t h e above expressions by assuming that, in a good transis


tor, (W /L ) 1 a n d hence cosh (W /L ) is close to unity. Since V
kT/q, the
unity in E q . (8.48) a n d t h e cosh t e r m in E q . (8.50) b e c o m e negligible and Eqs. (8.48)
and (8.50) m a y b e written w i t h p ( 0 ) r e p l a c i n g p exp (q V /kT),
as
B

EB

= (qAD p(0)/L )

Ep

coth (W /L )

(8.51)

sinh W /L )

(8.52)

= q A D p(Q)/(L

Cp

EB

The transport factor, , becomes

^ V^

S e c h

W P)

A further simplification is m a d e by expanding sech (W /L )


W /L
1 so that t h e transport factor b e c o m e s
B

a n d assuming

2
ing W

We conclude again that a transport factor of almost unity is achieved by mak


L . This is consistent with o u r earlier conclusion.
The injection efficiency has b e e n defined as

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

Analytical Relations for the Currents

247

(8.55)
DL N
p

nE

AE

The injection efficiency is increased mainly by a high emitter doping relative


to t h e doping of the base.
The c o m m o n - b a s e current gain, a, is the product of the transport factor and
the injection efficiency, and the common-emitter current gain, , is a / ( l - a ) . For
active m o d e operation, we label alpha as o^, and beta as ^ with the subscript refer
ring to forward. This is in contrast to the special region where the roles of the emit
ter a n d collector a r e interchanged and labeled inverse-active m o d e . For the inverse
active m o d e , alpha and beta will b e labeled a and ^.
We will n o w use the expressions in Eqs. (8.53) and (8.55) to determine an
expression for ^, defined as a /1 a , w h e r e a - -.
The expression for ^ becomes
R

_L

- f

nEW N

2\L j

DB

1_/WV|2D 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

For W = and /.,, = 34. )

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 common-emitter 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 integrated-circuit, 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

Bipolar Transistors I: Characteristics and First-Order Model

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

We n o w define the following:


D

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)

Ebers-Moll Model

Section 8.10

249

We n o t e that t h e expression for a in E q . (8.58(c)) is identical to the expres


sion found by taking the product of Eqs. (8.53) and (8.54) and letting W /L
1 so
Uu&\iuALW /L
=
W /L .
T h e general expressions for I , I , and I for a bipolar junction transistor using
Eqs. (8.57) and (8.58) b e c o m e
p

B=

CT

-I

E ~ C

EF ~ R

I =a I
c

()
(b)

C "

p)

+ hR

"

(8.59)

R)

()

These equations are k n o w n as t h e Ebers-Moll equations for t h e P N P transis


tor. E q u a t i o n s (8.59) also apply to t h e N P N transistor except that appropriate mod
ifications are m a d e to t h e terms defined in Eqs. (8.58). Recall that these equations
are valid for the currents' reference directions a d o p t e d (which are actual directions)
for a device operating in the active mode.
If we d e n o t e t h e prefactor of t h e exponential in E q . (8.58(a)) by I and the
similar t e r m of E q . (8.58(b)) by I , we obtain the following equality
ES

cs

ES F

= CS R

where I
=
D^jWg.
The Ebers-Moll equations for I
and (8.59) as

= SM

6 0

SM

ex

h = i s[ P

(q E /
V

and I

i] -

can t h e n b e written from Eqs. (8.58)

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 common-emitter
short-circuit (B to C) current gain and a as t h e common-base short-circuit 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
Q8-10 Identify the currents and voltages in the common-base and common-emitter connec
tions.
Q8-11 Why is the base width an important dimension in the quality of a transistor?

250

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

ES

~ ES

exp

I e of diode
r

CR

- cs exp (

^Tf^

l)

Figure 8.19 The Ebers-Moll model for the PNP transistor. Current directions are actual.
Q8-12
Q8-13
Q8-14
Q8-15

Write the Ebers-Moll equations for an NPN transistor.


Identify three physical parameters that contribute to a high beta transistor.
Use the Ebers-Moll equations to define a and ^,.
Use the Ebers-Moll equations to define a and .
p

HIGHLIGHTS

Subject to the assumption that there is no generation or recombination in the deple


tion regions, analytical relations for the components of the three currents of a BJT are
derived using diffusion currents expressions.
The above is accomplished by solving the continuity equation for the minority carriers
in each of the emitter, base, and collector subject to the boundary conditions in each
region.

The boundary conditions in the emitter and collector are functions of the emitter-base
and collector-base 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 emitter-base and collector-base 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 emitter-base and collector-base
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 Ebers-Moll relations are expressions for the emitter and collector currents as
functions of the emitter-base and collector-base voltages and in terms of a , a , I ,
and/ .
The Ebers-Moll relations are applicable in all modes of operation of the device.
En

Cn

En

Cn

REC

c s

EXERCISES
E8-3 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

Ebers-Moll 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

Ans: a) 0.99283, c) 1.61 X 10" A, d) 0.48, f) 3.3 X 10" A


13

13

To emphasize their generality, we will use t h e Ebers-Moll equations t o derive


Eqs. (8.13) and (8.15). In t h e active region of operation, V
is negative and
|V* | kT/q. Using this, solving for t h e first t e r m of E q . (8.61)(a) and replacing
for it in E q . (8.61)(b), we have
CB

CB

I =

0LpI

+I

cs

(1 - )

+ I

(8.63)

CBO

where I
has b e e n defined as t h e collector-to-base current with t h e emitter opencircuited (I = 0). E q u a t i o n (8.63) represents t h e common-base 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 common-emitter
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 collector-to-emitter current with t h e base opencircuited.
E q u a t i o n (8.64) represents t h e common-emitter 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

are defined by the prefactors in Eqs. (8.58)(a) and (b):

I=
m

cs

pPo

pPo

D n /L ]
nE

{)E

nE

D njL ]

nC

nC

Using the values of Example 8.2, we obtain


l5

= 18.6341 X 1 0 ~ A and l

ES

i5

c s

= 27.9637 X 1 0 ~ A .

The values for a and a are calculated to be


F

a, - 0.996 and a = 0.663.


The corresponding values of

and are

252

Chapter 8

Bipolar Transistors I: Characteristics and First-Order Model

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 Ebers-Moll 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

For the saturation region use V = 0.63V and V

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 Ebers-Moll 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

the recombination component of the base current

8.3 Determine for Problem 8.2.


8.4 A silicon NPN BJT has N
= 10 cm- , N = 10 cm- , and N
= 10 cm- .
Assume: i) the base width is much smaller than the diffusion length of minority carri
ers in the base, ii) the widths of the emitter and collector are much greater than the
diffusion lengths of minority carriers in the respective regions.

18

16

DE

a)

15

AB

DC

Sketch the minority carrier distributions in the base for saturation and cutoff
and identify, by cross-hatching, 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

Bipolar Transistors I: Characteristics and First-Order Model

8.7 For the device of Problem 8.6, given i


Determine:

= 10 s,

nE

_7

= 10 s, and T

_6

N C

= 10 s.

hp

hn

C)

hn

8.8 For the device of Problem 8.7, determine:

b)
c)
d)

the emitter efficiency


the transport factor
I

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

the component of I due to carriers crossing from emitter to base.


the charge density stored in the base as a function of the collector current
density.
Assume injection efficiency = 1 and I = 0.
E

8.10 The Ebers-Moll 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

8.13 A silicon NPN BJT has N


= 5x 10 cm- , N
j = . Determine
at = 300K for
17

DE

16

= 10 cm" , W = 5, and t

AB

pE

p C

a)

8.14 A silicon NPN BJT has N


= 10 cm- and = . Determine W so that the
transport factor is 0.995 at = 300K. Assume ( W / L ) 1.
8.15 The expression for E
of a BJT in saturation is given by
16

AB

C S A T

I C SATl

1 -

/ ,

where = I /I .
Use the Ebers-Moll 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)

carrier distribution in the base and in the new emitter,


Using the distribution of minority carriers, briefly
a

< < :

explain

why

8.21 A diode is formed in an integrated circuit by placing a short-circuit from collector to


base of a transistor. For such a diode:
a)
Sketch the minority carrier density in the base.
b)
Based on storage time considerations, briefly explain why such a connection
is preferable to other diode connections made from transistors.
8.22 A diode on an IC can also be formed between the base and emitter of a transistor
while the collector is open-circuited.
a)
Sketch the minority carrier density in the base.
b)
Using the carrier distribution and carriers' motions, explain how the collec
tor current becomes zero.
8.23 Use the Ebers-Moll equations to derive expressions for V and V of a PNP transis
tor in terms of I , I , I , $ , and .
EB

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 current-voltage 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 collector-base junction.
2. Multiplication of carriers in the collector-base 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

Effects of Limitations on Static Characteristics

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 small-signal 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 small-signal 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 gain-bandwidth 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 .

9.1 EFFECTS O F LIMITATIONS O N STATIC C H A R A C T E R I S T I C S


Experimentally observed characteristics for a P N P transistor in the common-base
and c o m m o n - e m i t t e r o u t p u t characteristics a r e shown in Fig. 9.1.
A comparison of t h e above characteristics with those shown in Fig. 8.12
reveals t h e following differences:

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 emitter-tocollector 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

A t voltages preceding b r e a k d o w n , t h e collector current increases rapidly with


t h e increase of t h e reverse bias, indicating a multiplicative effect.

A t a certain collector-to-base voltage, in t h e common-base characteristics,


and, at a corresponding emitter-collector voltage in t h e common-emitter
characteristics, b r e a k d o w n occurs. It is worth noting, from Fig. 9.1, t h e large
difference b e t w e e n t h e b r e a k d o w n voltage in t h e C-B connection,
BV ,
and that in t h e C-E connection,
BV .
CBO

CEQ

We will n o w explain t h e differences b e t w e e n t h e ideal and actual characteris


tics by relating these differences to t h e p h e n o m e n a listed in t h e introductory
section.

Increase of Collector Current w i t h V


Region

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 base-collector 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 emitter-base
depletion layer is fixed by V . T h e minority carrier density at the base side of the
collector-base 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 common-emitter 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

Effects of Limitations on Static Characteristics

259

Metallurgical boundaries

W (1V)-

Depletion layer
due to V

Depletion layer
forV =lV

5 C

W (30Y)

EB

Depletion layer

'W (effective) for V


B

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

In conjunction with the intersection of t h e constant I curves at V , as


shown in Fig. 9.3, we also observe an increase in t h e slope of the constant I curves
with an increase of I . To explain this change, we will derive an expression for the
conductance dI /dV
at a constant V . Since the base current, in most m o d e r n
B

EC

EB

260

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

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

By using E q . (9.1), we find


dI /dW
c

= -qAD p{0)/W

-I /W
C

E q u a t i o n (9.2) can then be written as


dl I
c

dV

= (-I /W )dW /dV

EC

(9.4)

EC

Since V is constant, o n e can assume that dW /dV


= dW /dV
because it
is V
that causes t h e change in t h e width of t h e base. A n increase in V
decreases
the base width, thus making dW /dV
< 0 and t h e dl / dV
positive.
The student m a y w o n d e r as to what would b e the effect of increasing V
on
the base width and the collector current. A n increase of V decreases the depletion
layer width at the E - B junction, thus increasing the width of the base while it
increases the hole density in t h e base at t h e junction. T h e increase of t h e base width
d u e to V , for practical values of V , is negligible, whereas the increase of hole
density at t h e base side of the junction and the electron density at the emitter side of
the junction lead to an increase of the collector and base current respectively.
Since V
is constant and I is constant, E q . (9.4) predicts an increase of I
with V
in the active region, for a constant I . It also predicts that at higher values
of V , thus higher values of I and I , t h e slope of the I V
curves is larger.
This p h e n o m e n o n of an increase in the collector current with an increase of
V , as shown in Fig. 9.3, was first described by Early and it is k n o w n as the Early
effect. This effect is also k n o w n as base-width modulation and V is the base width
modulation factor or the Early voltage. Extrapolation of the characteristics of Fig.
9.3 back intersects the V
axis at V , where V = I /(dI /dV ).
Substitution of
E q . (9.4) in the expression for V gives
EB

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)

Referring to the Ebers-Moll equation for I , assuming operation in the active


region so that the second t e r m is neglected, and replacing apI by I , E q . (8.61(b))
is modified to include the Early effect. Thus, for a P N P transistor
c

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.

