ebook download CMOS Past, Present and Future 1st Edition - eBook PDF all chapter

CMOS Past, Present and Future 1st
Edition - eBook PDF

Go to download the full and correct content document:
https://ebooksecure.com/download/cmos-past-present-and-future-ebook-pdf/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Electric Utility Resource Planning: Past, Present and

Future 1st edition - eBook PDF
https://ebooksecure.com/download/electric-utility-resource-
planning-past-present-and-future-ebook-pdf/
(eBook PDF) American Foreign Policy: Past, Present, and

Future Eleventh Edition
http://ebooksecure.com/product/ebook-pdf-american-foreign-policy-
past-present-and-future-eleventh-edition/
Understanding Mobility as a Service (MaaS): Past,

Present and Future 1st Edition David A. Hensher - eBook
PDF
https://ebooksecure.com/download/understanding-mobility-as-a-
service-maas-past-present-and-future-ebook-pdf/
Advanced Biofuel Technologies: Present Status,

Challenges and Future Prospects 1st Edition- eBook PDF
https://ebooksecure.com/download/advanced-biofuel-technologies-
present-status-challenges-and-future-prospects-ebook-pdf/
Production Processes of Renewable Aviation Fuel:
Present Technologies and Future Trends 1st edition -
eBook PDF
https://ebooksecure.com/download/production-processes-of-
renewable-aviation-fuel-present-technologies-and-future-trends-
ebook-pdf/
(eBook PDF) Criminological Theory: Past to Present:

Essential Readings 6th Edition
http://ebooksecure.com/product/ebook-pdf-criminological-theory-
past-to-present-essential-readings-6th-edition/
(eBook PDF) Traditions & Encounters: A Global

Perspective of the Past: From 1500 to the Present: 2
5th Edition
http://ebooksecure.com/product/ebook-pdf-traditions-encounters-a-
global-perspective-of-the-past-from-1500-to-the-present-2-5th-
edition/
CMOS VLSI Design A Circuits and Systems Perspective

(4th Edition)
http://ebooksecure.com/product/cmos-vlsi-design-a-circuits-and-
systems-perspective-4th-edition/
Pan-genomics: Applications, Challenges, and Future

Prospects 1st edition - eBook PDF
https://ebooksecure.com/download/pan-genomics-applications-
challenges-and-future-prospects-ebook-pdf/
CMOS Past, Present, and Future
Related titles
Germanium-Based Technologies: From Materials & Devices
(ISBN 978-0-08-044953-1)
Sputtering Materials for VLSI and Thin Film Devices
(ISBN 978-0-8155-1593-7)
High Performance Silicon Imaging: Fundamentals and Applications of CMOS and CCD
sensors
(ISBN 978-0-85709-598-5)
Millimeter-Wave Digitally Intensive Frequency Generation in CMOS
(ISBN 978-0-12-802207-8)
Woodhead Publishing Series in Electronic
and Optical Materials
CMOS Past, Present,

and Future
Henry H. Radamson
Jun Luo
Eddy Simoen
Chao Zhao
An imprint of Elsevier
Woodhead Publishing is an imprint of Elsevier
The Officers’ Mess Business Centre, Royston Road, Duxford, CB22 4QH, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
The Boulevard, Langford Lane, Kidlington, OX5 1GB, United Kingdom
Copyright © 2018 Elsevier Ltd. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic
or mechanical, including photocopying, recording, or any information storage and retrieval system, without
permission in writing from the publisher. Details on how to seek permission, further information about
the Publisher’s permissions policies and our arrangements with organizations such as the Copyright
Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/
permissions.
This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical treatment
may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating
and using any information, methods, compounds, or experiments described herein. In using such
information or methods they should be mindful of their own safety and the safety of others, including
parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume
any liability for any injury and/or damage to persons or property as a matter of products liability, negligence
or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained
in the material herein.
Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-08-102139-2 (print)
ISBN: 978-0-08-102140-8 (online)
For information on all Woodhead publications

visit our website at https://www.elsevier.com/books-and-journals
Publisher: Matthew Deans

Acquisition Editor: Kayla Dos Santos
Editorial Project Manager: Gabriela D. Capille
Production Project Manager: Joy Christel Neumarin Honest Thangiah
Cover Designer: Christian J. Bilbow
Typeset by SPi Global, India

Contents
Contributors ix
Preface xi
Acknowledgments xiii
1 Basics of metal–oxide–semiconductor field-effect transistor (MOSFET) 1

H.H. Radamson
1.1 Introduction 1
1.2 Basics of MOSFET’s operation 4
1.3 Figures of merit of MOSFETs 9
1.4 Evolution of the MOSFET structure 12
References 16
2 Scaling and evolution of device architecture 19

H.H. Radamson, E. Simeon, J. Luo, G. Wang
2.1 Introduction 19
2.2 Dimension and architectural scaling 20
2.3 Lithography for downscaling 26
2.4 Electron-beam lithography (EBL) 30
2.5 Strain engineering 31
2.6 Impact of scaling 31
2.7 Beyond CMOS and beyond Si CMOS 33
References 36
3 Strain engineering 41
H.H. Radamson
3.1 Introduction 41
3.2 Basic definitions of strain type and design 41
3.3 Strain design in MOSFETs 44
3.4 Methods to induce strain in the CMOS structure 47
3.5 Embedded Si1 yCy(eSi1 yCy) for nMOS 54
References 64
Further reading 67
4 High-κ dielectric and metal gate 69

C. Zhao, X. Wang, W. Wang
4.1 The scaling of MOSFET devices 69
4.2 SiO2 and poly-Si gate electrode 70
vi Contents
4.3 High-κ dielectric and metal gate 74

4.4 Hafnium-based high-κ dielectrics 81
4.5 Integration of HfO2 dielectric with compatible metal gate electrodes 90
4.6 High-κ dielectric and metal gate electrodes for FinFETs 93
4.7 Summary 97
References 97
Further reading 103
5 Channel materials 105

H.H. Radamson, E. Simeon
5.1 Introduction 105
5.2 High-mobility channels 105
5.3 SiGe channel 107
5.4 Ge channel 108
5.5 GeSn channel 110
5.6 III-V channel 111
5.7 Two-dimensional channel materials 114
References 118
6 Challenges in ultra-shallow junction technology 125

E. Simoen, H.H. Radamson
6.2 Basics of dopant activation and diffusion 127
6.3 Shallow boron junctions in silicon 132
6.4 N-type diffusion in germanium 139
6.5 Summary and outlook 146
References 147
Further reading 155
7 Advanced contact technology 157

J. Luo, K.P. Jia
7.2 Characterizations of extremely low ρc 160
7.3 Ohmic contacts on Si/SiGe substrate 164
7.4 Ohmic contacts on Ge/III-V 180
7.5 Ohmic contacts on potential CMOS channel materials 197
7.6 Summary 201
References 201
8 Advanced interconnect technology and reliability 215

Y. Li, C. Zhao
8.2 Copper interconnect integration 216
8.3 Low-k dielectric characteristics and classification 221
8.4 Copper interaction with silicon and dielectrics 227
8.5 Metal diffusion barriers 228
Contents vii
8.6 Reliability of copper metallization 231

8.7 Reliability of advanced intermetal dielectrics 233
8.8 Reliability statistics and failure models 239
8.9 The future of interconnect: beyond Cu and low-k 242
8.10 Summary 243
References 243
Final words 249

Acronyms 251
Index 255
This page intentionally left blank
Contributors
H.H. Radamson Institute of Microelectronics of Chinese Academy of Sciences

(IMECAS); University of Chinese Academy of Sciences (UCAS), Beijing, PR China
K.P. Jia Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

University of Chinese Academy of Sciences (UCAS), Beijing, PR China
Y. Li Imec, Leuven, Belgium
J. Luo Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

