Transistor Circuit Theory

TRANSISTOR
CIRCUIT THEORY
AND DESIGN
TRANSISTOR
CIRCUIT THEORY
AND DESIGN
John Franklin Pierce
Professor of Electrical Engineering
University of Tennessee
Knoxville, Tennessee
Consultant, Instrumentation and Controls Division

Oak Ridge National Laboratory
Oak Ridge, Tennessee
CHARLES E. MERRILL BOOKS, INC., Columbus, Ohio.

Copyright @ 1963, by Charles E. Merrill Books, inc.,
Columbus, Ohio. All rights reserved. No part of this
book may be reproduced in any form, by mimeograph
or any other means, without the written permission of
the publishers.
Library of Congress Catalog Card Number: 63 :15292
First Printing December, 1963

Second Printing . January, 1965
PRINTED IN THE UNITED STATES OF AMERICA

To
Edythe,
Virginia,
and
George
PREFACE
The principal objectives of this book are (1) to present the theoretical
aspects of transistor circuit analysis, (2) to reduce this theory to practice,
and (3) to present the material so clearly that the intended reader can
understand at least seventy-five per cent of the book without outside
assistance. Accomplishing these objectives without sacrificing the high-
level material that should be in a good text has been a difficult but a
rewarding task for both the author and his students.
The first objective was accomplished by including a discussion of both
the physical operation of semiconductor devices and the general theory of
linear active networks. In the first chapter, the reader will obtain a clear,
concise picture of the atomic structure and the characteristics of semi-
conductors; current conduction also is included. Chapter 2 contains both
a discussion of the physical concepts of semiconductor diodes and the
mathematical approach to characterizing these diodes as circuit elements.
Several diode circuits are included at this point to illustrate how diodes are
used. In Chap. 3, the diode discussion is extended to junction transistors.
General transistor equations are introduced, and these equations are
correlated with the physical operation of the transistor for various bias
conditions. These three chapters provide a firm physical and mathematical
foundation for the circuit work that follows.
Chapters 4 through 7 form a section on basic transistor circuit analysis.
The two-port black box is discussed briefly in Chap. 4. This discussion is
followed by a development, using impedance parameters, of the equivalent-
T circuit. This circuit is then extended to the three most useful transistor
connections. Theoretical equations are derived for all three configurations.
These equations are reduced to the approximate forms that are so widely
used in industry. A parallel study of the hybrid parameters and hybrid
equivalent circuits is contained in Chap. 5. These discussions are restricted
to the mid-frequency region. With the study of bias and bias stability in
vii
viii PREFACE
Chap. 6, the reader is in a position to design multistage circuits for

applications in the low- and mid-frequency range. Using the basic material
in Chap. 3, the inherent high-frequency limitations of transistors are
discussed in Chap. 7. Then the hybrid-n equivalent circuit is developed
and interpreted. By the end of this chapter, the reader has extended his
knowledge of transistor circuits to include the complete frequency analysis
of RC-coupled multistage amplifiers.
The remaining chapters arc devoted to four special classes of amplifiers.
Feedback amplifiers are very important as functional building blocks in
practical circuits. Therefore. a thorough discussion of several different
types offeedback circuits is included in Chap. 8. Tuned amplifiers are con-
sidered in Chap. 9. Low noise amplifiers. normally neglected in electronic
texts but often very important in industrial and experimental circuits, are
treated in Chap. 10. Finally, power amplifiers with their associated thermal
problems are illuminated in Chap. II. The transistor circuit theory
included in this book, while not exhaustive. is "ufficient to form a firm
basis for work in transistor CIrcuit design and ror higher level theoreti-
cal work.
The second and third objectives listed above were achieved by a liberal
use of examples and problems, most of which were obtained from industrial
circuits. The theory was not considered to be clear unless it could be used
to solve practical problems. At the end of almost every section, the reader
will find one or more clearly worked out examples illustrating the important
points in that section. Examples are usually followed by problems with
answers and then by problems without answers. This arrangement enables
the reader to lise his study time most efficiently. At the end of each chap-
ter the reader will find addibonal problems that will reinforce his newly
acquired knowledge and also extend his appreciation of the functional
relationships. The reader will find that most of the examples and problems
have a refreshingly practical flavor. After completing the book, the diligent
student will find that he has an excellent practical and theoretical perspec-
tive of the transistor circuit field.
The third objective was also accomplished by using the material
in several sections of a transistor course at the University of Tennessee.
Both the style of presentation and the content were changed until the
student indicated that he could easily understand the material. This
approach required a brief review of network theorems prior to their use in
the text and also some repetition of the basic material in various parts of
the book. The practicing engineer will also appreciate this approach.
Many other teaching aids are incorporated into this text to facilitate
the learning process. The author's sincere wish is that this book will make
a significant contribution both as a text and as a self-study technical book
for practicing engineers, scientists, and technical managers.
PREFACE ix
The author wishes to express his appreciation to the many people, too
numerous to name here, who have contributed to this book. Mr. B. J.
Moore should receive special recognition as the co-author of Chaps. 2, 4,
5 and part of 6. Mr. E. J.: Kennedy wrote the final draft of Chap. 6.
Messrs. T. V. Blalock and R Nlitt made several contributions to Chap. 10,
and Mr. J. W. Waller helped prepare some. of the material in Chap. 11.
Thanks are due to Professors Frank F. Carden, Basil L. Cochrum, Alfred
A. Pandiscio and Howard E. Tompkins for their critical reviews and valu-
able suggestions. The author wishes to thank Mrs. Edythe Pierce and Mrs.
YirginiaKnies for typing the manuscript, and Mr. John M. Haberman,
Technical Editor (Production), and Mr. Richard Summers, Art Director,
for the excellent editorial and art work during the production of this book.
J. F.P.
CONTENTS
Chapter I
SEMICONDUCTOR PHySiCS .
1.1. Energy levels in the isolated atom ~. . . . I

1.2. Energy levels in crystals. . . . . . . . . . . . . . . . . . . . . . . 5
1.3. Electrical characteristics of a conductor. . . . . . . . . 8
104. Electrical characteristics of a semiconductor ..... 11
1.5. Electrical characteristics of an insulator . . . . . . . . . 15
1.6. Semiconductors with impurities . . . . . . . . . . . . . . . . 17
1.7. Diffusion current 21
Chapter 2
JUNCTION DIODES......................... 25
2.1. The potential barrier. . .. . .. . .. .. . .. . . .. . .. .. . . 25

2.2. Rectification................................. 30
2;3. Electrical characteristics of diodes. . . . . . . . . . . . . . 37
204. Speed of response . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.5. Diode equivalent circuit ,.. 46
2.6. Breakdown (Zener) diodes ~.. . . . .. . . . . . . 48
2.7. Tunnel diodes. . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 57
2.8. The backward diode. . . . . . . . . . . . . . . . . . . . . . . . . . 67
xi
xii CONTENTS
Chapter 3
THE JUNCTION TRANSISTOR. 72
3. I. Static characteristics 73
3.2. Transistor equations 79
3.3. Physical operation 82
3.4. Transistor amplifiers 89
3.5. Transistor types 92
Chapter 4
SMALL-SIGNAL LOW-FREQUENCY
AMPLIFIERS USING IMPEDANCE
PARAMETERS ... 103
4.1. Black box theory 103

4.2. T-equivalent circuits 113
4.3. Variation of T-parameters 132
4.4. Cascaded amplifiers 136
Chapter 5
SMALL-SIGNAL LOW -FREQUENCY
AMPLIFIERS USING HYBRID PARA-
METERS . 148
5.1. The general hybrid equivalent circuit 148

5.2. The common-base amplifier. . 153
5.3. The common-emItter amplIfier ISS
5.4. The common-collector amplifier 161
5.5. Hybrid-impedance rdationships 165
5.6. Amplifier performance charactenstics 168
Chapter 6
D-C BIAS AND BIAS STABILITy....... 177
6.1. Selection of the operating point for linear amplifiers 177

6.2. Effect of J,u on the operating point 180
6.3. Effect of V HI; on the operatmg point . . . . . . .. 184
6.4. Effect of h FE on the operating point .. 186
6.5. Effect of E{'{'on the operating point . . . . . . .. 188
6.6. Linear stabilization hy stabIlity factor an,i1ysis. . .. 189
6.7. Nonlinear compensallon 201
6.8. Differential compensation . .. 211
CONTENTS xiii
Chapter 7
HIGH.FREQUENCY EQUIVALENT
CiRCUiTS 223
7.1. Major high-frequency effects in transistors , 223

7.2. The modified T-equivalent circuit. . . . . . . .. . . . . .. 226
7.3. The hybrid-n equivalent circuit. . . . . . . . . . . . . . . . . 227
7.4. Characteristics of hybrid-n parameters. . . . . . . . .. 234
7.5. High-frequency parameter relationships. . . . . . . .. 237
7.6. Application of hybrid-n parameters ;...... 240
Chapter 8
NEGATIVE FEEDBACK AMPLIFIERS .. 255
8.1. Single-stage feedback. . . . . . . . .. . . . . . . . . . . . . . . .. 255

8.2. Two-stage feedback amplifiers. . . . . . . . . . . . . . . . .. 270
8.3. Multistage feedback amplifiers. . . . . . . . . . . . . . . .. 285
Chapter 9
TUNED AMPLIFIERS ;............... 293
9.1. Tuned circuits 293

9.2. Single-tuned amplifiers. . . . . . . . . . . . . . . . . . . . . . .. 297
9.3. Frequency selectivity. . . . . . . . . . . . . . . . . . . . . . . . . . 300
9.4. Cascaded stages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
9.5. r-f amplifiers 306
9.6. Neutralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 307
Chapter 10
LOW NOISE AMPLIFIERS 312
10.1. Origin and classification of noise . . . . . . . . . . . . .. 312

10.2. Noise models. . . . . . . . . . . . . . . .. . . . . . . . . . . . . .. 321
10.3. Figures of merit. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 331
lOA. Low noise amplifier design. . . . . . . . . . . . . . . . . . . 337
10.5. Low noise amplifier examples 341
xiv CONTENTS
Chapter II
POWER AMPLIFIERS 349
II. I. Power transistor characteristics. . . . . . . . . . . . . . .. 349

11.2. The thermal problem .. . . . . . . . . . . . . . . . . . . . . .. 351
11.3. The distortion problem. . . . . . . . . . . . . . . . . . . . . .. 354
11.4. Class A power amplifiers. . . . . . . . . . . . . . . . . . . .. 359
11.5. Thermal stability. . . . . . . . . . . . . . . . . . . . . . . . . . .. 370
11.6. Push-pull power amplifiers............... ..... 374
Appendix I
REFERENCES.. . .. . .. 387
Appendix II
LIST OF SyMBOLS..... .. .... .. .. .. . .. .... 395
Appendix III
TABLES...................................... 401
I. Table of elements 402

1/. Useful constants . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 403
1/I. Physical values for intrinsic germanium and silicon 403
IV. Matrix conversions. . . . . . . . . . . . . . . . . . . . . . . . . .. 404
V. Circuit performance equations. . . . . . . . . . . . . . . .. 404
VI. Parameter interrelationships. . . . . . . . . . . . . . . . . .. 406
Appendix IV
FREQUENCY RESPONSE ANALYSIS. 407
INDEX 413
Chapter 1
SEMICONDUCTOR
PHYSICS
Transistors are made from semiconducting materials. As the name

implies, semiconductors are neither good conductors nor good insulators.
In this chapter our first goal is to develop an understanding of the conduc-
tion process in solids and the factors that influence this process. We will
then study the electrical characteristics of specific semiconductor materials
normally used in diode and transistor work.
This study of semiconductor physics is intended to be a qualitative
review of the present theory. The reader is referred to the many excellent
books on this subject for a more rigorous treatment.
1.1. ENERGY LEVELS IN THE ISOLATED ATOM
A study of the current conducting process in solids must begin with the
atomic structure of solids. Let us consider first the isolated atom. The
hydrogen atom is the easiest one to picture, because it consists of a single
electron moving in an orbit around a nucleus. The nucleus contains one
proton. The positive charge of this proton neutralizes the negative charge
on the electron.
Our first important generalization about this isolated atom is that, from
a macroscopic point of view, it is charge neutral.
From a microscopic point of view, a force of attraction or coulomb
force exists between these two charges that tends to pull the electron into
the nucleus. But the electron is moving around the nucleus and, having
mass, the centrifugal force tends to make the electron move away from the
nucleus. An energy equilibrium results when the coulomb force equals the
centripetal force. One might assume that an orbit of any diameter could be
maintained simply through an adjustment of the electron's velocity to
2 SEMICONDlCTOR PHYSICS [~1.1]
make the two forces equaL This assumption. however, is not valid,
because the electron has wave characteristics as well as mass. Oversimpli-
fied, if a stable condition is Lo exist, the electron must move in an orbit
so that the wave can retrace its path. The spaces between stable energy
levels, i.e., allowable orbits, are called forbidden regions, because electrons
cannot stay in these regions for any appreciable period of time.
-0.378 ev
Fig. 1.1. The energy-level diagram for the hydrogen atom.
Schrodinger's wave equation indicates that each atom has several dis-
crete orbits in which an electron may stay for a period of time. These
possible orbits are called energy shells or bands. As shown in Fig. 1.1, the
K.band is nearest the nucleus. The next shells are L, M, N, 0, etc. in that
order. Energy is required to move an electron from the K-shell (quantum
number n = 1) to the L-shell (quantum number n = 2) or to any higher
shell. This energy is often quantified in terms of electron volts: One
electron volt (ev) is the energy acquired by an electron in falling through
a difference of potential of one volt.
As indicated in Fig. 1.1, the K-shell of the hydrogen atom is associated
with an energy level of -13.58 ev. This means that the electron must
acquire an energy of at least 13.58 ev in order to move to an orbit with an
[~l.l] SEMICONDUCTOR PHYSICS 3
infinite radius or to escape altogether the attraction of the nucleus. Con-
versely, an electron starting at infinity and falling into the K-band of the
isolated hydrogen atom would lose an energy equivalent of 13.58 ev. The
zero energy reference is located at infinity, and the minus sign indicates a
loss of energy as the electron moves closer to the nucleus. Energy lost by
the electron when moving from a given shell to a lower shell appears in the
form of electromagnetic radiation. In considering electric current conduc-
tion, we are directly interested in the energy that must be transferred to an
electron in order to move it from a given shell to a higher shell.
Our second generalization is that the isolated atom has many possible
electron orbits and that these orbits exist only at discrete energy levels.
In atoms having more than one electron, the orbits are more complex.
For example, lithium has three electrons, but only two of these are in the
K-band; the other is in the L-band. Since electrons tend to occupy the
lowest possible energy level, the question naturally arises as to why one of
the three electrons normally occupies an energy state in the L-band. Each
energy state is uniquely identified by four quantum numbers. * No two
electrons in an isolated atom can have the same set of quantum numbers.
The K-band for any atom has only two energy states; therefore, no more
than two electrons can occupy this energy level at the same time. In
lithium, two electrons fill all of the available states in the K-band. The
third electron does occupy the next lowest energy state, which is in the
L-band. Table I in Appendix III shows how the energy bands fill up as the
atoms become more complex. Note that when the L-band has eight
electrons in it, the next electron must establish an orbit in the M-band
which has 18 energy states.
When eight electrons are in the M-band, the next electron must go to
the N-band although ten of the 18 states are not occupied. This "octet
theory" of eight electrons in the outer or valence band is illustrated by all
of the inert gases. An important point in our semiconductor theory is that
when eight electrons are established in an outer band, the next electron
must go to the next higher energy band. Table I further indicates that after
one or two electrons are in the next higher band, the lower band may accept
more electrons until all its states are filled.
An isolated atom for each element has K, L, M, N, etc. energy bands.
Corresponding bands have the same number of energy states. However, the
corresponding bands for atoms of different elements do not have the same
These quantum numbers are as follows: n is the primary quantum number and
identifies the band or shell; i.e., n = I for K-band, n = 2 for L-band, etc. The angular-
momentum quantum number is /. The quantum number that specifies the orientation
of an orbit is mI. The spin number is m8, which has a value of either +! or -I. All
quantum numbers except m. are integers. The values of / may range from 0 to (n - I);
m, may range from -/ to +/. Essentially, nand / determine the energy of the state,
whereas m and m. determine the number of energy states at each level.
4 SEMICONDUCTOR PHYSICS [~l.1]
average energy level. For example, the K-band for indium has a smaller
radius (and therefore a lower energy level) than the K-band for germanium.
The 49 protons in the nucleus of the indium atom exert a greater attractive
force than the 32 protons in the nucleus of germanium on the two electrons
in the K-band. This greater force causes the electrons to rotate in a smaller
orbit. In fact, all of the bands for indium have smaller radii than the
corresponding bands for germanium.
One final point about the isolated atom should be noted. Each energy
band may be made up of more than one level. These different levels are
designated by the letters S, P, D, F, etc. Each level has a specific value for
the second quantum number, l. For example, carbon has six electrons.
Two are in the K-band and four in the L-band. Of these latter four
electrons, two are in the S-level and two in the P-Ievel, which has a slightly
higher energy.
Example 1.1. Find the mmimum amount of energy required to