Effects of Limitations on Static Characteristics

Section 9.1

261

Carrier Multiplication a n d B r e a k d o w n
A t t h e collector-base 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 collector-to-base
with t h e third subscript indicating that t h e third terminal, t h e emitter, is open-cir
cuited. O n t h e common-emitter characteristics, b r e a k d o w n occurs at a smaller volt
age, labeled BV ,
referring to t h e collector-to-emitter b r e a k d o w n voltage with
the o p e n base.
A t collector-to-base 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 electron-hole 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 reverse-biased 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
base-collector 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

The multiplication factor has b e e n d e t e r m i n e d empirically and referred to


in C h a p t e r 7. It is modified for a P N P B J T t o be
M=l/[l-(V /BV y)
BC

(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 collector-base voltage with t h e emitter open.
Consequently, with I = 0, E q . (9.7) becomes
CB0

I = MI
C

(9.9)

CB0

T h e current approaches a very high value as V


> BV'
and > . T h e
transistor current can b e limited, even at b r e a k d o w n , by inserting a resistor in
series in t h e external collector-base circuit so that t h e power dissipated in t h e B J T
is below t h e r a t e d power. B r e a k d o w n in t h e common-base connection is shown in
Fig. 9.4(a).
The collector current with I = 0 (also labeled 1^) is considerably smaller
than t h e reverse current of a P N junction diode having properties that are identical
to those of t h e collector-base region. T h e reason for this is that w h e n I is zero, t h e
hole density gradient at t h e edge of t h e depletion layer in t h e base is zero, thus
reducing t h e slope of t h e hole density slope at = W . O n t h e o t h e r hand, in inte
grated circuits, diodes are m a d e by shorting t h e base to t h e emitter of a B J T ( V =
BC

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 common-emitter 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

Effects of Limitations on Static Characteristics

263

C o m m o n emitter characteristics are shown in Fig. 9.4(b). By replacing I in E q . (9.7)


by (I + I ) and solving for I , we obtain
E

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)

= 0, the collector current becomes

I = MI /{i-Ma )
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)

N e a r b r e a k d o w n , the collector current for n o n z e r o values of I , calculated at


the same V , is larger as I is increased. A t values of V away from b r e a k d o w n ,
where = 1, E q . (9.11) indicates that t h e collector to emitter current with t h e base
o p e n is I , which is considerably larger than I , a n d is given by I /(l
a ).
Relative magnitudes of BV
and I
for the common-base connection and the
corresponding quantities for t h e common-emitter connection are shown in Fig. 9.5.
B

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

Bipolar Transistors II: Limitations, Switching, and Models

F r o m E q . (9.13) and Fig. 9.5, we n o t e that the b r e a k d o w n voltage for a B J T in


the c o m m o n - e m i t t e r connection is m u c h smaller than that in the common-base con
nection. We have plotted in Fig. 9.4 the common-base and common-emitter charac
teristics for an N P N B J T operating in the active region. We assumed that the B J T
has = 79, V = 50V, = 4, and BV
= 80V The b r e a k d o w n voltage in the
c o m m o n - e m i t t e r characteristics, BV , is calculated to occur at V
= 26.75V The
mechanism responsible for this difference results from the current-amplifying p r o p
erty of the c o m m o n - e m i t t e r connection. A s V b e c o m e s large and multiplication of
carriers takes place at the collector junction, a hole by the avalanche process pro
duces an electron-hole pair with t h e primary and secondary holes moving into the
collector. The electron is swept into the base by the electric field at the junction. To
preserve charge neutrality, the excess electrons swept into the base invite excess
hole injection from the emitter. The new holes diffuse to the collector junction,
resulting in an increase of the collector current, thus causing a regenerative process.
It is i m p o r t a n t to realize that the change in the slope of the common-emitter
characteristics of Fig. 9.4(b) is a consequence of two p h e n o m e n a . A t voltages, V ,
of about 10V in Fig. 9.4(b), the increase in t h e slope is a result of the Early effect.
Beyond that, the collector current increases due mainly to the multiplication of car
riers, at the reverse-biased collector junction, as b r e a k d o w n is approached.
The following example illustrates the calculation of b r e a k d o w n voltages.

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

where V, = V - V and for |vj V V. = \VJ = V,


Since % = ^ , we solve the above equation for V and since N
u

m a x

V, =

& e/2qN

9
BV
b) BV

CBO

CEO

= BV ,/(lW >
CHf

br

10
L

10

L F J

11.8 X 8.85 X 10~ _ .


3
i s - -73V
14

-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 collector-base junction
extends, because of t h e reverse bias, so far into the base that it reaches the emitter-

Section 9.2

Effects at Very Low and High Injection

265

base junction before avalanche b r e a k d o w n can occur. This p h e n o m e n o n is k n o w n


as punchthrough.
A s soon as t h e collector-base depletion layer reaches t h e emitterbase depletion layer, t h e emitter and collector regions of a P N P transistor are
joined by o n e depletion layer providing a highly conductive p a t h from emitter to
collector, which can lead to currents that d a m a g e the BJT. A t p u n c h t h r o u g h , the
base is depleted of mobile carriers and loses its controlling action on the collector
with a corresponding reduction in barrier height for hole injection from the emitter,
as shown in Fig. 9.6. The current increases rapidly and is limited by the external cir
cuit resistance and whatever small resistances are offered by the emitter and collec
tor bulk materials. In most transistors, avalanche b r e a k d o w n precedes
p u n c h t h r o u g h . However, p u n c h t h r o u g h m a y precede avalanche b r e a k d o w n if the
base has very small width or relatively low doping.
9.2 EFFECTS AT V E R Y L O W A N D H I G H I N J E C T I O N
T h e curves illustrating t h e changes in I and I , over a b r o a d range of changes in
V , as shown in Figs. 9.7 and 9.8, can b e divided into three regions.
In the m o d e r a t e collector current level (ideal region) of the curves, b o t h I
and I are diffusion currents and vary with a slope of exp (q V /kT).
A t lower val
ues of V , the base current is d o m i n a t e d by the c o m p o n e n t that accounts for
recombination in t h e emitter-base depletion layer and which is larger t h a n the diffu
sion c o m p o n e n t .
In the third region, which corresponds to values of V
that are greater than
those of the ideal region, two effects cause a reduction in the current gain ^.. O n e of
these effects occurs at the emitter junction and t h e other is at the collector junction.
A t t h e emitter junction and at high values of I , conductivity modulation of the base
B

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

Bipolar Transistors II: Limitations, Switching, and Models

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 collector-base 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 forward-biased emitter-base 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 emitter-base 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

Effects at Very Low and High Injection

267

A t m o d e r a t e current levels for operation in t h e active m o d e (V

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

A t low current levels, the expression for I is unchanged, but I increases to T


d u e to the additional recombination current. T h e expression for the recombination
current in the depletion layer of a forward-biased diode, given by E q . (7.2), is modi
fied for t h e emitter-base junction with V
kT/q to b e c o m e
C

EB

ho

e x

i^pj

P 1 E NKT

(9.15)

w h e r e W is the width of t h e depletion layer and is the lifetime of electrons in t h e


layer.
The n e w expression for I' (shown in Fig 9.7) at low current levels b e c o m e s
0

= (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

H i g h - L e v e l Injection a n d t h e Kirk Effect a t t h e Base-Collector


Junction
The Kirk effect is t h e p h e n o m e n o n that results in the widening of t h e base region
and a corresponding slower increase of I as V
is increased b e y o n d the m o d e r a t e
region. T h e base current, however, continues to increase at t h e m o d e r a t e rate.
We consider first an N P N B J T in which the collector current is high. A t this
value of collector current, the mobile carriers' (electrons) space-charge density is n o
longer negligible w h e n c o m p a r e d to the fixed-dopant densities; in particular, to the
negatively charged acceptor ions in t h e collector-base space-charge layer. The
resulting space-charge density in the base side of the C-B space-charge layer is
increased. T h e space-charge layer on t h e collector side is so heavily d o p e d that it is
C

EB

268

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

200

100

Figure 9.8

10mA

1mA

I,c

Plot of common-emitter current gain, fi , versus I at a fixed | V \.


f

CE

not affected. To maintain zero space-charge density through t h e layer, t h e layer on


the base side must shrink and thus cause an increase of the width of the base leading
to a reduced slope in the electron distribution in the base. Hence, a reduction in the
collector current.
A few analytical relations for this effect will help the explanation. We start by
writing Poisson's equation in the depletion region as
d%/dx = l/e[q

N(x) -

I /v(x)A ]
c

(9.18)

where N(x) = (Np - N ) in t h e space-charge region (negative acceptor ions in the


base side and positive donor ions in t h e collector side). The velocity v(x) is that of
the electrons, A is the emitter area, and I /v(x)A
represents the mobile electron
density of t h e collector current.
For a constant, V , and by neglecting the built-in voltage w h e n c o m p a r e d to
V , the relation b e t w e e n V
and the electric field is
A

CB

CB

cn

(9.19)
where (x x ) is t h e width of t h e space-charge 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)

The relevance of I will be shown later in this section.


Next, we consider a second B J T structure whereby a lightly-doped epitaxial
layer separates t h e base from the highly d o p e d buried layer of the collector, and the
o

Section 9.2
+

Effects at Very Low and High Injection

269

configuration is N P N ~ N as shown in Fig. 8.1c. A n increase of the space charge


density, caused by an increase of collector current, subtracts from the positive epi
taxial d o n o r ions. To balance the charges, the depletion layer in the epitaxial side
tends to expand with minimal change in the base side. A t the same time, and m o r e
importantly, an increase of I results in an increase of the voltage that is d r o p p e d in
the u n d e p l e t e d part of the epitaxial region, causing a reduction of t h e applied
reverse bias b e t w e e n the base and epitaxial layer. The depletion layer tends to
shrink and eventually collapse. A t a sufficiently high current, this junction becomes
forward-biased, causing a quasi-saturation effect that increases t h e electron density
at this junction in the base. This reduces t h e slope of the electron distribution in the
base and hence reduces the collector current.
In b o t h situations described above, for t h e N P N ~ N BJT, t h e effective end
result is a widening of the width of the base w h e n I exceeds t h e specific critical
value / .
c

A n expression for t h e collector-base space charge region width has b e e n


obtained* as
x, co
(9.21)
x, CB

where x
E q . 9.20.

c o

is the space charge width at I

= 0 and I is the critical current defined by

High-Level Injection a t t h e Emitter-Base J u n c t i o n


Two p h e n o m e n a take place in t h e emitter and base regions that tend to cause I to
increase at a lower rate in t h e high-injection regime than t h e m o d e r a t e regime. First,
the increasing minority carrier concentrations in t h e base causes an equal increase
in the majority carrier concentrations. The increase in the electron density in the
base of the P N P transistor is reflected by an increase in N
of E q . (8.55), thus lead
ing to a reduction in the injection efficiency and hence a reduction in the current
gain.
The second effect is t h e current "crowding" u n d e r the emitter, as shown in Fig.
9.9. The lateral m o t i o n of the base-current carriers causes a potential drop, due to
the base resistance, in t h e lateral direction. Because of this, different parts of the
emitter are biased to different voltages, giving rise to uneven injection. T h e greater
part of the injection occurs from the edge of t h e emitter that is nearest to t h e base,
causing a localized concentration in the emitter current. This is labeled emitter
crowding. The current density at that edge exceeds the average value so that con
ductivity m o d u l a t i o n of the base occurs earlier t h a n it would have if t h e current was
spread out over t h e whole emitter area.
In Sec. 9.2, we have confined our discussions to the differences that exist in t h e
active mode of the B J T b e t w e e n the ideal and actual characteristics. Switching of a
c

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 common-emitter 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
Q9-1 Briefly define the significance of the multiplication factor M.
Q9-2 Why is punchthrough of concern in modern BJTs?
Q9-3 To increase the punchthrough voltage, the base doping can be increased. What is the
negative effect of an increase in the base doping?
Q9-4 Explain the negative effect of reducing the base doping on the base resistance.
Q9-5 What is the effect on the emitter of high base resistance at high current?
Q9-6 Explain why BV
is larger than
BV .
Q9-7 Briefly explain the shape of the curve $ versus I in Fig. 9.8.
CB0

CEO

HIGHLIGHTS
The collector-base voltage has dramatic effects on the common-emitter 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 reverse-bias at the collector-base junction is
so large that the depletion layer in the base extends to the emitter-base junction.
As the collector-base 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
lector-base 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
E9-1

Punchthrough takes place in a certain NPN silicon transistor when \V \ = 30V.


Assume the collector doping is much greater than that of the base where N =
10 cm . The relative dielectric constant is 11.8. Determine the approximate value
of the base width when \ V \ = 0.
CB

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 short-circuit in t h e O N position and an open-circuit 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 common-emitter 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 open-circuit. 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

Bipolar Transistors II: Limitations, Switching, and Models

(a)

(b)

EC

(c)

Figure 9.10 Switching of a transistor (a) circuit, (b) BJT characteristics and load line, and (c)
base current drive.

saturation modes. It is t h e storage of minority carriers in the base of a transistor


operating in t h e active and saturation regions that accounts for t h e diffusion capaci
tance.
The time delay that occurs in switching may also b e related to t h e transit time
of t h e minority carriers that flow from emitter to collector through the base.
S t o r e d C h a r g e a n d Transit T i m e
First, let us relate the magnitude of the collector current to the minority carrier
charge stored in the base. We assume a linear distribution of minority carriers, as
shown in Fig. 9.12, for operation in the active region.

Section 9.3

Transistor Switching

273

Figure 9.12 Distribution of excess


minority carriers in the base of a PNP
transistor.

