Escolar Documentos
Profissional Documentos
Cultura Documentos
CONTENTS
Preface
xiii
PA R T
Materials 1
Chapter 1
Crystal Structures, Planes, and Directions 39 Summary 41 Reading List 42 References 42 Review Questions 42 Problems 43
Chapter 2
Homogeneous Semiconductors
48
in Crystals
49
1.3.1 The Bohr Model for the Hydrogen Atom 5 1.3.2 Application to Molecules: Covalent Bonding 11 1.3.3 Quantum Numbers and the Pauli Exclusion Principle 13 1.3.4 Covalent Bonding in Crystalline Solids 14
Conduction Band Structure 56 Valence Band Structure 58 Intrinsic Semiconductors 59 Extrinsic Semiconductors 62
2.6.1 Donors 62 2.6.2 Acceptors 66
2.7
67
69
23
2.8
1.6.1 The Free Electron in One Dimension 23 1.6.2 The de Broglie Relationship 25 *1.6.3 The Free Electron in Three Dimensions 26 1.6.4 The Quasi-Free Electron Model 27 1.6.5 Reection and Tunneling 32
2.9
Fermi-Dirac Statistics
73
1.7
Energy
76
vi
Contents
3.7 3.8
93
94
2.12.1 Impurity-Induced Band-Gap Narrowing 94 2.12.2 Apparent Band-Gap Narrowing 97 2.12.3 Carrier Concentrations in Degenerate Semiconductors 99
Minority Carrier Diffusion Lengths Quasi Fermi Levels 149 Summary 152 Reading List 154 References 154 Review Questions 154 Problems 155
147
2.13 Summary
100
Chapter 4
Reading List 103 References 103 Review Questions 103 Problems 104
Chapter 3
Nonuniform Composition 170 Graded Doping and Graded Composition Combined 173 Summary 175 Reading List 175 References 175 Review Questions 176 Problems 176
Supplement to Part 1
Materials 179
Supplement 1A
Recombination
131
3.5.1 Band-to-Band Generation and Recombination 133 3.5.2 Two-Step Processes 133
3.6
133
182
Contents
vii
PA R T
2
239
Electrons Mechanics
185 187
Diodes 235
Chapter 5
Prototype pn Homojunctions
5.1 5.2
(Quantitative)
253
S1A.7 Summary 210 S1A.8 Review Questions S1A.9 Problems 211 Supplement 1B
211
5.3.1 Energy Band Diagram at Equilibrium (Step Junction) 253 5.3.2 Energy Band Diagram with Applied Voltage 256 5.3.3 Current-Voltage Characteristics of pn Homojunctions 263 5.3.4 Reverse-Bias Breakdown 284
5.4
215
5.5
Transient Effects
294
Bound States 219 S1B.3 Carrier Freeze-out in Semiconductors 222 S1B.4 Phonons 223
*S1B.4.1 Carrier Scattering by Phonons 228 S1B.4.2 Indirect Electron Transitions 230
5.6 5.7
301
Summary 232 Reading List 232 References 232 Review Questions 232 Problems 233
Built-in Voltage 302 Junction Width 302 Junction Current 303 Junction Breakdown 304 Capacitance 305 Transient Effects 305
5.8 Reading List 305 5.9 Review Questions 306 5.10 Problems 306
viii
Contents
Chapter 6
350
311
S2.3.1 Junction Capacitance in a Prototype (Step) Junction 350 S2.3.2 Junction Capacitance in a Nonuniformly Doped Junction 352 S2.3.3 Varactors 353 S2.3.4 Stored-Charge Capacitance of ShortBase Diodes 354
6.3
Semiconductor Heterojunctions
317
6.3.1 The Energy Band Diagrams of Semiconductor-Semiconductor Heterojunctions 317 6.3.2 Effects of Interface States 327 *6.3.3 Effects of Lattice Mismatch on Heterojunctions 329
Diodes
356
S2.4.1 Tunneling Through Schottky Barriers 357 S2.4.2 Barrier Lowering in Schottky Diodes Due to the Image Effect 359
361
6.4
Metal-Semiconductor Junctions
331
6.4.1 Ideal Metal-Semiconductor Junctions (Electron Afnity Model) 331 6.4.2 Inuence of Interface-Induced Dipoles 331 6.4.3 The Current-Voltage Characteristics of Metal-Semiconductor Junctions 334 6.4.4 Ohmic (Low-Resistance) Contacts 337 6.4.5 I -Va Characteristics of Heterojunction Diodes 339
S2.5.1 The Use of SPICE as a Curve Tracer 362 S2.5.2 Transient Analysis 365
PA R T
3
385
385
386
Capacitance in Nonideal Junctions and Heterojunctions 339 Summary 340 Reading List 340 References 340 Review Questions 341 Problems 341
The MOSFET
7.1 7.2
Supplement to Part 2
Diodes
346
346
7.3
7.2.1 Introduction to MOS Capacitors 7.2.2 MOSFETs at Equilibrium (Qualitative) 390 7.2.3 MOSFETs Not at Equilibrium (Qualitative) 392
MOSFETs (Quantitative)
403
S2.2.1 Case 1: Dielectric Relaxation Time for Injection of Majority Carriers 347 S2.2.2 Case 2: Injection of Minority Carriers 349