move an electron (a) from the K-shell to the L-shell and (b) from
the O-shell to the P-shell 10 the isolated hydrogen atom.
Solution:
(a) EL - EK = - 3.39 - (- 13.58) = 10.19 ev
(b) Ep-Eo= -O.387-(-0.543)=0.165ev
In the first case an energy of 10.19 ev must be transferred to the

electron in order to move it from the K-shell to the next highest
energy shell. An energy of less than thts amount would not move
the electron out of the K -shell. If the electron reaches the L-shell,
it may remain there for a short period of time. Electrons tend to
go to the lowest energy level available and eventually an electron
would fall back to the K-shell.
Example 1.2. When the electron in Ex. 1.1 falls from the L-band
to the K-band, how much energy does it give up in the form of
radiation?
Solution: The energy released is equal to the energy required to

move the electron from the K-band to the L-band; i.e., 10.19 ev.
The frequency of the quantum of radiation is given by
f= Efh
where E is the energy absorbed or emitted
h is Planck's constant = 6.62 x 10-27 erg-sec
= 4.18 x 10-15 ev-sec
[~1.2] SEMICONDUCTOR PHYSICS 5
10.19 15
f= 1 -15 = 2.44 x 10 cps
4.18 x 0
The wavelength of the radiated X ray is given by
c 3 X 1010 5
A.=]= 2.44 X 1015 = 1.23 x 10- em
Problem 1.1. The ionization potential of a neon atom is approximately

21.5 ev. Find the minimum velocity of an electron that will ionize this
gas. (Ans. 2.74 x 108 em/sec)
Problem 1.2. When an electron returns to its normal orbit, an energy of
21.5 ev will be released. Find the wavelength of the X ray produced.
1.2. ENERGY LEVELS IN CRYSTALS
Crystals are solids in which the atoms are arranged in an orderly

manner. Atoms in a crystalli.re bound together by two forces: electrostatic
and quantum-mechanical. This binding occurs in at least three ways:
ionic, metallic, and covalent.
An example of ionic binding is found in sodium chloride or table salt.
From Table I of Appendix III, we note that chlorine has seven electrons in
its outermost occupied (or valence) band, whereas sodium has only one. A
chlorine atom may take the electron from a sodium atom to form an
"octet" of electrons in its outer orbit, forming a very stable condition. In
this process, however, the sodium atom is left with a net positive charge
and the chlorine atom with a net negative charge, and the two atoms are
attracted to each other by an electrostatic force. If the two atoms move too
close to each other, their filled bands will overlap; but, from Pauli's
exclusion principle, we know that two filled bands cannot overlap. As the
two atoms move close to each other a separation distance will be reached
wherein the quantum mechanical repulsion force becomes equal to the
electrostatic attraction force. At this point a stable separation distance
exists, and the total energy of the two atoms is a minimum.
The atoms are extremely close together in some crystals. This close
proximity produces some displacement of the energy levels in the valence
region. Because of this condition, the electrons in the valence band
can move from atom to atom with very little additional energy. The
resulting electron sharing helps to produce the force that binds the atoms
together. A crystal composed of a given number of atoms has a lower
energy than an equal number of isolated atoms. This type of binding is
called metallic binding, because it is found so frequently in metals.
Covalent binding is very important in semiconductor work. Atoms
having four valence electrons tend to form crystal structures in which one
Other
bands
tI 0\
14 states {
empty
3D
3P M-
>- band
Cl 4 states { 35 }
I
L..
QJ filled
C
I.LI
Forbidden gap c:
~
"ll
~
8 states{~ "
<Ill
filled ~ ~ 2P L-
} band
~ I Forbid~n gap 25
2 states I ' 15 K-
filled :~Normal spacing in crystal band
Distance between atoms
Fig. 1.2. Band spreading and splitting as a function of the atomic spacing for silicon. Note that the lower bands do not spread at normal
spacing (not to scale). -
~
CO'>
~
atom shares an electron with each of its four neighbors. This sharing
effectively gives each atom eight electrons in its valence band and results
in a closed covalent "octet." Electrons in this "octet" are hundreds of
times more difficult.to dislodge than the same electrons would be in the
isolated atom. Examples of materials that form covalent bonds are carbon
(diamond), germanium, and silicon. The latter two materials are frequently
used in semiconductor diodes and transistors.
When atoms are bound together in a crystal, two important effects are
produced on the energy levels of the various atoms. First, the individual
levels are shifted to produce energy bands instead of discrete energy levels.
A band may be made up of many energy states so close together that it
would be impractical to try to distinguish individual states. At the higher
energy levels, these allowed energy bands may cover an energy range of
one electron volt or more.
The second effect is band splitting. As atoms move closer together,
some of the individual states may move to higher energy states and some
to lower ones. Band spreading and band splitting as a functionef the
distance between atoms are illustrated in Fig. 1.2.
The electrical characteristics of a crystal are determined by the manner
in which the energy bands spread, split, and overlap. The following
example illustrates how the levels of the isolated atom are modified by
bringing atoms close together.
Example 1.3. How do the energy levels in an atom of silicon

change as one atom moves close to other atoms of the same type?
Solution: From Table I in Appendix III, we find that silicon has
the K- and L-bands filled and has four electrons in the M-band.
The 18 energy states of the M-band (or n = 3) appear in three
levels: two in the 3S-level, six in the 3P-level, and ten in the 3D-
level. Two of the four valence electrons fill the two 3S-levels and
the other two are in the lower 3P-levels.
If the two atoms are many atomic diameters apart, the energy
bands are very narrow, as shown in Fig. 1.2. When the separation
distance is two or three atomic diameters, the valence-band energy
levels begin to spread. The important point here is that the two
states in the 3P-level containing the two electrons move down,
while the other four empty states move up. At the normal atomic
spacing found in a silicon crystal, an energy gap (or forbidden gap)
of approximately 1.11 ev exists between an energy state containing
a valence electron and the next lowest unfilled energy state. This
energy gap is caused by the band splitting effect. Energy con-
ditions at normal atomic spacings are illustrated in diagram form
shown in Fig. 1.3.
8 SEMICONDUCTOR PHYSICS [~1.3]
Conduction
band
Valence
band
>-
~ Forbidden gap
Q)
c:
IJJ L-band
Forbidden gap
K-band
Fig. 1.3. Energy bands in silicon crystal. At absolute zero, all energy
states in the valence band are filled and all states in the conduction band
are empty.
1.3. ELECTRICAL CHARACTERISTICS OF A CONDUCTOR
As indicated in Sec. 1.2, the electrical characteristics of a solid are

directly related to its electron distribution and energy-band picture. When
a voltage is applied across a solid, current can flow only if current carriers
can effect transitions between energy states. In semiconductor work, we
recognize the existence of two types of current carriers: holes and electrons.
(lfhe hole will be considered in greater detail later. At this point in our
discussion, it is sufficient to state that a hole is a vacancy in the covalent
bond which under normal circumstances is occupied by an electron.
Electrically, the hole can be considered as a positive electron, and as such
it is a current carrier.)
In good conductors, such as gold, silver, and copper, electrons are the
dominant carriers. Let us study copper to determine why electrons are
readily available for the current carrying process and how they move be-
tween energy states. From Table I of Appendix Ill, we note that a copper
atom has 29 electrons. Two are in the K-shell, eight in the L-shell, eighteen
in the M-shell, and one in the N-shell. When copper atoms form a crystal,
the energy levels in the M- (n = 3) and N- (n = 4) shells spread, causing
the 3D-, 4S-, and 4P-bands to overlap. Each copper atom supplies ten
electrons in its 3D-level (D-level of the M-shell) and one in the 4S-level
(S-level of the N-shell) for a total of eleven electrons in the overlapped
energy region. Each atom also supplies ten energy states from the 3D-level,
two from the 4S-level and six from the 4P-level or a total of 18 available
energy levels. If the crystal contains x atoms, then this one band consists
of l8x different, but closely spaced, energy levels and llx electrons to
occupy these levels. Therefore, 7x unoccupied energy levels exist in this
band.
- - - - - - - - - - - - - - - - - - - - - - - - - - - fO
1 "0
C
0
,g
c
.~ +
u~
"0
C
0
U
_________ L
(a) (b) (e)
Distance~
Fig. 1.4. Motion of electrons in conduction band: (a) random motion

due to thermal energy; (b) uni-directional motion due to electric field;
(c) combined motion. Numerous collisions cause changes in the direction
of motion.
At a temperature of absolute zero, these electrons occupy the llx

lowest energy levels. These levels constitute the valence band. At room
temperature, some of these electrons receive additional energy from
phonons (packets of thermal energy) and move into many of the upper 7x
unoccupied states. From there they move randomly from one level to
another in this band as shown in Fig. l.4a. Since this band of energy is
shared by the entire crystal, * copper has a tremendously large number of
electrons per cubic centimeter that can wander randomly from one place
* These energy levels are coupled, making them interdependent.

to SEMICONDUCTOR PHYSICS [~1.3]
in the crystal to another while remaining within the energy limits of this
band.
Copper is a good conductor because it has a large number of electrons
whose normal energy leveh at absolute zero are located very close to
unfilled levels. If an electric field is developed in a piece of metal, electrons
in the higher levels are accelerated in a direction opposite the direction of
the field (in the direction of the positive terminal). As electrons gain energy
from the field, they move to higher energy le\ els as shown in Fig. lAb.
Since the available empty le\cls are extremely dose together, the electrons
may appear to move smoothly rather than in jumps. An electron does not
move very far before it has a collision. and loses some of its energy. This
loss of energy causes it to drop to a lower level (see Fig. lAb). Electron
drift, due to an applied field, is a series of start-stop motions. The total
movement of an electron is obtained by combinmg the random movement
and the drift as shown in Fig. 1.4c.
A good conductor is characterized electncally by its low value of
resistivity or its high value of conductance. The .:quation for conductance is
(1.1)
where (1 is drift conductance in mhos per centimeter

q is the charge on an electron (1.6 x :0-19 coulomb)
n is the concentration of electrons available for conduction in
electrons per cubic centimeter
p is the concentratIon of holes in holes per cubic centimeter
Iln is the mobility factor for electrons in centimeters per second
for a voltage gradient of one volt per centimeter (cm2/volt-sec)
IIp is the mobility factor for holes in centimeter per second for
voltage gradient of one volt per centimeter.
The mobility constant indkates the relative ease with which an electron
can move through the lattice and IS a function of the arrangement of the
atoms in the crystal, the dicicctnc constant, and the amount of vibration
exhibited by the atoms. For a particular matenal
Il = BT 3/2 ( 1.2)
where B is a constant for a given material
T is the absolute temperature ('K).
An electron may be accelerated to extremely high velocities in a very short period

of time because it has a large charge-to-mass ratio. However, the number of collisions
in a solid per unit of time is also ,cry large. For example, an electron moving in a solid
may have several million collision, per second and travel approximately 10-7 cm between
collisions.
[~I.4] SEMICONDUCTOR PHYSICS 11
This equation indicates that as temperature increases, the mobility con-

stant J1 decreases. Atomic vibrations in the lattice increase in amplitude
as temperature increases. As a result, electrons have more collisions in a
given period of time. Thus the average forward drift velocity of electrons
decreases as temperature increases, thereby producing a lower value of J1.
Example 1.4. The mobility constant for a sample of copper at

25 C is approximately 35 cm2/volt-sec, and the number of elec-
trons available for conduction is approximately 8.2 x 1022 /cm3
Find the conductivity and resistivity.
Solution: Since copper has no hole-type carriers, the conductivity is
(J = 1.6 x 10-19(8.2 x 1022 x 35) ~ 4.6 x 105 mho/cm
Resistivity is the reciprocal of the conductivity:
p = 1/(4.6 X 105) ~ 2.16 X 10-6 ohm-cm.
Problem 1.3. Find the drift velocity for electrons in a section of No. 18
wire (0.122 em in diameter) made from the copper in Ex. 1.4 when it
carries a current of one ampere. (Ans. 6.5 x 10-3 em/sec). [Hint: Find
the length of wire that contains one coulomb or 6.2 x 1018 free electrons.]
Problem 1.4. Find the voltage required to produce the one ampere of
current in 5000 feet of this wire. How much heat is dissipated in this
length of wire? What causes the temperature of the wire to increase?
Problem 1.5. Find the mobility constant for the sample of copper when
the temperature is increased 100 C. If the number of free electrons
increases to 8.5 x 1022/cm3, find the values for a and p at the higher
temperature. (Ans. fLn = 22.7)
1.4. ELECTRICAL CHARACTERISTICS OF A

SEMICONDUCTOR
A semiconductor has a higher value of resistivity than a conductor.