E. Simeon Imec, Leuven, Belgium
G. Wang Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

X. Wang Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

W. Wang Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

C. Zhao Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

Preface
We are in some way or other affected by the enormous impact of integrated electronics
and integrated circuits (ICs). They have changed virtually all aspects of our social and
professional lives. CMOS is the main unit in ICs and its design and structure have
undergone an evolution for many decades to follow the historical Moore’s Law.
This book is written in a notably understandable language for readers and provides
a good insight from the basics of CMOS to its development toward the advanced tran-
sistor structures.
The book is dedicated to the students and researchers in the field of electrical engi-
neering, who desire to learn more about the CMOS technology.
Acknowledgments
Professor Anders Hallen from KTH is acknowledged for his scientific discussions.
Many contributors from the Chinese Academy of Sciences have been involved
with this book. Our special gratitude to Dr. Huilong Zhu, Dr. Huaxiang Yin, and
Dr. Jiang Yan for their remarkable contributions and discussions to the all-last
high-k/metal gate integration part. We express our appreciation to Miss Dan Zhang,
Mr. Ningyuan Duan, Miss Shujuan Mao, Mr. Jinbiao Liu, and Miss Xuewei Zhao for
their valuable help.
Basics of metal–oxide–
semiconductor field-effect 1
transistor (MOSFET)
H.H. Radamson
Institute of Microelectronics of Chinese Academy of Sciences (IMECAS), Beijing, PR China;
1.1 Introduction
MOSFETs are the brick units of integrated circuits (ICs) which are used in many
devices in our daily life. They are being used for analog and digital applications.
The latter application is for a logic device where the transistor acts as a switch and
the performance is qualified with switching speed and switching energy. Since the
MOSFET structure has been downscaled in size according to Moore’s law, the density
and switching speed of transistors have enhanced exponentially where the switching
energy has reduced in a similar manner. This evolution of MOSFET structures finally
resulted that the planar design was abandoned to three-dimensional (3D) one provid-
ing a remarkably better control on the parasitic resistances and capacitances.
The idea of operational field effect transistors (FETs) extends to a US patent sub-
mitted by Julius Edgar in 1925; meanwhile, the discovery of MOSFETs by Dawon
Kahng and Martin Atalla at Bell labs emerged late in the 1950s. In the beginning,
the gate material in the transistors was usually metals, for example, Al, which led
to a “Metal Oxide Semiconductor” abbreviation for such devices. However, later,
heavily doped poly-Si was used as the standard gate material due to its high thermal
stability without reacting with gate oxide. Nowadays, the new research introduces the
metal gate material to be reused when SiO2 gate oxide was substituted with more
advanced high-k dielectrics.
The first ICs based on MOS were commercially announced by General Microelec-
tronics in 1964. Although this step was a large technological development, it took
almost a decade to solve the fundamental yield and reliability issues allowing
MOS as the main component in IC technology.
The main benefit of MOSFET in IC production has been that the transistor needs
practically no input current to manipulate the carrier current compared to the bipolar
transistor. The channel conductance can be controlled by an applied voltage to the gate
terminal in different operation modes.
MOSFETs operate through four-terminal contacts: gate (G), source (S), drain (D),
and substrate body (B) as shown in Fig. 1.1. The body terminal could be connected to
the source, resulting in a three-terminal device.
CMOS Past, Present and Future. https://doi.org/10.1016/B978-0-08-102139-2.00001-X

Copyright © 2018 Elsevier Ltd. All rights reserved.
2 CMOS Past, Present and Future
Metal Gate oxide

contact Spacer Gate
Extended
implantation SiGe S/D
GOX
Gate
Source Drain Shallow trench

Silicide
Si-fin isolation
Si substrate
Gate oxide Halo Transistor’s body

implantation
Circuit symbols
N-channel P-channel
MOSFET MOSFET
G G
S D S D
Fig. 1.1 Figure illustrates a cross section of 2D planar MOSFET (left) and 3D FinFET (right).
The circuit symbols have also been shown for nMOSFETs and pMOSFETs.
A complementary field effect transistor (CMOS) may contain an MOS transistor

(n-channel) where the drive carriers are electrons; meanwhile, the body is p-type
doped and a pMOS transistor (p-channel) where carriers are holes with an n-type
doped body.
The first microprocessors in the 1970s were all fabricated either from pMOS logic
or nMOS logic separately.
The coupling of these p- and nMOSFETS in CMOS configuration allows one high/
low switch which could be volume effectively packed in small chips with low-power
dissipation and heat generation. In CMOS, if both gates and both drains are connected,
then by applying an appropriate high-gate voltage the nMOSFET is active but the
pMOSFET does not conduct, and when a low-gate voltage is used the situation is
reverse. During the switching time when the voltage changes, the status is that both
MOSFETs conduct temporarily.
In principle, MOSFETs operate on the modulation of charge concentration through
an MOS capacitance formed between the gate and body electrodes when a dielectric
layer separates them. The carriers in the conducting channel may be inverted to elec-
trons for transistors with a p-type doped body (n-channel) or holes for n-type doped
ones (p-channel). The desired characteristics for a MOSFET are high Ion and low Ioff.
In traditional transistor processing, the S/D regions are formed by dopant implan-
tation or epitaxy growth, whereas the gate formation is started by deposition of gate
oxide and metal (or polycrystalline material) with nitride/oxide spacers. Further
adjustment of dopant profile below the gate region can be performed by a tilted angle
implantation (so-called Halo implantation). Such implanted pockets are necessary to
have better control over the carrier transport in the channel region. The typical MOS-
FET’s S/D formation was modified by introducing etch to create recess and later fill-
ing it with SiGe material to perform strain engineering. The reason for this
Basics of metal–oxide–semiconductor field-effect transistor (MOSFET) 3
modification was to improve the carrier transport in the channel region. Chapter 3
covers in detail the most important techniques to induce strain in the MOSFET
structure.
In transistors, the quality of gate oxide determines the leakage current. There is a
need to decrease the resistance in MOSFETs’ contacts and silicide layer is applied
prior to the metal contact for this purpose. Both the gate oxide and silicide layers were
modified in 90-nm technology node and beyond. For example, the gate oxide was
silicon oxide for many decades until, in the 45-nm technology node, it was replaced
by high-k material. In order to avoid any damage on high-k oxide due to high-
temperature annealing steps, the process flow of the transistor was modified to the
high-k-last process where these materials are deposited after thermal treatments. In
such cases, a dummy gate of silicon oxide is initially formed and the transistor is
processed. Afterward, the dummy gate is removed and the real gate containing
high-k materials is formed.
During the past 40 years, the MOSFET’s size has been shrunk to follow an inter-
national technological road map semiconductor (ITRS) to increase the density of tran-
sistors and to decrease the power consumption. In this evolution of MOSFET in 2003
in the 22-nm technology node, the conventional 2D transistor architecture was
replaced by 3D structure, the so-called trigate design [1]. Trigate transistors were later
renamed FinFETs, which were already defined by other research groups a few years
before [2]. The main structures of a (two-dimensional) 2D and (three-dimensional) 3D
MOSFET are defined in Fig. 1.1 as follows:
In order to study the performance of a transistor, different notations have to be
defined. The source, drain, and gate for transistors in Fig. 1.1 are written shortly as
S, D, and G, whereas other important notations are as follows:
VTH: Threshold voltage
VDS: Voltage between D and S terminals
Vov: Vov ¼ VGS VTH
Cox: Oxide capacitance
μe or μh: Channel mobility
kn: Coefficient of conductance
CGS, CGD: Capacitances between S, D, and G
rDS: Resistance between D/S (rDS ¼ 1/gDS)
VDD: Supply voltage
VGS: Voltage between G and S terminals
Vox: Voltage over oxide
VG: Gate voltage
gDS: Conductance between D/S
gm: Transconductance
Gint: Intrinsic gain (Gint ¼ gm/gDS)
RS, RD, and RG: Resistances of S, D, and G
In order to provide a better understanding of 2D-to-3D transition, the process flow of
these transistors is shown in Fig. 1.2. The major differences between 2D and 3D
processing are the formation of Si fin in FinFETs, S/D formation, and the
planarization step [3].
- Fin patterning vs planar active region patterning

- Oxide filling, planarization, and recessing S/D regions
- Doping to form well isolation
- Gate oxide growth, and dummy gate deposition, planarization and
patterning
- Doping to form S/D extensions
- Spacer deposition and patterning
- Epitaxy forming S/D regions (embedded SiGe and raised Si)
- Inter level dielectric (ILD) and chemical mechanical polish (CMP)
- Dummy gate removal
- Replacement of high-k and metal gate stack
- Self-aligned contact formation
- Back end of line
Fig. 1.2 Process flow for the bulk FinFETs and planar MOSFETs where additional processes
for FinFETs are highlighted.
1.2 Basics of MOSFET’s operation

Fig. 1.3A illustrates the band diagram of a typical planar MOSFET. In this figure, a
bending of energy bands (qφs) occurs at the Si side because of the work function dif-
ference between the gate material (metal or polycrystalline Si) (qψ g) and Si (qψ s), as
well as existence of any charges in the gate oxide and interfacial states. By applying a
gate voltage, the energy band at the SiO2/Si becomes flat where VG ¼ Vfb ¼ ψ g ψ s.
For a transistor, this is a flat band condition as shown in Fig. 1.3B.
In the flat-band condition, the surface electric field in the Si substrate is zero and
this makes the electric field in the oxide is also zero. The bandgap of SiO2 is 9 eV. The
vacuum level (E0) is defined as the energy required for electrons to be taken outside
the material. The difference between E0 and Ec is called electron affinity. This param-
eter is 4.05 eV for Si, whereas for SiO2 is 0.95 eV. In a MOSFET, the energy difference
between Ec of Si and SiO2 is 3.1 eV and this creates a barrier so-called Si–SiO2
cSi
3.1 eV qyg =4.0 eV
qySi = cSi + (Ec - Ef)
Ec
Ec Ef Ec
Ec Ef Ef Vfb
Ev Ef
Ev Ev
Ev
N+ poly gate Transistor’s body N+ poly gate Transistor’s body
(A) Gate oxide (B) Gate oxide
Fig. 1.3 (A) The energy band diagram of a p-type silicon substrate with a formed gate when no
applied gate voltage is applied and (B) with a gate voltage equal to Vfb.
electron energy barrier. In a similar way, the energy barrier for holes is defined and is
estimated to be 4.8 eV. Owing to these energy barriers, Holes and electrons are not
able to move through the SiO2 gate dielectric. It is also expected an energy barrier
of 3.1 eV at the interface between the polysilicon gate and SiO2.
In general, a relationship for the applied voltages in a MOSFET can be written as:
VG ¼ Vfb ϕs Vox (1.1)
where for a flat-band situation VG ¼ Vfb and ϕs ¼ Vox ¼ 0; therefore, a simple form of
Eq. (1.1) is written as: Vox ¼ VG Vfb.
To study the performance of a MOSFET, a special type of electrical measurement
can be performed when a MOS capacitance configuration is established. There are
three biasing modes when the gate voltages vary over a range of measurement: accu-
mulation, depletion, and inversion as shown in Fig. 1.4A–C.
VG < Vfb Vfb< VG < VTH VTH < VG
Gate Gate + + + + Gate