7.3.1 Long-Channel MOSFET Model with Constant Mobility 404 7.3.2 More Realistic Long-Channel Models: Effect of Fields on the Mobility 417 *7.3.3 Series Resistance 432
Contents
ix
Comparison of Models with Experiment 434 Summary 435 Reading List 438 References 438 Review Questions 438 Problems 439
8.11.2 MESFETs 476 8.11.3 Junction Field-Effect Transistors (JFETs) 481 8.11.4 Bulk Channel FETs: Quantitative 482
Chapter 8
Summary 485 Reading List 486 References 486 Review Questions 487 Problems 487
Supplement to Part 3
Field-Effect Transistors
S3.1 S3.2
491
Introduction 491 Comments on the Formulation for the Channel Charge Qch 491
S3.2.1 Effect of Varying Depletion Width on the Channel Charge 491 S3.2.2 Dependence of the Channel Charge Qch on the Longitudinal Field L 493
8.5
S3.3
8.6
Unity Current Gain Cutoff Frequency fT 463 *8.8 Short-Channel Effects 464
8.7
8.8.1 Dependence of Effective Channel Length on VDS 464 8.8.2 Dependence of Threshold Voltage on the Drain Voltage 466
8.9 MOSFET Scaling 467 8.10 Silicon on Insulator (SOI) 8.11 Other FETs 473
S3.7
Universal Relations for Low-Field Mobility 507 Measurement of VT 509 Alternative Method to Determine VT and lf Applicable to Long-Channel MOSFETs 513 MOS Capacitors 514
S3.7.1 Ideal MOS Capacitance 515 S3.7.2 The C-VG Characteristics of Real MOS Capacitors 520 S3.7.3 Parameter Analyses from C-VG Measurements 521
469
Contents
*S3.8
9.5
9.6 9.7 9.8 9.9 9.10 9.11 9.12 9.13 9.14 9.15 9.16 9.17
*S3.9
Device Degradation
530
*S3.10 Low-Temperature Operation of MOSFETs 535 *S3.11 Applications of SPICE to MOSFETs 538
S3.11.1 Examples of the Use of SPICE with MOSFETs 539 S3.11.2 Determining the Transient Characteristics of a CMOS Digital Inverter 543
Summary 545 Reading List 546 References 546 Review Questions 547 Problems 547
The Basic Ebers-Moll DC Model 583 Current Crowding and Base Resistance in BJTs 586 Base Width Modulation (Early Effect) 590 Avalanche Breakdown 594 High Injection 594 Base Push-out (Kirk) Effect 595 Recombination in the Emitter-Base Junction 597 Summary 598 Reading List 599 References 599 Review Questions 600 Problems 601
Chapter 10
PA R T
609
Introduction 557 Output Characteristics (Qualitative) 561 Current Gain 563 Model of a Prototype BJT 564
9.4.1 Collection Efciency M 567 9.4.2 Injection Efciency 568 9.4.3 Base Transport Efciency T 570
624
10.6
High-Frequency Transistors
625
10.7
628
Contents
xi
PA R T
5
675
632
10.8
635
Summary 638 Reading List 639 References 639 Review Questions 639 Problems 639
Generic Photodetector 675 Solar Cells 683 The p-i-n (PIN) Photodetector Avalanche Photodiodes 691
689
Supplement to Part 4
11.3
Light-Emitting Diodes
692
Bipolar Devices
642
Transistors (HBTs)
642
11.4
11.3.1 Spontaneous Emission in a ForwardBiased Junction 692 *11.3.2 Isoelectronic Traps 694 11.3.3 Blue LEDs and White LEDs 696 11.3.4 Infrared LEDs 696
Laser Diodes
11.4.1 11.4.2 11.4.3 11.4.4 11.4.5
702
Optical Gain 703 Feedback 706 Gain + Feedback Laser 709 Laser Structures 710 Other Semiconductor Laser Materials 714
11.5
Image Sensors
715
S4.5 Silicon Controlled Rectiers (SCRs) S4.6 Parasitic pnpn Switching in CMOS
654
11.6 11.7 11.8 11.9 11.10
Circuits
658 658
11.5.1 Charge-Coupled Image Sensors 715 11.5.2 MOS Image Sensors 717
S4.7.1 Parasitic Effects 661 S4.7.2 Low to Medium Currents 661 S4.7.3 High Currents 663
Summary 718 Reading List 719 References 719 Review Questions 719 Problems 720
to BJTs 664 Summary 669 References 670 Review Questions Problems 671
Appendices
Appendix A Constants
670
724 725
PREFACE