T h e collector current is related to t h e excess hole density by


I

= ~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)

Therefore, to change t h e collector current, it is necessary to change t h e charge


stored in the base. If the collector current is I , then a charge, Q , is accumulated or
r e m o v e d in time.
Let us now find a relation b e t w e e n a n d t h e approximate time it takes a hole
to travel from the emitter junction to the collector junction.
T h e incremental distance traveled by a hole in time dt is dx, given by
c

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)

The current of holes is


I = -qA
p

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

Bipolar Transistors II: Limitations, Switching, and Models


I = qAp'(x)

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

E q u a t i o n (9.29) is used in E q . (9.26) to obtain the transit time


W

'W (l-x/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

Charge Control Relations


The charge control equations represent a powerful tool used in t h e analysis of the
transient behavior of t h e transistor. They relate the excess charge stored in the base
to the base current and the collector current.
The general expression for the excess hole density charge stored in the base of
a P N P transistor, Q , is given by E q . (9.31)
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

We integrate E q . (9.32) from = 0 to = W

Jo

Jo

as

dx I
Jo

p'{x)qAdx=-\
o t

(9.32)

qA dJ

(9.33)

The first integral in E q . (9.33) represents t h e rate of change of the stored


charge, Q , with time; the second integral is the ratio of t h e stored charge to the life
time of holes; and the third integral b e c o m e s (i i ), where i = i and i = i
so that the difference is i (t). L o w e r case i's refer to instantiations.
The result of E q . (9.33) is the first charge control equation to be stated as
B

Cp

Cn

Cn

..) .Ml iMl


+

dt

(934)

Section 9.3

Transistor Switching

275

E q u a t i o n (9.34) is interpreted as indicating that t h e r a t e of change of stored


charge in t h e base dQ (t)/dt, is d e t e r m i n e d by two processes: t h e base current i (t),
which adds charge to t h e base, a n d by t h e process of recombination that removes
charge at t h e rate
Qh.
T h e stored charge, Q , is transferred totally to t h e collector every seconds
by t h e collector current where is t h e base transit time. If t h e charge is transferred
at this rate, a n d it is a steady-state charge, then it must b e replenished at t h e same
rate. These statements, which a r e valid in t h e active region of operation, are
expressed in equation form
B

i (t)

= Q (t)h

(9.35)

E q u a t i o n s (9.34) a n d (9.35) are k n o w n as t h e charge control

equations.

Turn-ON 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
(9-36)
dt

E q u a t i o n (9.36) is a first order differential equation requiring o n e b o u n d a r y


condition.
A t t = 0, Q = 0, so that t h e solution to E q . (9.36) is given as

G (0
B

=V ,

[1

"

(-'A,)]

(9.37)

The collector current for active m o d e operation is d e t e r m i n e d by t h e total


charge that must b e transferred to t h e collector every seconds. Using this defini
tion in E q . (9.35), the expression for t h e collector current becomes

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

If t h e value of i is such that it places i at the b o u n d a r y b e t w e e n t h e active


and saturation regions, such as point 1 in Fig. 9.13(a), then Q is just equal to < 2 as
d e t e r m i n e d from t h e area u n d e r plot 1 in Fig. 9.13(b). E q u a t i o n (9.38) can b e writ
ten in t h e steady-state as
B

S A T

If i , however, places i in d e e p saturation, then t h e value of i is d e t e r m i n e d


as ( V V )/R
and V
for a silicon transistor is approximately 0.2V. W e
observe that t h e collector current m a y b e assumed t o b e approximately constant
while t h e transistor is moving from t h e edge of saturation t o d e e p saturation. This
B

c c

ECSiCP

EC

276

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

collector current is /
, which is approximately equal to V /R
since V

V . O n c e t h e transistor is in saturation, an increase of i has negligible effect on the


value of i . H e n c e , the ratio of i to i is ^ only at the edge of saturation, any
increase in i drives the transistor d e e p e r into saturation and t h e ratio of i to i is
C S A T

CC

cc

C S A T

(a)

Section 9.3

Transistor Switching

277

less t h a n E v e n though t h e value of i m a y b e /


, Q continues to increase
b e y o n d <2 , as shown in Fig. 9.14.
The transistor is assumed to be t u r n e d - O N w h e n i is at /
, an expression
for which is obtained from E q . (9.38) at time i
as

C S A T

SAT

C S A T

Q N

V /R
CC

= ^

= ^

HONA,)]

[1 - exp

(9.40)

The t u r n - O N time is d e t e r m i n e d from E q . (9.40) by setting /


(neglecting V
) so that

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

Therefore, a fast t u r n - O N time is favored by a large i , a smaller /


(V /R ),
as well as a smaller .
B

CC

C S A T

Turn-OFF 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)

A s shown in t h e sketches of Fig. 9.15, the hole density decreases and Q


decreases. The collector current remains at /
until Q decreases to <2 , at
which time i decreases exponentially towards zero.
The t u r n - O F F time, defined as the time required to reduce the collector cur
rent to almost zero, is m a d e u p of two increments: the time it takes Q to reach
<2 , k n o w n as t h e storage time, t , and second, the time t , it takes the collector cur
rent to reach zero, or m o r e practically to a value of about 0.1 7
. T h e decrease of
stored charge and current are shown in Fig. 9.15.
A s s u m e that the base current is m a d e zero at a new t = 0. The solution to E q .
(9.42) for t > 0 is given by
B

C S A X

SAT

SAT

C S A T

Q (t)
B

= Q (0)

e-H

(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""> = /

T h e storage time, t , is found from


s

sat

C S A T

e~*P

(9.44)

278

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

QB(0

Figure 9.15 Sketches of Q , i , andp'(x) during the turn-OFF period as i is reduced to zero
instantaneously.
B

Solving for t , we obtain


s

t = ^ n(Q (Q)/Q )
s

(9.46)

SAT

O n c e t h e transistor returns to t h e active region at t = t, the collector current


decays exponentially towards zero and is given by E q . (9.44) as /
e~' p. The time,
tj, it takes the collector current to b e c o m e 0.1 /
is found from E q . (9.44) to be
2.3T .
h

C S A T

C S A T

We n o t e from E q . (9.46) that a small storage time is favored by a small i and a


smaller T^.The smaller i causes a smaller charge buildup and a small lifetime causes
the stored charge to b e r e m o v e d faster. A d d i n g gold to the base doping reduces the
lifetime.
We will n o w use an example to illustrate the t u r n - O N and t u r n - O F F time
delays.
B

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,

, we use Eq. (9.24) to determine


= Wl(2D

5mA

= Am

Section 9.3

279

Transistor Switching

(/1)

'is

200

200

t(ns)
Figure 9.E1

From Eq. (9.40) we determine Q

S A T

Base current pulse

as
T

GAT W = P
S

We now determine the turn-ON 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

= 200 X 10~ A, R = 1, and = 4 10- s


L

= 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

g (200ns) = 200 X 10 ~ X 0.25 X 10^ [1 - exp(-200 1~ /2.5 X 10~ )]


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

t = 2.5 X 10~ (27.5/20) = 79.6ns


s

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 metal-semiconductor 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

Bipolar Transistors II: Limitations, Switching, and Models

Q (pc)
B

50

27.5
20

i (mA)
B

12.5

127
Figure 9.E2

279.6

t ( n s )

Stored charge and collector current variations with time.

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

Small-Signal Equivalent Circuit


I +AI.
r

281

/
EC

,EC
AV

EB

L =
C

-M

(")

v,EB
Figure 9.16

cc

Biasing of a PNP transistor in the common-emitter connection.

the slope is used. The A C equivalent circuit is m a d e of lumped elements, such as


resistances and capacitances, in addition to a d e p e n d e n t current generator. The cur
rent generator models the amplification property of the transistor.
Just as in the derivation of the static characteristics, it is sufficient to study the
variations in two of the t h r e e currents caused by the variations in two of the t h r e e
voltages. We will consider the c o m m o n emitter circuit, which is the circuit of the
transistor most commonly used for amplification. This circuit, shown in Fig. 9.16, will
be assumed to be biased by the D C voltages, V
and V , to an operating point on
the static characteristics. In other words, the values of I , I , and V
are fixed.
Since we are causing the small-signal variations to take place on the slope at
the operating point, we are in effect linearizing the variations in the excitation and
in the response. Because of this the equivalent circuit is labeled a small-signal linear
equivalent circuit. We are in effect representing the relations between the varying
currents and voltages by linear elements in the equivalent circuit. Because of the
nonlinear shape of the static characteristics, a change in the location of the operat
ing point is accompanied by a change in the slope at the new point, resulting in
changes of the values of the circuit elements.
O u r task consists in determining the effects of changes in V
and V
on the
processes in the transistor and hence on the currents I and I .
To simplify our analysis, let us m a k e the following assumptions:
EB

cc

EC

EB

CB

a) O p e r a t i o n is in the active m o d e so that V is positive and V


the magnitudes of both are much greater than kT/q.
EB

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) Low-level injection of holes from the emitter into the base.

282

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

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>

In this section, we shall investigate t h e effects of changes in V


on the cur
rents in the transistor and we will derive equations for t h e changes in the currents in
terms of the changes in V .
Let us review the effects of an increase in V .
EB

EB

EB

1. The hole density in the base at the edge of the emitter-base depletion layer
increases. To maintain neutrality, the electron density in t h e base increases. We
assume a fixed V
at the reverse-biased C-B junction so that the excess hole
density at = W is zero. The hole density distribution is shown in Fig. 9.18.
CB

2. The gradient of the hole density increases t h r o u g h o u t the base.


3. T h e hole currents in the base at the emitter and collector junctions increase.
4. T h e electron current across the emitter-base junction increases.
5. T h e concentration of holes and electrons stored in the base increases.
6. Because of t h e increased density of holes in the base, there is a greater rate of
recombination.
7. The voltage barrier at t h e emitter-base junction decreases due to a reduction of
the density of ionized atoms in t h e depletion layer. The base supplies a transient
current to cover the ionized acceptors.
H o w then, do all these changes influence the base and collector currents in Fig.
9.17?
First, the base lead has to supply the additional electrons required by the
increased current of electrons, , crossing the emitter-base junction. Second, the

Bo+

Equivalent circuit

-o C

be

Figure 9.17 Representation of


incremental quantities in commonemitter circuit.

Section 9.4

Small-Signal Equivalent Circuit

283

base lead has t o supply m o r e electrons to recombine in t h e base, Al . Third, t h e


base lead has t o supply additional electrons to maintain neutrality in t h e base.
Fourth, t h e base supplies t h e electrons that cover t h e ionized donors in t h e deple
tion layer. T h e first two currents have t o b e continuously supplied, whereas t h e last
two are transient in nature. Fifth, t h e collector current increases as a result of t h e
increase in t h e gradient of t h e hole density in t h e base. W e will consider first t h e
steady-state current changes. T h e base current increment becomes
i =-M
b

= -(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
collector-base 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)

= qAD p' {Q)/W

(b)

(9.48)

By subtracting t h e terms in E q . (9.48(a)) from E q . (9.48(b)), t h e change in t h e


collector current becomes
AI

qAD Ap(0)/W
p

(9.49)

where () = ' ( ) and


(0) =

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

[exp (qAV /kT)


EB

AV )
EB

kT

qV
exp kT
E

~!]

- l]

By an expansion of t h e exponential in E q . (9.50) and assuming AV


smaller t h a n kT/q,we
can n o w show that

(9.50)
EB

to b e much

284

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models


(0) = (0)

(q/kT)

Pl

(9.51)

AV,EB

O u r last assumption serves as a general guideline t o t h e interpretation of


incremental changes. In o r d e r t o obtain t h e linear relation of E q . (9.51), t h e change
in AV
must b e m a d e smaller than kT/q or smaller than 26mV at r o o m tempera
ture. By expanding exp (qv /kT)
where v
AV , we have
EB

eb

eb

exp (qvJkT)

EB

= 1 + qvJkT

+ \
(^$)
2\kT

(9.52)

If v kT/q, E q . (9.50) reduces t o E q . (9.51). Thus, t h e small signal analysis


is valid only for v 2 6 m V at r o o m t e m p e r a t u r e .
The electron current across t h e emitter junction is given by E q . (8.33)
eb

eb

En

[exp (qV /kT)

(9.53)

- 1]

EB

We replace t h e exponential in E q . (9.53) by its equivalence,p[(0)/p , and we have


0

qAD n p[(0)
nE

*En

OE

( 9

'

5 4 )

^nEFO

By following t h e p r o c e d u r e of Eqs. (9.48), (9.49), and (9.50), t h e change in I ,


reflected by t h e change i n p ' ( 0 ) , Ap'(0), or Ap(0), is
En

<lAD n
nE

Ah

Ap(0)

0E

(9.55)

LePo

We n o w consider t h e change in t h e base recombination current. A l t h o u g h we


assumed a linear distribution of hole density, there is still recombination. The
recombination current is negligible c o m p a r e d t o t h e collector current b u t not negli
gible c o m p a r e d to t h e base current.
The base recombination current is
4 c = Qah

= 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)

where p[(0) = p [exp (q V /kT)


- 1].
Using t h e same p r o c e d u r e employed earlier, E q . (9.57) becomes
u

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

Small-Signal Equivalent Circuit

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

where ~r D m a y b e replaced by L and p p ' ( 0 ) .


We also find a n expression for i by using Eqs. (9.48), (9.49), and (9.51) as
p

c = fflcv

(9-61)

be

The factor multiplying v has units of conductance. In fact, it is k n o w n as t h e


transconductance
g a n d it is t h e ratio of t h e change in t h e collector current caused
by a small-signal change in t h e base-to-emitter voltage so that
be

= K K e= y y

(9-62)

where V = kT/q = 26mV at r o o m t e m p e r a t u r e .


T h e base current increment is written in terms of i by using Eqs. (9.60) and
(9.62) as
c

h = S v J^
m

= iJ^

(9.63)

where is a dimensionless quantity k n o w n as t h e small-signal common-emitter


current gain a n d d e t e r m i n e d at t h e operating point (V
and V ) for small varia
tions i n n ^ and i from Eqs. (9.60) and (9.62).
0

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

(9-65)

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 small-signal low frequency equivalent circuit for t h e B J T in
the common-emitter connection as shown in Fig. 9.19. This is a low-frequency 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 small-sig
nal current gain, .
We have defined earlier t h e common-emitter current gain, ^, t o b e I /I . T h e
symbol is t h e ratio of t h e collector small-signal current t o t h e base small-signal
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

Bipolar Transistors II: Limitations, Switching, and Models

Bo+

<
Figure 9.19 Low-frequency commonemitter equivalent circuit of the
transistor.

precise definition for . Since is t h e ratio of the incremental currents


( /AI )
and using I = &pI , we have
0

iji

= M /M
0

iji

dl,

(9.66)

If does not vary with I , t h e n t h e last t e r m in E q . (9.66) is ^..


