Table of Contents
Title Page
Copyright
Foreword
Preface
Contributors
Acknowledgments
Introduction: Historic Timeline
Part I: Basic Electron Devices
Chapter 1: Bipolar Transistors
1.1 Motivation
1.2 The pn Junction and its Electronic Applications
1.3 The Bipolar Junction Transistor and its Electronic Applications
1.4 Optimization of Bipolar Transistors
1.5 Silicon-Germanium Heterojunction Bipolar Transistors
References
Chapter 2: MOSFETs
2.1 Introduction
2.2 MOSFET Basics
2.3 The Evolution of MOSFETs
2.4 Closing Remarks
References
Chapter 3: Memory Devices
3.1 Introduction
3.2 Volatile Memories
3.3 Non-Volatile Memories
3.4 Future Perspectives of MOS Memories
3.5 Closing Remarks
References
Chapter 4: Passive Components
4.1 Discrete and Integrated Passive Components
4.2 Application in Analog ICs and DRAM
4.3 The Planar Spiral Inductor – A Case Study
4.4 Parasitics in Integrated Circuits
References
Chapter 5: Emerging Devices
5.1 Non-Charge-Based Switching
5.2 Carbon as a Replacement for Silicon and the Rise of Graphene Electronics and Moletronics
5.3 Closing Remarks
References
Part II: Aspects of Device and IC Manufacturing
Chapter 6: Electronic Materials
6.1 Introduction
6.2 Silicon Device Technology
6.3 Compound Semiconductor Devices
6.4 Electronic Displays
6.5 Closing Remarks
References
Chapter 7: Compact Modeling
7.1 The Role of Compact Models
7.2 Bipolar Transistor Compact Modeling
7.3 MOS Transistor Compact Modeling
7.4 Compact Modeling of Passive Components
7.5 Benchmarking and Implementation
References
Chapter 8: Technology Computer Aided Design
8.1 Introduction
8.2 Drift-Diffusion Model
8.3 Microscopic Transport Models
8.4 Quantum Transport Models
8.5 Process and Equipment Simulation
References
Chapter 9: Reliability of Electron Devices, Interconnects and Circuits
9.1 Introduction and Background
9.2 Device Reliability Issues
9.3 Circuit-Level Reliability Issues
9.4 Microscopic Approaches to Assuring Reliability of ICs
References
Chapter 10: Semiconductor Manufacturing
10.1 Introduction
10.2 Substrates
10.3 Lithography and Etching
10.4 Front-End Processing
10.5 Back-End Processing
10.6 Process Control
10.7 Assembly and Test
10.8 Future Directions
References
Part III: Applications Based on Electron Devices
Chapter 11: VLSI Technology and Circuits
11.1 Introduction
11.2 MOSFET Scaling Trends
11.3 Low-Power and High-Speed Logic Design
11.4 Scaling Driven Technology Enhancements
11.5 Ultra-Low Voltage Transistors
11.6 Interconnects
11.7 Memory Design
11.8 System Integration
References
Chapter 12: Mixed-Signal Technologies and Integrated Circuits
12.1 Introduction
12.2 Analog/Mixed-Signal Technologies in Scaled CMOS
12.3 Data Converter ICs
12.4 Mixed-Signal Circuits for Low Power Displays
12.5 Image Sensor Technologies and Circuits
References
Chapter 13: Memory Technologies
13.1 Semiconductor Memory History
13.2 State of Mainstream Semiconductor Memory Today
13.3 Emerging Memory Technologies
13.4 Closing Remarks
References
Chapter 14: RF and Microwave Semiconductor Technologies
14.1 III-V-Based: GaAs and InP
14.2 Si and SiGe
14.3 Wide Bandgap Devices (Group-III Nitrides, SiC and Diamond)
References
Chapter 15: Power Devices and ICs
15.1 Overview of Power Devices and ICs
15.2 Two-Carrier and High-Power Devices
15.3 Power MOSFET Devices
15.4 High-Voltage and Power ICs
15.5 Wide Bandgap Power Devices
References
Chapter 16: Photovoltaic Devices
16.1 Introduction
16.2 Silicon Photovoltaics
16.3 Polycrystalline Thin-Film Photovoltaics
16.4 III-V Compound Photovoltaics
16.5 Future Concepts in Photovoltaics
References
Chapter 17: Large Area Electronics
17.1 Thin-Film Solar Cells
17.2 Large Area Imaging
17.3 Flat Panel Displays
References
Chapter 18: Microelectromechanical Systems (MEMS)
18.1 Introduction
18.2 The 1960s – First Micromachined Structures Envisioned
18.3 The 1970s – Integrated Sensors Started
18.4 The 1980s – Surface Micromachining Emerged
18.5 The 1990s – MEMS Impacted Various Fields
18.6 The 2000s – Diversified Sophisticated Systems Enabled by MEMS
18.7 Future Outlook
18.8 Acknowledgment
References
Chapter 19: Vacuum Device Applications
19.1 Introduction
19.2 Traveling-Wave Devices
19.3 Klystrons
19.4 Inductive Output Tubes
19.5 Crossed-Field Devices
19.6 Gyro-Devices
References
Chapter 20: Optoelectronic Devices
20.1 Introduction
20.2 Light Emission in Semiconductors
20.3 Photodetectors