-------- +++++++ +++++++
Gate oxide Gate oxide Gate oxide
+ + + + -------- - - - -
-------- --------
--------
Accumulation Depletion Inversion
charge, Qacc
Ec
qVox fs = 2fB
3.1 eV
qVox Ei
Ec, Ef qfB
Ef
qVG Ev
Ev E0 qVG = qVT
Ec, Ef
qVG Depletion
qfs Ev region
Gate oxide Gate oxide Gate oxide
(A) (B) (C)
Fig. 1.4 (A) Accumulation, (B) depletion, and (C) inversion operation modes for a MOSFET.
1.2.1 Accumulation
This occurs when the gate voltage is more negative than the thermal potential for the
p-doped Si substrate. The charge density in the channel region is dependent on the
applied gate voltage. An electric field is constituted inside the oxide region where
the direction of the field is dependent on the amount of applied gate voltage.
Qacc
Vox ¼ (1.2)
Cox
1.2.2 Depletion
When the gate voltage is increased (VG is more positive for p-type and more negative
for n-type doped Si), the charge distribution changes and the energy bands bend down-
ward. As a result, a depletion region with a depth of Wdep is formed. At the surface, the
electron density (or hole density for n-type doped Si) is stills light. The voltage over
the gate oxide can be written as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Qacc qN a Wdep qNa2εs ϕs
Vox ¼ ¼ ¼ (1.3)
Cox Cox Cox
Therefore, φs can be written as
2
qN a Wdep
ϕs ¼ (1.4)
2εs
By inserting the Vox and ϕs expression in the original equation for gate voltage:
2
qN a Wdep qN Wdep
VG ¼ Vfb + ϕs + Vox ¼ Vfb + + a (1.5)
2εs Cox
This a quadratic equation and Wdep can be solved as a function of VG(Wdep).
1.2.3 Inversion
By further increase of VG, the energy band bends more and EF could be located in the
vicinity of Ec. This situation leads to a threshold where the p-type doped semicon-
ductor is inverted to N-type (or vice versa for n-type doped Si) at the interface of the
semiconductor and gate oxide. The threshold voltage is usually reached when the
surface electron (or hole) concentration (ns) approaches the same doping level as
in the bulk (Na). A simple rule can be applied to obtain threshold voltage and that
is when the surface potential satisfies the condition: ϕs(TH) ¼ 2ϕB (N-type body :
ϕs(TH) ¼ 2ϕB).
Therefore, the main equation of gate voltage for threshold voltage (VG ¼ VT) is as
follows:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
qNa2εs 2ϕB
VTH ¼ Vfb + 2ϕB + (1.6)
Cox
The gate voltage in the vicinity of threshold voltage is also known as the weak inver-
sion region.
1.2.4 Strong inversion

Strong inversion in MOS capacitance occurs when VG exceeds beyond VTH. In this
situation, both φs and Wdep do not change considerably from the depletion mode since
any small variation results in a large electron (or hole) density. This means that the
equation of VG is modified as follows:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Qdep Qinv qNa2εs 2ϕB Qinv
VG ¼ Vfb + 2ϕB ¼ Vfb + 2ϕB + (1.7)
Cox Cox Cox Cox
A summery of all operation modes (as described in Fig. 1.4) including strong inversion
is illustrated in Fig. 1.5.
In a transistor, no carrier transport occurs between the source and drain terminals
(ID ¼ 0) when no gate voltage is applied. As an example, in a MOS transistor, if the
gate terminal is set to a positive voltage and VGS > VTH, then a graded voltage is
formed across the channel region, leading to channel conductance, which becomes
1000 Fig. 1.5 Figure shows an

overview of different MOS
Accumulation operation modes.
100 Inversion
Flatband
Qs/QTH
10 Depletion
0.1
−20 −10 0 10 20 30 40
fs/VTH
large enough to allow electron transport in the channel. It is important to mention here
that the condition VDS < VGS VTH has to be satisfied in order to avoid a pinch through
a situation in the transistor. In the situation where VGS controls the conductance of the
channel and the channel operates as a variable resistor, conductance between the
source and drain (gDS) is written as

1 1
gDS ¼ ¼ kn VGS VTH VDS (1.8)
rDS 2
where rDS is the resistance between drain and source. As a result, the induced charge
(Q) in the channel region can be expressed as
Q ¼ Cox ðVGS VTH ϕs Þ (1.9)
The channel current can also be obtained by carrier mobility (μe), the formed electric
field (εy) along the channel direction (y), channel length (W), and charge (Q) in the
channel, according to
ID ¼ Wμe Qεy where εy ¼ dψ=dy (1.10)
Therefore, ID can be rewritten as follows:
ID dy ¼ Wμn Cox ðVGS VTH ψ Þdψ ) ID

W 1 2
¼ μn Cox ðVGS VTH ÞVDS VDS Þ (1.11)
L 2
or

W 1 2
ID ¼ μn ε0 εox ðVGS VTH ÞVDS VDS (1.12)
toc L 2
and the ID in saturation mode is obtained
μn Cox W
IDðsatÞ ¼ ðVGS VTH Þ2 (1.13)
2L
Eq. (1.12) shows that ID is a quadratic function of VDS with a maximum point at VTH.
This equation is used for transistor characterization where the channel mobility can be
obtained from the electrical measurements.
An nMOS transistor has three operating modes with the following conditions as
illustrated in Fig. 1.6:
1. Cutoff: It occurs when VGS < VTH and the channel current ID ¼ 0 A
This is the state when the transistor is in the turnoff mode. A more detailed study using
the Fermi-Dirac distribution shows that some electrons with thermal energy at the
source can move inside the channel region and flow to the drain. This causes a
ID VDS = VGS –VTH
Linear
(triode)
Saturation
VGS
VGS<VTH
Cutoff VDS
Fig. 1.6 I-V curves of an n-channel MOSFET in different operation modes [4].
subthreshold current which acts as an exponential function of VSG. The subthreshold

current is considered as a leakage current since ID has to approach to zero fast when the
transistor is switched off [5,6]:
VGS VTH
ID ffi ID0 e nV T (1.14)
where ID0 current, and the slope factor n is written as:
CD
n¼1+ (1.15)
Cox
where CD and COX are the capacitance of the depletion layer and oxide layer, respec-
tively. Since the subthreshold voltage I-V is exponentially dependent on threshold
voltage, it makes it vulnerable to any change in the transistor structure such as gate
oxide thickness, transistor body doping, and junction depth.
2. Triode or linear mode occurs when the condition is 0 < VDS < VGS VTH and the conduc-
tance is large and the carrier transport is established along the transistor channel with a cur-
rent of ID.
3. Saturation mode: It occurs when VGS > VTH and VDS VGS VTH. This is also a voltage con-
trolled current source and ID can be rewritten as ID ¼ (1/2) kn (VGS VTH)2.
It is worth mentioning here that all the above equations for currents and voltages for an
nMOS transistor can be written with a reversed sign for pMOS. This means that VTH,
VSG, VDS, and VOV become negative and, for example, conditions for saturation mode
become: VGS < VTH and VDS < VGS VTH.
1.3 Figures of merit of MOSFETs