A quick m e t h o d to d e t e r m i n e the p a r a m e t e r s , g , and
is to use E q . (9.62)
and the c o m m o n - e m i t t e r graphical characteristics, such as those shown in Fig.
8.12(a). A t t h e operating point, I is known, so that g is d e t e r m i n e d by using E q .
(9.62). Since for all practical purposes, is equal to ^, the ratio of I /I
at t h e
operating point of the characteristics is . resistance
is then \S /g .

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

The rate of change of stored charge is accompanied by a transient base cur


rent, AI , which by using E q . (9.51) and assuming that p p ' ( 0 ) is given by
BS

BS

= AQ /At
B

= qAW Ap(0)/At/2

= (qAW p[(0)

(q/kT)

AV /2At
EB

(9.67)

We introduce I from E q . (9.48)(a), replace the delta quantities by their incre


mental equivalents, and replace I by its g equivalent so that
c

'/,

A/

r B

e s

= / (Wg/2D ) ^^
c

= (Wl/2D )
p

Since the current in a capacitance is C dv/dt, we can write

' "^

(9.68)

Section 9.4

Small-Signal Equivalent Circuit

Figure 9.20

287

Small-signal equivalent circuit based on changes of V

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)

T h e r e is an additional c o m p o n e n t of transient base current, which u p o n an


increase of AV , is used to neutralize some of the acceptor ions in the depletion
layer of the emitter-base junction. This neutralization is required to reduce the
depletion layer width that accompanies an increase of V . This is a capacitive cur
rent defined by
EB

EB

dv,
i -C ~-f
bt

(9.71)

je

where C. is the emitter-base 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 collector-to-base 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 reverse-bias at the

288

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

collector-base junction h a s a n important, though secondary, effect on t h e static char


acteristics of t h e transistor. T h e increase of this voltage increases t h e collector cur
rent.
The complete-small signal equivalent circuit of t h e transistor in t h e commonemitter connection represents t h e changes in I and I brought about by changes in
V
and V . T h e immediate effect of an increase in t h e magnitude of V
(higher
reverse bias) is to widen t h e depletion layer at t h e collector-base junction and thus
decrease t h e effective width of t h e base, as was shown in Fig. 9.2.
You may w o n d e r why we did not consider base width modulation from the
emitter side w h e n V
changed. T h e reason is that t h e change in V is very small
c o m p a r e d with AV .
The change in t h e distribution of hole density brought about by base-width
m o d u l a t i o n has three consequences that a r e i n d e p e n d e n t of t h e effects of
AV^.We
consider that V
is fixed and d e t e r m i n e t h e effects of increasing V
(making it
m o r e negative).These effects a r e as follows.
c

EB

CB

CB

EB

EB

CB

EB

CB

1. Increase in collector current d u e t o t h e increase in t h e gradient of t h e hole den


sity in the base.
2. D e c r e a s e of total stored base charge, which requires a transient c o m p o n e n t of
base current.
3. D e c r e a s e of recombination rate, and hence a decrease in t h e base current,
which supplies electrons to recombine in t h e base.
4. T h e transient base and collector currents increase the density of ionized atoms
in t h e collector-base depletion layer.
T h e effects listed in (1) and (3) p r o d u c e steady-state changes in t h e currents,
whereas those in (2) a n d (4) a r e transient in nature.
Collector Current C h a n g e
The increase in collector current caused by a change in V can b e written in terms
of t h e change in t h e base width using t h e expression for t h e collector current qAD
P[(0)/W
BC

AI

AV

die AW

-qApKO)

dW AV

BC

AW

BC

AV

BC

"

A decrease in t h e base width is accompanied by an increase in t h e collector current.


By introducing I , E q . (9.72) becomes
c

A ^ . z k ^
AV
F r o m E q . (9.5) and for V

BC

BC

AV

(9 V3)
BC

= V , t h e expression for t h e Early voltage, V , is


EC

(9.74)

where V is a positive quantity since an increase in V for a P N P transistor causes a


decrease in W . Therefore, replacing AV
a n d AW with dV. and dW , we have
A

BC

BC

Section 9.4
U

Small-Signal Equivalent Circuit

289

= W I /V

BC

(9.75)

In terms of small-signal quantities and using E q . (9.62), we let i

cl

AI

ii = \ V ^ = V J / ^

(9-76)

where g = I /V and V is kT/q.


A n increase in the base-collector voltage decreases W and increases i T h e
expression in E q . (9.76) represents a ( d e p e n d e n t ) current generator from collector
to emitter, which d e p e n d s on t h e base- to-collector voltage.
We will include this current generator in t h e circuit together with t h e elements
resulting from t h e recombination base current change in t h e next section.
m

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)

is included in E q . (9.77) by using Eqs. (9.38) and (9.39)

BC

e c

AI
AV

rec

BC

AQ
i AV
p

_ - (7cTB)
AV

BC

BC

-AI

0

(9.78)

The negative sign has b e e n introduced t o indicate that an increase in V


s an increase in I whereas AI
decreases.
Using E q . (9.75) to replace AT /AV
with IJV_
in E q . (9.78) and AV
with
v , we have

BC

IEC

BC

BC

bc

AV

PO BC

PO A

We have expressed t h e change in recombination current in terms of t h e small


signal base collector voltage.

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

A s V increases, t h e charge stored in t h e base of a P N P transistor decreases


because of t h e n a r r o w e r base width. This corresponds to AQ being equal to
(/^Tg). T h e expression for t h e transit time, , from E q . (9.24) is WB/2D . Using
this in E q . (9.80), we have
BC

290

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

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?

(9-82)

The capacitance, C , connected from base to collector, is a very small fraction


of the base storage capacitance considering that kT/q is 26mV and V is in tens of
volts.
By using the current generators representing the effects of t h e collector cur
rent change, t h e base current change, and the storage capacitance, the circuit shown
in Fig. 9.20 is modified to include two additional current generators and a capaci
tance. The new circuit is shown in Fig. 9.21.
We will summarize below t h e changes that t h e circuit elements of Fig. 9.21
represent:
bu

change in the base current caused by a change in

change in t h e collector current caused by a change in

C.

capacitance of the base-emitter depletion layer

base charging capacitance caused by a change in

base charging capacitance caused by a change in

change in the base recombination current resulting from a

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

Figure 9.21 High-frequency small-signal equivalent circuit modified to include


effects of the changes of V .
CB

Section 9.4
g V v /V
m

bc

291

Small-Signal Equivalent Circuit

change in the collector current resulting from a change in

CB

We will n o w reduce the n u m b e r of current generators by inserting resistances


and we will also include parasitic elements to form the complete circuit.
By setting v = v + v , the current generator i given by E q . (9.76), is split
u p into
bc

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

The term, r , is k n o w n as the C E output resistance of the transistor.


The expression for the base current i given by E q . (9.79), may b e m o d e l e d by
the inclusion of a resistance from base to collector, d e t e r m i n e d as
Q

bV

r = f
u

(9.85)

( /

A collector-base junction or transition-capacitance, C., has to be included in


the circuit to account for the change in the width of the depletion layer as the collec
tor-base voltage changes. We identify, as well, the ohmic base resistance, r , and the
ohmic resistances of the emitter and collector as r and r , with the complete equiva
lent circuit shown in Fig. 9.22.
A l t h o u g h the circuit in Fig. 9.22 has b e e n obtained for a P N P transistor, it is
equally applicable for the N P N transistor.
Typical values for the elements calculated at I = 1mA, V
= 5 volts for a
transistor having = 100, V = 50 volts, and = 0.4ns are shown in the accompa
nying table.
b

EC

Small-signal elements of the CE BJT


Symbol

Name

Relevant Equation

transconductance
CE current gain
CE input resistance
CE output resistance
storage capacitance ( V )
E-B junction capacitance
storage capacitance ( V )
C-B 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

Bipolar Transistors II: Limitations, Switching, and Models

B'o

' <^

be

Cb + Cje -

n be

<C o

6E'
Figure 9.22 Complete high-frequency small-signal 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 short-cir
cuit current gain. (/), of the transistor. It is defined as the ratio of the current I in
a short-circuit 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 short-circuit, 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)

A t high frequencies, t h e magnitude of the imaginary part of t h e d e n o m i n a t o r


of E q . (9.86) is m u c h greater than unity, so that

where g = / ^ .
The symbol, f , is defined as the frequency at which the magnitude of the
short-circuit 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-

Figure 9.23 Circuit for calculating f of a transistor.


r

fr

If we define as 1 / , where = 2/ , we have

= ^

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)

The capacitances, C. and C , consist mainly of transition capacitances and


h e n c e d e p e n d , in accordance with E q . (7.16) on the voltages across the emitter-base
and collector-base depletion layers respectively. The storage capacitance, C , on the
other hand, d e p e n d s on the quiescent collector current through g and on the base
transit time. The base transit time, , is uniquely d e t e r m i n e d for a given transistor.
It is obvious that low values of t h e capacitances contribute to an improved
high frequency response. A high value of g is an indication of a high ratio of smallsignal collector current to a small signal input voltage. H e n c e , the f of a transistor is
a m e a s u r e of its current amplification performance at high frequencies. It is also
known as the figure of merit of a high-frequency transistor.
a

REVIEW QUESTIONS
Q9-8

To switch a transistor (PNP) from cutoff to saturation, what should happen to the

hole density in the base and how is this accomplished?


Q9-9 Explain clearly the meaning of each of the charge-control equations.
Q9-10 By what mechanism is the turn-ON time reduced?

294

Chapter 9

Bipolar Transistors II: Limitations, Switching, and Models

Q9-11 Why is the BJT small-signal circuit labeled linear?


Q9-12 Briefly define the significance of the frequency f

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 turn-ON time is enhanced by a large amplitude of
the base current, a smaller saturation collector current, and a small lifetime.
The turn-OFF 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 small-signal circuit is determined, which
represents the transistor for incremental changes about an operating point on the out
put characteristics.

The quality of a BJT as an amplifier is measured by the gain and high-frequency


response. Large gain is enhanced by a large transductance, g , and high-frequency
operation is improved by smaller storage and junction capacitances.
m

EXERCISES
E9-2

Calculate the value of

for a BJT operating at l = 2mA and whose = 100.


c

Ans: r = 1.3.
E9-3

A BJT operating at I = 1mA h a s / = 500 MHz, = 100, and C = C = lpF, deter


mine C .
c

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

The Gummel-Poon Model

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 Ebers-Moll 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 Ebers-Moll 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:

R e c o m b i n a t i o n in t h e emitter depletion layer at low values of emitter-base


bias, resulting in a decrease of current gain.

D e c r e a s e of common-emitter current gain at high values of collector current.

Non-zero slope in the active region of t h e common-emitter output character


istics, resulting from t h e Early effect.

T h e G u m m e l - P o o n relations are derived through a three-step process. First,


the Ebers-Moll relations for I and I a r e written in terms of I , ^, and . The
resulting equations are k n o w n as t h e " t r a n s p o r t version" of t h e Ebers-Moll rela
tions. T h e second step consists of modeling the increase of base current at low val
ues of emitter-base bias by a superposition of ideal diode and non-ideal diode
parameters, as d e t e r m i n e d from m e a s u r e m e n t s on the transistor similar to those
shown in the plot of Fig. 9.5. This constitutes the addition of four new parameters.
The third step is the replacement of I by an equivalent majority carrier charge
in the base. B o t h the effects of high injection and base-width modulation influence
the majority carrier charge in the base through the emitter and collector injected
carriers. These carriers are c o m p e n s a t e d by an increase in t h e majority carrier
charge n e e d e d to neutralize the injected minority carriers. F u r t h e r m o r e , the
changes in the emitter-base and collector-base voltages influence t h e ionized
charges t h r o u g h the changes in the depletion layers at the emitter and collector.
We refer to E q . (8.48) to obtain the expression for I , defined as the factor
multiplying t h e V
exponential as
B

EB

I =
s

qADfo/W,II

qAD nj

qADn

N W

qAN W

DB

DB

(9.91)

The d e n o m i n a t o r of E q . (9.91) may be defined as the majority carrier charge


in t h e base at equilibrium. Since the total charge varies due to t h e effects of high
injection and base-width modulation, we write E q . (9.91) in terms of a charge Q as
G

296

Bipolar Transistors II: Limitations, Switching, and Models

Chapter 9

where

= qA

n{x)dx

(9.93)

Jo
T h e symbol Q

GQ

is k n o w n as the G u m m e l n u m b e r and is defined as


QGO =

N dx

(9.94)

DB

Jo
Because of the effects of high injection and base-width modulation,
becomes
Q

= Q

G0

+ Q +Q +Q +Q
E

(9.95)

where
Q
Q

G0

and Q

Q and Q
p

= net majority carrier charge c o m m e n s u r a t e with space charge neu


trality
= immobile ion charges uncovered by the changes in the depletion
layer boundaries
= base majority carrier charge n e e d e d to neutralize the emitter and
collector injected minority carriers.