20.4 Integrated Optoelectronics
20.5 Optical Interconnects
20.6 Closing Remarks
References
Chapter 21: Devices for the Post CMOS Era
21.1 Introduction
21.2 Devices for the 8-nm Node with Conventional Materials
21.3 New Channel Materials and Devices
21.4 Closing Remarks
References
Index
This edition first published 2013
© 2013, John Wiley & Sons Ltd
Registered office
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom
For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com.
The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Library of Congress Cataloging-in-Publication Data
Guide to state-of-the-art electron devices / edited by Joachim N. Burghartz.
pages cm
Papers by members of the IEEE Electron Devices Society.
Includes bibliographical references and index.
ISBN 978-1-118-34726-3 (hardback)
1. Electronic apparatus and appliances. I. Burghartz, Joachim N. II. IEEE Electron Devices Society.
TK7870.G83 2013
621.3815′28--dc23
2012040303
A catalogue record for this book is available from the British Library.
Print ISBN: 9781118347263
Note: Dr. George E. Smith received the 2009 Nobel Prize in Physics for "the invention of an imaging semiconductor circuit—the CCD sensor". He is also a Celebrated Member of the IEEE Electron Devices Society.
No one can ever doubt that electronic devices provide the basis of the way large components of our civilization operate today. This book refreshingly provides both a well-organized history and state of the art of the many technologies involved, how they are connected and most importantly, how this evolution (and frequently revolution) took place. I am overjoyed to see that the IEEE Electron Device Society has taken on this challenging undertaking and succeeded so well. The editor has made excellent choices of authors to cover such a daunting challenge. I have known many of them personally and can attest to that claim.
The initiation of the EDS and its growth has coincided with this revolution and it has been the principal organization in creating conferences and providing publications in which people can present results and interact with others working in their field. The importance of this cannot be overemphasized since the trading of ideas, not to mention competition, between people in a field provides fertilizer for new ideas. Incidentally, it is not too much of an exaggeration to say that, at conferences, as much information has been traded in the hallways as in the technical sessions. The inception of this book is a continuation of that fine tradition of publicizing information.
Although each chapter in this book covers a separate subject, they all start with a historical and tutorial mix before attacking the current state of the art. This is very beneficial to both students and experts in a given field who wish to broaden their horizons. I especially applaud the use of the blue sidebars which explain terms and concepts which require no explanation for those in the field but are enigmas to those less knowledgeable.
George E. Smith
This book marks the 35th anniversary of the IEEE Electron Devices Society (EDS), a journey that began with the formation of the IRE Professional Group on Electron Devices 60 years ago. The major technical advancements in the field of Electron Devices are commemorated chronologically through the “video clips” at the bottom of the pages throughout the book. These clips represent snapshots of many pioneers whose aspirations and dedication led to the discovery of numerous device concepts and their implementation into practical use. This historical time line links well to the electronic booklet “50 Years of Electron Devices,” which is freely available on the EDS web site.
Although the invention and development of vacuum tube devices for communications and sensing predated the age of transistor, it was the advent of the solid-state “triode” in 1947, followed by the integrated circuits in the late 1950s, that has ceaselessly pushed new frontiers in computers, communications, and many other emerging areas in the several recent decades. The work described in this book has revolutionized the way we live and the way we think. As the continuous, insatiable demand for higher performing electronics drives the search for new materials and devices, the global electron devices community will again and again respond to new challenges with novel solutions.