his is a textbook on the operating principles of semiconductor devices. It is appropriate for undergraduate (junior or senior) or beginning graduate students in electrical engineering as well as students of computer engineering, physics, and materials science. It is also useful as a reference for practicing engineers and scientists who are involved with modern semiconductor devices. Prerequisites are courses in chemistry and physics and in basic electric circuits, which are normally taken in the freshman and sophomore years. This text is appropriate for a two-semester course on semiconductor devices. However, it can be used for a one-semester course by eliminating some of the more advanced material and assigning some of the sections as read only. The authors have attempted to organize the material so that some of the detailed derivation sections can be skipped without affecting the comprehension of other sections. Some of these sections are marked with an asterisk. This book is divided into ve parts: 1. 2. 3. 4. 5. Materials Diodes Field-Effect Transistors Bipolar Transistors Optoelectronic Devices
The rst four parts are followed by Supplements that, while not required for an understanding of the basic principles of device operation, contain related material that may be assigned at the discretion of the instructor. For example, the use of SPICE for device and circuit analysis is briey discussed for diodes, eldeffect transistors, and bipolar transistors. While SPICE is normally taught in courses on electric circuits, it is useful to know the origin of the various parameters used to characterize devices. This material on SPICE is relegated to supplements, since not all schools cover SPICE in courses on electron circuit analysis and such courses may be taught before, concurrently with, or after the course on semiconductor devices. Part 1, Materials, contains four chapters and two supplements. The rst two chapters contain considerable review material from the prerequisite courses. This material is included since it is used extensively in later chapters to explain the principles of device operations. Depending on the detailed content of the prerequisite courses, much of these chapters can be relegated to reading assignments.
xiii
xiv
Preface
The level of quantum mechanics to be covered in a course like this varies widely. In this book some basic concepts are included in the main chapters of Part 1; those wishing to cover quantum mechanics in more detail will nd more extensive material in Supplement A to Part 1. The basic operating principles of large and small devices of a particular type (e.g., diodes, eld-effect transistors, bipolar junction transistors, photodetectors) are the same. However, the relative importance of many of the parameters involved in device operation depends on the device dimensions. In this book the general behavior of devices of large dimensions is treated rst. We treat, in each case, prototype devices (such as step junctions and long-channel eld-effect transistors), from which the fundamental physics can be learned, and then develop more realistic models considering second-order effects. These secondorder effects can have signicant inuence on the electrical characteristics of modern, small-geometry devices. The instructor can go into as much depth as desired or as time permits. Topics treated that are typically omitted in undergraduate texts are:
The differences between the electron and hole effective masses as used in density-of-state calculations and conductivity calculations. The differences in electron and hole mobilities (and thus diffusion coefcients) if they are majority carriers or minority carriers. The effects of doping gradients in the base of bipolar junction transistors (and/or the composition in heterojunction BJTs) on the current gain and switching speed. Band-gap reduction in degenerate semiconductors. While this has little effect on the electrical characteristics of diodes or eld-effect transistors, its effect in the emitter of bipolar junction transistors reduces the current gain by an order of magnitude. The velocity saturation effects due to the longitudinal eld in the channel of modern eld-effect transistors with submicrometer channel lengths reduces the current by an order of magnitude compared with that calculated if this effect is neglected.