The high-injection effects are m o d e l e d using the charge control relations and
the charge-control time constants and , while base-width 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'

C o m p l e t e numerical analysis using the required p a r a m e t e r s is easily carried


out by using S P I C E .
PROBLEMS
9.1 An NPN silicon BJT has N
= KFcm" , N = 10 cm- , and W = 1.2. It is
operating in the active region at V = 0.6V and = 300K. Determine:
(a)
the change in the width of the base as V changes from 1.5V to 6.5V DC.
(b)
the corresponding change in the collector current.
9.2 A silicon NPN BJT has N
= 10 cm" , N
= 1 0 c m , W = 0.6, V = 0.7V,
and V = 5V. Given D = 20cm /s and = 5 X 10~ s. When V is increased to
10V, the minority carrier diffusion current in the base increases by 20 percent.
Determine, at = 300K,
(a)
the base doping.
(b)
the Early voltage V .
9.3 For problem 9.1, determine the Early voltage, V .
3

16

DB

AC

BE

CE

18

DE

16

CB

DC

nB

BE

CB

Chapter 9

Problems

297

9.4 A silicon NPN BJT has N = 10 cm" , N


= 10 cm- , and W = 0.25.
Determine, at = 300K,
(a)
the punchthrough voltage.
(b)
the average value of the electric field intensity at punchthrough.
To increase the punchthrough voltage, the base doping can be increased. At what cost
would this be?
9.5 An NPN silicon BJT at = 300K has heavy collector doping and N = 10 cm- .
Given W = .. Determine:
(a)
the breakdown voltage for active operation in the common-base mode. The
breakdown field in silicon is 3 X 10 V/cm.
(b)
the punchthrough voltage.
9.6 A silicon PNP BJT at = 300K has N
= 2x 10 cm- , N = 4X 10 cm" , W =
2, D = 11.2cm /s, and = . and A = 10 cm . Determine the output resis
tance at I = and V = -10V.
9.7 For a BJT operating in the common-base active region, derive an expression for the
output resistance r in terms of I , W , L , and dW /dV
The output resistance is
defined by l/r
=
dI /dV \l .
9.8 Show that the transport factor of a BJT is given approximately by
17

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)

the turn-ON time.


the storage time.
Use the Ebers-Moll expression for I in the active region of operation of an
NPN BJT and assume a small-signal voltage u ( ) is superimposed on
V . By expanding the exponential show for u V , the small-signal collec
tor current is given by g i.
For what maximum approximate value of . is the small-signal equivalent cir
cuit valid?
c

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

Bipolar Transistors II: Limitations, Switching, and Models

(d)

20

16

9.13 A GaAs PNP BJT has N = 10 cm~ , N


= 10 cm- , D = 30cm /s, = ,
A = 10 cm , and W = 12. At = 300K, and for I = 2mA in the active region,
determine:
(a)
the base storage capacitance C .
(b)
the emitter junction capacitance.
9.14 For Problem 9.13, given = 100, and C = 0.25pF, determine, at = 300K,
(a)
r
(b)
f
AE

_3

DB

pB

chapter 10
JUNCTION FIELD-EFFECT
TRANSISTORS

10.0 I N T R O D U C T I O N
Field-Effect 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 silicon-based
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 metaloxide-semiconductor 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
conductor-semiconductor surface or at a metal-semiconductor surface, and we
m

299

300

Chapter 10

Junction Field-Effect Transistors

distinguish the latter by labeling it t h e M E S F E T , leaving t h e J F E T designation to


the P N junction device.
In this chapter, we will study the P N junction J F E T . In later chapters, we will
consider t h e M E S F E T and t h e M O S F E T . We begin our study with t h e J F E T for two
reasons. First, the J F E T represents a natural transition from the B J T to the M O S
F E T and, second, it is a simpler device to fabricate and to analyze. It must be men
tioned h e r e that it is not nearly as widely used as t h e M O S F E T .

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
two-gate 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 gate-channel junction is reverse-biased 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 drain-source 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 gate-to-source 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 drain-source 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 reverse-bias across the gate-channel 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 drain-source voltage, t h e width of the channel and
therefore the effective cross-sectional area of the channel n o r m a l to the carrier flow
is changed.
The change in the cross-sectional 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

Construction and Operation

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 one-gate
model.

conductance modulation. Because of t h e voltage applied b e t w e e n drain and source,


the drift current of electrons moving from source to drain is d e t e r m i n e d by O h m ' s
law, it being t h e ratio of t h e drain-source voltage to t h e resistance of the channel
b e t w e e n drain and source.
We can conclude from the above that, for a fixed drain-source voltage, by mak
ing the P gate m o r e negative with respect to t h e source, and hence with respect to
the channel, t h e width of the depletion layers increases, the width of the channel as
+

302

Chapter 10

Junction Field-Effect Transistors


oV

-IV
R = resistance of channel
R

<D1

R = , > \
2

-oV

ID2 <

-oV

D2
IDI

= -AV

R - R > R
3

fD3

ID3 < ID2

(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 gate-source volt
age. In fact, t h e drain-source 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 cross-sectional 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

Construction and Operation

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 pinched-off
at the drain end.

304

Chapter 10

Junction Field-Effect Transistors

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 pinched-off 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

Construction and Operation 305

Section 10.1

part of t h e channel that is pinched-off, we label as the isolation region.


The increased drain voltage above V ,
V - V ,
appears across t h e
depleted isolation region so that the voltage d r o p across the unpinched part of the
channel is V .
A b o v e pinchoff, a strong longitudinal (along the channel) electric field
appears across the isolation region. This field originates on the positively ionized
donors in the depleted isolation region and terminates on electrons in t h e channel.
Such a strong electric field forces electrons to move from the tip of t h e remaining
channel through the isolation region and into the drain. This field attracts all elec
trons that arrive at t h e tip. T h e n u m b e r of electrons that are available at t h e point of
j u n c t u r e of t h e channel with t h e isolation region is d e p e n d e n t u p o n the voltage drop
(V - V ). Therefore, the current depends on this voltage difference.
If we assume that the length of the pinched-off section of the channel at the
drain end is negligibly small c o m p a r e d to the channel length, L , and since the volt
age, V^ , appears across L, then it is reasonable to conclude that the drain current
for V > V
remains constant at the value it has w h e n V = V . We remind the
r e a d e r that V and V are m e a s u r e d with respect to the source, which we have
assumed is grounded.
For V = 0, pinchoff occurs at t h e drain e n d at a certain V . For negative val
ues of V , and since the channel gets pinched-off at a fixed channel-gate voltage,
pinchoff occurs at lower values of V and consequently the drain current at pinchoff
is smaller than that at V = 0, as shown in t h e sketch of the characteristic curves of
Fig. 10.5. A n appropriate question at this time is: W h a t combination of V and V
causes the onset of pinchoff?
A t pinchoff, the width of t h e depletion layer b e c o m e s equal to a, half the chan
nel width, shown in Fig. 10.1(b). The channel-to-gate voltage at pinchoff, V , is
d e t e r m i n e d from the expression for the width of the depletion layer given by E q .
(5.30). In fact, this voltage is equal to the magnitude of the gate voltage, for V = 0,
at which the channel is completely pinched off.
Since pinchoff occurs at t h e drain end first at V = V , the voltage drop
from channel to gate at the pinchoff point, d e t e r m i n e d by Kirchhoff s voltage law,
becomes (V
- V ). We therefore define a pinchoff voltage, V , as
SAT

SAT

SAT

SAT

SAT

SAT

SAT

CG

SAT

SAT

V = V

S A T

-V

(10.1)

We n o t e that the built-in 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

Junction Field-Effect Transistors

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 gate-channel junction is operating as a reverse-biased 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 built-in voltage of the gate-channel. 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 forward-biased diode b e t w e e n the gate and
D

SAT

SAT

Section 10.2

Current-Voltage Characteristic Equation 307

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)

w h e r e V is the built-in voltage d r o p from channel to gate across the depletion


layer, N is the channel doping, N (gate) N , and is the dielectric constant of
t h e semiconductor.
This drain-to-gate voltage, at which t h e channel is completely pinched-off, is
t h e pinchoff voltage V , so that
bi

V = qN aV(2s)
p

= V

SAT

+ V - V
bi

Figure 10.6 NJFET structure with symbols used in derivation.

(10.3)

308

Chapter 10

Junction Field-Effect Transistors

The expression for t h e resistance of the channel is given by


R

( ) = p- =
A

q[x N 2[a
n

- W(x)]Z

(10.4)
v

where is t h e mobility of t h e electrons, W(x) is the width of the depletion region of


each of the top and b o t t o m P N junctions at x, and 2Z[a - W(x)] is t h e crosssectional area of the channel n o r m a l to the direction of m o t i o n of electrons from
source to drain. We n o t e that t h e resistance of t h e channel varies with because the
width of t h e depletion layer d e p e n d s on x.
The current-voltage relations we will derive are idealized principally because
of t h e following assumptions:

1 . T h e channel length, L , is assumed to b e so large that even in the saturation


region of t h e characteristics, where high values of longitudinal electric fields
may exist, t h e mobility of the electrons is assumed to b e constant.
2. The rate of change of the transverse electric field, d% /dy, is m u c h greater than
d%Jdx, the r a t e of change of the electric field in t h e direction of m o t i o n of the
electrons, which we label the x-direction. To obtain an exact expression for the
depletion layer width, we n e e d to use Poisson's equation in two dimensions
since t h e r e are variations in the electric field in b o t h the and y directions. The
assumption, d% /dy d%Jdx, implies that the change in the depletion layer
width is a function of t h e voltage b e t w e e n gate and channel only, and it permits
one to use the one-dimensional Poisson equation in place of the two-dimen
sional equation. This assumption, which simplifies the mathematics consider
ably, is k n o w n as the gradual-channel
approximation.
y

D e r i v a t i o n of Current-Voltage 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)

where dV is t h e voltage d r o p across an increment dx in t h e longitudinal direction.


T h e depletion layer width at is found from E q . (5.30) by invoking the gradualchannel approximation as
x

W(x) = [2s(V

bi

- VJ/qNj^

(10.7)

where V is the built-in voltage across t h e channel-gate 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

Current-Voltage Characteristic Equation 309

The voltage, V , is t h e voltage d r o p from a point in the channel to the source.


Using E q . (10.8), E q . (10.7) is written as
x

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)

E q u a t i o n (10.12) shows t h e d e p e n d e n c e of the drain current on the applied


voltages, V and V , on the built-in voltage V , and on t h e physical p a r a m e t e r s of
the device. U p o n rearranging terms, it b e c o m e s
D

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)

where G = 2q^ ZaN /L


and V = qN a /2.^he
t e r m G represents the conduc
tance of a channel having width 2a, d e p t h Z , and length L. E q u a t i o n (10.14) is valid
only in the linear region, u p to pinchoff, as either V , V , or both are allowed t o
vary. B e y o n d pinchoff, for V > V , the drain current is assumed to saturate at the
value it has at pinchoff. T h e drain voltage at pinchoff, V , is d e t e r m i n e d from E q .
(10.2) to b e (V - V + V ). This can also b e verified from E q . (10.14) by finding
dI ldV
and setting it equal to zero. T h e characteristic curve for cutoff is found by
letting V
equal zero, which gives V = V - V at I = 0.
()

SAT

S A T

hi

SAT

bi

In the preceding derivation, we have used an idealized two-gate mode. The


resulting equation also applies to t h e one-gate model, whose structure is shown in
Fig. 10.1(d). Effect of the gate voltage is illustrated in Fig. 10.7. We now apply the
foregoing relations in t h e following example.

310

Chapter 10

Junction Field-Effect Transistors


V <0

V >0

U'(.v)

P~

X
x-O

x-L
(a)

(b)

Figure 10.7 Single-gate model showing (a) depletion layer and (b) pinchoff.