As we mentioned earlier, MOSFETs may be used as a switch or amplifier. N-channel
transistors turn on when VGS > VTH; meanwhile, p-FETs are on if VGS < VTH. The tran-
sistors in the logic ICs are contacted in a cascaded structure where the output signal of
one gate works as the input for the next gate. The cascaded transistors need a resto-
ration of the signals to guarantee that a logic gate can reliably distinguish whether an
input signal is on logic state 1/0. As an example, a NAND component consists of
CMOS logic gates where an n-FET is connected to the ground and a p-FET is con-
nected to the positive supply voltage VDD. Such an Si CMOS circuit has excellent per-
formance of signal restoration offering low static power dissipation. This point is
crucial for complex logic ICs since power dissipation and the generated self-heating
are main problems. Thus, an FET-based successor for Si CMOS logic needs transistors
which are able to switch off fast. One of the figures of merit (FOM) of a transistor is
how the switching occurs, thus Ion/Ioff ratio has to be high. MOSFETs have to show an
Ion/Ioff ratio within 104–107 for logic application. Another FOM for MOSFETs is the
subthreshold swing (S) during switching. This FOM is measured in the mV/decade
and expresses how the VGS has to be changed to make a change in ID by a factor
of 10 as shown in Fig. 1.7.
The subthreshold voltage is defined as

dVG kB T εSi tox
S¼ 1+ (1.16)
d log 10 ID q log 10 e εox wd
In Eq. (1.16), εSi is the dielectric constant of silicon (11.9) and wd is the drain depletion
layer width. T is the absolute temperature, kB is Boltzmann’s constant, q is the electron
charge in absolute value, and e is the exponential number. It is clear from Eq. (1.16) that
S has a lower limit value of about 0.06 V/decade at room temperature. It therefore con-
stitutes a lower limit for the threshold voltage of scaled logic transistors, as a gate volt-
age swing of a few kBT/q is required to switch the device from off to on.
If a dc VGS is applied to the transistor on an on-state operating mode, then a small
RF signal will have to be amplified. This is the situation when a small variation of VG
occurs; the number of carriers in the channel as well as the drain current is changed.
The amplification behavior of MOSFET is characterized in terms of the unilateral
power gain (U), the intrinsic gain (Gint), and the current gain (h21). In general, all these
parameters are dependent on frequency as shown in Fig. 1.8.
Iog ID
ION
Subthreshold swing, S
[mV/dec ]
IOff
VG
VTH VDD
Fig. 1.7 Figure illustrates the ID(VG) curve to estimate the subthreshold voltage.
50 Fig. 1.8 The frequency

Measured h21 dependency of small-signal
U (−20 dB/dec)
Measured U current gain (h21) and
40
unilateral power gain U in an
Gain h21 and U (dB)
RF MOSFET [7].
fmax = 140 GHz
30
20 h21(−20 dB/dec)
fT = 110 GHz
10
0
1 10 100
Frequency (GHz)
Intrinsic part Fig. 1.9 Figure illustrates a small

signal equivalent circuit for an FET.
RG Cgd RD
G D
gmvGS
Cgs rds
Ri
Rs
The most important FOM of RF transistors are the cutoff frequency (fT) and the
maximum frequency of oscillation (fmax).fT is defined as the frequency when the mag-
nitude of h21 has dropped to unity (0 dB) and at fmax when the unilateral power gain is
unity. Both fT and fmax determine the frequency limits where the transistor’s amplify-
ing performance is degraded. The frequency characteristics of MOSFETs are
described by the small signal equivalent circuit as shown in Fig. 1.9.
There are both intrinsic and extrinsic parts for a transistor. The intrinsic part covers
the gate and the channel region and it contains CGS, CGD, gm, rDS (see the notations in
Fig. 1.1) together with an intrinsic resistance Ri as shown in the figure. The extrinsic
elements are RG, RS, and RD. The fT and fmaxfor the intrinsic part are obtained from the
following expressions [8]:
gm
fT ð intÞ ¼ (1.17)
2π ðCGS + CGD Þ
gm
fmax ð intÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffi (1.18)
4π CGS gDS Ri
For RF transistors, high fT and fmax are mainly desired and the transistor can be
designed for this purpose.
1.4 Evolution of the MOSFET structure

MOSFETs have continuously downscaled to follow an ITRS. Scaling and evolution of
device architecture are discussed in detail in Chapter 2, but in this chapter the focus
will be on the carrier transport in the channel and short channel effect.
The size shrinking of the transistors resulted that the carrier transport in nanoscale
transistors becomes more different compared to the classical ones. The classical
MOSFET equations are modified for ballistic carrier transport which is explained
by carrier injection at the source [9]. The drain current in the saturation mode is
given by
IDðSatÞ ¼ W hvð0ÞiCox ðVG VTH Þ (1.19)
where hv(0)i is the average velocity of carriers at the source side and Cox(VG VT) is
the charge density. When the channel length becomes smaller, hv(0)i can be set to
thermal velocity (vT) since the carriers are thermally injected from source [10]. vT
is given by
sffiffiffiffiffiffiffiffiffiffiffiffi
2kB TL
vT ¼ (1.20)
πm∗t
where TL denotes the lattice temperature. The drain current in nanotransistors can be
rewritten as
sffiffiffiffiffiffiffiffiffiffiffiffi
2kB TL
Id ¼ WCox ðvG vTH Þ (1.21)
πm∗t
This indicates that the drive current is increased by reducing the effective mass of the
carrier (or inducing strain) [11].
In order to develop the discussion to a more detailed level, more knowledge about
how MOSFETs function has to be provided. Fig. 1.10A and B illustrates how during
the evolution of the MOSFET structure the parasitic capacitance and resistances are
changed. In general, in an optimized MOSFET, all the undesired resistances and
capacitances have to be minimized [5].
The downscaling of transistors leads to an increase in leakage current as a result of
the smaller gate length. For industrial applications, CMOS technology has to ensure a
negligible dissipated power in the standby state. Therefore, the leakage power of
8 1400
Intrinsic channel
capacitance 7 Channel 1200
Resistance (Ohm-mm)
Capacitance (F/mm)
resistance
6 1000
5 Parasitics 800
dominate!
Total parasitic 4 600
capacitance 3 Parasitic 400
resistance
2 200
1 0
120 100 80 60 40 20 0 120 100 80 60 40 20 0
Technology node (nm) Technology node (nm)
(A) (B)
Fig. 1.10 (A and B) The evolution of the parasitic components over different technology nodes.
transistors has to be controlled for logic application. In principle, there are no limita-
tions for reducing the size of transistors but practically the leakage current rises for
planar transistors with a gate length of 20 nm [6].
In nanotransistors, the body of transistor is very thin causing the external resistance
becomes comparable in magnitude to the channel resistance.
The rapid increase in parasitics in nanotransistors is due to the small space between
neighboring devices (a few tens of nanometers) to increase the density. In the chip, the
S/D and contact size are also scaled down to be compatible with the transistor size.
The S/D contacts and the gate are a few nanometers apart; also, the contact size will
lead to magnification of the parasitic problems.
The problem with parasitic resistances is a fundamental issue for integration of
nanotransistors in logic circuits. This occurs due to the small S/D volume compared
to the transistor body which is 5–10 nm thick [12,13]. This problem is even more
severe in integration of carbon nanotubes or graphene ribbons as channel material
where the transistor body is only one atomic layer [14].
Another issue with downscaling of MOSFETs is the control of the carrier transport
through the channel. This problem occurs when the drain voltage decreases source-to-
channel barrier height and allows an off-state leakage. This phenomenon is known as
drain-induced barrier lowering (DIBL) or short channel effect (SCE).
Although control of SCE is a large problem for 2D transistors, this issue had
no impact on 3D transistors. For almost 40 years, MOSFETs were manufactured
in planar form until 2003 when Intel introduced 3D design, the so-called multigate
field effect transistor (MuGFET). MuGFETs are fully depleted transistors and they
are categorized to include double gate, Trigate, FinFETs, π-gate, Ω-gate, and the
gate-all-around (GAA) transistors. Gate control was continually improved in
the MuGFET group starting with double gate to the GAA transistors as shown
in Fig. 1.11. The benefits of MUGFETs compared to the 2D MOSFETs are as
follows:
1. The problem with doping of the body compared to MOSFETs is solved
2. MuGFETs have a remarkably higher drive current
3. The problem of the random dopant fluctuation is reduced
Gate Gate Gate 1 Gate
Bulk-Si SOI Gate 2
Planar UTB planar Double gate Fin FET
Gate Gate Gate Gate
GAA Ω-gate Pi-gate Tri-gate
Fig. 1.11 Schematics of different transistor designs starting from the classical to advanced
three-dimensional transistors.
The two important points for the performance of MUGFETs are the thickness of BOX
to control the transport through the channel and the doping level in the body for a
designed threshold voltage. At first, by reducing the BOX thickness, DIBL is
decreased. The strategy of BOX thickness in downscaling of the transistors is dem-
onstrated in Fig. 1.12A and B. This is crucial to increase the control of carrier transport
as well as high drain current [15].
Following this design, FinFETs were introduced. The distinguishable difference of
a FinFET structure with other transistors is that the body of the device is a thin fin of Si
(SiGe, Ge, or III–V material) where its thickness is the effective channel length of the
transistor. Therefore, all categories of these transistors are also known as multigate
FETs (MUGFETs).
In principle, MUGFET shave a number of parallel nanowires (the so-called fingers)
with a joint gate electrode. In such structures, the drive current drive could be
increased by increasing the number of fingers of the transistor. The most benefit of
3D transistors is that by wrapping around the transistor gate, a better control over elec-
trical transport through the channel is provided. Therefore, the SCE, which is an
important problem for small gate lengths, is reduced, resulting in a decrease in the
leakage current in the transistor. This leakage phenomenon emerges drastically for
gate lengths smaller than 20 nm [6].
The performance of FinFETs depends strongly on physical dimensions (length,
width, and height), and shape as well as on the crystal orientation of the fin and body
doping concentration.
The width of the fin is directly coupled to the short channel effect, whereas the
height is limited to the etch process. Fig. 1.13 shows the impact of the fin dimensions
on DIBEL. The figure shows that careful attention has to be paid on one side on the
link of fin width to gate length of transistor but also the fin height [16]. The optimum
shape is obtained when the width and height of fins have to be less than a half and one-
fifth of the gate length, respectively.
Impact of BOX thickness tsi reduction with Lg scaling