This book was compiled from contributions from many volunteer leaders of the Society. It is our sincere wish that its technical content will serve as a bridge between the last sexagenary cycle and the next.
Paul Yu
IEEE EDS President 2012–2013
Electronics and power electronics have grown to be indispensable technologies supporting our lifestyle, our health and our safety, social interaction and security. None of today's industries would be viable without electronic communication, automation and control. Through electronics people stay connected, get supported in their professional life, and can enjoy leisure entertainment. Transportation and energy supply depend on electronics as well. Electron devices are the foundation of electronics and power electronics. They enable all kinds of electronic signal processing and allow for switching and steering electrical energy.
This book features a concise guide to state-of-the-art electron devices. It is written by 67 specialists who are members of the IEEE Electron Devices Society (EDS). In 21 chapters they share their expert view on a particular group of electron devices or device aspects. The chapters not only illustrate the broad variety of electron device and device aspects but they are also a mirror of the diversity within the EDS. There are contributions from industry, academic and government institutions. Authors come from all five continents – a true world class team which includes the top-level industry manager as well as young engineer, the renowned university professor, and the young academic. The editor therefore tried to keep the apparent differences in style and language intact to reflect the rich diversity in regional background and affiliation.
Most of the authors are members of one of the 14 Technical Area Committees (TACs) in EDS, which assist the Executive Committee (ExCom) and Board of Governors (BoG) of EDS with their expertise in decision making and strategy processes. The current TACs in EDS are listed in Table 1.
Table 1 Technical Area Committees of the IEEE Electron Devices Society
Compact Modeling | CM |
Compound Semiconductor Devices and Circuits | CSDC |
Device Reliability Physics | DRP |
Electronics Materials | EM |
Microelectromechanical Systems | MEMS |
Nanotechnology | NT |
Optoelectronic Devices | OD |
Organic Electronics | OE |
Photovoltaic Devices | PD |
Power Devices and ICs | PDIC |
Semiconductor Manufacturing | SM |
Technology Computer Aided Design | TCAD |
Vacuum Devices VLSI Technology and Circuits | VLSI |
The Institute of Electrical and Electronics Engineers (IEEE), the world's largest professional association, has its roots in the American Institute of Electrical Engineers (AIEE; founded 1884) and the Institute of Radio Engineers (IRE; founded 1912), which merged in 1963 to form the IEEE. The IRE already paid attention to the significance of electron devices by establishing an ‘Electron Tube and Solid-State Devices Committee’ in 1951 shortly after the invention of the transistor. In 1952 the committee's name was changed to ‘IRE Professional Group on Electron Devices’. With the merger of AIEE and IRE an ‘IEEE Electron Devices Group’ was established in 1964, which in 1976 became the ‘IEEE Electron Devices Society (EDS)’.
Today, we look back at the 35-year history of the IEEE Electron Devices Society and at its foundation in the IRE 60 years ago: two great reasons to celebrate and provide the members and potential new members of EDS with this concise guide to state-of-the-art electron devices at a very affordable price. This was made possible by substantial sponsorship through EDS and by dedicated volunteer contributions.
The book is organized in three parts, of which Part II (5 chapters) and Part III (11 chapters) are closely aligned with the TACs of EDS (see Figure 1). Part I of the book introduces in five chapters the fundamentals of electron devices. Sidebars are used in all chapters to define important figures-of-merit or definitions for a particular electron device and to offer an easy entry into the topic to the novice.
Figure 1 Relationship of the Technical Area Committees in the IEEE Electron Devices Society and their organization in Parts II and III of the book
A highlight of the book is the comprehensive timeline, featuring the historical milestones of electron devices in chronological order in three eras, after 1976, between 1952 and 1976, and before 1952. These eras mark the time periods of EDS, of electron devices in the IRE prior to EDS, and of the early electron devices, respectively.
Besides the actual contributors many people have helped ‘behind the scenes’ to turn the idea of this anniversary book into reality. They are duly acknowledged in the Acknowledgments section.