This increase in resistance is due primarily to the scarcity of current
carriers. For example, a cubic centimeter of silicon has approximately one
current carrier for each four trillion (4 x 1012) carriers in copper. This
tremendous difference in the number of current carriers can be explained
in terms of the energy band.
When silicon atoms form a crystal, the M-band splits, as shown in
Fig. 1.2. At normal atomic spacing the filled covalent levels are separated
from the nearest empty levels by a forbidden region or gap of approxi-
mately I.II ev. Similar conditions are present in carbon and germanium.
Forbidden gaps formed by band splitting are approximately 7 ev in carbon
(diamond) and 0.69 ev in germanium.* A comparison of the energy levels
for normallattic spacing in copper and silicon are shown in Fig. 1.5. As
stated in Sec. 1.3, current can flow only if current carriers can easily
effect transitions between states. Since no empty states exist at the cova-
lent level in silicon, an electron must move across the forbidden gap
before it can help conduct current. The electron can move easily from
one energy level to another in the conduction band, because this band
contains many closely spaced empty levels.
(0) (b)
Fig. 1.5. Energy-level diagram at nonnal atomic spacing for (a) copper
and (b) silicon (not to scale).
A covalent electron in silicon must receive at least 1.11 ev of energy

fr6m some external sourcet before it can jump across the forbidden gap
to the conduction band. When an electron does jump the gap and move
into the conduction band, it leaves a vacancy in the covalent band that is
called a hole. A hole has a strong affinity for an electron on the bases both
of charge attraction and of quantum mechanical energy. Using the mass-
energy transformation, we can treat the hole as a mass having a positive
charge. A hole can move from one atom to another in the valence band
Values ranging from 0.69 to 0.785 are quoted in the literature. The magnitude is
a function of temperature.
t This source is usually the thermal energy present at operating temperatures. Other
sources are high energy particles and rays.
[~I.4] SEMICONDUCTOR PHYSICS 13
by capturing a covalent electron from an adjacent atom. The atom that
loses the electron gains the hole. By this method a hole moves along a
random path from atom to atom while remaining in the covalent band as
shown in Fig. 1.6a. When an electric field is developed in the crystal, a
drift is superimposed on this random motion, as shown in Fig. 1.6b. A
Fig. 1.6. Hole movement in a crystal: (a) random motion and (b) motion
under the influence of a field. A hole originates at the impurity atom
designated by the dashed circle and moves from one atom to another as
indicated by the arrows.
comparison of hole flow with electron flow shows that electrons flow in a
direction opposite that of the field, whereas holes flow in the same direction
as that of the field.*
* Hole flow is in the same direction as conventional current.

While electrons are in the conduction band of silicon, their movement
is somewhat similar to electrons in the conduction band of copper.
However, the electrons remain in the conductIon band for only a short
period of time, called the jree-time or lifetime of a free electron, which
varies from less than one microsecond to more than one millisecond. At
the end of its free-time, the electron may drop back directly into a hole in
the covalent band or it may be trapped momentarily in the forbidden
region before filling a hole.
The number of electrons in the conduction band of intrinsic or pure
material is equal to the number of holes in its valence band at any instant
of time, because electrons and holes are always generated and recombined
in pairs. Such electron-hole pairs are constantly being generated and ex-
tinguished in a semiconductor crystal, and although the exact number of
pairs present in a sample varies from one instant to the next, the average
number remains relatively constant. The average number of electrons in
the conduction band of an intrinsic material at any time may be found
from the equation
(1.3)
where n is the average number of electrons In the conduction band for
j
intrinsic material
T is the absolute temperature in Kelvin degrees
k is Boltzmann's constant = 1.38 x 10-23 joules per centigrade
degree
g is the width of the forbidden gap in electron volts
q is the charge of an electron
A is a constant for a given material
At room temperature (300 K) pure germanium has approximately
2.39 x 1013 and silicon has approximately 1.52 x 1010 electrons per cubic
centimeter in the conduction band. An equal number of holes are present
in the valence band. The increase in the width of the forbidden gap from
0.69 to 1.11 ev is primarily responsible for the large difference between the
number of free electrons for the two materials.
Problem 1.6. Find the constant A in Eq. (1.3) for (a) germanium and
(b) silicon. Note that kTlq at 300" K is 0.0258 ev. (Ans. (b) 6.14 x 1015)
From the preceding discussion, it is now clear that two types of

current carriers, holes and electrons, are present in a semiconductor. In
general, holes have a lower mobility constant for a given material than
do electrons. For example, for intrinsic silicon I1n is 1350cm2/volt-sec and
IIp is 480 cm2/volt-sec when T is 300" K. Corresponding values for in-
trinsic germanium are 3900 cm2jvolt-sec and 1900cm2Jvolt-sec. Although
the number of holes and electrons are equal in intrinsic material, the lower
mobility constant for holes causes the hole current to be considerably less
than the electron current.
Example 1.5. Find the conductivity and resistivity of pure ger-

manium at 300 K.
Solution: From Eq. (1.1), the conductivity of intrinsic material
(fi is
(fi '"" q(niJln + PiJlp)
Since the number of holes in pure germanium equals the number
of electrons, then
(fi = qn;(Jln + Jlp)
= 1.59 x 10-19 x 2.39 x 1013(3900 + 19(0)
= 0.022 mho/em
Note that owing to the difference in the mobility constants,
approximately two-thirds of the current is carried by electrons
and one-third is carried by holes.
The resistivity is
Pi = 1/(fi = 1/(0.022) = 45.5 ohm-em
Problem 1.7. Find the conductivity and resistivity for intrinsic silicon
at 300 K. (Ans. (Tt = 4.45 X 10-6 mho/em; Pt = 2.25 X 105 ohm-em)
Problem 1.8. Find and compare the resistivity values for silicon and
germanium at a temperature of 200 C. [Hint: First find nt' J1-n, and fLp
at 200 C, then use Eq. (1.1) to find (Tt']
1.5. ELECTRICAL CHARACTERISTICS OF AN INSULATOR
The major difference between semiconductors and insulators is in the

width of the forbidden gap between the valence and conduction bands.
The width of the gap for insulators may be several electron volts, whereas
the width of the gap for semiconductors is usually in the order of one ev.
The energy comparison is illustrated in Fig. 1.7. Only a few electrons
obtain an energy greater than 5 ev from thermal sources at room tempera-
ture. If the conduction and valence band are separated by an energy gap
greater than a few electron volts, very few electrons (or holes) are available
for current conduction.
If the temperature of a semiconductor or an insulator is drastically
increased, many more current carriers are available, as indicated by
Eq. (1.3). Although the mobility constants for a material decrease at
higher temperatures, the increase in the number of available carriers more
than offsets this effect and causes the resistivity to decrease. Thus, semi-
conductors and insulators usually exhibit a negative temperature co-
efficient. Good conductors, however, usually have a positive temperature
coefficient. In conductors, as in semiconductors, the mobility constant
decreases as temperature increases. Although the number of current
Conduction band
7 ev gap
Conduction band
t-_c_o_nd_u_c...,t
...
io_n_b_a_n_d_
1
4'
0.69: ev gap 1.11 ~v gop
>- Valence band Valence band Valence band

~
QI
...--------
C
UJ
> >
(a) (c)
Fig. 1.7. A comparison of the energy levels for three tetravalent elements
In crystal fonn: (a) germanium and (b) silicon as semiconductors; (c)
diamond (carbon) as an Insulator.
carriers does increase with temperature, the mobility constant of a con-

ductor decreases by a greater factor and produces a decrease in conduc-
tivity or an increase in resistivity with an increase in temperature. Several
compounds exhibit a zero temperature coefficient over a temperature
range. In these materials, the mobility constants decrease by the same
factor as the number of available current carriers increase.
A sharp distinction between conductors, semiconductors, and insu-
lators does not exist. Silver has a resistivity of approximately 10-7
ohn-cm, whereas quartz has a value of approximately 1019 ohm-em. Any
value of resistivity between these two may be obtained. Semiconductors
usually fall in the range extending from 10- 2 to 109 ohm-em. Conductors
usually have values below this range, and insulators have values higher
than 109 ohm-cm.
Example 1.6. A material has a forbidden gap between the valence

band and the conduction band of 2.58 ev. The constant A in
Eq. (1.3) is 10 x 1016. Find the number of current carriers
available at room temperature (300 K). 0
Solution: From Eq. (1.3), the number of free electrons is

ni = (10 x 1016)(3003/2)e-2.S8/(2xO.02S8)
= 5.18 x 102/eso = 10-1

or one electron and one hole for each ten cubic
centimeters.
Problem 1.9. Find the resistivityof the material in Ex. 1.6if /L n = 4000
cm2/volt-secand /Lp = 2000 cm2/volt-sec. (Ans. 1.045 x 1016)
Problem 1.10. Find the resistivityof the material in Ex. 1.6 when the
temperature is increasedto 2000 K.
1.6. SEMICONDUCTORS WITH IMPURITIES
The resistivity of a semiconductor can be drastically reduced by the

addition of small amounts of certain types of impurities. This is called
doping. Impurities added to silicon and germanium are usually penta-
valent (five valence electrons) or trivalent (three valence electrons). When
a relatively small number of antimony atoms (five valence electrons) is
added to germanium, four of the five electrons enter the covalent bond.
The fifth valence electrons on each pentavalent atom is loosely attached
to the atom. In terms of energy, these electrons occupy a state located
very close to the bottom of the conduction band, as shown in Fig. 1.8a.
An energy of only 0.01 ev is required to raise the fifth electrons into the
conduction band, whereas 0.69 ev is required to break the covalent bond
and elevate any of the other electrons into the conduction band. At room
temperature, almost all of the impurity atoms supply one current carrier
each.
The addition of small amounts of pentavalent impurities to silicon or
germanium produces N-type semiconductor material. Because they add
electrons to the conduction band without adding holes in the valence band,
pentavalent impurities are often called "donors."
The addition of trivalent impurities produces P-type semiconductor
material. Trivalent impurities are often called "acceptors," because they
add holes in the valence band without adding free electrons in the con-
duction band. The three valence electrons of a trivalent impurity may
enter the covalent bond. Since four electrons should be contributed by
each atom to complete the covalent "octet," the bond is not complete. In
effect, a hole is produced. If an electron should fall into this hole, it will
be bound tightly into the covalent bond, even though the inpurity atom
then has an extra electron or a net negative charge. In terms of conduction,
the addition of trivalent impurity atoms creates unfilled energy levels close
to the filled valence levels, as shown in Fig. 1.8b. Thermal energy available
at room temperature enables the valence electrons from other atoms to
move into these empty states. When this happens, empty states are avail-
able in the valence band, and the conduction of current can take place in
the valence band with holes acting as current carriers .
{==-.----
0.6, ev I
0.69 ev
0.0' ev
~-~ 00000000001
T
( 0) (b)
Fig. 1.8. Energy states for germanium doped with (a) pentavalent and
(b) trivalent Impurities.
A comparison of a cubic centimeter of intrinsic, P-type, and N-type

semiconducting material is summarized below:
1. Each block is charge neutral, because each atom in the block was
charge neutral when the solid was formed.
2. Intrinsic material has an equal number of electrons in the conduction
band and holes in the valence band.
3. P-type material has fewer electrons, and N-type material has more
electrons in the conduction band than holes in or near the valence
band.
4. With equal temperature values and light doping, the product of the
number of holes and conduction electrons is the same in all three
cases. This product is approximately 5.7 x 1026 for germanium and

2.3 x 1020 for silicon at 300 K. We may conclude that P-type
material has more holes and fewer free electrons than intrinsic
material and that N-type has more free electrons and fewer holes
than intrinsic material.
5. The mobility constants vary with the degree of doping. For
example, in silicon, the following p. values have been measured
(cm2/volt-sec).
Doping N-type P-type
atoms/cm3 /Ln /Lp /Ln /Lp
1X 1016 1380 460 1280 460

5x 1015 980 390 870 410
1x 1016 830 360 720 390
2x 1016 700 315 560 370
A comparison of these values with those for intrinsic silicon may

explain why the literature has many different values for the mobility
constants.
6. The generation rate for electron-hole pairs is approximately the
same for both intrinsic and doped semiconductors.
Example 1.7. Find the average number of holes and electrons in

a cubic centimeter of silicon at 300 K if the sample is doped with
one impurity donor atom for each 107 intrinsic atoms.
Solution: Silicon contains approximately 5 x 1022 atoms per
15
cubic centimeter. Therefore, 5 x 10 impurity atoms are added
to the sample in obtaining the desired doping. Assuming that each
impurity atom contributes one electron, then the total number
of electrons is approximately the sum of the intrinsic and donor
electrons or
The number of holes is
nt = 2.35 xx 1015
p,:::::;;
10 20
:::::4.6 x 10
4
This last number is smaller than that for intrinsic silicon because,
with many more electrons present, the probability of recombina-
tion is greater.
Problem 1.11. Find the resistivity of the sample of silicon in Ex. 1.7.
Assume that I-'n = 1135 and I-'p = 410 em2/volt-sec. (Ans. Approx. 1.1
ohm-em)
20 SEMICONDUCTOR PHYSICS [~1.6)
o = hole
- = electron
--
Current direction Current direction
..
+ +
(0 ) (b)
Fig. 1.9. Current flow by tbe mecbanism of drift In an impurity semi-

conductor: (a) N-type and (b) P-type.
b b b
I
I
I' ',', " , '
I
, II.
(/)
':'",
I
',,',"",,
,. ,',
" ,',",
, j/,
I,
1'1'
r
1,'1
I '
" I
I
,
,
I. '.
.,'
" ,; , ,.1,"
, I, ,I'
,'1
"
II' I I'"
I ',': 't' " I '1 ' ,

" , ,
,
I" '
.,
1, I" I,
, I,
,
I , ' I' " "1' , , : , '. ~,I ' , I.' ,

:', ',', , I'"
I I
I I
a Q a
(a) 1=10
Fig. 1.10. Current flow by diffusion of carriers from an area of high

concentration to an area of low concentration: (a) Initial distribution;
(b) distribution a short time later; (c) eventual distn"bution.
[~l.7] SEMICONDUCTOR PHYSICS 21
Problem 1.12. Find the current in the samesamplecube of dopedsilicon
when 0.1 volt is applied to opposite faces of the cube. What part of
this current is carried by holes?
As indicated in these examples, N-type semiconductors will have a

high concentration of free electrons and a negligible number of holes
moving randomly throughout the crystal structure. If suitably designed
metal electrodes are attached to the crystal and if a voltage is applied, a
net drift toward the positive terminal will be superimposed upon the ran-
dom electron motion, as shown in Fig. 1.9a.
Holes are the majority carriers in P-type semiconductors. When a
voltage is applied across the crystal, holes that are in random thermal
motion drift toward the negative electrode and are annihilated by free
electrons in the metal electrode. The uncovered negative, immobile ions
near the positive terminal force electrons into the metal electrode; i.e.,
holes are injected into the crystal at the positive electrode. Thus, the
crystal remains electrically neutral. It is significant to notice that, even
though holes carry the current in the P-type semiconductor, the current
flow in the electrodes and connecting wires consists of electron movement.
1.7. DIFFUSION CURRENT
The previous sections were concerned with current flow or the drift of
current carriers under the influence of an electric field. Another type of
current flow that is extremely important in transistor work is the diffusion
of current carriers when no field is present.
The magnitude of current flow is the net number of charges passing a
given cross section of the conductor in a unit time. If the charge carriers
are uniformly distributed and move randomly, the average diffusion
current is zero, since the number of current carriers moving across a given
cross section in one direction is equal to the number moving across the
same plane in the opposite direction.
Let us next consider the case in which the charge carriers are not
uniformly distributed. Assume that a quantity of charge has been dumped
or injected into one end of a crystal as shown in Fig. 1.10. The process by
which this is accomplished will be discussed in the next two chapters. The
probability that a current carrier near the plane a-b will move to the left is
equal to the probability that it will move to the right. However, there are
more carriers on the left than on the right. Therefore, on the average,
more carriers cross the plane from left to right than from right. to left.
The result is a net flow of current across plane a-b. Eventually the current
carriers will be uniformly distributed, as shown in Fig. 1.1Oc. Until this
condition exists, diffusion current will flow.
22 SEMICONDUCTOR PHYSICS [~l.7]
Mathematically the current at any cross section, a-b, as shown in
Fig. 1.10 is
i = (qDA) (-slope of density curve) (1.4)
where q is the charge on an electron or hole
D = (jJkT/q) is the diffusion constant of carriers in square centi-
meters per second.
A is the cross-sectional area in square centimeters.
In Fig. 1.10 a given quantity of charge has been injected into the end of
the bar at 1 = 0, so that charge distribution is shown for three different
instants; 11 occurs a relatively short time after 1o, and 12 a very long time
later. The current as a function of time decays to zero on an exponential
curve. *
Even more important in our future work is the case in which we con-
tinue to inject carriers into the left end and remove them from the right
end. For this case, if we assume that there are no recombinations of holes
and electrons in the material, the diffusion current must equal the external
current.
In order to maintain this diffusion current, a nonuniform distribution
of charge must be established along the bar. This situation is similar to a
metal bar that is heated with a torch on one end and attached to a heat
sink on the other end. The torch end will be much hotter than the sink
end and a temperature gradient will exist along the bar. Equilibrium con-
ditions are established when the quantity of heat accepted by the torch
end of the bar per unit time is equal to the quantity delivered to the sink
per unit time, assuming no radiation from the bar. The temperature
gradient along the bar will increase until this condition exists. In the same
way, charge can flow through a crystal by diffusion at a given rate, pro-
vided the charge distribution can increase the probability of net charge
movement to a sufficient magnitude.
The important points here are:
1. Current can flow in a field-free region.