300 11
Impact of UTBOX exp. FDSOI (EOT ~ 1.7 nm) BOX = 7.5 nm 18 nm 38 nm
10
250
LG = 18 nm 9
EOT = 1.7 nm 14 nm
200
DIBL (mV/V)
TBOX = 145 nm 8 12 nm
Tsi (nm)
24 nm
150 7
22 nm
6 BOX = 10 nm
100
Tox > 1.2 nm
TBOX = 10 nm 5
with GP
50
4 BOX = 50 nm
1020cm−3
0 3
0 5 10 15 10 20 30
TSOI (nm) Gate length (nm)
Fig. 1.12 (Left) The impact of oxide thickness of SOI wafer on transistors’ DIBL whereas
(right) the correlation between Si cap layer thickness with transistors’ gate lengths.
Tri-Gate FET
Lg relaxed fin dimensions
Gate WSi > Lg/2; HSi > Lg/5
Lg
FinFET
hSi Si Narrow fin
WSi ~ Lg/2
wSi hSi Si Gate

Body dimensions Pfin
1.5 required for wSi
DIBL = 100 mV/V
HSi / Leff
1.0 Tox = 1.1 nm Lg
0.5
Gate
hSi Si
UTB FET
0.0 wSi Ultra thin SOI
0.5 1.0 1.5 2.0
WSi / Leff HSi ~ Lg/5
Fig. 1.13 Correlation of FinFETs’ width and height to the gate length and the impact on DIBIL.
The shape of fins depends on processing and it determines the transistor’s perfor-
mance. For example, a triangle fin is usually desired not only due to its high structural
strength for manufacturing, but also because such fins have a higher area-to-volume
ratio compared to rectangular ones.
Since its conduction in Si depends on the crystal orientation, the optimum mobility
is obtained in (100) and (110) direction for nFETs and pFETs, respectively [17]. From
a practical point of view, when the fins are formed in // or ┴ to the wafer flat, the tran-
sistor channel is oriented along (110) planes, otherwise a rotation at 45 degrees makes
a (100) oriented channel as shown in Fig. 1.14 [17].
Another random document with
no related content on Scribd:
the act.
"Not at all, I'm just issuing a fair warning that the signs that say:
'DANGER! HIGH-VOLTAGE!' are not there for appearance."
"Sounds like a threat to me."
"Have I threatened you? It sounds to me as though I were more than
anxious for your welfare. Any threat of which you speak is utterly
without grounds, and is a figment of your imagination; based upon
distrust of Venus Equilateral, and the personnel of Venus Equilateral
Relay Station."
Kingman shut up. He went down the list, marking off items here and
there. While he was marking, Channing scribbled a circuit and listed
the parts. He handed it over as Kingman finished.
"This is your circuit?" asked the lawyer skeptically.
"Yes."
"I shall have to ask for an explanation of the symbols involved."
"I shall be happy to present you with a book on essential radio
technique," offered Channing. "A perusal of which will place you in
possession of considerable knowledge. Will that suffice?"
"I believe so. I cannot understand how; being uncertain of your steps
a few minutes ago, you are now presenting me with a circuit of your
intended experiment."
"The circuit is, of course, merely symbolic. We shall change many of
the constants before the day is over—in fact, we may even change
the circuit."
"I shall require a notice before each change so that I may pass upon
the legal aspects."
"Walt," said Don, "will you accompany me to a transparency
experiment on the Ninth Level?"
"Be more than glad to," said Walt. "Let's go!"
They left the office quickly, and started for Joe's. They had not
reached the combined liquor vending and restaurant establishment
when the communicator called for Channing. It was announcing the
arrival of Barney Carroll, so instead of heading for Joe's, they went
to the landing stage at the south end of the station to greet the
Visitor.
"Barney," said Don, "of all the companies, why did you pick on Terran
Electric?"
"Gave us the best deal," said the huge, grinning man.
"Yeah, and they're getting the best of my goat right now."
"Well, Jim and I couldn't handle anything as big as the power
transmission set-up. They paid out a large slice of jack for the
complete rights. All of us are well paid now. After all, I'm primarily
interested in Martian artifacts, you know."
"I wonder if they had lawyers," smiled Walt wryly.
"Probably. And, no doubt, the legals had a lot to do with the fall of the
Martian civilization."
"As it will probably get this one so wound up with red tape that
progress will be impossible—or impractical."
"Well, Barney, let's take a run up to the lab. We can make paper-talk
even if Brother Kingman won't let us set it to soldering iron. There
are a lot of things I want to ask you about the tube."
They sat around a drawing table and Channing began to sketch.
"What I'd hoped to do is this," he said, drawing a schematic design.
"We're not interested in power transmission, but your gadget will do
a bit of voltage amplification because of its utter indifference to the
power-line problem of impedance matching. We can take a relay
tube and put in ten watts, say, across ten thousand ohms. That
means the input will be somewhat above three hundred volts. Now, if
our output is across a hundred thousand ohms, ten watts will give us
one thousand volts. So we can get voltage amplification at the
expense of current—which we will not need. Unfortunately, the relay
tube as well as the rest of the system will give out with the same kind
of power that it is impressed with—so we'll have amplification of
driver radiation. Then we'll need a detector. We haven't been able to
get one either yet, but this is a start, providing that Terran Electric will
permit us to take a deep breath without wanting to pass on it."
"I think you may be able to get amplification," said Barney. "But to do
it, you'll have to detect it first."
"Huh?"
"Sure. Before these darned things will work, this in-phase anode
must be right on the beam. That means that you'll require a feedback
circuit from the final stage to feed the in-phase anodes. Could be
done without detection, I suppose."
"Well, for one thing, we're going to get some amplification if we
change the primary anode—so. That won't permit the thing to handle
any power, but it will isolate the output from the input and permit
more amplifications. Follow?"
"Can we try it?"
"As soon as I get Terran Electric's permission."
"Here we go again!" groaned Walt.
"Yeah," said Don to Barney, "now you'll see the kind of birds you sold
your gadget to."
They found Kingman and Farrell in conference. Channing offered his
suggestion immediately, and Kingman looked it over, shaking his
head.
"It is not permitted to alter, change, rework, or repair tubes owned by
Terran Electric," he said.
"What are we permitted to do?" asked Channing.
"Give me your recommendation and I shall have the shop at Terran
Electric perform the operation."
"At cost?"
"Cost plus a slight profit. Terran Electric, just as Venus Equilateral, is
not in business from an altruistic standpoint."
"I see."
"Also," said Kingman severely, "I noticed one of your men changing
the circuit slightly without permission. Why?"
"Who was it?"
"The man known as Thomas."
"Charles Thomas is in charge of development work," said Channing.
"He probably noted some slight effect that he wanted to check."
"He should have notified me first—I don't care how minute the
change. I must pass on changes first."
"But you wouldn't know their worth," objected Barney.
"No, but Mr. Farrell does, and will so advise me."
Wes looked at Channing. "Have you been to the Ninth Level yet?"
"Nope," said Channing.
"May I accompany you?"
Channing looked at Farrell critically. The Terran Electric engineer
seemed sincere, and the pained expression on his face looked like
frustrated sympathy to Don. "Come along," he said.
Barney smiled cheerfully at the sign on Joe's door. "That's a good