Joachim N. Burghartz
The editor wishes to acknowledge the EDS Executive Committee (ExCom) and the EDS Board of Governors (formerly the Advisory Committee (AdCom)) for supporting this anniversary book project and their vote for financial sponsorship, making it possible to offer this book to EDS members at a very low price. Christopher Jannuzzi (EDS Executive Director) stepped in with hands-on support and helped at various ends, particularly in working with the IEEE History Center, New Brunswick, New Jersey from where we received numerous images used in the historical timeline.
The editor, on behalf of all contributors, considers it a great honor that this book is introduced by a foreword of Nobel Laureate George E. Smith. His kind words helped us to complete this ambitious project.
Joachim Deh from IMS CHIPS in Stuttgart helped with editing some of those and other images used in the timeline and in Chapter 4 and with arranging copyright permissions for such images. Fernando Guarin (EDS Secretary) carried out a critical review of parts of the book prior to the copyediting by the publisher. Peter Mitchell, Laura Bell, Liz Wingett, Richard Davies, and Genna Manaog from John Wiley & Sons as well as Sangeetha Parthasarathy from Laserwords Pvt Ltd. are acknowledged for arranging a smooth copyediting and publication process; particular thanks goes to Peter Mitchell for his patience in negotiating the terms of the publication contracts. Advance acknowledgments go to Bin Zhao (EDS Vice President Meetings), Xing Zhou (EDS Vice President Regions & Chapters), Jean Bae (EDS Executive Office), again Chris Jannuzzi, and all organizers of EDS-sponsored meetings and EDS chapter chairs involved in distributing the book within the EDS community.
The authors of Chapter 6 gratefully acknowledge Yi-Hsiang Huang of National Taiwan University and Noah Jafferis of Princeton for expert assistance with graphics.
The authors of Chapter 7 give credit to Gennady Gildenblat, professor at Arizona State University, who provided a great amount of support to that chapter without being a co-author. His credentials in the field of compact modelling are reflected in the historical timeline. The authors of Chapter 8 gratefully acknowledge Claudio Fiegna, University of Bologna, for many helpful discussions.
The authors of Chapter 15 would like to thank Dr. Phil Krein and Professor Emeritus Nick Holonyak of the ECE Department of the University of Illinois at Urbana-Champaign for their assistance in identifying records and references chronicling early developments in power semiconductors.
The authors of Chapter 16 would like to extend a special acknowledgement to Dr. Tyler Grassman of The Ohio State University for his key contributions to both the technical content and integration of information throughout the chapter.
The authors of Chapter 19 gratefully acknowledge the valuable contributions from Heinz Bohlen (Communications and Power Industries, USA, retired), Ernst Bosch (Thales Electron Devices GmbH, Germany), Gregory Nusinovich (Institute for Research in Electronics and Applied Physics, University of Maryland, College Park, USA), and Edward Wright (Beam-Wave Research Inc., USA).
Electron devices go back a long time: from the vacuum tube to the inventions of the transistor and the integrated circuit, through 40 years of scaling microelectronics to the exciting possibilities brought about by the current investigations on emerging research devices. The history of electron device applications is also captured here. Looking back in time means learning from the past so that future progress can be made more efficiently. It also means taking pride in the pioneers' achievement and viewing them as role models empowering young electron device engineers.
A historic timeline runs through this entire anniversary book, and marks key milestones of electron device development and applications in more than 1000 slides. Landmarks of world history and other technical breakthroughs place those milestones into historical perspective. The book can, thus, be read in two ways; chapter-by-chapter, or along the timeline of device history.
Click here to view the historic timeline
Part I
Click here to view the historic timeline
Chapter 1
In terms of its influence on the development of modern technology and hence, global civilization, the invention of the point contact transistor on December 23, 1947 at Bell Labs in New Jersey by Bardeen and Brattain was by any reckoning a watershed moment in human history [1]. The device we know today as a bipolar junction transistor was demonstrated four years later in 1951 by Shockley and co-workers [2] setting the stage for the transistor revolution. Our world has changed profoundly as a result [3].
Interestingly, there are actually seven major families of semiconductor devices (only one of which includes transistors!), 74 basic classes of devices within those seven families, and another 130 derivative types of devices from those 74 basic classes (Figure 1.1) [4]. Here we focus only on three basic devices: (1) the pn homojunction junction diode (or pn junction or diode), (2) the homojunction bipolar junction transistor (or BJT), and (3) the special variant of the BJT called the silicon-germanium heterojunction bipolar transistor (or SiGe HBT). As we will see, diodes are useful in their own right, but also are the functional building block of all transistors.