While the major emphasis is on silicon and silicon devices, the operation of compound semiconductor devices, alloyed devices (e.g., SiGe, AlGaAs) and heterojunction devices (junctions between semiconductors of different composition) are also considered because of the increased performance that is possible with such band-gap engineering. Many of the seminal publications on semiconductor devices cited in the references at the end of each chapter through 1990 are reprinted in Semiconductor Devices: Pioneering Papers, edited by S. M. Sze, World Scientic Publishing Co., Singapore, 1991. Fabrication, while an important part of semiconductor engineering, is often skipped in the interest of time. This material is introduced in Appendix C, and can be assigned as read-only material if desired.
Preface
xv
ACKNOWLEDGEMENTS
We would like to thank, rst and foremost, our spouses Bill and Claire for their love, support, patience, and help. We are also grateful to the anonymous manuscript reviewers for their comments and suggestions, as well as the staff at McGraw-Hill for all their help. We thank our students for valuable feedback on the manuscript. Finally, we would like to thank all the companies and individuals that provided photographs and data for this book. Anderson & Anderson
P A R T
Materials
emiconductors form the basis of most modern electronic systems (e.g., computers, communication networks, control systems). While there are applications for other materials in electronics (e.g., magnetic materials in hard drives), this book concentrates on electronic devices that are based on semiconductors. Understanding the operation and design of semiconductor devices begins with an understanding of the materials involved. In Part 1 of this book, we investigate the behavior of electrons in materials, starting with the atoms themselves. Then we progress to electrons in crystalline semiconductors. We will see that classical mechanics does not provide a complete picture of electron activity in solids. In principle one should instead use quantum mechanics to predict the electrons behavior, but the application of quantum mechanics is not as simple as the more familiar classical or Newtonian mechanics. We will therefore introduce pseudo-classical mechanics, which modies familiar classical equations to account for some quantum mechanical effects. Some basic quantum mechanical concepts important for the understanding of device operation are covered in Chapter 1. (A more thorough discussion is contained in Supplement A to Part 1, found after Chapter 4.) In Chapter 2, we cover pseudo-classical mechanics, which allows us to predict the reaction of electrons to complicated elds, while using simple and intuitive pseudo-classical equations. The use of pseudo-classical mechanics will also allow us to draw and use energy band diagrams. These diagrams are indispensable for understanding and predicting the motion of the electrons and holes, and thus the current in semiconductors. In Chapter 3, we will see that conductivity of semiconductors is controlled by the number of charge carriers available to carry current. The charge carriers in semiconductors are electrons and holes. Their numbers are controlled by the concentrations of impurity elements that are intentionally added to the material. The carrier concentrations also depend on temperature, and if light is shining on the sample, the concentrations can also vary.
1
PA R T 1
Materials
It will emerge that there are two major forms of current in semiconductors, drift current and diffusion current. Drift current is caused by the presence of an electric eld, whereas diffusion current arises when the carrier concentrations vary with position. Chapter 4 covers nonhomogeneous semiconductor materials, in which the doping or the material composition itself may vary. These variations can lead to internal electric elds that can enhance device performance. Most modern semiconductor devices have regions of such nonhomogeneous materials. The Supplements to Part 1 contain additional topics relevant to semiconductor materials, including a more detailed discussion of quantum mechanics, the statistics of electrons in bound states, and phonons. We will start with electrons in atoms.
C H A P T E R
PA R T 1
Materials
quantum mechanics is presented. A somewhat greater discussion of quantum mechanics appropriate to some of the electronic processes in semiconductor devices is presented in Supplement A to Part 1, after Chapter 4. The key to understanding semiconductors is to appreciate the physical interpretation of the mathematical results. Physical understanding is emphasized in this book.