one, 'Best Bar in Twenty-seven Million Miles, Minimum!' What's the
qualification for?"
"That's as close as Terra ever gets. Most of the time the nearest bar
is at Northern Landing, Venus; sixty-seven million miles from here.
Come on in and we'll get plastered."
Farrell said, "Look, fellows, I know how you feel. They didn't tell me
that you weren't going to be given permission to work. I understood
that I was to sort of walk along, offer suggestions and sort of prepare
myself to take over some research myself. This is sickening."
"I think you mean that."
"May I use your telephone? I want to resign."
"Wait a minute. If you're that sincere, why don't we outguess 'em?"
"Could do," said Wes. "How?"
"Is there any reason why we couldn't take a poke to Sol himself?"
"You mean haul power out of the sun?"
"That's the general idea. Barney, what do you think?"
"Could be—but it would take a redesign."
"Fine. And may we pray that the redesign is good enough to make a
difference to the Interplanetary Patent Office." Channing called Joe,
"The same. Three Moons all around. Scotch," he explained to the
others, "synthesized in the Palanortis Country."
"Our favorite import," said Walt.
Joe grinned. "Another tablecloth session in progress?"
"Could be. As soon as we oil the think-tank, we'll know for sure."
"What does he mean?" asked Barney.
Joe smiled. "They all have laboratories and draftsmen and
textbooks," he said. "But for real engineering, they use my
tablecloths. Three more problems and I'll have a complete tablecloth
course in astrophysics, with a sideline in cartooning and a minor
degree in mechanical engineering."
"Oh?"
"Sure. Give 'em a free hand, and a couple of your tubes and a
tablecloth and they'll have 'em frying eggs by morning. When I came
out here, they demanded a commercial bond and I thought they
were nuts. Who ever heard of making a restaurateur post a bond? I
discovered that all of their inventions are initially tinkered out right
here in the dining room—I could steal 'em blind if I were dishonest!"
Joe smiled hugely. "This is the only place in the System where the
tablecloths have been through blueprint machines. That," he said
confidentially to Barney, "is why some of the stuff is slightly garbled.
Scotch mixed with the drawings. They have the cloths inspected by
the engineering department before they're laundered; I lose a lot of
tablecloths that way."
Joe left cheerfully amid laughter.
The Three Moons came next, and then Don began to sketch.
"Suppose we make a driver tube like this," he said. "And we couple
the top end, where the cathode is, to the input side of the relay tube.
Only the input side will require a variable-impedance anode, coupled
back from the cathode to limit the input to the required value. Then
the coupling anodes must be served with an automatic-coupling
circuit so that the limiting power is passed without wastage."
Barney pulled out a pencil. "If you make that automatic-coupling
circuit dependent upon the output from the terminal ends," he said,
"it will accept only the amount of input that is required by the power
being used from the output. Over-cooling these two anodes will
inhibit the power-intake."
"Right," said Wes. "And I am of the opinion that the power available
from Sol is of a magnitude that will permit operation over and above
the limit."
"Four million tons of energy per second!" exploded Walt. "That's
playing with fire!"
"You bet. We'll fix 'em with that!"
"Our experience with relay tubes," said Farrell slowly, "indicates that
some increase in range is possible with additional anode-focussing.
Build your tube-top with an extra set of anodes, and that'll give us
better control of the beam."
"We're getting farther and farther from the subject of
communications," said Channing with a smile. "But I think that we'll
get more out of this."
"How so?"
"Until we get a chance to tinker with those tubes, we won't get ship-
to-ship two ways. So we'll gadgeteer up something that will make
Terran Electric foam at the mouth, and swap a hunk of it for full
freedom in our investigations. Or should we bust Terran Electric
wholeheartedly?"
"Let's slug 'em," said Walt.
"Go ahead," said Wes. "I'm utterly disgusted, though I think our
trouble is due to the management of Terran Electric. They like legal
tangles too much."
"We'll give 'em a legal tangle," said Barney. He was adding circuits to
the tablecloth sketch.
Channing, on his side, was sketching in some equations, and Walt
was working out some mechanical details. Joe came over, looked at
the tablecloth, and forthright went to the telephone and called
Warren. The mechanical designer came, and Channing looked up in
surprise. "Hi," he said, "I was just about to call you."
"Joe did."
"O.K. Look, Warren, can you fake up a gadget like this?"
Warren looked the thing over. "Give me about ten hours," he said.
"We've got a spare turnover driver from the Relay Girl that we can
hand-carve. There are a couple of water-boilers that we can strip, cut
open, and make to serve as the top end. How're you hoping to
maintain the vacuum?"
"Yes," said Wes Farrell, "that's going to be the problem. If there's any
adjusting of electrodes to do, this'll take months."
"That's why we, on Venus Equilateral, are ahead of the whole ding-
busted Solar System in tube development," said Don. "We'll run the
thing out in the open—and I do mean open! Instead of the tube
having the insides exhausted, the operators will have their envelopes
served with fresh, canned air."
"Like a cartoon I saw somewhere," grinned Walt. "Had a bird in full
armor tinkering with a radio set. The caption was: 'Why shield the
set?'"
"Phooey," said Warren. "Look, Tom Swift, is this another of the
Franks' brainchildren?"
"Tom Swift?" asked Wes.
"Yeah. That's the nom de plume he invents under. The other guy we
call Captain Lightning."
"Oh?" asked Farrell. "Do you read him, too?"
"Sure," grinned Warren. "And say, speaking of comics, I came upon
an old, old volume of Webster's International Dictionary in a rare-
edition library in Chicago a couple of months ago, and they define
'Comic' as amusing, funny, and ludicrous; not imaginative fiction.
How things change."
"They do."
"But to get back to this goldberg, what is it?"
"Warren," said Channing soberly, "sit down!" Warren did. "Now,"
grinned Channing, "this screwball gadget is an idea whereby we
hope to draw power out of the sun."
Warren swallowed once, and then waved for Joe. "Double," he told
the restaurateur. Then to the others he said, "Thanks for seating me.
I'm ill, I think. Hearing things. I could swear I heard someone say that
this thing is to take power from Sol."
"That's it."
"Um-m-m. Remind me to quit Saturday. This is no job for a man
beset by hallucinations."
"You grinning idiot, we're not fooling!"
"Then you'd better quit," Warren told Don. "This is no job for a bird
with delusions of grandeur, either. Look, Don, you'll want this in the
experimental blister at south end? On a coupler to the beam-turret
so that it'll maintain direction at Sol?"
"Right. Couple it to the rotating stage if you can. Remember, that's
three miles from south end."
"We've still got a few high-power selsyns," said Warren, making
some notations of his own on the tablecloth. "And thanks to the guys
who laid out this station some years ago, we've plenty of unused
circuits from one end to the other. We'll couple it, all right. Oh,
Mother. Seems to me like you got a long way off of your intended
subject. Didn't you start out to make a detector for driver radiation?"
"Yup."
"And you end up tapping the sun. D'ye think it'll ever replace the
horse?"
"Could be. Might even replace the coal mine. That's to be seen.
Have you any idea of how long you'll be?"
"Make it ten hours. I'll get the whole crew on it at once."
"Fine."
"But look. What's the reason for this change in program?"
"That's easy," said Don. "First, we had a jam session. Second, we've
come to the conclusion that the longest way 'round is the shortest
way home. We're now in the throes of building something with which
to dazzle the bright-minded management of Terran Electric and thus
make them susceptible to our charm. We want a free hand at the
transmission tubes, and this looks like a fair bit of bait."
"I get it. Quote: 'Why buy power from Terran Electric? Hang a
Channing Power Beam on your chimney pot and tap the sun?'
Whoa, Mazie. Bring on the needle, Watson. Hang out the flags, fire
the cannon, ring the bells; for Venus Equilateral is about to hang a
pipeline right into four million tons of energy per second! Don, that's
a right, smart bit of power to doodle with. Can you handle it?"
"Sure," said Channing with a wave of his hand, "we'll hang a fuse in
the line!"
"O.K.," said Warren, sweeping the tablecloth off the table like Mysto,
the Magician, right out from under the glasses. "I'll be back—wearing
my asbestos pants!"
Wes Farrell looked dreamily at the ceiling. "This is a screwy joint," he
said idly. "What do we do for the next ten hours?"
"Red Herring stuff," said Channing with what he hoped was a
Machiavellian leer.
"Such as?"
"Making wise moves with the transmission tubes. Glomming the
barrister's desk with proposed ideas for his approval; as many as we
can think of so that he'll be kept busy. We might even think of
something that may work, meanwhile. Come, fellow conspirators, to
horse!" Channing picked up his glass and drained it, making a wry
face. "Rotten stuff—I wish I had a barrel of it!"
Channing surveyed the set-up in the blister. He inspected it carefully,