Figure 1.1 The transistor “food chain” showing all major families of semiconductor devices.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
Surprisingly, all semiconductor devices can be built from a remarkably small set of materials building blocks (Figure 1.2), including [4]:
Click here to view the historic timeline
Figure 1.2 The essential building blocks of all semiconductor devices.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
Why do we actually need transistors in the first place? Basically, because nature attenuates all electrical signals. By this we mean that the magnitude of all electrical signals (think “1s” and “0s” inside a computer, or an EM radio signal from a cell phone) necessarily decreases as it moves from point A to point B, something we call “loss”. When we present an (attenuated) input signal to the transistor, the transistor is capable of creating an output signal of larger magnitude (i.e., “gain”), and hence the transistor serves as a “gain block” to “regenerate” (recover) the attenuated signal in question, an essential concept for electronics. In the electronics world, when the transistor is used as a source of signal gain, we refer to it as an “amplifier.” Amplifiers are ubiquitous to all electronic systems.
Click here to view the historic timeline
Not only can the transistor serve as a wonderful nanoscale sized amplifier, but importantly it can also be used as a tiny “regenerative switch”; meaning, an on/off switch that does NOT have loss associated with it. Why is this so important? Well, imagine that the computational path through a microprocessor requires 1 000 000 binary switches (think light switch on the wall – on/off, on/off) to implement the complex digital binary logic of a given computation. If each of those switches even contributes a tiny amount of loss (which it inevitably will), multiplying that tiny loss by 1 000 000 adds up to unacceptably large system loss. That is, if we push a logical “1” or “0” in, it rapidly will get so small during the computation that it gets lost in the background noise. If, however, we implement our binary switches with gain-enabled transistors, then each switch is effectively regenerative, and we can now propagate the signals through the millions of requisite logic gates without excessive loss, maintaining their magnitude above the background noise level.
In short, the transistor can serve in one of two fundamental capacities: (1) an amplifier or (2) a regenerative switch. Amplifiers and regenerative switches work well only because the transistor has the ability to produce gain. So a logical question becomes, where does transistor gain come from? To answer this, first we need to understand pn junctions.
Virtually all semiconductor devices (both electronic and photonic) rely on pn junctions (a.k.a., “diodes”, a name which harkens back to a vacuum tube legacy) for their functionality. The simplest embodiment of a pn junction is the pn “homojunction”, meaning that within a single piece of semiconductor (e.g., silicon – Si) we have a transition between p-type doping and n-type doping (e.g., p-Si/n-Si). The opposite would be a pn heterojunction, in which the p-type doping is within one type of semiconductor (e.g., p-GaAs), and the n-type doping is within another type of semiconductor (e.g., n-AlGaAs).
Click here to view the historic timeline
As shown in Figure 1.3, to build a pn junction we might, for instance, ion implant and then diffuse n-type doping into a p-type wafer. The important thing is the resultant “doping profile” as one moves through the junction (ND(x) – NA(x), which is just the net doping concentration). At some point in the doping transition, ND = NA, and we thus have a transition between net n-type and net p-type doping. This point is called the “metallurgical junction” (x0 in Figure 1.3) and all of the important electrical action of the junction is centered here. To make the physics easier, two simplifications are typically made: (1) Let us assume a “step junction” approximation to the real pn junction doping profile, which is just what it says, an abrupt change (a step) in doping occurring at the metallurgical junction (Figure 1.3). (2) Let us assume that all of the dopant impurities are ionized (one donor atom equals one electron, etc., an excellent approximation for common dopants in silicon at 300 K).
Figure 1.3 Cartoons of a pn junction, showing doping transition from n-type to p-type.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
So, how does a pn junction actually work? The operation of ALL semiconductor devices is best understood at an intuitive level by considering the energy band diagram, which plots electron and hole energy as a function of position as we move physically through a device. An n-type semiconductor is electron rich (i.e., majority carriers), and hole poor (i.e., minority carriers). Conversely, a p-type semiconductor is hole-rich and electron-poor. If we imagine bringing an n-type and p-type semiconductor into “intimate electrical contact” where they can freely exchange electrons and/or holes from n to p and p to n, the final equilibrium band diagram shown in Figure 1.4 will result. Note, that under equilibrium conditions, there is no NET current flow across the junction.