2. In a crystal this is called diffusion current.
3. Before diffusion current can flow, a nonuniform distribution of
charge must be established.
4. A charge distribution with a linear slope will produce a constant
current flow (assuming no recombinations).
5. Diffusion current cannot be increased any faster than the necessary
charge distribution can be established.
* In order to simplify the process, the assumption is made that no recombinations

take place.
Example. 1.8 Assuming that the slope of the charge distribution
curve at the plane a-b in Fig. 1.l0b is 1014 electrons/cm4 and that
the cross-sectional area is I cm2, find the instantaneous current
flow across the plane a-b. The block is germanium and has a
diffusion constant for electrons of 100 cm2/sec.
Solution: From Eq. (1.4)
i = 1.59 X 10-19 x 100 x 1 x (_1014) = -1.59 rna
or a current of 1.59 rna to the left.
Problem 1.13. Find the instantaneous current in Ex. 1.8 if the carriers
are holes. D p is approximately 50cm2/sec for germanium. (Ans. Approx.
0.8 rna to the right)
Problem 1.14. A steady diffusion current of 1 rna flows in a block of
silicon which has a cross-sectional area of 0.1 cm2 Find the charge
gradient that will produce this flow (a) if the carriers are electrons and
(b) if the carriers are holes. Dn is 35 cm2/sec, and Dp is 12.4 cm2/sec.
CHAPTER PROBLEMS
1.15. Find the shortest wavelength of light that can be transmitted

through silicon. [Hint: Find the wavelength corresponding to an
ionization energy of 1.11 ev.]
1.16. Number 20 copper wire has a diameter of 32 mils. If the mobility
constant is 35 cm2/volt-sec, and if8 x 1022 electrons per cm3 are avail-
able for conduction, find the average drift rate for electrons when a
drop of one volt per 1000 feet is measured. Find the resistance of 1000
feet of this wire. (Ans. 1.15 x 10-3 em/sec.)
1.17. Plot the average number of electrons per cubic centimeter in the
conduction band of intrinsic germanium as a function of temperature
from 300 K to 600 K. [Hint: Take the log of both sides of Eq. (1.3).
Plot log n as a function of l/T.] Assume that Eg = 0.69 at 300 K and
A = 2.85 X 1016
1.18. Repeat Prob. 1.17, using silicon. At what temperature will one
cubic centimeter of intrinsic silicon have as many electron-hole pairs
as intrinsic germanium has at 300 K. Assume that Eg = 1.11 at 300 K
and A = 6.14 X 1016
1.19. Plot the resistivity of intrinsic silicon over a temperature range
from 300 K to 600 K, (a) assuming the mobility factor is constant and
(b) correcting the mobility factor for changes in temperature. Assume
that P-n = 1350 cm2/volt-sec and P-p = 460 cm2/volt-sec at 300 K.
1.20. Find the number of impurity atoms that must be added to 1 cms
of intrinsic silicon in order to convert it to (a) 10 ohm-cm P-type silicon
and to (b) 10 ohm-cm N-type silicon. Use intrinsic p. values.
1.21. If a donor-type impurity is added to the silicon sample in Prob.
1.20, find the number of free electrons and the number of holes in 1 ems
of the 10 ohm-cm material. (Ans. 4.96 x 105 holes/emS)
1.22. A circular rod 1 cm long with a cross-sectional area of 0.1 cm2 is
made from the 10 ohm-cm N-type silicon in Prob. 1.21. What is the
average electron-drift velocity when a 6-volt battery is placed across
the ends of the bar?
1.23. An electron gun is used to inject electrons into the face of a cube
of intrinsic germanium one centimeter on each side. The opposite face
is grounded. Find the slope of the charge distribution when a steady
current of 100 p.a flows by diffusion. (Ans. 6.3 x 1012 e/cm4)
1.24. Find the number of electrons required to establish the charge
gradient for the sample in Prob. 1.23 in excess of the number in intrinsic
germanium at room temperature.
Chapter 2
JUNCTION DIODES
In Chap. I we investigated the electrical characteristics of crystals.

A junction diode is formed when a piece of N-type semiconductor is
joined to a piece of P-type material in such a way that the crystal lattice is
continuous across the junction. The electrical characteristics of a diode
are radically different from those of either component material. For
example, P-type or N-type material will conduct equally well in either
direction, but a diode will conduct well in only one direction. This dif-
ference is attributed to the nature of the region in the neighborhood of
the junction. In this chapter we will focus our attention first on the junc-
tion and its characteristics. Then we will treat diodes as circuit elements.
Finally, special types of diodes and their applications will be discussed.
The junction diode is an important circuit element. Equally important
is the fact that junction diodes are the basic building blocks for almost all
semiconductor devices. A good understanding of P-N junctions is neces-
sary in all semiconductor work.
2.1. THE POTENTIAL BARRIER
Single crystal P-N junctions may be produced by several processes.

Three of the most frequently used methods are:
I. a growing process in which the impurity concentration is varied
while a crystal is being "pulled" as grown;
2. an alloying process in which the impurity is melted into a block of
semiconductor;
3. a diffusion process in which the impurity gas diffuses into a piece
of semiconductor material held at a high temperature.
The process employed determines to some extent the characteristics of a
diode. For example, an alloy junction is rather abrupt while a grown
25
26 JUNCfION DIODES [~2.1]
junction is not. Our immediate problem is to develop the theory of the
P.N junction. In this development, an abrupt junction will be assumed;
therefore, we can expect some minor deviations from the theory for the
gradual junctions.
+8+8-8 +
8+8+++8
-8+ ~ 1:+
+ 8+ _ 8+
Fig. 2.1. Current carriers in P-type and N.type materials. Note that
electrons are majority carriers on tbe N.side and holes are majority
carriers on the P.side. Circles represent charged immobile impurity
atoms.
Assume that a block of N-type semiconductor is to be joined to a block

of P-type material, as shown in Fig. 2.1, to form a single crystal P-N
junction. Prior to the formation of the junction, each block is charge
neutral, since each block contains an equal number of electrons and pro-
tons. The N-type material contains a large number of free electrons or
majority carriers and an almost equal number of immobile, ionized atoms,
each with one plus charge. This block also contains a few holes or minority
carriers. In both blocks the product of the majority and minority carriers
is nt,
as explained in Chap. 1.
P Depletion N
region
+ +- + ee (f)(f)
- +
+ + + ee (f)(f) +
+ -
+
+ ee (f)<t>
+ +
+ ee
e
Et>(f)
(f) +
+ + +
I _ _ I
F R
Fig. 2.2. A potential barrier formed at the junction by an array of

charged impurity atoms.
JUNCTION DIODES 27
When the junction is formed, majority carriers on both sides diffuse
across the junction. Since each block was initially charge neutral, an elec-
tron moving to the P-side leaves a positive ion on the N-side and adds a
negative charge to the P-side. In a similar manner, a hole moving into
the N-side adds a positive charge to the N-side and leaves a negative
charge on the P-side. The net result of this diffusion current is to produce
a large number of atoms with positive charges (uncovered impurity atoms)
close to the junction on the N-side and an equal number of negative
charges (also impurity atoms) on the P-side, as shown in Fig. 2.2. A field
is established between the two groups of charges that reduces the diffusion
of majority carriers across the junction. The voltage associated with the
field is called the potential barrier.
Minority carriers on either side of the junction enter the picture as
soon as a field is produced. Minority carriers (holes on the N-side or
electrons on the P-side) that wander into the field set up by the charged
immobile atoms are immediately swept to the other side of the junction.
Thus, electrons are forced to move to the N-side, and holes to the P-side.
This action reduces the net charge on each side and thereby reduces the
magnitude of the potential barrier. The flow of minority carriers across
the junction constitutes drift current and is often called reverse current, lR'
The potential barrier increases in magnitude as more impurity atoms
are uncovered on each side of the junction. The region containing these
uncovered (charged) atoms is called the depletion region, because it con-
tains very few current carriers. As the depletion region extends farther
into the material on each side of the junction, more minority carriers are
available for reverse current. Finally, a point is reached where the dif-
fusion or forward current is equal in magnitude to the drift or reverse
current and equilibrium is established. *
Equilibrium conditions are shown in Fig. 2.3. Very briefly these con-
ditions may be summarized as follows:
1. Immobile impurity atoms with charges are arrayed on each side of
the junction-positive charges on the N-side and negative charges
on the P-side.
2. A field exists between these unlike charges.
3. A voltage called the barrier potential is associated with this field.
The magnitude, under normal conditions, of this voltage is approxi-
mately 0.3 volt for germanium and 0.7 volt for silicon.
4. Majority carriers (holes on the P-side and electrons on the N-side)
can cross this barrier only if they have energies equal to or greater
This process may also be treated in terms of energy. The average energy level or
Fermi level of the N-type material is higher than that of the P-material. When the
junction is formed, the two energy levels must eqyalize. This is accomplished by moving
electrons into the P-region and holes into the N-region until the average energy levels
(or Fermi levels) are equal.
28 JUNCTION DIODES [~2.1]
P - region Junction N- region

(-)
(+)
(-)
(+)
(a) Concentration of immobile ions
Electrons
nN
3.68 x 1015/cm3
p
11xN Holes
Electrons 1.55 x 1011/ cm3
np 9
3.25 )(10 / cm3
(c) Net charge distribution

I
Hole
energy 1
.
I
I ~~ 0.36 volts
-~- ---_!_---
(d) The potential barrier
Fig. 2.3. EquiUbrium conditions for a typical unbiased diode. (Adapted

from Electronic and Radio Engineering by F. E. Terman. Copyright,
1955, McGraw-Hill.)
[~2.l] JUNCTION DIODES 29
than 0.3 ev for germanium or 0.7 ev for silicon and are moving in
the proper direction. * Many majority carriers diffuse across the
junction.
5. An equal number of minority carriers formed in or near the de-
pletion region are drifting, under the influence of the space charge,
in the opposite direction.
6. When an electron (majority carrier) leaves the N-region and moves
into the P-region, it becomes a minority carrier. An electron in
P-type material does not travel far before it combines with a hole.
In a similar manner, a hole moving into the N-side from the P-side
quickly combines with an electron. The average distance that
minority carriers travel before the total number is reduced to l/f.
of the original number is called the diffusion distance, L.
7. Regions beyond several diffusion distances from the junction are
essentially undisturbed and may be treated simply as N-type or
P-type materials.
An equation for finding the value of the potential barrier is
kT PN kT np
VB= --In-= --In- (2.1)
q Pp q nN
where PN is the concentration of holes in the N-region, Pp is the concen-
tration of holes in the P-region, np is the concentration of electrons in the
P-region and nN is the concentration of electrons in the N-region.
Example 2.1. A germanium diode has 3.68 x 1015 electrons/cm3

in the N-region and 1.75 x 1017 holes/cm3 in the P-region. What
is the potential barrier across the unbiased junction?
Solution: The electron concentration in the P-region is
np=-=
n; 5.76 x 10
5
26
1017=3.3 x 109e1ectrons/cm3
Pp 1.7 x
kT np 6
Then VB = - - In - = - 0.026 In (0.896 x 10- ) = 0.362 volt
q nN
Problem 2.1. Find the magnitude of the potential barrier for a silicon
diode which is doped one part in 106 on the P-side and one part in
5 x 107 on the N-side. Silicon has 4.96 x 1022 atoms/em3 (Ans.
Approx. 0.68 volt.)
Problem 2.2. A germanium diode is made of 10 ohm-cm P-type material
and 1 ohm-em N-type material. What is the magnitude of the potential
The energy of the moving carrier is t mv2 The energy associated with the velocity
component perpendicular to the barrier must be equal to or larger than the above values.
barrier at room temperature if II-n is 3900cm2jvolt-secand 11-1' is 1900
cm2jvolt-sec? [Hint: Assumethat the resistivityis controlled solelyby
majority carriers.]
In a sense, the potential barrier acts as an energy filter or a current

equalizer. The potential barrier in the unbiased diode will adjust the for-
ward or diffusion current so that it is equal to the reverse or drift current.
If the drift current changes for any reason, the magnitude of the barrier
will change, also. For example, an increase in temperature will increase
the number of minority carriers generated; therefore, the reverse current
will increase. The forward current component will also increase, because
more majority carriers have sufficient energy to cross the potential barrier.
However, the reverse current increases more than the forward current.
In order to equalize the two currents, the potential barrier decreases
approximately two millivolts per centigrade degree rise. * The amount of
doping also has an effect on the magnitude of the barrier voltage. Minority
carriers must be generated in or near the depletion region to take part in
reverse current flow. Decreasing the width of the depletion region will
reduce the reverse current and allow the barrier voltage to increase. An
increase in the degree of doping decreases the width of the depletion region
and increases the barrier potential.
2.2. RECTIFICATION
The junction diode is a nonlinear device because it conducts current

better in one direction than in the other direction. In the circuit shown in
Fig. 2.4 the diode is "foward biased" and will conduct current easily. The
battery polarity and the polarity of the potential barrier established within
the diode are opposite. Most of the battery voltage is dropped across the
junction. The barrier is lowered and the forward or diffusion current in-
creases. If the net potential barrier approaches zero, the diffusion current
becomes enormous, and the diode may be destroyed. t
When the battery polarity is reversed in the circuit shown in Fig. 2.4,
the diode is "reverse biased." With reverse biasing, the polarities of the
potential barrier and the external source are the same. The external voltage
raises the potential barrier and decreases the forward or diffusion current.
As in the case of a forward bias, virtually all of the external reverse-bias
* This temperature effect is treated more fully in Chap. 6.