as did the others. When he spoke, his voice came through the
helmet receivers with a slightly tinny sound: "Anything wrong? Looks
O.K. to me."
"O.K. by me, too," said Farrell.
"Working in suit is not the best," said Don. "Barney, you're the bright-
eyed lad, can you align the plates?"
"I think so," came the muffled booming of Barney's powerful voice.
"Gimme a screwdriver!"
Barney fiddled with the plate-controls for several minutes. "She's
running on dead center alignment, now," he announced.
"Question," put in Wes, "do we get power immediately, or must we
wait whilst the beam gets there and returns?"
"You must run your power line before you get power," said Walt. "My
money is on the wait."
"Don't crack your anode-coupling circuit till then," warned Wes. "We
don't know a thing about this; I'd prefer to let it in easy-like instead of
opening the gate and letting the whole four million tons per second
come tearing in through this ammeter!"
"Might be a little warm having Sol in here with us," laughed
Channing. "This is once in my life when we don't need a milliameter,
but a million-ammeter!"
"Shall we assign a pseudonym for it?" chuckled Walt.
"Let's wait until we see how it works."
The minutes passed slowly, and then Wes announced: "She should
be here. Check your anode-coupler, Barney."
Barney advanced the dial, gingerly. The air that could have grown
tense was, of course, not present in the blister. But the term is just a
figure of speech, and therefore it may be proper to say that the air
grew tense. Fact is, it was the nerves of the men that grew tense.
Higher and higher went the dial, and still the meter stayed inert
against the zero-end pin.
"Not a wiggle," said Barney in disgust. He twirled the dial all the way
'round and snorted. The meter left the zero pin ever so slightly.
Channing turned the switch that increased the sensitivity of the
meter until the needle stood halfway up the scale.
"Solar power, here we come," he said in a dry voice. "One-half
ampere at seven volts! Three and one-half watts. Bring on your
atom-smashers! Bring on your power-consuming factory districts.
Hang the whole load of Central United States on the wires, for we
have three and one-half watts! Just enough to run an electric clock!"
"But would it keep time?" asked Barney. "Is the frequency right?"
"Nope—but we'd run it. Look, fellows, when anyone tells you about
this, insist that we got thirty-five hundred milliwatts on our first try. It
sounds bigger."
"O.K., so we're getting from Sol just about three-tenths of the soup
we need to make the set-up self sustaining," said Walt. "Wes, this in-
phase anode of yours—what can we do with it?"
"If this thing worked, I was going to suggest that there is enough
power out there to spare. We could possibly modulate the in-phase
anode with anything we wanted, and there would be enough junk
floating around in the photosphere to slam on through."
"Maybe it is that lack of selectivity that licks us now," said Don, "Run
the voltage up and down a bit. There should be D. C. running around
in Sol, too."
"Whatever this power-level is running at," said Barney, "we may get
in-phase voltage—or in-phase power by running a line from the
power terminal back. Move over, boys, I'm going to hang a test clip
in here."
Barney's gloved hands fumbled a bit, but the clip was attached. He
opened the anode-coupler once again, and the meter slammed
against the full-scale peg.
"See?" he said triumphantly.
"Yup," said Channing cryptically. "You, Bernard, have doubled our
input."
"Mind if I take a whack at aligning it?" asked Wes.
"Go ahead. What we need is a guy with eyes in his fingertips. Have
you?"
"No, but I'd like to try."
Farrell worked with the deflection plate alignment, and then said,
ruefully: "No dice, Barney had it right on the beam."
"Is she aligned with Sol?" asked Channing.
Walt squinted down the tube. "Couldn't be better," he said, blinking.
"Could it be that we're actually missing Sol?" asked Don. "I mean,
could it be that line-of-sight and line-of-power aren't one and the
same thing?"
"Could be," acknowledged Wes. Walt stepped to the verniers and
swung the big intake tube over a minute arc. The meter jumped once
more, and Channing stepped the sensitivity down again. Walt fiddled
until the meter read maximum and then he left the tube that way.
"Coming up," said Channing. "We've now four times our original try.
We now have enough juice to run an electric train—toy size!
Someone think of something else, please. I've had my idea for the
day."
"Let's juggle electrode-spacing," suggested Wes.
"Can do," said Walt, brandishing a huge spanner wrench in one
gloved hand.
Four solid, futile hours later, the power output of the solar beam was
still standing at a terrifying fourteen watts. Channing was scratching
furiously on a pad of paper with a large pencil; Walt was trying
voltage-variations on the supply-anodes in a desultory manner;
Barney was measuring the electrode spacing with a huge vernier
rule, and Wes was staring at the sun, dimmed to seeable brightness
by a set of dark glasses.
Wes was muttering to himself. "Electrode-voltages, O.K. ... alignment
perfect ... solar power output ... not like power-line electricity ... solar
composition ... Russell's mixture—"
"Whoooo said that!" roared Channing.
"Who said what?" asked Barney.
"Why bust our eardrums?" objected Walt.
"What do you mean?" asked Wes, coming to life for the moment.
"Something about Russell's Mixture. Who said that?"
"I did. Why?"
"Look, Wes, what are your cathodes made of?"
"Thorium, C. P. metal. That's why they're shipped in metal containers
in a vacuum."
"What happens if you try to use something else?"
"Don't work very well. In fact, if the output cathode and the input
dynode are not the same metal, they won't pass power at all."
"You're on the trail right now!" shouted Channing. "Russell's Mixture!"
"Sounds like a brand of smoking tobacco to me. Mind making a
noise like an encyclopedia and telling me what is Russell's Mixture?"
"Russell's Mixture is a conglomeration of elements which go into the
making of Sol—and all the other stars," explained Don. "Hydrogen,
Oxygen, Sodium, and Magnesium, Iron, Silicon, Potassium, and
Calcium. They, when mixed according to the formula for Russell's
Mixture, which can be found in any book on the composition of the
stars, become the most probable mixture of metals. They—Russell's
Mixture—go into the composition of all stars, what isn't mentioned in
the mix isn't important."
"And what has this Russell got that we haven't got?" asked Walt.
"H, O, Na, Mg, Fe, Si, K, and Ca. And we, dear people, have Th,
which Russell has not. Walt, call up the metallurgical lab and have
'em whip up a batch."
"Cook to a fine edge and serve with a spray of parsley? Or do we cut
it into cubes—"
"Go ahead," said Channing. "Be funny. You just heard the man say
that dissimilar dyno-cathodes do not work. What we need for our
solar beam is a dynode of Russell's Mixture so that it will be similar
to our cathode—which in this case is Sol. Follow me?"
"Yeah," said Walt, "I follow, but, brother, I'm a long way behind. But
I'll catch up," he promised as he made connection between his suit-
radio and the station communicator system. "Riley," he said, "here
we go again. Can you whip us up a batch of Russell's Mixture?"
Riley's laugh was audible to the others, since it was broadcast by
Walt's set. "Yeah, man, we can—if it's got metal in it? What, pray tell,
is Russell's Mixture?"
Walt explained the relation between Russell's Mixture and the
composition of Sol.
"Sun makers, hey?" asked Riley. "Is the chief screwball up there?"
"Yup," said Walt, grinning at Don.
"Sounds like him. Yeah, we can make you an alloy consisting of
Russell's Mixture. Tony's got it here, now, and it doesn't look hard.
How big a dynode do you want?"
Walt gave him the dimensions of the dynode in the solar tube.
"Cinch," said Riley. "You can have it in two hours."
"Swell."
"But it'll be hotter than hell. Better make that six or seven hours. We
may run into trouble making it jell."
"I'll have Arden slip you some pectin," said Walt. "Tomorrow morning,
then?"
"Better. That's a promise."
Walt turned to the rest. "If any of us can sleep," he said, "I suggest it.
Something tells me that tomorrow is going to be one of those days
that mother told me about. I'll buy a drink."
Walt opened the anode-coupler circuit, and the needle of the output
ammeter slammed across the scale and wound the needle halfway
around the stop pin. The shunt, which was an external, high-
dissipation job, turned red, burned the paint off of its radiator fins,
and then proceeded to melt. It sputtered in flying droplets of molten
metal. Smoke spewed from the case of the ammeter, dissipating in
the vacuum of the blister.
Walt closed the coupler circuit.
"Whammo!" he said. "Mind blowing a hundred-amp meter?"
"No," grinned Don. "I have a thousand-amp job that I'll sacrifice in
the same happy-hearted fashion. Get an idea of the power?"
"Voltmeter was hanging up around ten thousand volts just before the
amp-meter went by."
"Um-m-m. Ten thousand volts at a hundred amps. That is one million
watts, my friends, and no small potatoes. To run the station's
communicating equipment we need seven times that much. Can we
do it?"
"We can. I'll have Warren start running the main power bus down
here and we'll try it. Meanwhile, we've got a healthy cable from the
generator room; we can run the non-communicating drain of the
station from our plaything here. That should give us an idea. We can
use a couple of million watts right there. If this gadget will handle it,
we can make one that will take the whole load without groaning. I'm
calling Warren right now. He can start taking the load over from the
generators as we increase our intake. We'll fade, but not without a
flicker."
Walt hooked the output terminals of the tube to the huge cable
blocks, using sections of the same heavy cable.
Warren called: "Are you ready?"
"Fade her in," said Walt. He kept one eye on the line voltmeter and
opened the anode-coupler slightly. The meter dipped as Warren
shunted the station load over to the tube circuit. Walt brought the line
voltage up to above normal, and it immediately dropped as Warren
took more load from the solar intake. This jockeying went on for
several minutes until Warren called: "You've got it all. Now what?"
"Start running the bus down here to take the communications load,"
said Don. "We're running off of an eight hundred thousand mile
cathode now, and his power output is terrific. Or better, run us a
high-tension line down here and we'll save silver. We can ram ten
thousand volts up there for transformation. Get me?"
"What frequency?"
"Yeah," drawled Channing, "have Chuck Thomas run us a control
line from the primary frequency standard. We'll control our frequency
with that. O.K.?"
"Right-o."
Channing looked at the set-up once more. It was singularly
unprepossessing, this conglomeration of iron and steel and plastic.
There was absolutely nothing to indicate that two and one-third
million watts of power coursed from Sol, through its maze of anodes,
and into the electric lines of Venus Equilateral. The cathodes and
dynode glowed with their usual dull red glow, but there was no
coruscating aura of power around the elements of the system. The
gimbals that held the big tube slid easily, permitting the tube to rotate
freely as the selsyn motor kept the tube pointing at Sol. The supply
cables remained cool and operative, and to all appearances the set-
up was inert.
"O.K., fellows," said Channing, "this is it—"
He was interrupted by the frantic waving of Kingman, from the other
side of the air lock.
"I feel slightly conscious-stricken," he said with a smile that showed
that he didn't mean it at all. "But let us go and prepare the goat for
shearing."
Kingman's trouble was terrific, according to him. "Mr. Channing," he
complained, "you are not following our wishes. And you, Mr. Farrell,
have been decidedly amiss in your hobnobbing with the engineers
here. You were sent out as my consultant, not to assist them in their
endeavors."
"What's your grief?" asked Channing.
"I find that your laboratory has been changing the circuits without
having previously informed me of the proposed change," complained
Kingman. "I feel that I am within my rights in removing the tubes
brought here. Your investigations have not been sanctioned—" He
looked out through the air lock. "What are you doing out there?"
"We have just succeeded in taking power from the sun," said Don.
He tried to keep his voice even, but the exultation was too high in
him, and his voice sounded like sheer joy.
"You have been—" Kingman did a double-take. "You what?" he
yelled.
"Have succeeded in tapping Sol for power."
"Why, that's wonderful!"
"Thank you," said Don. "You will no doubt be glad to hear that Wes
Farrell was instrumental in this program."
"Then a certain part of the idea is rightfully the property of Terran
Electric," said Kingman.
"I'm afraid not," said Don. "Dr. Farrell's assistance was not
requested. Though his contribution was of great value, it was given
freely. He was not solicited. Therefore, since Terran Electric was not
consulted formally, Dr. Farrell's contribution to our solar power beam
can not be considered as offering a hold on our discovery."
"This is true, Dr. Farrell?"
"I'm afraid so. You see, I saw what was going on and became
interested, academically. I naturally offered a few minor suggestions
in somewhat the same manner as a motorist will stop and offer
another motorist assistance in changing a tire. The problem was
interesting to me and as a problem, it did not seem to me—"
"Your actions in discussing this with members of the Venus
Equilateral technical staff without authorization will cost us plenty,"
snapped Kingman. "However, we shall deal with you later."
"You know," said Farrell with a cheerfully malicious grin, "if you had
been less stuffy about our tubes, they might be less stuffy about my
contribution."
"Ah, these non-legal agreements are never satisfactory. But that is to
be discussed later. What do you intend to do with your invention, Dr.
Channing?"
Channing smiled in a superior manner. "As you see, the device is
small. Yet it handles a couple of million watts. An even smaller unit
might be made that would suffice to supply a home, or even a
community. As for the other end, I see no reason why the size might
not be increased to a point where it may obsolete all existing power-
generating stations."
Kingman's complexion turned slightly green. He swallowed hard.
"You, of course, would not attempt to put this on the market
yourself."
"No?" asked Channing. "I think you'll find that Venus Equilateral is as
large, if not larger, than Terran Electric, and we have an enviable
reputation for delivering the goods. We could sell refrigerators to the
Titan colony, if we had the V-E label on them and claimed they were
indispensable. Our escutcheon is not without its adherents."
"I see," said Kingman. His present volubility would not have jogged a
jury into freeing the armless wonder from a pick-pocketing charge.
"Is your invention patentable?"
"I think so. While certain phases of it are like the driver tube, which,
of course, is public domain, the applications are quite patentable. I
must admit that certain parts are of the power transmission tube, but
not enough for you to claim a hold. At any rate, I shall be busy for the
next hour, transmitting the details to Washington, so that the
Interplanetary Patent Office may rule on it. Our Terran legal
department has a direct line there, you know, and they have been
directed to maintain that contact at all costs."
"May I use your lines?"
"Certainly. They are public carriers. You will not be restricted any
more than any other man. I am certain that our right to transmit
company business without waiting for the usual turn will not be
contested."
"That sounds like a veiled threat."
"That sounds like slander!"
"Oh, no. Believe me. But wait, Dr. Channing. Is there no way in
which we can meet on a common ground?"
"I think so. We want a free hand in this tube proposition."
"For which rights you will turn over a nominal interest in solar
power?"
"Forty percent," said Channing.
"But we—"
"I know, you want control."
"We'd like it."
"Sorry. Those are our terms. Take 'em or leave 'em."
"Supposing that we offer you full and unrestricted rights to any or all
developments you or we make on the Martian transmission tubes?"
"That might be better to our liking."
"We might buck you," said Kingman, but there was doubt in his
voice.
"Yes? You know, Kingman, I'm not too sure that Venus Equilateral
wants to play around with power except as a maintenance angle.
What if we toss the solar beam to the public domain? That is within
our right, too."
Kingman's green color returned, this time accompanied with beads
of sweat. He turned to Farrell. "Is there nothing we can do? Is this
patentable?"
"No—Yes," grinned Farrell.
Kingman excused himself. He went to the office provided for him and
began to send messages to the Terran Electric Company offices at
Chicago. The forty-minute wait between message and answer was
torture to him, but it was explained to him that light and radio crossed
space at one hundred and eighty-six thousand miles per second and
that even an Act of Congress could do nothing to help him hurry it.
Meanwhile, Channing's description tied up the Terran Beam for
almost an hour at the standard rate of twelve hundred words per
minute. Their answers came within a few minutes of one another.
Channing tossed the 'gram before Kingman. "Idea definitely
patentable," said the wire.