Figure 1.4 Energy band diagram of a pn junction at equilibrium.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
Click here to view the historic timeline
We might logically wonder what actually happened inside the junction to establish this equilibrium condition. When brought into contact, the n-type side of the junction is electron rich, while the p-type side is electron poor. That is, there is a large driving force for electrons to diffuse from the n region to the p region. Recall, that there are in fact two ways to move charge in a semiconductor: (1) drift, whose driving force is the electric field (voltage/length), and (2) diffusion, whose driving force is the carrier density gradient (change in carrier density per unit distance). The latter process is what is operative here. Once in electrical contact an electron moves from the n-side to the p-side, leaving behind a positively charged donor impurity (ND+). Note, that far away from the junction, for each charged donor impurity there is a matching donated electron, hence the semiconductor is charge neutral. Once the electron leaves the n-side, however, there is no balancing charge, and a region of “space charge” results. The same thing happens on the p-side. Hole moves from p to n, leaving behind an uncompensated acceptor impurity (NA−) behind. This resultant charge “dipole” produces an electric field, pointing from + to − (to the right in this case). How does that induced field affect the diffusion-initiated side-to-side transfer of charge just described? It opposes the diffusive motion of both electron and holes via Coulomb's law. Therefore, in a pn junction the diffusion gradient moves electrons from n to p and holes from p to n, but as this happens a dipole of space charge is created between the uncompensated ionized dopants, and an induced electric field opposes the further diffusion of charge. When does equilibrium in the pn junction result? When the diffusion and the drift processes are perfectly balanced and the net current density is zero.
Click here to view the historic timeline
The pn junction in equilibrium consists of a neutral n region and a neutral p region, separated by a space charge region of width W. This structure forms a capacitor (conductor/insulator/conductor), and pn junctions have built-in capacitance which will partially dictate their switching speed. The electric field in the space charge region (for a step junction) is characteristically triangular shaped, with some peak value of electric field present. There is a built-in voltage drop across the junction, and, thus, from the energy band diagram we see that there is a potential barrier for any further movement of electrons and holes from side-to-side. This barrier to carrier transport maintains a net current density of zero, and the junction is by definition in equilibrium.
If one wanted to get current flowing again across the junction, how would this be done? Well, we must unbalance the drift and diffusion mechanisms by lowering the potential barrier to the electron and hole transport, and we can do this trivially by applying an external voltage to the n and p regions such that the p region (anode) is more positively biased than the n region (cathode). As shown in Figure 1.5, this effectively lowers the side-to-side barrier, drift no longer balances diffusion, and the carriers will once again start diffusing from side-to-side, generating useful current flow. This is called “forward bias”. What happens if we apply a voltage to the junction of opposite sign? (i.e., p region more negatively biased than the n region). Well, the barrier the carriers experience grows, effectively preventing any current flow, a condition called “reverse bias” (Figure 1.5).
Figure 1.5 The pn junction under both forward and reverse bias, showing the resultant current–voltage characteristics.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
The pn junction thus forms a solid-state switch (a.k.a. the “diode”). Consider: Apply a voltage of one polarity and current flows. Apply a voltage of the opposite polarity and no current flows; an on/off switch. Shockley shared the Nobel Prize with Bardeen and Brattain largely for explaining this phenomenon, and of course by wrapping predictive theory around it which led to the demonstration of the BJT. The result of that particularly elegant derivation is the celebrated “Shockley equation” which governs the current flow in a pn junction
1.1
where A is the junction area, V is the applied voltage, Dn,p is the electron/hole diffusivity (Dn,p = μn,p kT), Ln,p is the electron/hole diffusion length, and IS is the junction “saturation current” which collapses all of these factors into a single (measurable) parameter.
Observe, that all of the parameters in the Shockley equation refer to the minority carriers. If we build our junction with the n and p doping the same, then the relative contributions of the electron and hole minority carrier currents to the total current flowing will be comparable (to first order). Let us look closer at the operation of the junction. Under forward bias, electrons diffuse from the n-side to the p-side, where they become minority carriers. Those minority electrons are now free to recombine and will do so, on a length scale determined by Ln, and thus as we move from the center of the junction out into the neutral p-region, the minority electron population decreases due to recombination, inducing a concentration gradient as we move to the p-side, which drives a minority electron diffusion current. The same thing is happening with holes on the opposite side of the junction, and these two minority carrier diffusion currents add to produce the total forward bias current flow. What is the actual driving force behind the forward bias current in a pn junction? Recombination in the neutral regions, since recombination induces the minority diffusion currents. Alas, simple theory and reality are never coincident, and there a finite limits to the voltages that can be applied to the diode, and how much current can be passed through it and how much voltage can be applied across it [3].