E X A M P L E 10.1
18

16

An channel silicon JFET has: N = 10 cm


N = 10 cm " , a channel length of 25,
= 250, and a = . Assume - 1300cm /V-s and determine:
a) The pinchoff voltage V .
b) The conductance G .
c) The built-in voltage.
d) The drain current at V = 0 and V = 4V.
e) V for V = - 2 V and for V = 0.
A

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

~ "

b) As defined in Eq. (10.14),


2q\i ZaN
n

GO -

19

10- S

kT
c) From Eq. (5.17), V

= l"

bi

G = 4.16

_ 2 X 1.6 10" X 1300 250 X t 0 " " 10"


~
25 X 10"

(N N /nf)
A

1q34
= .026 ^7^20 = 0.838V

d) By using Eq. (10.14), we have


I = 4.16
D

10

2 X 7.65 /0.838 + 4\yi


3
{ 7.65 I

153 /0.838\3/2
3 \ 7.65

6.73mA

Section 10.2

Current-Voltage Characteristic Equation 311

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
pinched-off 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 turn-OFF 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)

Since the current in the saturation region has b e e n assumed to b e i n d e p e n d e n t


of V , we d e t e r m i n e t h e expression for that current, 7 , by substituting the expres
sion for V
from E q . (10.2), (V + V - V ), in E q . (10.14) to yield
D

S A T

S A T

^SAT

bi

(10.16)

^ 0

3 V

V
P

E q u a t i o n (10.16) shows the d e p e n d e n c e of the drain current in t h e saturation


region on the voltages V and V , and on the physical properties of the device rep
resented by G and V .
O n the characteristic curves of Fig. 10.5, the cutoff, linear, and saturation
regions are clearly identified. The locus of the b o u n d a r y b e t w e e n the linear and sat
uration regions is shown by t h e dotted line, which represents the locus of all the
pinchoff points.
Q

bi

REVIEW QUESTIONS
Q10-1 For a fixed V , explain how the drain current in an N-channel JFET is reduced as the
gate voltage is made more negative.
Q10-2 For a fixed gate voltage, explain how the drain current is increased as the drain volt
age is increased.
Q10-3 Briefly define and explain in an equation the significance of the pinchoff voltage.
Q10-4 Explain the reason for using the gradual channel approximation.
Q10-5 What is the difference between the pinchoff voltage and the threshold voltage?
Q10-6 Explain how there is drain current after the channel is pinched off.
D

312

Chapter 10

Junction Field-Effect Transistors

HIGHLIGHTS

The JFET is basically a 3-terminal 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 N-channel device, is biased positively with respect to the source. The
gate, which in the N-channel JFET is a region, is biased negatively with respect to
the source and sits on top of the channel. By controlling the gate-to-drain 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
cross-sectional 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 built-in 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 pinched-off at
the drain end and the current reaches its saturation value 7 , corresponding to a
drain voltage V .
G

SAT

SAT

EXERCISES
E10-1 An N-channel double-gate 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 built-in
voltage and b) the pinchoff voltage.
15

18

Ans: a) V = 0.816V

b) V = 1.72V.

hi

E10-2 An N-channel double-gate silicon JFET operating at 300K has N = 10 " ,


N = 10 cm- , a = 0.5, L = 25, = 0.05cm, and = 1200cm /V - s.
Determine: a) the drain current for V = V with the gate connected to the source, b)
the gate voltage at which, for all V , the transistor is OFF.
16

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

defined as the slope of the I

- V

relations at a certain

^ d I

/ d V

(10.17)

By differentiating E q . (10.14) at constant V , we obtain t h e d e p e n d e n c e of g


and V as
G

on V

S = G {1 - [(V
d

bi

+ VD

V )/V r]
G

(10.18)

Section 10.3

Channel Conductance and JFET Transconductance

In the linear region and for values of V

( V V ), E q . (10.18) becomes

bi

(;

8 -G [l-[(V -V )/V r}
d

bi

313

(10.19)

E q u a t i o n (10.19) represents t h e slope of the characteristics near t h e origin, which


d e p e n d s o n F E T constants a n d V . Thus, t h e device may b e viewed as a variable
resistor whose resistance is controlled by V . Obviously, in t h e saturation region,
and since the current is constant, the channel conductance is zero.
The transconductance, g , relates t h e change in t h e drain current t o t h e
change of t h e gate voltage at constant V . In t h e saturation region, w e find g by
differentiating J
in E q . (10.16), with respect t o V at constant V , so that
(;

S A T

dip
dV Vr

- (^V/

constant

10 2

( - )

This expression is identical t o t h e equation for t h e channel conductance given by


E q . (10.19). It is to b e observed that the largest value of g and t h e largest value of
g a r e obtained w h e n V = 0. The transconductance, g is an important p a r a m e t e r
as it is a m e a s u r e of the voltage gain obtained w h e n using a B J T or a J F E T as an
amplifier.
In t h e following example, we illustrate the application of some of the relations.
m

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

V = 0.838V. G = 4.16 X 10~ S, and V = 7.65V


bl

By using Eq. (10.18) at V = 0 and V = 0, we find


G

= 4.16

10"

= 2.78

10" S

The channel resistance r - l/g = 360


b) By using Eq. (10.16), we have for the saturation current at V = - 2 V
d

4 A T = -W X 10"

S A T

I 7.65

= 3.58mA

c) By using Eq. (10.20), we determine the transconductance at V = 2V


G

= 4.16

10-

1 -

/2.838\o.5
= 1.62
7.65

X 10-*$

314

Chapter 10

Junction Field-Effect Transistors

10.4 S E C O N D A R Y EFFECTS
Channel-Length 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 base-width m o d u
lation of t h e B J T and is k n o w n as channel-length
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 gate-channel
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

Figure 10.8 Measured characteristics of an NJFET, indicating a nonzero slope in


the saturation region and breakdown.

Section 10.5

Small-Signal Equivalent Circuit

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 drain-source 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 current-voltage 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 small-signal low-frequency 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)

E a c h of the variables in E q . (10.22) is t h e sum of its value at t h e operating


point plus a small incremental change. We illustrate this by the drain current i ,
expressed as
D

316

Chapter 10

Junction Field-Effect Transistors


h

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)

By expanding the first t e r m on the right-hand side of E q . (10.24) and subtract


ing t h e second term, E q . (10.24) b e c o m e s

^^

1!

. ...

III.
v

VD

ds

VG

+ higher order terms

(J.U.Z3)

E q u a t i o n (10.25) includes non-linear terms represented by the higher o r d e r


terms. It can be linearized by neglecting the non-linear terms, as they are m u c h
smaller than the linear terms.
U p o n neglecting t h e higher o r d e r terms in E q . (10.25), it becomes
h

= 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
small-signal low-frequency 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 low-frequency
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 high-frequency 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 high-frequency
equivalent circuit of t h e J F E T . This circuit is m a d e u p of t h e low-frequency circuit
plus the capacitances.

Figure 10.9 Small-signal linear lowfrequency equivalent circuit of the JFET.

- \

Section 10.6

Figure of Merit of the JFET

317

5
Figure 10.10 High-frequency equivalent circuit of the JFET.

In t h e high-frequency equivalent circuit of t h e F E T shown in Fig. 10.10, two


capacitances are shown: C from gate to source and C from gate to drain.
Because of t h e reverse-biased P N junction, a distributed junction capacitance
exists all along t h e channel. It is, however, distributed unevenly due to the shape of
the depletion layer. In the equivalent circuit, we have arbitrarily divided this capaci
tance into two equal parts; t h e gate-to-source capacitance and the gate-to-drain
capacitance. It is actually the gate-to-channel capacitance.
T h e effect of t h e capacitances is to limit the high-frequency response. T h e dis
tributed capacitance of the channel is very difficult to d e t e r m i n e and o n e must
therefore obtain an estimate of t h e frequency limit by approximating t h e total
capacitance and the total channel resistance. In order to change t h e gate voltage,
o n e must charge this distributed capacitance through the associated channel resis
tance.
For t h e one-gate model, the capacitance across the P N junction at zero gate
voltage and at pinchoff has area equal to ZL and separation of average value of a/2,
assuming a gradual linear channel depth. H e n c e , the capacitance C is
gd

C = 2sZL/a

(10.27)

This capacitance replaces t h e distributed capacitance, and hence must be


located at t h e center of the channel. In this case, it charges t h r o u g h half t h e channel
resistance. The channel has length L and average area, Za/2. H e n c e , half the chan
nel resistance is given by

Za/2

q\i, ZaN
n

T h e charging time constant T is therefore


c

= RC

T h e limiting frequency f can


t

(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

Junction Field-Effect Transistors

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 short-circuit 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 short-circuit
current gain, / //., where / is the R M S value of the current through a short-circuit
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

The m a g n i t u d e of t h e current gain is unity at the frequency given by


fr = ~ s ~

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 one-junction 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,

The above expression is also labeled t h e gain-bandwidth


product, f , is also k n o w n
as the cutoff
frequency.
C o m p a r e d to the carrier transit time, the R C time constant given by Eq.
(10.29) is t h e feature that sets the high-frequency limit. The transit time of a carrier,
however, is d e p e n d e n t u p o n the i m p o r t a n t p a r a m e t e r s shown in E q . (10.33). The
transit time from source to drain, assuming a constant drift velocity % and assum
ing a uniform field so that % = V /L, has the very approximate form given by the
ratio of the channel length to the velocity as
T

where V is the drain-source 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 high-frequency 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 high-frequency 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 gain-bandwidth 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

High-Frequency Limitations

319

decreases for a fixed conduction velocity as L is reduced. These two effects


account for the presence of L in Eqs. (10.29) and (10.30). The smallest value of
L that one can use is limited by two factors:
2

i) Technological difficulties of fabrication have so far limited the lowest value


of L achievable.
ii) A t very low values of L , and hence a high value of % , the mobility ceases to
b e a constant and begins to decrease with an increase of V .
x

2. The use of semiconductors with a high mobility of carriers increases f . A higher


mobility implies a higher velocity and less transit time for t h e carriers.
r

3. W h e n we examine E q . (10.33), we n o t e that if the channel doping is increased,


the high-frequency response is enhanced. This is true so long as the channel
conductivity is not t o o high (high mobility is incompatible with high conductiv
ity).

REVIEW QUESTIONS
Q10-7 At values of V greater than V
and for short-channel JFETs, two phenomena
work in opposite directions, one to increase the current and another to decrease it.
Explain.
Q10-8 What is the major source of capacitance in the equivalent circuit of the JFET?
Q10-9 Why is the gate-to-source connection open-circuited in the small-signal 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 channel-length modulation.
Breakdown at the reverse-biased diode between gate and channel takes place beyond
saturation when the drain-to-gate 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 emitter-base 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 reverse-biased 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
E10-3 A double-gate N-channel silicon JFET operating at 300K has N = 1 0 c m ,
N = I0 cm" , and a = 1.5. Determine: a) the built-in 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

Junction Field-Effect Transistors

Ans: a) V = 0.894V
b) w' = 2.32

c) W = 2

bi

E10-4

a) Determine an expression for the maximum value of g of a NJFET as V is


changed. Neglect the built-in voltage,
b) Compare the ratio of g to the current at maximum g to that of a BJT. For the
NJFET, use the drain current, and for the BJT, use the collector current.
m

Ans: a) g = G
b) For FET gjl
m

= 3/V

and for BJT gjl

= 38.6

E X A M P L E 10.3
'

. / "

For the silicon NJFET of Example 10.1, determine:


a) The capacitance C .
b) The gain-bandwidth product f
r

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 =
=
~

C = 1.305pF. The corresponding charging resistance is calculated from Eq. (10.28) to be


480.7ohms.
b) From E q . (10.30)
14

14

4veL

19

16

1.6 X 10" X 1300 X 10 X 10


4 X 11.8 X 8.854 X 625 X 10
-14

f = 253MHz
T

PROBLEMS
Unless otherwise indicated, all devices are double-gate silicon at = 300k.
10.1 An N-channel 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 N-channel silicon JFET has N = 5x 10 cm" , N = 10 cm" , and a = 1.2.
Determine:
a)
the built-in voltage.
b)
the pinchoff voltage.
10.3 A P-channel 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)

the built-in voltage.


the pinchoff voltage.
the gate bias required to make the width of the undepleted channel equal to
0.15 at V = 0.
10.4 An N-channel silicon JFET has jV = 5 x 10 ~ , N = 10 cm- and a = 0.3.
Determine:
a)
the built-in voltage.
b)
the value of V that will cause pinchoff at the drain for V = 2V.
c)
the width of the undepleted channel for V = 1.5V, V = 0.
10.5 Use the data of Prob. 10.4 to determine an expression for the drain resistance of a
JFET in terms of G at V = 0 and
D

16

fl

18

-v/4

V =

-V /9

0
C = -2V
10.6 An N-channel silicon JFET has N = 10 cm- , N = I0 cm- , a = lm, L = 25,
and = 1mm. Determine:
a)
the built-in 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

for V < 0, V > 0

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 P-channel 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 N-channel silicon JFET has = 24, L = 4, a = 1.2, N = 10 cm" ,
N = 5 X 10 cm- . Use = 1200cm /V-s.
G

DS

19

15

At V = 3V, determine:
a)
the drain current at saturation.
b)
the drain voltage at saturation.
10.10 An N-channel silicon JFET is to be used as a voltage-variable resistor at very low
values of V . The channel has resistivity of 5 ohm-cm, = 1200cm /F-5,
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

Junction Field-Effect Transistors

10.11 An N-channel silicon JFET has N = 5 X 10 cm- , N = 10 cmr , and a = 1.2.


Determine the ratio Z/L to realize 7
= 2mA at V = 0. Assume the channel resis
tivity is 3 ohm-cm.
15

17

S A T

10.12 An N-channel silicon JFET has V = 4V, V = 0.8F and I


a)
G,
b)
the saturation current at V = 2V.
c)
the transconductance at V = 2V.
p

bi

DSS

= lmA(V

= 0).

chapter 11
METAL-SEMICONDUCTOR
JUNCTIONS AND
DEVICES

11.0 I N T R O D U C T I O N
Metal-semiconductor 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 N-type 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

Metal-Semiconductor Junctions and Devices


Metal

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

Energy-Band Diagrams of Metal and N-Semiconductor

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 N-type 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.

Thermal Equilibrium Conditions of M e t a l a n d


N-Semiconductor after C o n t a c t S c h o t t k y Barrier
A s a result of t h e net transfer of electrons from the semiconductor to the metal, a
depletion layer consisting of positively ionized d o n o r atoms is established in t h e sil
icon. This is coupled with a layer of excess electrons lining the surface of the metal
in contact with the semiconductor. A n electric field, directed from semiconductor to
metal is thus built and a potential barrier is established. This is accompanied, just as
h a p p e n s in the region of a P N junction diode, by bending of t h e energy levels of
the conduction b a n d and of the valence b a n d in silicon in the region of the depletion
layer. Since t h e electron affinity is a constant of the solid, the bending of the conduc
tion b a n d is accompanied by an identical bending of the vacuum level in t h e semi
conductor. B o t h in the metal and in the semiconductor, the locus of constant E is
not changed, it is still the energy that an electron has when it is free of the solid.
The bending of the conduction b a n d level of the semiconductor is largest at
the surface of contact with the metal where the electric field is largest. Since the
Q

326

Chapter 11

Metal-Semiconductor Junctions and Devices


Semiconductor

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.