(a)
(b)
(c)
Figure 1.1 (a) The Thompson model of an atom, in which the positive charge is
uniformly distributed in a sphere and the electrons are considered to be negative point charges embedded in it; (b) the Bohr model, in which the positive charge is concentrated in a small nucleus and the electrons orbit in circles; (c) the Wilson-Sommerfeld model, which allows for elliptical orbits.
CHAPTER 1
In 1916, Wilson and Sommerfeld generalized the Bohr model to apply it to any physical system in which a particles motion is periodic with time. This modication allows for the possibility of elliptical orbits, as shown in Figure 1.1c.
where r is the distance between the two charges and 0 = 8.85 1012 farads/ meter is the permittivity of free space (because there is only free space between the nucleus and the electron). The expression at the far right-hand side of Equation (1.1) is obtained by recognizing that the hydrogen nucleus has only one proton, so Q 1 is equal to +q = 1.602 1019 coulombs, the elemental charge, and the charge of the electron Q 2 is equal to q. The resulting negative sign in Equation (1.1) indicates that the force is attractive. We now have an expression for the attractive (centripetal) force between the two particles, and we recall from classical mechanics that the force F on a particle is equal to minus the gradient of the potential energy, or F = E P = d EP dr (1.2)
In the last expression, the gradient is taken in the r direction, and E P is the potential energy of the electron at position r. Equation (1.2) with the aid of (1.1) can be rewritten as d E P = d E P (r) = Fdr = q 2 dr 40r 2 (1.3)
One can integrate both sides to obtain E P , but there will be a constant of integration. The actual value of the potential energy is arbitrary (as is the choice of
PA R T 1
Materials
the constant), since the value of the potential energy depends entirely on ones choice of reference. We can choose a convenient reference by noting that the Coulomb force at innite distance is zero. It makes sense for this case, then, to choose r = as a reference point, so we dene the potential energy at r = as the vacuum level, E vac : E P (r = ) = E vac (1.4) This is the energy required to free the electron from the inuence of the nucleus, essentially by moving the electron innitely far away from it. If the electron is innitely far from the nucleus, it cannot really be considered part of the atomit is now a free electron in vacuum. Now we can solve Equation (1.3) for a given value of r:
E vac
d EP =
EP r
q 2 dr 40r 2
(1.5)
where E P is the electron potential energy at some distance r from the nucleus. Integrating both sides and rearranging, we obtain E P = E vac q2 40r (1.6)
Figure 1.2 shows a plot of the r dependence of EP. From Equation (1.1), and since the force is equal to minus the gradient (slope) of the potential energy, we see that the force on the electron is directed toward the nucleus, or the Coulomb force is centripetal. Since the electron is revolving in a circle of radius r around the nucleus, we know from Newtonian mechanics that its centrifugal force is equal to F= mv 2 r
F Evac EP
(1.7)
F Electron energy
r r 0
for an electron in the vicinity of a single positive point charge. The electron is considered to be a point charge.
CHAPTER 1
For the atom to be stable, the net force on the electron must be zero. Equating our previous expression for the centripetal force due to the Coulomb attraction [Equation (1.1)] to the centrifugal force [Equation (1.7)], we can write q2 mv 2 =0 r 40r 2 (1.8)
Bohr also postulated that the integral of the angular momentum around one complete orbit is an integer multiple of Plancks constant h:
2
P d =
0
mvr d = nh
(1.9)
where n is an integer. Since the orbit is assumed circular in the Bohr model, r is a constant, and so are the potential energy E P and the speed v. Therefore, the integral becomes 2mrv = nh There is a solution for each integer value of n, so we write mvn rn = n h = nh 2 (1.11) (1.10)
Here we have introduced a new symbol; it turns out that engineers and physicists (and now you) use the quantity h/2 so much that there is a special character for it, h , pronounced h-bar. The subscripts n in Equation (1.11) indicate the partic ular orbital radius or speed associated with a specic quantum number n. If we simultaneously solve Equations (1.8) and (1.11), we can derive an expression for the Bohr radius of the nth state, where by state we mean the properties associated with a particular value of n: rn = 40 n 2 h 2 2 mq (1.12)
Our primary goal, however, is to nd the energies associated with these states. We know that the total energy of a system is equal to the kinetic energy plus the potential energy. The kinetic energy of the nth energy level is 1 2 mq 4 E K n = mvn = 2 2(40 )2 n 2 h 2 (1.14)
PA R T 1
Materials
For the nth energy level, we can nd rn from Equation (1.12) and use that in Equation (1.6) to write for the potential energy E Pn = E vac mq 4 mq 4 E Pn = E vac 2(40 )2 n 2 h 2 (40 )2 n 2 h 2 (1.15)
We say that the energy is quantized. It can have only discrete values associated with the quantum number n.