Click here to view the historic timeline
So, what makes the junction so useful? Well, as stated, it makes a nice on/off switch with low loss when forward biased, and it can provide very good electrical isolation when reverse biased. In power electronics the diode would be said to provide a “blocking” voltage, not allowing current flow in reverse bias up to some finite, and often huge, applied reverse voltage (hundreds to even thousands of volts). This is very useful. The diode can also function as a wonderful solid-state “rectifier”. Rectifiers are ubiquitous in power generation, conversion, and transmission, (e.g., to turn AC voltage into DC voltage). Finally, the diode can also emit and detect light, which is also extremely useful as a transducer for converting optical to electrical energy, and vice versa (see Chapters 16 and 20).
All of this said, however, the diode does NOT possess gain, and, thus, is insufficient for realizing complex electronic systems. From a transistor perspective, however, the pn junction can be used to make a tunable minority carrier injector, which, if cleverly employed, can indeed produce gain when carefully implemented within a transistor. Importantly, one can trivially skew the relative magnitudes of the minority carrier injection from side-to-side in a pn junction by making the doping levels on one side of the junction much more heavily doped than on the other side. Let us imagine that the n-doping is far larger than the p-doping. Fittingly, this is referred to as a “one-sided” junction. In this scenario, it can be easily shown that electrons make up most of the total current flow in forward bias in such a junction. If we wanted to use a pn junction under forward bias to enhance the “forward-injection” of electrons into the p-region, and suppress the “back-injection” of holes into the n-region, we could simply use an n++ − p− junction as an “electron injector”! This will lead us directly to the BJT, a transistor with gain.
Click here to view the historic timeline
The pn junction, as a two-terminal object, can be made to serve as an efficient minority carrier injector, but it does NOT possess inherent gain. This is the fundamental reason why we do not build microprocessors from diode-resistor logic. Diodes make excellent binary switches, but without a gain mechanism to overcome Nature's preference for attenuation, complex functions are not going to be achievable in practice. Let us imagine, however, that we add an additional third terminal to the device which somehow controls the current flow between the original two terminals. Let terminal 1 = the input “control” terminal, and terminals 2 and 3 have high current flow between them when biased appropriately by the control terminal. Then, under the right bias conditions, with large current flow between 2 and 3, if we could somehow manage to suppress the current flow to/from 1, we'd be in business. That is, small input current (1) generates large output current (from 2 to 3), and hence we have gain!
How do we do this in practice? Let us use two pn junctions, placed back-to-back, such that the control terminal (our #1; which we will call the “Base” terminal – B) is in the central p region, and the two high current flow path output terminals (our #2 and #3, which we will call the “Emitter” and “Collector” terminals – E, and C), are the two outside n regions (see Figure 1.6). Since the two central p regions are shared by both diodes, those can be coincident. That is, an n region separated from another n region by an intermediate p region actually contains two pn junctions.
Figure 1.6 (a) Schematic of the two back-to-back pn junctions that form a bipolar junction transistor; (b) the circuit symbol of both doping polarity types are also shown.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
Click here to view the historic timeline
Let us imagine forward biasing the emitter–base junction, and reverse biasing the collector–base junction, and then adding two more puzzle pieces: (1) We must dope the emitter very heavily with respect to the base, such that when we forward bias the emitter–base junction we have large electron flow from E to B and simultaneously suppress the hole flow from B to E (this is our tunable minority carrier injector!). (2) We must make the central base region VERY thin. Why? Well, if we don't, then the electrons injected from E to B will simply recombine in the base before they can reach the collector (to be collected and to generate the required large output current flow from E to C). Recall that the rough distance a minority carrier can travel before it recombines is given by the diffusion length (Ln,p). Clearly, we need the width of the p-type base region to be much, much less than this number; in practice, a few hundred nm is required for a modern BJT. The final result? We have created the npn BJT! (One could of course swap the doping polarities n to p and p to n and achieve the same result – a pnp BJT. We thus have two flavors of BJT, and this is often VERY handy in electronic circuit design.