Fermi energy level is fixed, t h e increase of energy separation b e t w e e n the conduc


tion b a n d and the Fermi level in the semiconductor n e a r the surface results in
reduced electron density at the surface with t h e metal. This is also the location in the
conduction b a n d where the electrons possess the highest energy.
After contact formation, the distribution of the density of electrons above the
b o t t o m of t h e semiconductor conduction b a n d is identical in both t h e metal and the

Section 11.1

Energy-Band Diagrams of Metal and N-Semiconductor

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 )

(ii-i)

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 built-in 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

Metal-Semiconductor Junctions and Devices


TABLE 11.1 Measured barrier heights for some metals on Si and
GaAs at 300K

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

Source: S.M. Sze, Physics of Semiconductor Devices, pp. 291-92,


copyright Wiley (1981). Reprinted by permission of John Wiley &
Sons, Inc.

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

We n o w consider conditions that ensue w h e n t h e equilibrium is disturbed by an


applied voltage. Since the built-in voltage at equilibrium appears only in t h e semi
conductor and since the metal cannot sustain any voltage drop, any voltage that is
applied appears entirely in the semiconductor. Therefore, the barrier height in the
metal, which appears in the metal b e t w e e n t h e Fermi level and the conduction b a n d
edge of the semiconductor at the surface, is unchanged. A s a consequence, the den
sity of electrons in the metal that have energies greater than this barrier is
u n c h a n g e d from its equilibrium value. This is shown in Figs. 11.3 and 11.4.

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.

Schottky Barrier Diode

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

A reverse bias increases the barrier height in the semiconductor, as shown in


Fig. 11.4, and reduces t h e current of electrons c o m p a r e d to that at thermal equilib
rium, which cross from t h e semiconductor to t h e metal. Because t h e metal does not
sustain a voltage drop, the barrier to electron flow from the metal to semiconductor
remains u n c h a n g e d and h e n c e the rate of the flow of charge from the metal to the
semiconductor is the same as it was at t h e r m a l equilibrium.
We conclude that with forward bias, the current from the metal to t h e semi
conductor, which results from electrons crossing from the semiconductor (labeled
the forward current I ), increases. With reverse bias applied, the current consists
only of the electrons that cross from t h e metal to the semiconductor and is a reverse
current.
m

330

Chapter 11

Metal-Semiconductor Junctions and Devices

Figure 11.5

1/Cj versus applied voltage.

Properties of Depletion Layer


The depletion layer in the semiconductor has properties that are similar t o t h e
region of a P N junction diode. Assuming an abrupt junction, where the charge den
sity in t h e depletion layer for 0 < < W is qN and the charge density and t h e elec
tric field are zero for > W, t h e expressions for the depletion layer width and the
electric field are obtained from Eqs. (5.30) and (5.20) respectively as
+

W = V(2e/qN )(V
D

_ V ) for N N

bi

(11.4)

%(x) = -(qN /s)(W-x)

(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

E q u a t i o n (11.7) can b e rewritten in the following form


(1/Cf)

= 2(V

hi

- V )/qeN
a

(11.8)

Assuming a uniform N t h r o u g h o u t t h e depletion layer, a plot of 1/Cj versus


the applied voltage, V , is shown in Fig. 11.5. The intercept o n t h e abcisse corre
sponds t o the built-in voltage and the slope of t h e line is a m e a s u r e of t h e doping
density, as shown by E q . (11.8).
D

E X A M P L E 11.1
ls

The silicon in an aluminum-silicon Schottky barrier diode has N = l() cm \ Determine the
built-in 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

Schottky Barrier Diode

331

(NJN )
D

= .0259 in

3 22 X lO ''
'
= 0.269V
5

The built-in 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

semiconductor. E n e r g y - b a n d diagrams for the metal and semiconductor before con


tact are shown in Fig. 11.6(a).
U p o n intimate contact, and because of the smaller work function of the metal,
electrons move from the metal into the semiconductor and the Fermi levels are
aligned.
The electrons that move into the semiconductor recombine with majority car
rier holes, leaving the negatively charged u n c o m p e n s a t e d acceptor ions in t h e semi
conductor to form a depletion layer whose width d e p e n d s u p o n the doping of the
semiconductor. O n the metal side, a layer of positive charge is formed at the inter
face with the semiconductor. A n electric field directed from the metal to the semi
conductor and an energy barrier are formed. The downwards bending of t h e energy
b a n d s in the semiconductor corresponds to an u p w a r d bending of t h e potential and
the built-in voltage, V , in the semiconductor is equal to ( - ) . The contact is
rectifying and the barrier from the metal to the semiconductor valence band, q<\>' , is
seen from Fig. 11.6(b) to b e (q x + E q ). This is t h e energy separation
b e t w e e n the aligned Fermi level and the surface of the valence b a n d in silicon, in
contact with t h e metal. Thus, the barrier, q V , is t h e energy that holes in the bulk of
the region n e e d to m o v e into the metal.
H e r e , we remind t h e reader that the electrostatic potential barrier for hole
motion is opposite to t h e direction of the barrier in the electron energy b a n d dia
gram. Therefore, the highest energy level of a hole in t h e semiconductor is at the
surface where the valence b a n d curve meets the metal surface. While holes at that
surface have t h e highest energy, they are not n u m e r o u s because of the larger sepa
ration of this level from the Fermi level c o m p a r e d to the separation experienced by
holes in t h e bulk of the semiconductor.
bi

332

Chapter 11
Metal

Metal-Semiconductor Junctions and Devices

Vacuum level

Semiconductor

lis

(a)

(b)

-qN

(c)

W
(d)

Figure 11.6 Metal P- semiconductor


with < (a) before contact,
(b) thermal equilibrium after contact,
(c) depletion layer charge density, and
(d) electric field.

A comparison b e t w e e n t h e energy b a n d diagrams for a rectifying metal-N


semiconductor contact and a metal-P semiconductor contact is shown in Fig. 11.7.
W h e n a voltage, V , is applied that m a k e s the semiconductor positive with
respect to the metal (forward bias), the barrier to hole flow from the semiconductor
to t h e metal is reduced. The reduction of barrier height is shown in Fig. 11.7. W h e n
the bias is reversed, making the metal positive, the barrier to hole flow from the

Section 11.3

Current-Voltage Characteristics of N-Semiconductor Schottky Diode


N-type semiconductor

333

P-type semiconductor

(c)
Figure 11.7 Energy band diagrams for metal N-semiconductor (<I> > ) and
metal semiconductor ( < ) at (a) equilibrium, (b) forward bias, and
(c) reverse bias.
m

semiconductor to t h e metal is increased and the barrier in t h e metal at the interface


is u n c h a n g e d so that t h e current is negligibly small.
In conclusion, it is important to n o t e that in b o t h t h e metal-N rectifying con
tact and t h e metal-P rectifying contact, t h e forward current is due to the injection of
majority carriers from t h e semiconductor to the metal. This is in contrast to t h e P N
junction diode w h e r e t h e currents consist of the injection of minority carriers.

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
N-SEMICONDUCTOR SCHOTTKY DIODE
Simplifying A s s u m p t i o n s
Before proceeding with analytical relations for t h e current-voltage 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

Metal-Semiconductor Junctions and Devices

3. A t t h e semiconductor surface, the average electron velocity c o m p o n e n t


directed toward the metal has b e e n shown* to b e approximately v /4, where v
is t h e t h e r m a l velocity.
tb

(h

4. A certain n u m b e r of electrons in t h e metal possess sufficient t h e r m a l energy to


cross the barrier and move into the semiconductor.

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

where E is t h e conduction b a n d edge in t h e bulk of the semiconductor.


The electron density in t h e bulk of t h e semiconductor, away from the surface,
which for n o r m a l doping is equal to t h e density of the d o n o r atoms N , is found
from
c

= N exp [-(E
c

- E )/kT]

(11.11)

where the relation b e t w e e n and N is defined by E q . (3.18).


By substituting E q . (11.10) in E q . (11.9), we can write
c

n = N [exp (-qVJ/kT]
s

[exp -{E

By substituting t h e expression for N

- E )/kT]

(11.12)

from E q . (11.11) in E q . (11.12), we write

n = N exp(-qV /kT)
s

(11.13)

Since all electrons at t h e surface of the semiconductor in the conduction b a n d


are assumed to cross into the metal with a m e a n velocity of magnitude v J4, the
current of electrons that cross from the semiconductor to the metal (directed from
metal to semiconductor), I , at t h e r m a l equilibrium is given by
t

ms

I^qAnpJA

(11.14)

where A is the cross sectional area of the junction and v is t h e t h e r m a l velocity of


the electrons in the direction from the semiconductor to the metal.
th

*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

Current-Voltage Characteristics of N-Semiconductor Schottky Diode

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

exp {-q<B /kT)

(11.17)

W h e n a forward bias is applied to t h e diode and since all t h e applied voltage,


V , appears in t h e semiconductor, the expression for the surface density of electrons
given by E q . (11.13) can be written as
a

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

n = [N exp (-q<i> /kr)]


5

[exp qVjkT]

(11.19)

In t h e expression for the current of electrons that cross from t h e semiconduc


tor to t h e metal, given by E q . (11.14), we replace n by its equivalence from Eq.
s

336

Metal-Semiconductor Junctions and Devices

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

kT) exp (qVjkT)

(11.20)

The current of electrons from the metal to the semiconductor at t h e r m a l equi


librium, with reverse bias or with forward bias, is always the same and is given by
E q . (11.15). The total current is the sum of t h e two currents, the current of electrons
that cross from the metal given by E q . (11.15) with n replaced by its equivalence
from E q . (11.9) and t h e current of electrons that cross from t h e semiconductor
given by E q . (11.20) so that
s

I = 4

+ ' s m = qA(vj4)N

exp

(11.21)

(-q<S> /kT)
B

The m e a n t h e r m a l velocity, defined by E q . (4.1), is proportional to


, from E q . (3.20), is proportional to T , so that we can write

112

and

/ = ART [exp
where R is qv N

(-<& /kJ)]
B

[exp(qVjkT)

- 1]

(11.22)

/4.

th

E q u a t i o n (11.15) can b e written as


I = I [txp(qVjkT)-l]

(11.23)

where for reverse bias I = I and I is given by


s

/ = ART
s

exp {-qiS> /kT)

(11.24)

The t e r m R, k n o w n as t h e Richardson constant, depends on t h e electron effec


tive mass and other physical constants and has units of Acm~ K~ . The value of R
depends on t h e particular semiconductor used and has an approximate value of
1 1 0 A c m ~ K ~ in N-type silicon and an approximate value of & 4 c m ~ K r for N-type
GaAs.
2

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 N-Si, 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 = = l|xs.
16

19

16

Section 11.4

Comparison of Schottky Diode with P N Diode

337

a) Calculate the Schottky diode current at a forward bias of 0.4V at 300K.


b) Determine the value of the forward bias that has to be applied to the PN junction diode to
deliver the same current as the Schottky diode.
Solution
a) From Table 11.1, the barrier is 0.85eV. Using 110Acm~ Kr for the Richardson constant, the
reverse saturation current of the Schottky diode is found from Eq. (11.24) as
2

I = " X 110 X (300) exp (-0.85/.0259) = 5.53 X 1 0 ~ A


s

For a forward bias of 0.4V, Eq. (11.23) predicts a current of 0.28mA.


b) For the P N diode, is calculated from Eq. (4.16) and I), is found to be 11.56 cm /s. Since the
diode current is essentially the hole injection current from to N, the saturation current is
given by qD p A/L ,
where L = 3.4 X 10 cm and /<.. = lo'cm .so that / = 5.44 X 10 A .
To obtain a current of 0.28mA. the diode forward voltage has to be 0.64V.
+

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 low-voltage, high-current rectifiers.
The lower cut-in 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)

Current mainly due to


electrons from
semiconductor to metal

from metal
Figure 11.8 Current-voltage characteristics of PN and Schottky-barrier diodes
and symbol for SBD.

338

Chapter 11

Metal-Semiconductor Junctions and Devices

to a factor of 10. A s we have seen in discussing the switching of a B J T in saturation,


there is a large injection of minority carriers into the base. This injection is responsi
ble for slowing the switching. With the Schottky diode turning on at a very low for
ward voltage, t h e collector-base junction is clamped at about 0.4V and the Schottky
diode provides a bypass (bypassing the base) for the excess minority carriers.
A n o t h e r major advantage of the Schottky diode lies in the fact that the cur
rents are majority carrier currents and hence t h e storage capacitance associated
with minority carrier currents does not exist. Therefore, the transient response of
the diode is d e t e r m i n e d only by the RC product, w h e r e C. is the depletion layer
capacitance and R is the diode series resistance. Consequently, w h e n the diode is
switched from a forward ( O N ) state to a reverse ( O F F ) state, the current tends to go
to zero as soon as the driving voltage is switched to the reverse value. The reverse
recovery time is therefore d e p e n d e n t u p o n RC. This product is usually about four
orders of magnitude smaller t h a n that of a corresponding P N junction diode.
T h e reverse recoveries of a P N and a Schottky diode are compared in
Fig. 11.9.
The main disadvantage of the Schottky diode is a consequence of the p r o p e r t y
that permits a low cut-in voltage or that permits a high current for a low voltage. The
j

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

Comparison of Schottky Diode with P N Diode

339

disadvantage is the large current in t h e reverse direction, which is approximately


four orders of m a g n i t u d e greater t h a n that of a P N junction diode.
+

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 metal-semiconductor
contact.
Q l l - 5 Why is there no diffusion (storage) capacitance in a Schottky-barrier diode?
Q l l - 6 Why is the Schottky-barrier 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

Metal-semiconductor contacts are either rectifying or ohmic. Rectifying contacts per


mit current in one direction only, whereas ohmic contacts allow easy current flow in
both directionsfrom metal to semiconductor and from semiconductor to metal.
Whether a contact is rectifying or ohmic is determined by the relations between cer
tain energy levels in the semiconductor and in the metal.
The energy levels of relevance are: The Fermi levels in both the metal and the semi
conductor, and the affinity level in the semiconductor, which is the energy separation
between E and E .
The work function is defined as the separation between E and E We note that for a
metal, E is, for all practical purposes, located at E , whereas for a semiconductor, the
work function varies with the doping because the energy level E is determined by the
doping.
For a metal-N semiconductor contact, wherein the metal work function is greater than
that of the semiconductor, application of a bias that makes the metal positive with
respect to the semiconductor causes electrons to move from the semiconductor to the
metal and very few electrons to move in the opposite direction, so that a current
results from metal to semiconductor. Reversing the bias does not change the height of
the barrier that electrons in the metal face, so that the same number cross as at for
ward bias. However, very few electrons cross from the semiconductor to the metal and
hence a small current results. The contact is thus rectifying and the device is a
Schottky-barrier diode.
Q

Schottky-barrier diodes may also be formed from a metal and a semiconductor


when the metal work function is smaller than that of the semiconductor. For both the
and semiconductors, the contacts with the metal are ohmic when the work func
tion differences are opposite to those of the rectifying contacts.