t A forward bias repels majority carriers toward the junction. As these carriers
approach the junction, they decrease the width of the depletion region and lower the
potential barrier. The increased current resulting from the bias is due almost completely
to diffusion. The drift current resulting from the interaction of majority carriers and
the voltage source is small and is usually neglected in comparison to the diffusion current.
[~2.2] JUNCTION DIODES 31
voltage is dropped across the junction. If the reverse bias is increased

beyond about 0.1 volt, the diffusion current is reduced to a negligible value.
o
17
1.75 X 10 holes/cm3: :
~
o
I
M
'15
3.68 x 10 electrons/cm3
nN
, I
3.25X10
9
elf~ctrons/cm3 lL ~55 xl 011 holes/cm3
P
N
np.~
r I
Fig. 1.4. A forward.biased diode: (a) circuit (b) typical carrier concen.
trations for a forward-biased diode. (Adapted from Electronic and Radio
Engineering by F. E. Terman. Copyright, 1955, McGraw-Hill.)
As indicated above, the reverse current is primarily a function of the

junction temperature. A comparison of the characteristics of biased and
unbiased junctions is shown in Fig. 2.5. If we assume that no recombina-
tions of holes and electrons occur within the depletion region, the same
number of majority carriers that enter the depletion region on one side
of the junction will emerge as minority carriers from the depleted region
on the opposite side of the junction. The hole density (holes/cm3) at the
edge of the depletion region in the N-material owing to diffusion from the
P-material is given by the following equation when no external bias is
applied to the diode:
(2.2)
where, as shown in Fig. 2.3,

PXN is the number of holes per cubic centimeter existing at the
edge of the depletion region in the N-material
----- - \~ ~ ~
32 JL'NCTION DIODES [~2.2]
XN is the distance that the depletion region extends into the
N-material from the junction
pP is the equilibrium concentration of holes (holes/cm3) in the
P-material.
With zero external bias, equilibrium exists between the forward and re-
verse currents in the diode. For this condition we can obtain Eq. (2.3) by
altering Eq. (2.2) so that the concentration of holes at XN that diffuse
p N
Electron cmJ
I
~O.4Xl()5
Unbiased
i ________
1 _
~C'O.36V
1 energy I
~
lO.31xlcrC+
Forward
biased
-----rr-
Reverse
biased
(a) (b)
Fig. 2.5. A comparison of typical junction clIaracteristics:

(a) potential
barrier; (b) cbarge distribution and depletion regions. (Adapted from
Electronic and Radio Engineering by F. E. Terman. Copyright, 1955,
McGraw-Hill.)
across the junction is equal to the equilibrium concentration of holes

existing in the N-material, PN' or
(2.3)
[Notice that Eq. (2.1) may be derived from Eq. (2.3).] If the potential
barrier is changed by the addition of an external bias, * the diffusion current
We are assuming that the total bias voltage V appears across the junction.
and therefore the concentration of holes at X N will change. With an

external bias voltage V applied to the diode,
PXN = ppeq(V-VB)/kT = ppe-qVB/kTeqV/kT
then by using Eq. (2.3) we obtain an equation for the new concentration
of holes at XN, or
(2.4)
The change in the concentration as a result of the bias is the difference

between the total concentration with an external bias and the concentration
existing with no bias or
(2.5)
These holes that diffuse across the junction are surrounded by large
numbers of electrons. Under this condition, a hole is likely to recombine
with one of the electrons and disappear. The recombination rate is pro-
portional to the density of the holes at any point. Mathematically, the
concentration of holes Pd in the N-region at some distance d from the
junction is
(2.6)
where Pd is the concentration of holes at some distance d into the N-region

XN is the distance the depletion region extends into the N-region
Lp is the distance at which Pd equals lie of its value at XN
Similarly, the concentration of electrons nd at some point d in the
P-region can be determined. Knowing this concentration, we can find the
diffusion current at any point in the diode from Eq. (1.4). From Eq. (2.6)
we see that the slope of the concentration curve of minority carriers
decreases with distance from the junction. After about four diffusion
lengths, the minority carrier diffusion current has decreased to a negligible
value. However, in a closed circuit, the current is continuous. This means
that throughout the diode, the total current is constant, and as the diffu-
sion current due to minority carriers decreases, the drift current of
majority carriers increases. The conditions existing in a forward-biased
diode are shown in Fig. 2.6.
Diffusion current has a maximum value at the edge of the depletion
region, as shown in Fig. 2.6. The magnitudes of the hole diffusion current
and the electron diffusion current at the edge of the depleted zone in the
N-material and P-material, respectively, are
I = AqDpPN [eqV/kT _ 1] amp (2.7)

P"N Lp
_ AqDnnp [ qV/kT _ 1] (2.8)

and Inxp - L e amp
n
34 lUNCllON DIODES [~2.2)
Junction
It'"
p N
Hole
drift
current
"
II
II
_.Ij--_.
- --
drift current
Total diode : I /
II I I /
current I"xp :f I/ diffu sion Curren t
"ll. _ ----f
Electron
diffusion
current>,;
,;;;
I
1 I
I
I
x= p x=O X=Xn
-I Depl~tion .-
region
Fig. 2.6. Current components in a forward-biased P-N junction. (Adap-

ted from Electronic and Radio Engineering by F. E. Terman. Copy-
right, 1955, McGraw-Hill.)
+I
( 0)
Breakdown Is
-v ~ +v
---U----
I (b) -I
I
I
Fig. 2.7. Characteristic curve for a junction diode: (a) theoretical and
(b) actual. The actual reverse characteristic curve is different from the
predicted 1heoretical curve because of leakage current and breakdown.
(~2.2] JUNCTION DIODES 35
Since the total diode current is the sum of these two currents, we can write
the diode current as a function of bias voltage as
p] (eqV/kT - 1)
I = I
PXN
+ I nxp = qA [DpPN
L
+ Dnn
L
(2.9)
p n
If the bias voltage is made a large negative number (reverse bias), the
diffusion current will be reduced to almost zero, and the only current in
the diode will be the reverse saturation current Is' By applying this con-
dition to Eq. (2.9), the reverse current becomes
Is = Aq [DpPN + Dnnp] (2.10)

Lp Ln
Then we can write the expression for the total diode current as
I = Is(eqV/kT - 1) (2.11)
Equation (2.11) is the important diode equation and is shown in graphical

form in Fig. 2.7. In practice, diodes have characteristics .which agree
closely with the curve plotted from Eq. (2.11). The only departures from
+I
Diode
current
-v Is
voltage
Fig. 2.8. Graphical solution for Ex. 2.2.

the theoretical curve are (a) a slightly increased value of Is with reverse
bias due to current leakage across the surface of the junction and (b) the
phenomenon known as breakdown which usually occurs at high reverse
voltages.
Example 2.2. A germanium diode with a reverse curr~nt Is of 5 f.Ul

has a sine wave applied to its terminals. If the sine wave has a
peak amplitude of 0.2 volt, what will be the ratio of the peak
forward to the peak reverse currents?
Solution: A plot of the performance is shown in Fig. 2.8. The

peak forward current occurs at the positive peak of the input
voltage. From Eq. (2.11), the peak current is
IF = I.(eqv/kT - 1) = (5 x 1O-6)(e39(0.2) - 1)
~ 12.2 rna
The current that flows when V is at its negative peak value is
IR = (S x 1O-6)(e-39(0.2) - 1) ~ 5 X 10-6 amp
The ratio of the forward to reverse peak currents is
From this example we observe that if the bias voltage exceeds some
small value (approximatelyO.1 volt), the diode expressions can be simplified
to
when V is greater than + 0.1 volt, and
when V is less than - 0.1 volt.
500 .0.
Fig. 2.9. Circuit for Prob. 2.5.

Problem2.3. Repeat Ex. 2.2 if the peak value of the sine wave is 0.1
volt. Why is the value of the ratio different from that obtained in
Ex. 2.2? (Ans. Approx. 50)
Problem2.4. Find the maximumand minimum currents that flowin a
diode as a function of Is if a 0.05-volt-peaksine wave is superimposed
on a +0.1 d-c voltage across the diode terminals.
Problem2.5. Find the current that willflowin the circuit of Fig. 2.9 if
the reverse diode current I. is 10 /Laat room temperature. (Ans.
1= 4.08 X 10-4 a).
2.3. ELECTRICAL CHARACTERISTICS OF DIODES
Diodes are rather complex circuit elements. In this section we will

investigate the important electrical characteristics that determine the be-
havior of the junction diode in a circuit. These characteristics are: (a)
forward resistance, (b) reverse resistance, (c) junction capacitance, and
(d) speed of response.
2.3.1. Forward Resistance
The nonlinear voltage-current characteristics of a forward-biased diode,

as shown in Fig. 2.10, will not allow us to treat it as a fixed value of
resistance for different current levels. The static or d-c resistance may be
determined graphically by finding the inverse slope (i.e., Vt/lt) of a line
drawn from the origin to a point on the curve corresponding to the d-c
current through the diode, as shown in Fig. 2.10. The dynamic or a-c
resistance of a diode is also a function of the d-c current through the diode.
The a-c resistance may be detennined. graphically by finding the inverse
slope of a line drawn tangent to the curve at the d-c current in the diode.
Both the a-c and the d-c forward resistances are rather high for low values
of current and decrease in value as the d-c current increases.
If the bulk resistance of a diode is neglected, the a-c and d-c resistance
may be obtained analytically. As indicated above, the dynamic resistance
may be found from the slope of the diode characteristics. In equation
form, this slope is
rd = dVfdI (2.12)
By taking the derivative of the diode equation (2.11) and inverting the
results, the dynamic resistance is obtained. The derivative of Eq. (2.11) is
dI = Isq eqVlkT qI
(2.13)
dV kT ~ kT
38 /UNCTION DIODES [~2.3)
The dynamic resistance is the inverse of Eq. (2.13) or
kT 0.025
'4=- ~--ohmat3OO0K (2.14)
qJ I
At room temperature (300 K) and one milliampere of d-c current, the
0
dynamic resistance is about 25 ohms. If the d-c current is increased to
I
I
+I I
'~_1__ = t:. V: Dynamic resistance

Slope t:.I
_1_ :J1.: Static

Slope.Ii resistance
-V +V
Fig. 2.10. Graphical determination of the static and dynamic resistance

ofa diode.
2 rna, the a-c resistance decreases to 12.5 ohms. This simple equation
should be memorized since it will be very useful not only in working
with diodes but also in working with transistors.
In practice, the dynamic resistance is slightly higher than the value
predicted by Eq. (2.14), owing to the bulk resistivity of the P- and N-
materials. This deviation is usually so slight that it can be neglected,
especially in the case of diodes made of low-resistivity materials. In either
case, the relationship between dynamic resistance and current is very
close.
Example 2.3. Calculate the values of the static and dynamic for-
ward resistance of a junction diode conducting a d-c current of
2 rna if the reverse saturation current is 2 JUl.
Solution: The first step is to find the voltage which causes this
amount of current to flow. Using the diode equation, and the
given values for forward and reverse currents,
2 x 10-3 ~ 2 x 1O-6e39v
or
from which v ~ 0.177 volt
T1).e static resistance is the ratio of total voltage to the total

current.
V 177 X 10-3
R = - = ---- = 88.5 ohms
s I 2 X 10-3
The dynamic resistance is found by changing V by some small
amount, A V, and finding the change in current, AI. The ratio of
A V to AI is the dynamic resistance. Let A V = 0.003 volt so that
the new voltage V2 is 0.18 volt. Then
12 ~ 2 x 1O-6e39(0.18) ~ 2.2 x 10-3 amp
AV 0.180 - 0.177
and rd ~ AI ~ (2.2 _ 2.0)10-3 ~ 15 ohms
Using Eq. (2.14), we find that the dynamic resistance is
rd = 25/2 = 12.5 ohms

The agreement between the two answers depends upon the mag-
nitude of the voltage change assumed and the accuracy of the
mathematical calculation.
Problem 2.6. A diode and resistor are connected in series with a 0.2-volt
d-c voltage. The diode has a reverse current of 3 ,."a. Find the resistance
value needed so that 0.1 volt is dropped across this resistor. (Ans.
Approx. 685 ohms)
Problem 2.7. Using the diode equation, calculate the dynamic resistance
of a diode for currents of 1 rna, 2 rna, 5 rna, and 10 rna if f. = 5,.a.
2.3.2. The Reverse Resistance
Diodes are frequently used to isolate circuits from voltage waveforms

that have a given polarity or that exceeds a given magnitude. Under these
conditions, it is important to know the reverse resistance of the diode in
order that the degree of isolation can be predicted. Since the reverse biased
diode is essentially a constant current source, the d-c reverse resistance,
RR' will be a function of the voltage across the diode, or
RR = V/IR
40 JUNCTION DIODES (~2.3]
The a-c resistance of a reverse-biased diode is rather high. The mag-
nitude of the a-c resistance may be determined graphically by finding the
inverse slope of the characteristic curve at the operating voltage.
Example 2.4. Find the values of the a-c and doc reverse resist-
ances for a reverse bias of 5 volts and 10 volts, respectively. The
reverse characteristic curve is shown in Fig. 2.11.
-v -lOv - 5v o
I
I
I
I I
4-- -~-- -2.0
I -2.5
-3.0
Ileokoge
-I (microamperes)
Fig. 2.11. Reverse diode characteristics for Ex. 2.4.
Solution: The reverse current IR is composed of two components,

the reverse saturation current I, and the junction leakage current
IL Since the saturation current is constant beyond about -0.1
volt, the slope of the characteristic curve and, hence, the a-c
resistance are determined solely by the leakage resistance. From
the characteristic curve we see that the a-c resistance is almost
constant over a large range of operating voltage. The a-c resistance
at both operating points is the slope of the curve or
~v 5
rr ~ A1 ~ 0 6 ~ 10 X 106 ohms
Ll 0.5 x I -
The doc resistance is
5
Rr = 6 ~ 2 X 106 ohms at -5 volts
2.5 x 10-
10
and Rr = 3 X 10-6 ~ 3.33 X 106 ohms at -10 volts
Problem 1.8. A resistor (R = 106 ohms) and a diode are connected in

series with a 22t-volt battery. The diode is reverse biased. Assuming
[92.3] JUNCTION DIODES 41
that the diode leakage current is zero, what is the maximum value of I.
that can be tolerated if the voltage across R must always be less than
10 volts? To what diode reverse resistance does this correspond? (Ans.
RR;;:::: 1.25 megohms)
2.3.3. Junction Capacitance
The junction diode, like most other active electrical components, has
some capacitance that affects its operation. The effect may be unimportant
at low frequencies or with doc currents. The junction capacity becomes
very important in high frequency applications and in fast switching cir-
cuits, because this capacitance determines the maximum usable frequency
of operation and the waveform distortion obtained. Since the trend in
solid state circuitry is to higher frequencies, it is important to understand
the characteristics of the junction capacitance and to be able to predict
how it will vary with the operating conditions of the diode.
Like the dynamic resistance, the junction capacitance is a function of
the doc current. For this reason when we determine the junction capaci-
tance, we will actually have an incremental value or a value for small
signals operating around some point on the characteristic curve.
Two capacitance values are associated with a P-N junction. They are
(a) the transition capacitances CT, which is common to both forward and
reverse bias conditions and (b) the diffusion capacitance CD, which is
significant only when the diode'is forward biased.
The depletion region contains very few current carriers, but it does
contain many trapped, charged impurity atoms. These charged, immobile
atoms set up an electric field. Assuming for a moment that no carriers
are present, the average distance between atoms with opposite charges and
the dielectric constant constitute an effective capacitance called the tran-
sition capacitance. * Although the capacitance is distributed and varies
with the width of the depletion region, it is somewhat similar to a parallel
plate capacitor if an average separation distance is used. For abrupt or
alloyed junctions, the transition capacitance is proportional to the inverse
square root of the doc voltage across the diode or
C _ dielectric constant x cross-sectional area

T - effective width of depletion region
(2.15a)
(for abrupt junctions)
where K is a constant for a given diode
VR is the total doc voltage across the diode (including both
thermal and external bias effects).
Also called the barrier or depletion-layer capacitance.