EXAMPLE 1.1
Find the energies and radii for the rst four orbits in the hydrogen atom.
Solution
mq 4 2(40 )2 n 2 h 2 (9.11 1031 kg)(1.60 1019 C)4 (2)(4)2 (3.1416)2 (9.85 1012 F/m)2 (1.05 1034 J s)2 1 n 1 n
2
1 n
(1.17) (2.18 10
17
J)
(13.6 eV)
Here a new unit of energy is introduced, the electron volt (eV). The electron volt is dened as the amount of energy acquired by an electron when it is accelerated through 1 volt of electric potential. To convert between S.I. (International System) units (joules) and electron volts, use 1 eV = 1.60 1019 joules Electron volts are not S.I. units, and therefore must be used with care in calculations. The Bohr radii can be calculated from Equation (1.12): rn = 40 n 2 h 2 (4)(3.1416)(8.85 1012 F/m)(1.05 1034 J s)2 = n2 mq 2 (9.11 1031 kg)(1.60 1019 C)2
2
(1.18)
rn = 0.053n nm The energies and Bohr radii of the rst four energy levels are given in Table 1.1. These energies and radii are plotted in Figures 1.3 and 1.4, respectively.
CHAPTER 1
Table 1.1 The rst four Bohr energies and orbital radii for the hydrogen atom
E E4 E3 E2 Electron energy
Ground state
E1
Evac
13.6 eV
atom. Higher energies occur increasingly close to each other, approaching the vacuum level.
r1
Radius
Figure 1.4 Radii of the rst four atomic orbits of the hydrogen atom,
10
PA R T 1
Materials
There are several things to notice in this example. First, the differences between the vacuum energy level and the allowed energies vary as 1/n 2 . Thus the higher the quantum number, and therefore the energy, the closer together (in energy) the energy levels are. Second, the Bohr radius varies as n 2 . This means that the higher the energy level, the farther the electron is from the atomic nucleus. If the electron has energy greater than E vac , the Coulombic force is not enough to keep the electron bound to the atom. In this case, a hydrogen ion (H+ ) is created as the electron leaves the atom. Also, notice that we do not give a number for the energy of a particular state, but rather we express the energies as so many electron volts from some reference level (in this case E vac ). It is pointless to say, This level is at 10 eV, since 10 eV could be anywhere, depending on your choice of reference. This point cannot be emphasized enough. Potential energy and thus total energy are arbitrary.
The energy of a state must always be expressed as an energy differencethe difference between the energy of the state and some known reference, for example, E vac E .
Finally, it should be pointed out that although the number of possible states is innite, once an electron occupies one of these states the hydrogen atom becomes neutral and other electrons are not attracted.
EXAMPLE 1.2
Consider a particle in a one-dimensional universe, oscillating in the parabolic potential energy shown in Figure 1.5. This represents an approximation to an electron in a modern quantum well laser.1 The potential energy function is a parabola, and the particle is a simple harmonic oscillator. Explain the motion of the particle using the energy diagram, paying attention to where the kinetic energy is largest, smallest, and the directions of the forces.
Solution
Conservation of energy dictates that the particle must remain at a constant energy. Thus, it oscillates back and forth at this particular energy. When the particle is in the center, it has the smallest potential energy and the largest kinetic energy, and thus the largest velocity. Recall that the force is equal to the negative of the slope of the potential energy [Equation (1.2)]. Therefore, as the particle travels through the center, e.g., to the right, the force is to the left (the slope is positive to right of center). Thus, the particle decelerates. It continues to slow as it moves to the right. The total energy is constant, but since the potential energy is increasing, the kinetic energy decreases. When the particle gets to the edge, it will have zero kinetic energy. The particle stops for an instant, but the force is still accelerating it to the left. It picks up speed until it passes through the center, at which point the force is to the right, and the particle begins to slow once again.
1
It is called a quantum well because the potential energy forms a well with quantized energy states.