For the same applied voltage, the Schottky-barrier diode (SBD) conducts considerably
larger forward and reverse currents than a PN junction diode.

340

Metal-Semiconductor Junctions and Devices

Chapter 11

The PN junction diode is a minority carrier device, whereas the SBD is a majority car
rier device.

EXERCISES
15

-3

E l l - 1 An SBD formed on N-silicon is operating at 300K. Given N = 2x 1 0 c m , the


affinity of silicon is 4.15eV, and the metal work function is 4.9eV. Determine: a) the
built-in voltage, b) the barrier height and c) the width of the depletion layer with
D

0.
Ans:

a) V = 0.5V
bi

b) barrier height = 0.75V


c)W=

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 metal-semiconductor junction
permits easy current flow w h e n t h e junction is forward-biased. 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 metal-semiconductor junctions is in the forma
tion of non-rectifying 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

Non-Rectifying Ohmic Contacts

341

Figure 11.10 Energy band diagrams for metal N-semiconductor contact with
<E> < (a) before contact, (b) after contact, (c) negative bias on the
semiconductor, and (d) positive bias on the semiconductor.
m

and semiconductor w h e n isolated and w h e n in contact are shown in Fig. 11.10(a)


and Fig. 11.10(b).
U p o n intimate contact and, because of t h e smaller work function of t h e metal,
electrons flow from the metal to t h e semiconductor until t h e r m a l equilibrium is
reached and the Fermi levels are aligned. Negative charges of electrons accumulate
on the semiconductor surface adjacent to t h e metal and the metal surface b e c o m e s
positively charged. After the alignment of the Fermi levels, a potential d r o p equal to
<& - is developed across the semiconductor and the energy b a n d s in the semi
conductor b e n d downwards at the surface.
The electrons that crossed from the metal to the semiconductor settled in the
semiconductor at the junction. N o depletion region exists in the semiconductor
since there is n o barrier to electron flow from the semiconductor to the metal nor in
the opposite direction.
We observe an important difference b e t w e e n t h e ohmic contact and the
Schottky contact, b o t h m a d e from a metal and an semiconductor, in that the
energy b a n d s b e n d d o w n w a r d in the ohmic contact while they b e n d u p w a r d in
the Schottky contact. The difference is manifested also in that the semiconductor
charge consists of free electrons in t h e ohmic contact while positive d o n o r ions
accumulate at the surface of the rectifying contact. In the ohmic contact and
because of the presence of excess electrons in the semiconductor surface, the
majority carrier concentration is e n h a n c e d n e a r the surface c o m p a r e d to the bulk.
Since n o barrier is developed in the semiconductor, an applied bias appears across
the neutral semiconductor, as shown in Fig. 11.10(c) and (d). A s a follow u p to the
previous discussion, and for a metal-P semiconductor contact with > , there is
m

342

Chapter 11

Metal-Semiconductor Junctions and Devices

n o barrier to hole flow from t h e semiconductor to t h e metal and a negligible barrier


to hole flow in the opposite direction. We conclude that such a contact is ohmic.
Tunnelling a t a M e t a l - S e m i c o n d u c t o r Contact
A m o r e commonly used m e t h o d for establishing an ohmic contact b e t w e e n a metal
and a semiconductor is to m a k e the contact favorable to the tunnelling of electrons
in b o t h directions. The contact is favorable w h e n electrons are not required to climb
a barrier, as they can cross directly from the metal to t h e conduction b a n d of t h e
semiconductor and also cross in the reverse direction.
A t the end of E x a m p l e 1.4 we considered an electron located in Region 1, hav
ing energy W separated from Region 3 by Region 2, which has a barrier of height
greater t h a n W. We concluded that there is a finite probability for the electron to
cross into R e g i o n 3 without going over the barrier, but r a t h e r by "tunnelling"
through the barrier of R e g i o n 2. T h e probability of tunnelling d e p e n d s on the width
of t h e middle region.
The n a r r o w e r t h e depletion layer at a junction, the higher is the probability of
tunnelling. For two semiconductor regions in contact, the width of the depletion
layer, given by E q . (5.28), is d e t e r m i n e d mainly by the doping of the m o r e weakly
d o p e d region. The higher t h e doping of b o t h regions, the smaller t h e width of the
depletion layer W. A t a metal-N semiconductor contact, the depletion layer exists
only in t h e semiconductor. To obtain a small W, the semiconductor is d o p e d very
highly and possibly to the degree of degeneracy at which the Fermi level is m o v e d
into the conduction band. For tunnelling to take place, it is necessary for t h e con
duction b a n d of one material to b e located, energy-wise, opposite empty states in
the conduction b a n d of t h e o t h e r material.
A t t h e r m a l equilibrium, t h e Fermi levels of the metal and semiconductor are
aligned. With reverse bias applied, the Fermi level of t h e semiconductor is
depressed so that some electrons in t h e conduction b a n d of the metal are at the
same energy level at, and across, the tunnel, from empty states in t h e conduction
b a n d of the semiconductor, as shown in Fig. 11.11(a). This results in the tunnelling of
these electrons and easy current flow from the semiconductor to the metal.
W h e n a forward bias is applied, making the metal positive with respect to t h e
semiconductor, electrons in the conduction b a n d of the semiconductor are opposite
empty states of the conduction b a n d of the metal, which m a k e s it possible for elec
trons to tunnel across. T h e bending of t h e energy b a n d s w h e n forward bias is
applied is shown in Fig. 11.11(b).
We conclude from the above discussion that a tunnelling ohmic contact is
obtained if the semiconductor is heavily doped. Normally, the doping of t h e semi
conductor is d e t e r m i n e d by other considerations, such as its function in a device. To
form a tunnelling contact to a lightly d o p e d semiconductor, a thin heavily doped
layer of the same conductivity is formed over t h e lightly d o p e d semiconductor. For
example, an ohmic contact to a lightly d o p e d semiconductor can b e formed by
depositing a heavily d o p e d N layer over the surface. The heavily d o p e d layer is
formed by ion implantation, or epitaxy. Such a contact generally possesses a linear
current-voltage characteristic, as shown in Fig. 11.12(b).
+

Section 11.5
Metal

Non-Rectifying Ohmic Contacts

343

Semiconductor
Electron tunnels

-c

(a)

Electron tunnels

Figure 11.11 Tunnelling at a metal


semiconductor junction with (a) reverse
bias and (b) forward bias.

(b)

Metal

/V -region

-type
semiconductor

(a)

(b)
+

Figure 11.12 (a) Band diagram for a metal-N -N~ ohmic contact and (b) currentvoltage characteristics of a Schottky barrier diode and an ohmic contact.

344

Chapter 11

Metal-Semiconductor Junctions and Devices

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 gain-bandwidth 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 semi-insulating 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 ohm-cm. It is thus k n o w n as semi-insulating gallium arsenide. Devices
and circuit-connections m a d e on substrates of semi-insulating 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

Semi-innsulting GaAs substrate


Mesfet

Schottky diode

Figure 11.13 Cross section of a MESFET and a Schottky barrier diode.

Section 11.6

The MESFET

345

diode. In these implants, silicon is the d o p a n t of choice and it behaves as N-type in


G a A s w h e n introduced by ion implantation. A density of 1 0 c m ~ is used for the
channel and the diode base while densities of 1 0 c m ~ are used in t h e drain, the
source, and t h e N region of the diode. T h e N region of t h e diode is n e e d e d to
m a k e a good ohmic contact to the base.
Following the implantation and to cure t h e d a m a g e caused by the implanta
tion, t h e wafer is annealed at a t e m p e r a t u r e of about 850C. A p r o b l e m arises in this
process. Annealling G a A s at this t e m p e r a t u r e leads to its decomposition since
arsenic evaporates at 600C ( G a A s melts at 1238C). To prevent this, G a A s wafers
are protected with a thin layer of silicon nitride prior to t h e implantation.
O n c e the implantation and annealing are completed, again by using photolith
ography, ohmic contacts are m a d e to the drain, the source, and the diode base. A
sequence of electron b e a m evaporation of A u / G e / N is used for the metal contacts,
and t h e contacts are then annealed for several minutes at 400C. A l u m i n u m is used
for the Schottky diode barrier using electron b e a m evaporation. The device has N
source and drain and N~ channel.
A major advantage of M E S F E T s is their usefulness at very high frequencies
because channel lengths of t h e o r d e r of and less can b e fabricated. In o r d e r to
m a k e it possible to deposit ohmic metallic lines of this dimension, special m e t h o d s
must b e used. O n e of the m o r e c o m m o n methods, k n o w n as t h e lift-off method, p r o
vides a powerful technique for defining a high resolution p a t t e r n for large scale
integrated circuits. In this process, aluminum is used for the gate of t h e M E S F E T .
Lift-off is accomplished, as shown in Fig. 11.14, by first spreading a thin film of
photoresist over t h e areas of the gate that are not to be covered by the metal. In this
way, photoresist is used in a reverse role to that used in C h a p t e r 6 in order to trans
fer patterns o n t o substrates. The photoresist is then developed as is normally done.
The thin aluminum film is then deposited over the resist, as shown in the figure.
Finally, the photoresist is r e m o v e d by a solvent, which does not affect the aluminum,
so that the metal covers only the gate.
12

15

Photoresist

Silicon wafer

Metal film

Metal
Figure 11.14 Illustrating the lift-off
technique.

346

Chapter 11

Metal-Semiconductor Junctions and Devices

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 (normally-off) type
and depletion (normally-on) 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 depletion-type 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 N-channel device. In the e n h a n c e m e n t device,
the built-in voltage causes the channel to be completely closed at zero gate-source
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 built-in 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

Schottky-Barrier gate
Source

1
,\'+
11

/V-channel

a
Semi insulting GaAs substrate

Figure 11.15 Depletion-type device


showing the depletion region.

hi

Section 11.6
V

The MESFET

(11.25)

347

a qN /2e
D

If (with V = 0) the built-in voltage causes the depletion layer to extend


across only p a r t of t h e channel width (as in the depletion type), and since the
required voltage that will close the channel is V , then a negative voltage must be
applied to t h e gate to close t h e channel b e t w e e n source and drain. This voltage is
labeled t h e threshold voltage, V , and we can write
D

(11.26)
However, w h e n t h e built-in 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 built-in 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 drain-source
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

Metal-Semiconductor Junctions and Devices


V =0
V><0

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 gate-source 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 current-voltage 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 Current-voltage 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 current-voltage characteristics shown in Fig. 11.17, we will intro
duce the effect of the drain-source 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 depletion-type 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

where a positive V and a negative V cause t h e depletion layer width to increase.


The higher t h e value of V , the less is the magnitude of V that will pinch the chan
nel at the drain.
For V = 0, the channel b e c o m e s pinched-off at the drain end w h e n
D

= 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 pinched-off 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

Metal-Semiconductor Junctions and Devices

given device. It is obvious that for V = 0, t h e channel is completely closed when


V = V and, consequently, t h e drain current is zero, as shown in Fig. 11.16.
In a n enhancement-type
device, the built-in voltage, V , is equal to o r greater
than t h e pinchoff voltage, V^.The drain current begins to flow w h e n V exceeds V
and V > 0. The relation for t h e threshold voltage is t h e same for both the depletion
and e n h a n c e m e n t - t y p e devices, as given by E q . (11.26) and r e p e a t e d here as
D

bi

V =V -V
T

(11.29)

E q u a t i o n (11.27) applies t o t h e enhancement-type device as well as indicating


that, at values of V > V and as V is increased (V > V > 0), t h e value of t h e
drain voltage at pinchoff increases and thus t h e current is larger, as shown in Fig.
11.17.
The current-voltage relations that we derived in Chapter 10 for t h e J F E T are
equally applicable to t h e long-channel M E S F E T , provided that t h e expression for
G is o n e half the value. This is to account for the reduction of t h e channel width
from 2a t o a. E q u a t i o n (10.14) for t h e linear region and E q . (10.16) for saturation
are r e p e a t e d h e r e as
G

2Vp(V ,

V \V2

Vr.

VP

'SAT

^ 0

v..

+ v</