For graded junctions which have a gradual change from P-type to N-type
materials, the transition capacitance varies as the inverse cube root of the
d-c voltage or
(for graded junctions) (2.15b)
The diffusion capacity, unlike the transition capacity, is a function of

the density of the charge carriers stored in both the N-region and the
P-region. If the total charge stored in the diode can be found, the value
of the diffusion capacitance may be determined from the general equation
C = dQJdV (2.16)
The charge stored in a diode which has a junction area A is
Q ~ qA[L"PN + Lnnp]eqYlkT coulombs (2.17)
and the diffusion capacitance, by Eq. (2.16), is
C = qA [qL P + qL n ]eqY/kT (2.18)

D kT I' N n P
Equation (2.18) is rather cumbersome. Some simplification results from

the use of the minority carrier relationship
L=~ (2.19)
where L is the diffusiop length in centimeters

D is the diffusion constant in square centimeters per second
't is the carrier lifetimes in seconds.
Substituting Eq. (2.19) into Eq. (2.18), we obtam
C ~ qA [qDpPN 't + qDnnp T ]eqY/kT (2.20)

D kT L I' L n
I' n
which from Eq. (2.9) can be written in the more recognizable form of
(2.21)
where II' is the maximum hole current in the N-region [Eq. (2.7)]
In is the maximum electron current in the P-region [Eq. (2.8)]
'tp is the average lifetime of a hole in the N-region
'tn is the average lifetime of an electron in the P-region.
The total current is the sum of II' and In, each of which will increase or
decrease as the diode current increases or decreases. The diffusion capaci-
tance will vary directly with the diode current. To keep CD low, the diode
should be operated at low d-c current levels and the minority carrier
lifetimes should be as low as possible. *
Let us now study the transient current produced when the bias on a
conducting diode is suddenly reversed. As shown in Fig. 2.12, the current
does not immediately reach a steady state condition. The stored charge
flows out of the diode in a reverse direction and produces an undershoot,
Time
fO I
I
I
Fig. 2.12. The current in a conducting diode that is suddenly reverse

biased at to. The initial reverse current is large because of the stored
charge in the diode.
i.e., a rather large reverse current. The sudden current reversal and under-
shoot can be treated as the discharge current that flows from a condenser
when the external voltage is reversed. Fig. 2.13 shows the minority carrier
densities (a) with forward and (b) with reverse bias conditions. From this
figure, we can see that the minority carrier densities must change drastic-
ally with the change in bias. Before the bias is reversed, holes in the
N-region diffuse away from the junction. When the bias is reversed,
these extra holes are eliminated in two ways: They may diffuse into the
N-type material and recombine, or they may return to the P-side by
recrossing the junction. The reverse current shown in Fig. 2.12 is produced
by carriers that return to the side of the junction from which they started.
A minority carrier does not necessarily change directions when it
collides with an atom in the crystal. However, unless such a collision
does occur, the carrier cannot readily change its direction of travel.t
Therefore, the time between collisions determines the minimum time for
The lifetime indicated here is more or less an average lifetime determined by the
lifetime of carriers on the surface of the junction and by the lifetime of carriers in the
volume of the diode. The relationship of the three is
1 I I
-=-+-
'T T8 Tv
t If the carriers slowly diffuse further away from the junction and recombine, the
effect would be inductive rather than capacitive.
carriers to change directions when the bias is reversed. Once the carrier
has lost its momentum, it will be under the influence of the new reverse
voltage and will be drawn back acrosss the depletion region. The average
time between collisions is approximately one picosecond (10-12 sec).
N- type
P-type II N-type
--~~
ll--- (-)
ll---:.-p (+)
--... -...::---IT/I'
",
(-)
(0) (b)
Fig. 2.13. A comparison of carrier concentrations near the jtmction for

(a) forward bias and (b) reverse bias. Note the cbange in minority carrier
concentration in the vicinity of the junction. Holes are represented by ( +)
and electrons by ( - ). (Adapted from Electronic and RtuJio Enlfineering
by F. E. Terman. Copyright, 1955, McGraw-Hill.)
Almost immediately after the bias reversal, a large number of minority

carriers adjacent to the junction are swept back across by the barrier.
The reverse current is quite high momentarily due to the large number of
carriers available, but rapidly decreases to Is.
Example 2.5. A germanium diode made of 10 ohm-cm N-type

material and I ohm-cm P-type material has lifetimes of 10-6 sec
for holes and 0.5 x 10-6 sec for electrons. What is the diffusion
capacitance at a current of t rna if the junction area is 0.01 cm2?
Solution: Assume that the diffusion constants in germanium are
46.8 for holes and 98.8 for electrons. The diffusion lengths are
Lp =.J DpTp = ,,46.8 X 10-6 = 0.0068 cm
and Ln = .J DnTn = " 98.8 x 0.5 x 10-6 = 0.007 cm

(~2.3] JUNCTION DIODES 45
Assuming further that the conductivity is due only to the majority
carriers, we can find the concentration of carriers:
1
Pp>::::!--
qppJ.lp
1 I
or
PP >::::! ppqJ.lp = (1.0)(1.59 x 10-19)(1900)
= 3.3 x 1015 holes/cm3
Then n --------
n; 5.69 x 1026
p - PP - 3.3 X 1015
= 1.72 X 1011 electrons/cm3
These corresponding terms for the N-material are

nN = 1.61 x 1014 electrons/cm3
and PN = 3.53 X 1012 holes/cm3

The maximum hole diffusion.current Ip is
39V
I = AqD p PN 6
P Lp
(0.01)(1.59 x 10-19)(46.8)(3.53 x 1012)639V
3
6.8 X 10-
= 3.87 X 1O-5639v ma
39V
AqD n e
and I = n P
n Ln
(0.01)(1.59 x 10-19)(98.8)(1.72 x 1011)e39V

7 x 10-3
= 3.87 x 10-6e39v ma
The ratio of hole current to electron current is 10 to I, and
therefore I>::::! Ip The diffusion capacitance CD is
CD>::::! qI p7:p/kT 40(1.0
>::::! x 10-3)(10-6) 0.04 J.lf
>::::!
Problem 2.9. A silicon diode made of P-type material doped to one

part in 105 and N-type material doped to one part in 107 has carrier
lifetimes of T p = 10-7 sec and Tn = 10-8 sec. What is the diffusion
capacity if the current is 5 rna? Silicon has 5 x 1022 atoms per cubic
centimeter. (Ans. 0.019 J.lf)
Problem 2.10. If the diffusion capacity of a junction diode is 0.0002 jkf
at 1= Ima, at what frequency will the reactance of the diode be equal
46 JUNCTION DIODES [~2.4)
to the dynamic resistance of the diode? If C T = 50 picofarads, what
will this frequency be if I is increased to 2 rna? (Ans. 28.3 mc)
Problem 2.11. Repeat Prob. 2.10 when I is decreased to 0.5 rna. What
do you conclude about the product of ruC as a function of the d-c
current in a junction diode?
2.4. SPEED OF RESPONSE
Switching times as short as one or two nanoseconds (10-9 sec) are

now available in silicon diodes. These diodes have special surface treat-
ments and small amounts of foreign impurities added to speed up the
recombination rate of minority carriers. This shortens the lifetime r of
minority carriers, thus reducing the diffusion capacitance. The relation-
ship between stored charge or diffusion capacitance and the speed of
response of a diode can be explained by noting that the diode cannot
conduct until a minority carrier gradient has been established. This con-
dition was first observed in Chap. I where, from the diffusion current
equation repeated here, the current that flows IS
1= qDA (- slope of density curve)
A finite amount of time is required for the "slope" term in the diffusion
current equation to reach its final amplitude. The amount of time neces-
sary to establish this charge distribution is indicative of the speed of diode
operation. A fast diode req uires very little additional stored charge in
order to produce an increase In current. For thIS reason, the stored charge
and, consequently, the diffusion capacitance are reduced as much as
possible for fast diodes.
2.5. DIODE EQUIVALENT CIRCUIT
If the electrical characteristics of the diode are to be utilized to the

fullest extent in circuit theory and circuit design, these characteristics must
be incorporated into an "equivalent circuit" composed of familiar circuit
elements. Section 2.3 discussed most of the more important diode charac-
teristics. In this section these characteristics will be combined into an
equivalent circuit for small signal applications. This circuit is shown in
Fig. 2.14. Rp and RN are the ohmic resistances of the P- and N-regions,
respectively. CD and C T are the diffusion capacitance and the transition
capacitance. The dynamic junction resistance is rD' RSL is a resistance to
account for the surface leakage across the junction. These circuits can be
simplified because of the relative magnitudes of some of these parameters.
In Fig. 2.14a, for example. the parallel combination of RSL and rD is
approximately rD because RSL is so large compared with rD' In addition,
Rp and RN can often be neglected for l<;>wresistivity materials. For the
reverse bias circuit, the bulk resistances can always be. neglected with
respect to the large value of RSL' Simplified diode equivalent circuits for
forward-bias and reverse-bias conditions are shown in Fig. 2.15.
Cr
(0 )
Fig. 2.14. Small-signal equivalent circuits for a diode under (a) forward-
bias conditions and (b) reverse-bias conditions.
(0 )
Fig. 2.15. Simplified diode equivalent circuits for small signals: (a)
forward bias and (b) reverse bias.
These equivalent circuits are rather general in nature. Magnitudes

associated with the circuit components must be obtained for the particular
doc operating point used. Values may be obtained from manufacturers,
specification sheets or by the methods covered earlier in this chapter.
Problem 2.12. Twoidenticaldiodesare connectedin parallel. How will

the combined dynamic resistance of the two parallel diodes compare
with the dynamic resistance of a singlediode if the total circuit current
is the same in both cases?
Problem 2.13. In the circuit of Fig. 2.16, show that the a-c voltage
across R increaseswith frequencyand finallylevelsoff ata value equal
to E .
Problem 2.14. In Fig. 2.16 plot the approximate output voltage Vo if
the signal e. is a square wave having a peak-to-peak amplitude E, Label
all maximum and minimum values. Remember that the voltage cannot
change instantaneously across a capacitor that is in series with a resistor.
2.6. BREAKDOWN (ZENER) DIODES
If the reverse bias on a junction diode is made large enough, the diode
will break down; i.e., a condition will be reached in which the reverse
current will increase rapidly with a small increase in junction voltage. A
-v v breakdown
-I
Fig. 2.17. Reverse-bias curve for a junction diode showing breakdown.

reverse characteristic curve is shown in Fig. 2.17. Special diodes are

available with breakdown voltages ranging from three or four volts to more
than one hundred volts. The factors contributing to the breakdown phe-
nomenon are not completely understood, but in general it can be explained
on the basis of two effects: (a) Zener (i.e., high field emission) and (b)
avalanche multiplication. Although the name "Zener diode" is the com-
monly used term, most of the breakdown diodes derive their characteristics
from a combination of the two effects.
2.6.1. Zener Breakdown
When both sides of a P-N junction are highly doped, the width of
the depletion becomes very small. For abrupt junctions the total width
of the depletion region W is
(2.22)
where Xp and XN are the widths of the depletion region in the P-

material and in the N-material, respectively
V is the voltage across the barrier
Keo is the dielectric constant
Na is the impurity density (acceptor atoms) in the P-
region
Nd is the impurity density (donner atoms) in the N-
region
The field intensity cC in the depletion region is
(2.23)
where
V is the barrier voltage plus the external reverse bias.
When the field intensity becomes sufficiently large (l05 to 106 volts/cm),
electrons in the valence band of the P-material can be pulled directly across
the junction into the conduction band of the N-materia1. This process is
called Zener breakdown or internal field emission.
The path for Zener current carriers is different from the path of normal
reverse-current carriers. In the latter case, electron-hole pairs are formed
with electrons moving from the valence band to the conduction band in
the P-material. The electrons then slide down the potential barrier to
the conduction band in the N-materia1. In the Zener case, electrons tunnel
through the forbidden gap from the P- to the N-material. Tunneling is
a quantum-mechanical process in which an electron moves from one side
of a junction to the other when the electron did not have sufficient energy
50 JUNCTION DIODES (~2.6)
to cross the forbidden gap on one side of the junction. For very high
doping densities, the barrier potential is sufficient to produce tunneling.
With slightly lower densities, an external bias voltage is required to give
the electrons the necessary energy to tunnel through the barrier. Tunnel-
ing will be considered in some detail in Sec. 2.7.
True Zener breakdown occurs only when the width of the depletion
region is in the order of 100 A or less. Since the width of the forbidden
gap decreases as temperature increases, electrons tunnel through the junc-
tion for lower external voltages at higher temperatures. This temperature
dependence gives the true Zener diode a negative temperature coefficient.
2.6.2. Avalanche Breakdown
Another factor that produces breakdown is the avalanche effect. As

indicated in our previous discussion, the reverse diode current is com-
posed of the drift of minority carriers across the junction. In moving
across the depletion region, a minority carrier may have thousands of
collisions with the lattice atoms. Although the electric field tends to ac-
celerate the carriers, the numerous collisions repeatedly slow them down
and cause them to lose the energy acquired from the field. As a result, the
carriers drift with an average velocity instead of being rapidly accelerated
toward the junction. If the voltage across a junction is increased, a value
can be obtained for which a carrier acquires enough energy between
collisions so that it ionizes an atom upon impact. Now instead of a single
carrier, there are three carriers: the original (assuming it is not lost by
recombination) and an electron-hole pair. Since all three carriers are
swept out of the junction by the field, the reverse current is increased. TIns
increase multiplied by the many thousand collisions for each carrier pro-
duces a large increase in current. This phenomenon is called avalanche
breakdown and occurs in transistors as well as in diodes. Because the
breakdown occurs very rapidly, this effect is useful in transistor circuits
in which fast rise times are desired.
The avalanche potential is dependent upon the resistivity and type of
material. P-type material will break down at lower potentials than N-type
material because the minority carriers in P-material (electrons) have a
higher mobility (cm2jvolt-sec) than the minority carriers in N-material
(holes). This means that for a constant applied field, an electron will drift
at a higher velocity than a hole. Since the energy conveyed to an atom
upon collision is dependent upon the carrier velocity (E = tmv2), an
electron acquires sufficient energy to ionize an atom at a lower voltage
than a hole, so that electrons and P-type material have the lower avalanche
potential.
If the temperature increases, the mobilities of holes and electrons de-
crease, because more collisions occur in a unit distance of travel. With less
distance between collisions, a higher voltage is required to produce break-
down. Diodes with breakdown voltages above approximately six volts
have positive temperature coefficients, whereas diodes with lower voltages
have negative coefficients. This observation further serves to strengthen
the theory that both Zener and avalanche effects are present in reverse
breakdown, with the Zener effect dominant in the lower voltage diodes
and the avalanche effect dominant in the higher voltage diodes.
Example 2.6. If the electric field required to cause breakdown is