Click here to view the historic timeline
Consider now how the BJT actually works: (1) The reverse-biased CB junction has negligible current flow. (2) The forward-biased EB junction injects (emits) lots of electrons from E to B, that diffuse across the base without recombining (because it is thin) and are collected at C, generating large electron flow from E to C (current). BUT, due to the doping asymmetry in the EB junction, while a large number of electrons get injected from E to B, very few holes flow from B to E. Forward electron current is large, but reverse hole current is small. That is: small input base current; large output collector. Gain! This is otherwise known in electronics as “current gain” (or β).
How do we make the BJT? Well, as might be imagined it is more complex than a pn junction, but even so, the effort is worth it. Figure 1.7 shows the simplest possible variant. Figure 1.7 also superposes both the equilibrium and forward-active bias energy band diagrams, with the carrier minority and majority carrier distributions, to help connect the pn junction physics to the BJT operation. Within the band diagram context, here is intuitively how the BJT works. In equilibrium, there is a large barrier for injecting electrons from the emitter into the base. Forward bias the EB junction and reverse bias the CB junction, and now the EB barrier is lowered, and large numbers of electrons are injected from E to B. Since B is very thin, and the CB junction is reverse biased, these injected electrons will diffuse across the base, slide down the potential hill of the CB junction, and be collected at C, where they generate a large electron current flow from E to C. Meanwhile, due to the doping asymmetry of the EB junction, only a small density of holes is injected from B to E to support the forward bias EB junction current flow. Hence, IC is large, and IB is small. Gain! A different visualization of the magnitudes of the various current contributions in a well-made, high gain, BJT, are illustrated in Figure 1.8.
Figure 1.7 Basic structure and operational principles of the bipolar transistor.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
Figure 1.8 Sketch of (a) the relative current contributions of the bipolar transistor and (b) the resultant current–voltage characteristics.
Reproduced with permission from Cressler, J. D.; Silicon Earth: Introduction to the Microelectronics and Nanotechnology Revolution; 2009, Cambridge University Press
Shockley's theory to obtain an expression for β is fairly straightforward from basic pn junction physics (although you have two different ones to contend with obviously), provided you make some reasonable assumptions on the thickness of the base (base width Wb << Lnb). For the output and input currents under forward-active (amplifier) bias, we obtain:
1.2
1.3
Click here to view the historic timeline
where the “b” and “e”, or “B” and “E”, subscripts stand for base and emitter, respectively. Interestingly, the current gain does not to first-order depend on bias voltage, the size of the junction, or even the bandgap! We finally obtain,
1.4
Click here to view the historic timeline
Clearly, the current gain is a tunable parameter, giving us great flexibility in design. A common way to plot the BJT current–voltage characteristics is shown in Figure 1.8, where linear IC is plotted versus linear VCE, as a further function of IB. Since IC is larger than IB, the gain is implicit here. This plot is known as the output “family” or “output characteristics”. We use the output family to define the three regions of operation of the BJT: (1) “forward-active” (EB junction forward-biased; CB junction reverse-biased); (2) “saturation” (both EB and CB junctions forward-biased), and (3) “cut-off” (both EB and CB junctions reverse biased). As indicated, forward-active bias is typically for amplifiers, and as we will see, switching between cutoff and saturation will make an excellent regenerative digital switch!
Click here to view the historic timeline
How fast can transistors switch states (on to off)? The current speed record for a bipolar transistor digital switch is less than 10 picoseconds (0.000000000010 seconds – 10 trillionths of a second!). What limits that speed? Intuitively, the speed is limited by the time it takes the electrons to be injected from the emitter, transit (diffuse across) the base, and then be collected by the collector. In other words, a transistor can't be faster than it takes the charge to move through it. In most transistors, step two is the limiting one, and the so-called “base transit time” (τb) sets the fundamental speed limit on how fast the BJT can switch. A first-order base transit time expression can be easily derived,
1.5
Hence, the smaller τb is, the faster the BJT can switch. Clearly, making Wb as small as possible gives us a double benefit. It helps increase the current gain, yes, but even more importantly, it makes the transistor faster – quadratically!
gm