ebook download CMOS Past, Present and Future 1st Edition - eBook PDF all chapter

Uploaded by

Copyright:

Available Formats

You might also like

ebook download CMOS Past, Present and Future 1st Edition - eBook PDF all chapter

Uploaded by

Document Information

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

ebook download CMOS Past, Present and Future 1st Edition - eBook PDF all chapter

Uploaded by

Copyright:

Available Formats

CMOS Past, Present and Future 1st

Edition - eBook PDF

Electric Utility Resource Planning: Past, Present and

(eBook PDF) American Foreign Policy: Past, Present, and

Understanding Mobility as a Service (MaaS): Past,

Advanced Biofuel Technologies: Present Status,

(eBook PDF) Criminological Theory: Past to Present:

(eBook PDF) Traditions & Encounters: A Global

CMOS VLSI Design A Circuits and Systems Perspective

Pan-genomics: Applications, Challenges, and Future

CMOS Past, Present,

Copyright © 2018 Elsevier Ltd. All rights reserved.

Library of Congress Cataloging-in-Publication Data

For information on all Woodhead publications

Publisher: Matthew Deans

Typeset by SPi Global, India

1 Basics of metal–oxide–semiconductor field-effect transistor (MOSFET) 1

2 Scaling and evolution of device architecture 19

4 High-κ dielectric and metal gate 69

4.3 High-κ dielectric and metal gate 74

5 Channel materials 105

6 Challenges in ultra-shallow junction technology 125

7 Advanced contact technology 157

8 Advanced interconnect technology and reliability 215

8.6 Reliability of copper metallization 231

Final words 249

H.H. Radamson Institute of Microelectronics of Chinese Academy of Sciences

K.P. Jia Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

Y. Li Imec, Leuven, Belgium

J. Luo Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

E. Simeon Imec, Leuven, Belgium

G. Wang Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

X. Wang Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

W. Wang Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

C. Zhao Institute of Microelectronics of Chinese Academy of Sciences (IMECAS);

CMOS Past, Present and Future. https://doi.org/10.1016/B978-0-08-102139-2.00001-X

Metal Gate oxide

Source Drain Shallow trench

Gate oxide Halo Transistor’s body

A complementary field effect transistor (CMOS) may contain an MOS transistor

- Fin patterning vs planar active region patterning

1.2 Basics of MOSFET’s operation

N+ poly gate Transistor’s body N+ poly gate Transistor’s body

(A) Gate oxide (B) Gate oxide

VG ¼ Vfb ϕs Vox (1.1)

VG < Vfb Vfb< VG < VTH VTH < VG

Gate Gate + + + + Gate

Gate oxide Gate oxide Gate oxide

(A) (B) (C)

Therefore, φs can be written as

This a quadratic equation and Wdep can be solved as a function of VG(Wdep).

1.2.4 Strong inversion

1000 Fig. 1.5 Figure shows an

Q ¼ Cox ðVGS VTH ϕs Þ (1.9)

ID ¼ Wμe Qεy where εy ¼ dψ=dy (1.10)

Therefore, ID can be rewritten as follows:

ID dy ¼ Wμn Cox ðVGS VTH ψ Þdψ ) ID

and the ID in saturation mode is obtained

ID VDS = VGS –VTH