2 x 105 volts/em, what reverse bias must be applied to the diode
in Ex. 2.1 to cause it to break down? What is the electric field
intensity (in volts per centimeter) across the depletion region with
no external bias?
Solution: The barrier thickness can be obtained for abrupt

junctions from the equation
Xp + XN = [ ----2VKo ] 1/2 [(nN)1/2

- + (pp)1/2]
- (2.24)
q(nN + pp) PP nN
where Xp is the penetration of the depletion region into the

P-region
XN is the penetration of the depletion region into the
N-region
V is the total voltage across the junction (the contact
voltage plus any external bias voltage)
K is the relative dielectric constant (K = 16 for german-
ium and 12 for silicon)
e0 is the dielectric constant for free space = 8.85 x 10-14
coulomb/volt-em
q is the charge on an electron or hole
=1.59 x 10-19 coulomb
PP is the majority carrier density in P-material
nN is the majority carrier density in the N-material.
The side of the diode with the lightest doping will contain the
larger part of the total depletion region. From Ex. 2.1, the carrier
densities for this diode are nN = 3.68 x 1015 andpp = 1.75 x 1017
The contact potential VB is 0.36 volt. The electric field intensity
across the junction is
V (Vl12
Iff ---- =,------,..---=-----,._=_ volts/em (2.25)
= Xp + XN [ 2Ko ] 1/2 [(nN)1/2 + (pp)I/2]
q(nN + pp) PP nN
52 JUNCTION DIODES [92.6]
For the unbiased diode,
(0.36)1/2
~ = 7.04 x 10 - 5 = 8540 volts/cm
The voltage required for breakdown is
Vt2 = If B [_ 2Kfo ] tl2 ["(nN)t/2 + (pp)1/2]

q(nN + pp) PP nN
= (2 x 105)(7.04 x 10-5):::::: 14.1
or V = 199 volts
The required external bias is the difference between this voltage
and the contact voltage, or
Vbias = 199 - 0.36 ~ 199 volts
Problem 2.15. Calculate the breakdown voltage for (a) a germanium

diode made of 0.1 ohm-cm P-material and 1.0 ohm-em N-material
(b) a germanium diode made of 0.01 ohm-cm P-material and 0.1 ohm-cm
N-material. Assume that the required electric field for breakdown is
3 x 105 volts/em, fLn 2000 cm~!volt-sec, and fLp = 1200 cm2/volt-sec
at 25 C. (Ans. (a) 520 volts and (b) 55 volts)
Problem 2.16. Repeat Prob. 2.15 for 0.1 ohm-em N-type material and
1.0 ohm P-type germanium.
Rs
r Ir
.-
I !Iz -.JL
Eo-c(unregulated) ~
+
Zener diode
R.
1
16(regulated)
L
Fig. 2.18. A simpleZener-clioderegulatedvoltagesupply.
2.6.3. Breakdown Diode Characteristics and Applications
Break<lown diodes have many applications. Only a few of the more

common ones will be discussed here.
A major application of "Zener" diodes is in voltage regulating circuits.
A simple circuit is shown in Fig. 2.18. In this circuit, the Zener current
changes as the load current changes so that the total current and the load
[g2.6] JUNCTION DIODES 53
voltage will remain approximately constant. In addition, if the input
voltage Vi changes, the Zener voltage will tend to maintain a constant
voltage across the load. This application. is illustrated in the following
example and problems.
Example 2.7. The nominal Zener voltage in Fig. 2.18 is 10 volts

at a current of 5 rna. What value of series resistance is required
if Vi is 28 volts, and RL is constant at WOO ohms?
Solution: If Iz is 5 rna and IL is 10 rna, the total current will be
IT = Iz + IL = 15 rna
V; - Vz 18 volts k'i h
Then Rs = --- = = 1.2 1 0 ms
IT 15 rna
Problem 2.17. A 10-volt Zener diode is used in a regulating circuit as

shown in Fig. 2.18. The lO-volt nominal value will be obtained if the
Zener current is maintained between 5 rna and 40 rna.
(a) If the supply voltage is 22t volts, what value of R. should be used
when the load varies between 250 ohms and 2 kilohms? (Ans.
278 ohms)
(b) If the load current is constant at 30 rna, how much can the battery
voltage decay with age before it must be discarded using the value
of R. calculated above? (Ans. !1E= 2.77 volts)
(c) Find the range for R. if the input voltage is constant at 22t volts
and the load changes from 750 ohms to 1500 ohms. (Ans. 268
to 682 ohms)
Problem 2.18. The positions of the diode and R. in Prob. 2.17 are inter-
changed.
(a) If R. is infinite (i.e., an open circuit), what are the limiting values
on RL if Vi = 22t volts and Iz varies between 5 rna and
40 rna?
(b) If RL is constant at 500 ohms, what is the lower limit on R.?
Some of the characteristics that are important in Zener applications

are power handling ability, dynamic resistance, and temperature coefficient.
These characteristics will be considered in that order.
The maximum power that a breakdown diode can dissipate is deter-
mined by the junction temperature. This power is
Pmax = ImaxVZ
and is normally stated for an ambient temperature of 25 C. As the
ambient temperature rises, the diode power rating must be decreased to keep
the junction temperature within a safe limit. A typical derating scale is
shown in Fig. 2.19. Usually the maximum power rating is derated a certain
number of milliwatts per centigrade degree rise in the ambient temperature.
Problem 2.19. A one-watt, lO-volt Zener diode is used in Fig. 2.20 for
a shunt voltage regulator. The diode is to be derated according to the
curve in Fig. 2.19.
Q)
~
0 100
~
0
=u
00
ELO 75
;:,(\J
E-
'x 0
oc
-.0
0a.
-
E.!:?
Q)';;;
50
C>lJ)
0'-
_'U
25
c ...
Q)Q)
~ ~
Q)O
Cl.a.
a
a 25 50 75 100 125 150 175
Ambient temperature (OC)
Fig. 2.19. A typical Zener-diode temperature-derating scale.
(a) If RL is 300 ohms, what are the lower limits on R. at 50 C and

100 C? (Ans. 335 and 518 ohms)
(b) If R. is 500 ohms and RL is 2 kilohms, how much can the tem-
perature increase above 25 C without overheating the diode?
Rs
48v 10v
1w

It is not unusual for Zener diodes to be used in circuits containing

both a-c and d-c voltages. Under these conditions, it is often desirable to
remove the a-c from the output Removing the ripple in the output
voltage of a d-c power supply is a good example. The extent to which this
can be accomplished depends upon the dynamic resistance of the diode at
the operating point. For good regulation, the dynamic resistance should
be as low as possible. Experimentally, the dynamic resistance is deter-
mined by measuring the a-c voltage developed across the diode when a
small a-c current is superimposed on the d-c current at the operating point
+I
-v +v
E
,4
I I
l I
:~ 1i
I r.=--
~v,
I ___ JI 1 ~ 1i
~V1
2 :
I
I~I2
I
I
, I
, I
L_J -I
AV2
Fig. 2.21. Dynamic resistance of a breakdown diode at two operating

points.
The resistance can be obtained from the characteristics of the diode by

obtaining AE/.M at the operating point. This procedure is illustrated for
two operating points in Fig. 2.21.
The voltage across a breakdown diode is dependent upon temperature.
If the voltage increases with an increase in temperature, the device has a
positive temperature coefficient; if the voltage decreases, the temperature
coefficient is negative. Diodes with breakdown. voltages below approxi-
mately 4 volts usually have negative coefficients, in agreement with true
-...J
Zener theory. Diodes with breakdown voltages above approximately 6
volts usually have positive temperature coefficients, in agreement with
avalanche theory. Diodes between these two groups can have either
positive or negative temperature coefficients, depending upon the reverse
current. By choosing the proper reverse current, it is possible to obtain a
Reverse bios (volts)
-7 -0
o
-25
~
-..~
o
..
Q)
fII
Co Q)
E_
Q)e -50 Q)
._--~
-Q)
Q)'-
> .~
:::
(/lQ)
Co
E
o
-..
00 E
0..0
-75 e
Q)
~
o
Gl
~
, -100 ~
Q)
0::
,
I
-125
-+25C
--+100C
-150
Fig. 2.22. Typical Zener-diode temperature-coefficient curves.
point of zero temperature coefficient for these "in-between" diodes. For

some regulation and reference applications a zero temperature coefficient
is very desirable. Typical characteristic curves for the three cases are
shown in Fig. 2.22.
Problem2.20. A lO-voltZener diode has a temperature coefficientof
+ 10 mvrC at a reversecurrent of 20 rna. The dynamic resistance at
this current is 25 ohms. The measured voltage at 25 C and 20 rna is
10.018 volts.
(a) The diode is to be placed in an oven and maintained at a constant
temperature of 75 C. According to Fig. 2.19, how much must
the power handling capabilities be derated?
(b) What will be the diode voltage at the elevated temperature if the
current remains essentially constant?
(c) The input voltage is 38 volts. What resistance will be necessary
in series with the diode when T = 25 C? How much will the
diode voltage vary if the input changes :l: 1 volt? (Ans. 40 per
cent, and 10.518 volts)
Problem 2.21. The series resistor in an unloaded Zener regulated

voltage supply is 2000 ohms. What is the maximum dynamic resistance
allowable if the ratio of change in output voltage to change in input
voltage is to be less than lo?
Problem 2.22. A resistor and a breakdown diode in series are sup-

plied from a constant 12-volt source. The resistor has a value of
800 ohms at 25 C. What must the diode voltage be so that the
current through the resistor is 8 rna? If the resistance has a posi-
tive temperature coefficient of 0.000625tC, what must the diode
coefficient be in order for the resistor current to be independent of
temperature?
2.7. TUNNEL DIODES
The tunnel or Esaki diode is a highly doped, alloyed junction diode

which exhibits a negative resistance over part of its operating range.
Typical doping is in the order of 1019 to 1020 impurity atoms per cubic
centimeter, as compared with 1013 to 1017 impurity atoms per cubic centi-
meter for ordinary junction diodes. The width of the depletion region in
a tunnel diode is approximately 10-6 em compared to approximately
4
10- em for an ordinary diode. The highly doped materials and the cor-
responding narrow depletion region cause the internal potential field across
the depletion region to be as large as 105 to 106 volts/em. This large field
is instrumental in producing a negative resistance region in the V-I
characteristics of the tunnel diode. A typical V-I curve for a germanium
tunnel diode is shown in Fig. 2.23.
One of the most interesting features of the tunnel diode is the negative
resistance region in the characteristic curve. This feature, together with its
speed of response, makes it useful both as a high frequency amplifier and
as an oscillator. The heavy doping used in tunnel diodes makes them less
sensitive to temperature variations and radiation damage than ordinary
diodes and transistors.
58 1UNCTION DIODES [~2.7]
1.0
~fJI
I
I
..
Q)
Q)
I
0- I :-r
E
0 I
I
.E 0.5
-....
c:
Q)
I
I
I
:J
I
U
Iv -,-- ------
I
o 100 200 300 400 500

Vp Vv VF
voltage (millivolts)
Fig. 2.23. A germanium tunnel-diode V-I characteristic curve.
2.7.1. Tunneling Current and Negative ResistGnce
The negative resistance characteristic is obtained from the quantum-

mechanical tunneling effect mentioned earlier in the discussion of Zener
diodes. This topic is too complicated to be discussed in detail here.
Instead, we will proceed with a simple intuitive explanation of the tunneling
effect.
As indicated above, when two pieces of doped semiconductor materials
are joined to form a diode, their mean energy levels (the Fermi levels) are
equal if the diode is unbiased. In ordinary junction diodes, this level is in
the forbidden region between the conduction band and the valence band
as shown in Fig. 2.24a. In the tunnel diode, however, there is an overlap,
A, ofthe two energy bands due to the narrow junction and the high degree
of doping as shown in Fig. 2.24b. In this overlapped region, an electron
can move from the valence band in the P-material to the conduction band
in the N-material, or vice versa, without a change in its energy, because
parts of the two bands are on the same level. When empty energy states
exist in the overlapped region d, the internal field is great enough (without
bias) for some electrons to change bands without a change in energy. This
[~2.7) JUNCTION DIODES 59
transition occurs very rapidly. When a large number of carriers are pre-
sent, we can assume that some will cross the junction. This movement of
electrons across the junction without a change in energy is known as
tunneling and the resulting current is called tunneling current. In the un-
biased diode the net current is zero so that the tunneling current Ieov of
electrons moving from the conduction band to the valence band (positive
conventional current) is equal to the tunneling current Iv_e of electrons
moving from the valence to the conduction band (negative conventional
current).* The tunneling currents are quite large, about 103 amp/cm2, in
p
N o N
A
Fermi level
-- - --
( b)
(a)
Fig. 2.24. Energy-band diagrams for thermal equilibrium in (a) an

ordinary junction diode and (b) a tunnel diode.
each direction at room temperature. In order to simplify the discussion

of the characteristics of a tunnel diode, we will start our discussion by
assuming a temperature of absolute zero.
At this temperature and with no external bias voltage, both components
of the tunneling current are zero, because all of the electrons are in the
We are neglecting the reverse current and the forward diffusion current in ordinary
diodes because the high degree of doping makes them relatively small.
P N P N
Conduction Conduction
bond bond
j~/1
(Empty) (Empty)
(Filled)
1111/11/1/
01~l~f;h
.
Ii"') (Filled)
I 1/11/11
0,;/)i;;
(0) (b)
p N P N
Conduction Conduction
bond bond
(Empty)
[77{TTlm
j~jj)j~~~//l'
(Filled) /111
(Empty)
'TTITT,zm
fj
. (Filled)
1'11111
Valence bond
/////////1////
(e)
Fig. 2.25. Tunnel-diode energy-band diagrams for different forward-bias

conditions at absolute zero: (a) zero bias; (b) slight forward bias; (c)
biased to point of maximum forward tunnel current; (d) biased slightly
beyond point of maximum forward tunnel current; (e) forward biased until
tunnel current is again zero.
lowest energy states. Therefore, the overlapped energy bands have only
filled states opposite filled states, and empty states opposite empty states, as
shown in Fig. 2.25a. When a small forward bias is applied to the diode,
the energy levels of the N-side will be raised relative to those of the P-side,
as indicated in Fig. 2.25b, so that some of the filled states in the conduction
band are opposite empty states in the valence band. Consequently, some
of the electrons from the conduction band in the N-material will tunnel
across the junction to the valence band of the P-material, so that a current
flow in the forward direction results. Tunneling from the valence band to
the conduction band will not occur, because there are no empty states
with the proper energy level available in the conduction band. As the
forward bias is further increased, the current will increase until all of the
I I
v v
(0 ) (e)
Fig. 2.26. The forward current components of a tunnel diode: (a)

tunneling current; (b) diffusion current; (c) total current.
filled energy states in the conduction band are opposite empty states in
the valence band, as shown in Fig. 2.25c. At this point, the forward
tunnel current will reach a peak value, corresponding to Ip in Fig. 2.23,
and the bias voltage will be Vp When the forward bias is increased beyond
Vp, the overlap of filled energy levels in the conduction band and empty
energy levels in the valence band decreases, as shown in Fig. 2.25d. This
condition causes a decrease in current below the peak current Ip The
tunneling current will continue to decrease with increasing bias until a
point is reached at which the current due to tunneling is again zero, as
indicated in Fig. 2.25e; this condition corresponds to the valley current
Iy in Fig. 2.23.
The total forward current of the tunnel diode at room temperature
is the sum of the tunneling current and the diffusion current, as shown
in Fig. 2.26. The only departure from theory appears in the form of an
excess current in the valley region of the curve. At this low value of bias
voltage, the diffusion current is not large enough to account for the current
measured on actual diodes. The presence of this "excess" current is not
fully understood.
When the bias is reversed, tunneling occurs from the valence band to
the conduction band. Since the carriers involved in this case are also elec-
trons, the current with reverse bias will be in the direction opposite to
that of the forward current. This is mentioned to reemphasize the fact
that only electrons are involved in the tunneling process, regardless of the
direction of current flow. The energy-band diagram for reverse biasing is
shown in Fig. 2.27. The current increases rapidly with reverse bias. The
p N p N p N
Conduction
bond
(Empty)
~mrJI
/I II IliJI}
ffiI
volence
bond
/1/1/1/1
(0) ( b) (c)
Fig. 2.27. Energy-band diagrams for a reverse-biased tunnel diode at

absolute zero: (a) zero bias; (b) smaU reverse bias; (c) large reverse bias.
reverse current in a tunnel diode is the same as the reverse break-

down current in a true Zener diode, the only difference being that the
tunnel diode is doped so heavily that breakdown occurs at essentially
zero voltage. The reverse current as a function of reverse bias is shown in
Fig. 2.28.
At a temperature above absolute zero, the two components of tunnel
current, Iv_e and Ie_v, have non-zero magnitudes with zero bias. As the
forward bias is increased, the forward current increases toward a peak
value Ip and the reverse current decreases toward zero. A reverse bias
causes the forward current to decrease and the reverse current to increase.
The current components and the total current for a temperature greater
than absolute zero are shown in Fig. 2.29.
-v
-I
Fig. 2.28. The voltage-current relationship of a tunnel diode under

reverse-bias conditions.
+I
-v
,,
-I
Fig. 2.29. Tunneling currents for a temperature greater than absolute

zero.
The tunnel diode has a negatil'e resistance region extending from the
peak current Ip to the valley current Iv. This region is called a negative
resistance region because an mcrease in the voltage across the diode results
in a decrease in the diode current. The negative resistance region is im-
portant for several reasons. When the diode is operated in the negative
resistance region, it does not absorb power from an a-c signal across its
terminals as a positive resistance does. Instead, the tunnel diode will in-
crease the signal. This characteristic of the tunnel diode makes it useful
as an amplifier. Furthermore, this region of the diode characteristics is
unstable. With the proper circuitry this instabihty can be made to produce
useful pulses or sinusoidal oscillations. Since carrier storage is not involved
in the current-conduction process, the current in a tunnel diode can be
changed extremely fast. ThIs characteristic makes the tunnel diode very
valuable in high-speed switching circuits and high-frequency oscillators.
2.7.2. Tunnel-Diode E.quivale.nt Circuit
The small-signal equivalent circuit for a tunnel diode operating in the

negative resistance region is shown in Fig. 2.30 where L. is the series lead
Ls
o 1)
I
c -L -r
Fig. 2.30. Tunnel-diode small-signal equivalent circuit,
inductance, R. the series lead and bulk resistance, C the diode transition
capacitance, and - r the dIOde negative resistance. Typical values are
10 nanoheneries for L.one ohm for R., 10 picofarads for C, and 100 ohms
for r.
The time constant rC, which determines many of the operating
characteristics of the tunnel diode, is strongly dependent upon the degree
of doping in the semiconductor material. Measured values vary with
doping as shown in Table 2. I _
When tunnel diodes arc incorporated into circuits, the equivalent
circuit is altered to include the source and load impedances and any in-
ductance and capacitance intentionally or unintentionally added to the
circuit. The new or modified equivalent circuit is shown in Fig. 2.31,
Table 2.1
Doping r(ohms) "c (nanoseconds)
2.4 x 1019 90 4.5

3.6 X 1019 4.5 0.9
4.8 X 1019 1 0.05
where LT represents the total circuit inductance, RT the total circUit

dissipative resistance, C the total capacitance, and - r the diode negative
resistance.
Ctotal -r
Fig. 2.31. Tunnel-diode circuit with added resistance, inductance, and

capacitance.
Example 2.8. At low frequencies the reactive components in the

equivalent circuit can be neglected. Also, in comparison with the
diode negative resistance, the bulk and the lead resistance can
usually be discarded. Making these assumptions, find the RMS
load current in Fig. 2.32 when the input current is 0.0025 milli-
amperes. The diode negative resistance determined from the slope
of the characteristic curves is 100 ohms and the load resistance
is 95 ohms.
Solution: By the current-splitting theorem used in circuit analysis,
the load current is
IT(-r) (0.0025)(-100) .
IL = R + (-r) = 95 _ 100 = 20(0.0025)
L
= 0.05 ma RMS.
ID~
T.O. -r
(a) (b)
Fig. 2.32. Tunnel-diode circuit for Ex. 2.8: (a) circuit and (b) simplified
equivalent circuit.
T.O.
!
Vi =10mv
-r=100.o.
T
Observe that the load current is actually larger than the input
current. The increase in current is attributed to the tunnel diode.
T.O.
-r: 130.0.
Rp
1
11,.: 5 mv
I
[~2.8] WNCTION DIODES 67
Problem 2.23. Making the same assumptions as in Ex. 2.8 for the circuit
shown in. Fig. 2.33, find (a) the RMS output voltage when the input
voltage is 10 mv; (b) the input impedance of the circuit; (c) the RMS
voltage across the diode. (Ans. (b) 5 ohms)
Problem 2.24. The circuit in Fig. 2.34 is the simplified low frequency
equivalent of a compound type tunnel-diode circuit. Find the output
voltage of this circuit (a) when Rp = <Xl and (b) when Rp = 1250.
2.8. THE BACKWARD DIODE
The backward diode is a'. semiconductor diode which is doped to a

level just less than that necessary to cause tunneling. With a forward bias
the diode current is a forward diffusion c~rrent similar to the current in
lmo +I- - - - - - - - - -
-v lOOmv +v
500mv
-lmo
-I
Fig. 2.35. Backward-diode characteristic curve.
an ordinary junction diode. With a small reverse bias, the electric field
across the junction becomes large enough to produce tunneling and break-
down. A typical backward-diode characteristic curve is shown in Fig. 2.35.
The backward diode has a lower voltage drop in the reverse direction
than in the forward direction. This characteristic is useful when backward
-------------------------------- ----

diodes are used in conjunction with tunnel diodes, because the signal levels
are small enough so that the backward diode can block in the forward
direction and conduct in the reverse direction. Normal diodes are difficult
to use with tunnel diodes because the high conductance region occurs at
voltages above tunnel-diode signal levels. Backward diodes, with their
high reverse conductance, do not have this drawback.
CHAPTER PROBLEMS
2.25. Sketch the output waveshape for the circuit in Fig. 2.36 assuming
that the diode has an infinite reverse resistance (a) when ep > E and
(b) when ep < E.
WI'.---ll
R
tn 0
o~ ~ r
2.26. Draw the total V-Icharacteristics for the circuits shown in Fig. 2.37
if the diodes have infinite back resistances and if the forward resistances
are negligible with respect to the passive circuit elements. Label all
important points and slopes in the curves.
,K , K
E 2K 2K
-=-6V -=-6V
(0) (b)
Fig. 1.37. Circuits for Prob. 2.26.

2.27. Find the output for the input conditions shown in Fig. 2.38 for
the diode OR gate. Assume that the diodes are ideal; i.e., if a diode
conducts, it is a short circuit.
200n
e,
T
r
tK
(b)
Fig. 2.38. Circuit for Prob. 2.27: (a) input voltages and (b) circuit.
2.28. The circuit in Fig. 2.39 is an AND gate. Sketch and label the
output for the inputs shown. Assume ideal diodes.
OV
e,
-12V
I I
IOV I
I I
I
I
I
.I
I
I
2K 1~
I I
-12V I
+
71 T2 73 74 T5 76
To T7 Ta 12 V-=-
-.1 I
(a) ( b)
Fig. 2.39. Diode AND gate for Prob. 2.28: (a) imputs and (b) circuit.
2.29. The diode in Fig. 2.40 is intended to deliver positive pulses to the
load R. Determine the time constants of the leading and trailing edges
of the output pulse. In the diagram, CD is the diode capacitance (5 pf),
rD is the diode resistance (200 ohms), r. is the bulk resistance (50 ohms),
Co is the output capacity (50 pf), and Ro is the load resistance (1 kilohm).
Make reasonable assumptions; for example, CD is negligible during the
"on" time of the diode. Find the width of the narrowest pulse which
will allow the output to reach 98 per cent of full amplitude. (Ans.
'TriBe = 0.01 fl-sec; 'Trail = 0.05 fl-sec)
CD
t
e,.
Co
I
(0) (b)
Fig. 2.40. Circuit for Prob. 2.29: (a) schematic diagram and (b) equiva-
lent circuit.
-I' =150 T.D.
Fig. 2.41. Tunnel-diode circuit for Prob. 2.31.

[~2.8]. JUNCTION DIODES 71
2.30. Calculate the breakdown voltage for a germanium diode made
of 0.5 ohm-em P-materialand 0.1 ohm-em N-material. Assume that
a field of 3 x 105 volts/em is required for breakdown, fl-n = 3900 em2/
volt-sec, and fl-p= 1900 em2/volt-seeat25 C.
2.31. Find the current through the load resistor in Fig. 2.41 (a) if Rs =0
and (b) if Rs = 48. What value of Rs is required to obtain an output
current of 400 fl-a?
2.32. A diode is made with 10 ohm-em P-type material and 1 ohm-

em N-type material. Assuming that Dn = Dp ~ 80 em2/see, Lp = Ln =
1O-2cm, A = 0.001 em2, fl-n = 3900 cm2/volt-sec, fl-p = 1900 em2/volt-
sec and n; = 5.7 x 1026, find the current that flows when a forward
bias of 0.15 volt is applied to the diode. (Ans. 1= 0.93 rna)

Transistor Circuit Theory

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Transistor Circuit Theory

Uploaded by

Copyright:

Available Formats

TRANSISTOR

Consultant, Instrumentation and Controls Division

CHARLES E. MERRILL BOOKS, INC., Columbus, Ohio.

First Printing December, 1963

PRINTED IN THE UNITED STATES OF AMERICA

Chap. 6, the reader is in a position to design multistage circuits for

1.1. Energy levels in the isolated atom ~. . . . I

2.1. The potential barrier. . .. . .. . .. .. . .. . . .. . .. .. . . 25

4.1. Black box theory 103

5.1. The general hybrid equivalent circuit 148

6.1. Selection of the operating point for linear amplifiers 177

7.1. Major high-frequency effects in transistors , 223

8.1. Single-stage feedback. . . . . . . . .. . . . . . . . . . . . . . . .. 255

9.1. Tuned circuits 293

10.1. Origin and classification of noise . . . . . . . . . . . . .. 312

II. I. Power transistor characteristics. . . . . . . . . . . . . . .. 349

I. Table of elements 402

Transistors are made from semiconducting materials. As the name

1.1. ENERGY LEVELS IN THE ISOLATED ATOM

Fig. 1.1. The energy-level diagram for the hydrogen atom.

Example 1.1. Find the mmimum amount of energy required to

(a) EL - EK = - 3.39 - (- 13.58) = 10.19 ev

(b) Ep-Eo= -O.387-(-0.543)=0.165ev

In the first case an energy of 10.19 ev must be transferred to the

Solution: The energy released is equal to the energy required to

Problem 1.1. The ionization potential of a neon atom is approximately

1.2. ENERGY LEVELS IN CRYSTALS

Crystals are solids in which the atoms are arranged in an orderly

Example 1.3. How do the energy levels in an atom of silicon

1.3. ELECTRICAL CHARACTERISTICS OF A CONDUCTOR

As indicated in Sec. 1.2, the electrical characteristics of a solid are

Fig. 1.4. Motion of electrons in conduction band: (a) random motion

At a temperature of absolute zero, these electrons occupy the llx

* These energy levels are coupled, making them interdependent.

where (1 is drift conductance in mhos per centimeter

An electron may be accelerated to extremely high velocities in a very short period

This equation indicates that as temperature increases, the mobility con-

Example 1.4. The mobility constant for a sample of copper at

Resistivity is the reciprocal of the conductivity:

p = 1/(4.6 X 105) ~ 2.16 X 10-6 ohm-cm.

1.4. ELECTRICAL CHARACTERISTICS OF A

A semiconductor has a higher value of resistivity than a conductor.

A covalent electron in silicon must receive at least 1.11 ev of energy

* Hole flow is in the same direction as conventional current.

From the preceding discussion, it is now clear that two types of

Example 1.5. Find the conductivity and resistivity of pure ger-

1.5. ELECTRICAL CHARACTERISTICS OF AN INSULATOR

The major difference between semiconductors and insulators is in the

>- Valence band Valence band Valence band

carriers does increase with temperature, the mobility constant of a con-

Example 1.6. A material has a forbidden gap between the valence

Solution: From Eq. (1.3), the number of free electrons is

= 5.18 x 102/eso = 10-1

1.6. SEMICONDUCTORS WITH IMPURITIES

The resistivity of a semiconductor can be drastically reduced by the

A comparison of a cubic centimeter of intrinsic, P-type, and N-type

cases. This product is approximately 5.7 x 1026 for germanium and

1X 1016 1380 460 1280 460

A comparison of these values with those for intrinsic silicon may

Example 1.7. Find the average number of holes and electrons in

The number of holes is

Fig. 1.9. Current flow by tbe mecbanism of drift In an impurity semi-

I ',': 't' " I '1 ' ,